2005 Symposium Proceedings 4

Document Sample
2005 Symposium Proceedings 4 Powered By Docstoc


The expertise of the entire South African membrane research base should be harnessed and co-
ordinated 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


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 symposium/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.

                                                                                    WISA-MTD 2005

                           A WORD FROM THE CHAIRPERSON

It has been two years since we met as “membrainologists”. In the meantime the WISA Conference 2004
came and is gone. Between our Vereeniging symposium and the Cape Town Conference a lot of research in
membrane technology took place. There is obviously no better way to share your findings and research
triumphs than with your membrane technology family. It is through a symposium like this that we refresh
ourselves of why we are in this terrain. Firstly, to assure our students that there are other membranologist
except those in their campuses. Secondly, to asset how far we have transgressed as a division and as
researchers so that we improve in our endeavors of bringing South Africa quality water. Finally, to
strengthen our relations with industry, communities and other relevant membrane stakeholders. It is through
this activity that we remind ourselves of the importance of volunteerism and serving in either branches or
divisions of Water Institute of Southern Africa. The latter being the reason why I wish to thank the outgoing
committee for having served the division so diligently in the past two years. In the same tone, I wish to thank
all the Water community that has made their resources available for us to hold our meetings. A big thank you
to all sponsors that has made this activity a success. In the spirit of building expertise, sharing knowledge
and improving quality of life, I, on behalf of the committee, wish to thank you for your participation and
hope that you will enjoy this 6th WISA-MTD Symposium and Workshop.

                                                                                              WISA-MTD 2005

                                 TECHNICAL PROGRAM

Sunday 13 March 2005
16:30 - 19:00     Early Registration

18:30 for 19:00   Meet & Greet

Monday 14 March 2005
7:30              Registration

8:30              Opening Address & Welcome by Division Chairman

8:45 - 9:30       Plenary Lecture 1: A critique of critical flux, critical concentrations, critical Peclet numbers
                  and critical pressure J. Howell

Short break
Session 1 (Joe Modise): Waste Water
9:40 - 10:00      Papers 1: Semi Dead-End Ultra filtration for large scale wastewater Re-use. Ir. S. van Hoof

10:00 - 10:20     Paper 2: Treatment of spent cutting oil with micro filtration, ultra filtration and membrane
                  distillation. I. Eggberry, JJ. Schoeman and CF. Schutte

10:20 - 10:40     Paper 3: The Activated Carbon – Micro filtration process: A novel technology for the
                  treatment of industrial effluents. PB. Persadh, EP. Jacobs and VL. Pillay

Tea Break:
Session 2 (Marshall Sheldon): Waste Water and Membrane Bioreactor Systems
11:00- 11:20      Paper 4: Membrane processes for cooling water treatment in a fertilizer plant. O. Mahle
                  and SJ. Modise

11:20 - 11:40     Paper 5:       Design and performance of BNR Activated Sludge systems with rat sheet
                  membranes for solid-liquid separation. M. Ramphao, MC. Wentzel and GA. Ekama

11:40 - 12:00     Papers 6: Quantifying growth kinetics of Phanerochaete chrysosporium immobilised on a
                  vertically oriented polysulphone capillary membrane. SKO. Ntwampe and MS. Sheldon

13:00 - 15:00     Workshop Introduction to membrane processes for the water industries – J. Howell

Tea Break :       Annual General MTD meeting

17:00 - 18:00     Poster Presentations

19:00 for 19:30   Conference Dinner / Evening Function

                                                                                               WISA-MTD 2005

Tuesday 15 March 2005
Session chair: (Stephanie Marais)
8:30             Start of days proceedings

8:45 - 9:30      Plenary Lecture 2: Sonic methods for visualization of fouling and reduction of fouling
                 during membrane filtration of natural brown water - RD. Sanderson

9:35 - 10:00     Paper 7: Eskom Holdings, Tutuka Power Station SRO Project: Desalination of mine
                 effluent, cooling water and the utilisation of first stage permeate as feed to the
                 demineralisation plant. JD. Janse van Noordwyk
Short break
Session 3 (Stephanie Marais): Drinking water
10:10 - 10:30    Paper 8: Membrane Technology – The sustainable solution to drinking water provision in
                 developing economies. VL. Pillay and EP. Jacobs

10:30 - 10:50    Paper 9: A novel approach to drinking water treatment: Ozonation bio-membrane filtration
                 for organic carbon removal. T. Leiknes

10:50 - 11:10    Paper 10: An ultra filtration systems without pumps: Exploiting natural heads. EP. Jacobs,
                 VL. Pillay and S. Victor
Tea Break:
Session 4 (Gerhard Offringa): Inorganics / Fouling
11: 35 - 11:55   Paper 11: Mechanistic study of active transport of copper(II) from aqueous medium using
                 LIX 984 as a carrier across a tubular supported liquid membrane. M. Aziz:

11:55 - 12:15    Paper 12:    Ultrasonic fouling index meter: A novel instrument for easy detection of
                 membrane fouling. RD. Sanderson, W. Scharff, J. Schiller, DK. Hallbauer, MB. Mbanjwa,
                 SK. Sikder, FJ. Reineke, DS. Mclachlan and DA. Keuler

12:15 - 12:35    Papers 13: Wavelets-based ultrasonic method of visualising fouling in micro filtration of
                 beer-brewing wastewater. SK. Sikder, MB. Mbanjwa, DA. Keuler, DS. McLachlan, FJ.
                 Reineke, RD. Sanderson

12:35 - 12:55    Paper 14:    Hydrophilisation of capillary ultra filtration membranes with the use of a
                 branched poly(ethylene oxide)-block-polysulphone copolymer. SP. Roux, EP. Jacobs, AJ.
                 van Reenen, C. Morkel and M. Meincken.

Closing of conference by new chairman


                                                                                             WISA-MTD 2005

                          POSTER PRESENTATIONS

 Poster 1:   Immobilisation and biofilm development of Phanerochaete Chrysosporium on polysulphone
             and ceramic membranes. MS Sheldon and K Mohamed

Poster 2:    A mathematical model of flow behaviour inside a capillary membrane bioreactor.
             B Godongwana and MS Sheldon

Poster 3:    The recovery of copper by tubular supported liquid membrane. M Aziz and S Mneno

Poster 4:    Development and evaluation of silicone rubber membranes as aerators for membrane
             bioreactors. XP Mbulawa, EP Jacobs and VL Pillay

Poster 5:     The development of sustainable salt sinks with the scope of producing value-added-
              products and re-usable water - Evaluation of an integrated membrane system for the
              recovery and purification of Magnesium Sulphate and Sodium Chloride from brine streams.
              L Mariah, C A. Buckley, D Jagan, E Drioli and E Curcio

                                                                                     WISA-MTD 2005

                                    PLENARY LECTURE 1

                                    CRITICAL PRESSURE.

                                                 J Howell
                      Emeritus Professor of Biochemical Engineering; University of Bath

This paper discusses the idea of a critical flux and how it has been developed by theoreticians and
experimentalists. Originally postulated as method of avoiding fouling it is now seen more as a
technique to limit fouling. Currently industry uses controlled fluxes at lower values than in the past
in many applications. They can operate many plants especially membrane bioreactors for extended
periods between chemical cleans, but they remain necessary. The paper notes these advantages and
limitations considers whether the concept of a critical flux which cannot be achieved in practice has
further value. In particular it is probable that the phenomenon described by the term is perhaps
better characterised by a critical membrane concentration or a critical Peclet number. Methods of
measuring the critical flux are briefly described and the concept of a sustainable flux introduced.

                                                                                          WISA-MTD 2005

                         SEMI DEAD-END ULTRAFILTRATION,
                                  IN THE 21ST CENTURY

                                       Ir. Stephan van Hoof
                                     Technology Manager, X-Flow

In recent years new developments in membrane filtration have led to a number of new applications
for this technology. One of these relatively new developments is the use of dead-end backwashable
Hollow Fibre Ultrafiltration. This has been widely applied in the area of potable water production as
a barrier to pathogens and viruses, surface water treatment, and for pre-treatment of Reverse
Osmosis (RO) for municipal effluent reuse and seawater desalination.

Final barrier potable water production
Regulations for potable water quality world-wide are becoming much more stringent, especially
with respect to microbial quality. With UF membranes a 6 log removal of bacteria and a 4 log
removal of viruses is feasible, dramatically increasing consumer and water company peace of mind.

Surface water treatment
A popular application of UF is surface water treatment, be it to produce irrigation water or potable
water or to feed an upstream Reverse osmosis plant, for the production of process water.

Sea water treatment
The Middle East and other arid regions have always been experiencing water shortages. To solve
this, membrane filtration is employed for desalination, next to thermal desalination. Many of these
sea water reverse osmosis plants are experiencing major fouling problems of their membrane
elements. Their conventional pre-treatment is simply not able to provide a sufficient waterquality.
Ultrafiltration can cope with the seawater quality, providing an excellent waterquality to feed any
reverse osmosis plant.

Re-use of effluent
The reuse of municipal effluent, otherwise typically discharged onto surface water, is a very
effective way to diminish demands on potable water resources. Reusing effluent for non potable
applications like irrigation or the production of process water saves potable water sources otherwise
                                                                                     WISA-MTD 2005

used for these non-potable applications. However, effluent can also be indirectly reused for potable
water production by supplementing existing potable water resources, like reservoirs and aquifers.
The ultimate reuse application, direct reuse of effluent for potable water production, is still only
executed in one location, Windhoek. Although technically not a problem, this application is still
facing scepticism worldwide, be it from a psychological and religious nature, not a technical one.

This paper will describe the Ultrafiltration technology responsible for the feasibility of these
applications. A number of plants will be discussed, amongst which the largest membrane based
effluent reuse plant in the world in Kuwait.

                                                                                     WISA-MTD 2005


                              J.J. Schoeman, C.F. Schutte and I. Eggberry
Water Utilisation Division, Department of Chemical Engineering, University of Pretoria, Fax: (012) 362-5089; email:

Small quantities of oil in water can affect biological treatment processes adversely. Therefore,
industrial effluents containing residual oil should be treated to remove the oil to the desired levels
prior to discharge into the municipal treatment system.                   Physical/chemical and membrane
technologies are available that can be used for the removal of residual and emulsified oil from
effluents(1). However, in many cases, these technologies cannot remove the oil concentration to the
desired levels. A combination, however, of either microfiltration (MF) or ultrafiltration (UF) with
membrane distillation (MD) should be able to remove the oil in the effluent to lower concentration
levels than any of the processes alone(2). Therefore, MF, UF and MD were evaluated for the
removal of oil from spent cutting oil.


MF treatment
Spent cutting oil was pretreated by passing the emulsion through a 70 micron bag filter prior to MF
treatment. The cutting oil was treated in batches by passing it through a Membralox MF ceramic
membrane (0,2 m2 membrane area) at an inlet pressure of 400 kPa. A compressed air backwash
was used for fouling control. Permeate flux and water recovery were measured as a function of
time. The COD and oil concentration levels in the feed and permeate were also determined.

UF treatment
The spent cutting oil was pretreated by passing it through a 70 micron bag filter prior to UF
treatment (batch mode) using polysulphone capillary membranes (40 000 MWCO - area 1 m2).
The permeate flux and water recovery were measured as a function of time as well as the COD,
BOD and oil concentrations. Tests were also conducted with tubular polysulphone UF membranes
(60 000 MWCO).
                                                                                                 WISA-MTD 2005

MD treatment
Permeate from the UF process was treated in a membrane distillation unit (polypropylene
membranes) (batch mode) and the flux and oil and grease concentrations were determined.

Results and discussion

MF treatment
The permeate flux remained fairly constant for the first six runs and varied between approximately
25,2 l/h to 15,6 l/h. Permeate flux was higher for the next three runs (30 l/h to 22,8 l/h) due to a
change in composition of the spent cutting oil.      It also appeared that no significant membrane
fouling took place. Permeate flux further increased with increasing feed temperature.

Typical COD and oil removal results are shown in Table 1.

Table 1:   COD and oil concentrations of the spent cutting oil feed, permeate and brine and
percentage removals.
Run        Feed                    Permeate              Brine                 Removal (%)
           COD         Oil         COD        Oil        COD        Oil        COD         Oil
           mg/l        mg/l        mg/l       mg/l       mg/l       mg/l
1          52 300      13 690      6 070      10         248 000    71 590     88,4        99,9
5          58 800      17 600      6 790      370        268 000    98 300     88,5        97,9
12         73 800      19 794      7 140      64                               90,3        99,7

Excellent COD and oil removals could be obtained with MF treatment of the effluent.

UF treatment
Permeate flux for five batch runs varied between approximately 1 730 l/m².d and 890 l/m².d
(capillary membranes).       It further appeared that membrane fouling could be controlled with
detergent cleanings.

COD, BOD and oil and grease concentration levels and percentage removals are shown in Table 2.

                                                                                      WISA-MTD 2005

Table 2: COD, BOD and oil and grease concentrations in the feed and permeate and percentage
                   Sample 1                               Sample 2
                   COD         BOD          Oil/grease    COD            BOD            Oil/grease
                   mg/l        mg/l         mg/l          mg/l           mg/l           mg/l
Feed               78 200                   14 200        59 500         -              14 200
Permeate           4 420       2 550        70            6 970          -              60
Removal (%)        94,4                     99,5          88,3                          99,6

Excellent COD and oil removals were again obtained.

Ultrafiltration using tubular polysulphone membranes has shown that the oil and grease in the UF
feed could be reduced from 1 140 mg/l to 39 mg/l (96,6% removal). This UF permeate was then
treated with MD.

MD treatment
The oil and grease concentration in the UF permeate could be further reduced from 39 mg/l to 8,4
mg/l with membrane distillation. Therefore, it appears that it should be difficult to reduce the oil
and grease concentration level to less than 2 mg/l (discharge requirement) with a combination of
MF or UF and MD. Physical-chemical pre-treatment followed by membrane treatment might give
the required concentration for discharge. This matter, however, needs to be investigated.

1.     José Manual Benito et al (2002): Design and construction of a modular pilot plant for the
       treatment of oil-containing wastewaters. Desalination 147 (2002) 5 - 10.
2.     M. Gryta, K. Karakulski and A.W. Morawski (2001): Purification of oily wastewater by
       hybrid UF/MD. Wat. Res. Vol. 35, No. 15, pp. 3665 - 3669.

                                                                                     WISA-MTD 2005


                                PB Persadh1, EP Jacobs2, and VL Pillay1

   Durban Institute of Technology, Depart of Chemical Engineering, S4 L1 Steve Biko Campus, Durban Institute of
  Technology, Mansfield Road, Durban, 4001 email: pillayvl@dit.ac.za; 2University of Stellenbosch, Chemistry and
                                 Polymer Sciences, Private Bag X1, Stellenbosch

Microfiltration is widely applied in solid-liquid separations. In general, conventional microfilters
are not expected to remove organic species, except for those associated with suspended solids and
colloids. Activated carbon has been widely used in potable water treatment for the removal of trace
levels of organics. Powdered activated carbon (PAC) is conventionally used by mixing it
intimatedly with the water. Organics adsorb onto the PAC which is then removed by conventional
filtration. A major drawback is that activated carbon has a finite adsorption capacity (saturation
limit). Hence its use in effluent treatment has been limited, since vast qualtities would be required to
remove the high organics loads in effluents.

The ACMF process is a novel development by the Water Technology Group, Durban Institute of
Technology. Firstly a PAC suspension is pumped into the microfilter, forming a very thin PAC
layer on the membrane (precoating). Thereafter the effluent to be treated is pumped into the
membrane. In trials on an untreated textile effluent, very exciting results were obtained. The
chemical oxygen demand (COD) was reduced by about 80 % (1600 mg/L to 200 mg/L) and the
turbidity was reduced from over 300 NTU to below 1 NTU. The permeate flux averaged 30 LMH.
This rejection and permeate flux was vastly superior to results obtained with the microfilter only,
and with a microfilter precoated with an "inactive" precoat (kaolin). Trials on other effluents
indicated similar trends - the ACMF process produced a significant COD removal, a permeate of
very low turbidity, and permeate fluxes significantly higher than that obtained with an "inactive"
precoat, or with an unprecoated tube.
The most remarkable observation was that there was no evidence of the PAC saturating, and hence
resulting in the breakthrough of COD. In runs of up to 220 hours, the permeate COD started off at a
low value, and remained substantially at this value. Experiments indicated that the total COD
removal was many orders of magnitude greater than the adsorption capacity of the PAC. A possible
explanation for this is that the organics and colloids adsorbed by the PAC are themselves forming a

                                                                                                WISA-MTD 2005

secondary separation layer on top of the PAC layer. Hence, although the PAC may saturate, this
secondary layer acts as a dynamic membrane and causes organics to be retained within the tube.
In overview, the ACMF process is a very promising one-step process for organics reduction and the
removal of suspended solids. The performance is superior to a microfilter without a precoat or a
microfilter with an "inactive" precoat, both from the point of view of rejection and permeate fluxes.
The removal of organics is seemingly superior to conventional PAC processes, due to the formation
of a secondary dynamic separation layer. As such, the process holds great potential in the
treatment/pretreatment of "difficult" industrial effluents.

                                                                                     WISA-MTD 2005

                                       FERTILIZER PLANT

                                           O Mahle , SJ Modise

   Chemistry Department, Faculty of Applied and Computer Sciences, P/Bag x021, Vaal University of Technology,
                                 Vanderbijlpark, 1900, RSA; mahleo@vut.ac.za

Most industrial production processes need cooling water for efficient and proper operation of
integral plant components. Cooling water systems control temperatures and pressures by
transferring heat from hot process fluids into cooling water, which carries the heat away. As this
happens, the cooling water heats up and must be cooled before it can be used again, or replaced
with fresh make-up water (Scholtz, 1995:7).

Typical effluent from a fertilizer plant largely depends on the type of fertilizer being manufactured,
and it varies significantly. Effluent from a fertilizer plant is composed of NO3-N; NH4-N; K+, F-,
Ca2+, CO32-, SO42- and H2PO4 of which levels are usually high. Also of concern is the high
conductivity in the process water after cooling water blow down (CWBD). This is so because the
water recycled in the plant gets concentrated in salts over time after use and reuse.

The dissolved salts need to be reduced to acceptable levels before the effluent can be discharged.
Membrane processes are used in many industries to treat such effluents, to enrich the products and
even for separation purposes. The main advantage the membrane processes have is they can be
easily combined with other separation processes (hybrid processing), and the fact that membrane
properties are variable and can be adjusted (Mulder, Tholen and Maaskant, 1997:10). These
processes can be used to remove or reduce these levels of the effluent constituents.

The purpose of this paper is to evaluate the ability of membrane processes (microfiltration,
ultrafiltration, nanofiltration and reverse osmosis), singly and collectively in treating the fertilizer
effluent from the cooling tower.

Keywords: membranes, separation process, pollution, fertilizer plant

Scholtz, C.C. (1995). The Monitoring of Biofouling and Biocorrosion in Cooling Water Systems,
M-Tech Dissertation, Vaal University of Technology, South Africa.

                                                                                               WISA-MTD 2005

Mulder, M., Tholen, J. and Maaskant, W. (1997) European Membrane Guide, Alinea, Netherlands.

                                                                               WISA-MTD 2005


                            M Ramphao, MC Wentzel and GA Ekama

            Water Research Group, Dept of Civil Eng., Univ. of Cape Town, Rondebosch 7701, RSA.

1.     Introduction
Effective solid-liquid separation in activated sludge (AS) wastewater treatment (WWT) systems is
an essential step because it has a major influence on effluent quality. Traditionally in AS systems
this step is accomplished in secondary settling tanks (SSTs). However, membranes in place of
SSTs for solids-liquid separation offer several potential advantages, including:
(1)    Insensitivity to sludge settleability and filamentous bulking; this is a significant advantage as
       biological nutrient removal (BNR) systems notoriously produce rather poor settling sludges
       (DSVI~150 mR/g) when aerobic mass fractions are low (< 60%)
(2)    Insensitivity to activated sludge flocculation characteristics and hydraulic shear in the
       reactor; membranes retain all solids which include free swimming bacteria.
(3)    SSTs are not required, a wastewater treatment plant (WWTP) footprint reduction.
(4)    Very high reactor concentrations of 15 to 20 gTSS/R (1.5 to 2%) resulting in reduced
       biological reactor volumes compared with conventional BNR systems with SSTs (a further
       footprint reduction).
(5)    Adjustable anaerobic, anoxic and aerobic zone sludge mass fractions for fixed zone volume
       fractions by changing the inter-reactor recycles to maximize biological N and P removal in
       conformity with influent N and P loads.
(6)    Production of possibly disinfected effluent for industrial or horticultural use; membranes
       with an nominal pore size of 0.4:m may be an effective barrier against bacteria and viruses
       (Churchouse and Brindle, 2003).
(7)    Possibly obviate waste activated sludge thickening when reactor concentrations are at the
       high end of the range (2% TSS).

Membrane activated sludge systems have been operated successfully both technically and
economically in the UK, Europe and Japan. However, application has been largely restricted to
COD (BOD5) and free and saline ammonia (FSA) removal (nitrification), i.e. mostly as fully
aerobic systems.    Membrane bioreactor (MBR) technology application to BNRAS systems is

                                                                                            WISA-MTD 2005

limited, and uncertainty exists as to the impact of the conditions induced by the MBRs on the
biologically mediated processes of nutrient removal.

2.     Impact of membranes n BNRAS system design
Installing membranes for solid-liquid separation into BNRAS systems makes a profound difference
not only to the design of the BNR system itself, but also to the approach to design of the whole
WWTP. Whereas the size of the biological reactor for SST BNR systems is governed by the
influent organic (COD) load and sludge age, for MBR BNR systems the size of the biological
reactor is governed by the peak wet weather flow (PWWF) or the oxygen transfer rate (OTR) of the
aeration system. With selected zone mass fractions and associated recycle ratios (for desired N&P
removal), the PWWF, i.e. hydraulic considerations, determines the surface area of membranes
required. This required surface area has associated with it a required aerobic zone volume and OTR
from the course bubble aeration for membrane scour and cleaning. If the OTR of the membranes
can meet the peak biological oxygen demand (i.e low wastewater strength), then the aerobic zone
(and hence the whole biological reactor) volume is governed by the volume required to
accommodate the membranes. However, if the membranes cannot meet the peak biological oxygen
demand, then the aerobic (and hence biological reactor) volume is increased to accommodate
additional fine bubble aeration. Under these circumstances (i.e. high wastewater strength), the
OTR of the membrane and additional aeration systems govern the volume of the aerobic zone and
hence the whole biological reactor. Accurate aeration information is therefore essential to correctly
size the reactor. With the volume of the biological reactor fixed by one of these two criteria, the
sludge age is determined so that the applied organic load generates sufficient sludge mass to meet
the aerobic zone MLSS concentration required by the membranes (12 to 18 gTSS/R for Kubota®
membranes considered here). The sludge ages usually are long (>20d) so achieving complete
nitrification is usually not a problem.

Of the design parameters, the aerobic mass fraction (fmaer) has the greatest impact on the membrane
reactor volume as a % of the equivalent SST system volume - the lower the fmaer, the greater the
%volume. However, increasing the fmaer is at the expense of biological N (anoxic mass fraction)
and P (anaerobic mass fraction) removal. Reducing the PWWF/ADWF (f q) ratio by flow balancing
significantly increases the treatment capacity per unit membrane surface area, the increase
depending on the extent to which the fq ratio is reduced. For high fq (>1.5) the indicated design for
the WWTP is extended aeration (no primary settling and long sludge age, >30d) with direct
discharge of waste activated sludge to dewatering so the high cost of membranes can be offset by
savings in sludge treatment costs. Treating flow balanced (fq<1.2) settled wastewater reduces the
                                                                                     WISA-MTD 2005

cost of the membranes per MR/d treated but the cost of primary and (possibly) WAS sludge
treatment has to be added (for details see Ramphao et al., 2004).

3.     Impact of membranes on performance of Bnr systems
Two laboratory scale UCT-type NDBEPR systems were set up (MBR and SST) to have identical
design parameters such as anaerobic (fmana = 0.126), anoxic (fmanx = 0.279) and aerobic (fmaer =
0.595) mass fractions, recycle ratios (aerobic to anoxic a = 2:1, anoxic to anaerobic r = 1:1 and
underflow to anoxic s = 1:1) and sludge age (20d). The only differences were the influent flow
(MBR 140 R/d and SST 15 R/d) and total reactor volume (MBR 74R and SST 25R) due to the
physical constraints imposed by the dimensions of the membranes, and the much higher reactor
concentrations in the MBR system. The membranes considered in this investigation were the
Kuboto® flat sheet ones.

The two systems were fed screened (1 mm mesh) raw sewage at 800 mgCOD/R from Mitchells
Plain WWTP, increased to 1000 mgCOD/R by adding 200 mgCOD/R sodium acetate to increase
BEPR. Ammonia also was added to increase TKN/COD to 0.10 and phosphorus to ensure no P
limitation. The sewage was collected in 2m3 batches and served as feed for about two weeks. To
date (Oct 2004) 21 sewage batches have been used and the systems have been operated for 397
days, during which the performance of the two systems was monitored to identify and quantify the
influence of the membranes (see Ramphao et al., 2004 for details).

3.1    System loading
Because the MBR system was operated at a higher reactor TSS concentration than the SST system
(a requirement for effective membrane scour) the COD mass loading per unit reactor volume was
3.2 times higher than for the SST system at 1867 mgCOD/(R.d) versus 592 mgCOD/(R.d). This
confirms the substantial reactor volume savings that can be achieved in MBR systems (advantage
4), but as indicated above, these volume reductions are strongly influenced by the PWWF/ADWF
ratio (fq) and the COD strength.

3.2    Trans-Membrane Pressure (TMP)
With time there was a very low gradual increase in TMP from 150 to 190 mm (0.10 mm/d) to
maintain a constant flux (0.235 m/d) which could not be reversed with water or chemical cleans.
This increase is so low to be of no concern on membrane life (700 mm in 20 years). The TMP

                                                                                  WISA-MTD 2005

appeared independent of sludge settleability and aeration intensity (floc shear) and confirms
advantages 1 and 2 above.

3.3     N and COD mass balances
Good N mass balances were obtained in the two systems (MBR 103.5% and SST 95.5%), indicating
accuracy of experimental and analytical techniques and reliability of data. Relatively low COD
mass balances were obtained (MBR 90.5% and SST 87.1 %). However, these COD mass balances
are similar to those obtained in earlier UCT and MUCT NDBEPR systems (84 - 90 %) and indicate
that the presence of an anaerobic reactor/zone may lead to a loss of COD that is not taken into
account in the calculated COD mass balance. The underlying mechanisms for the COD “loss” have
not been established in the past decade of BNR research.

3.4     System removals and effluent quality
The MBR system exhibited removals that were equivalent or superior to those produced by the SST
system, i.e. (i) COD removal 96% vs 94%, (ii) TKN removal 97% vs 97%, (iii) FSA removal 98%
vs 97%, (iv) total nitrogen removal, 74% vs 75% and (v) P removal 66% vs 54%. The largest
difference was in the P removal which was substantially higher in the MBR system (27.0 vs 21.5
mgP/R). The reason for this appears to be in the higher anoxic P uptake in the SST systems (39% vs
12%). The MBR system produced a solids free effluent with a COD of 35 mg/R. The SST system
produced an unfiltered and a 0.45 :m membrane filtered effluent COD of 73 and 53 mgCOD/R.
Therefore the membranes retain organics that would be considered “soluble” from a SST system.
This is accommodated in the BNRAS steady-state models as a reduced unbiodegradable soluble
COD fraction (fS,us) in the influent, determined as 0.036 and 0.058 respectively for the two systems.

The microbiological quality of the MBR system effluent also was superior to that of the SST system
- 0 versus 1000 CFU/100R. However, only one test has been performed to date and this needs to be
routinely tested in future to substantiate the results.

3.5     Anoxic P uptake BEPR
The nitrate load on the anoxic reactor (via the “as” recycle) appeared to be a determining factor in
stimulating anoxic P uptake BEPR, in agreement with the observations of Hu et al. (2002). At
present anoxic P uptake BEPR is not explicitly incorporated in the steady-state design procedures
for BEPR systems, as quantitative relationships linking the extent of P uptake to the system design
or operational parameters have not been established. The lower BEPR with anoxic P uptake can be
accommodated by reducing the P content of the PAOs (fXBGP), which were 0.259 and 0.232
                                                                                      WISA-MTD 2005

mgP/mgVSS for the MBR and SST systems. However, predicting a priori exactly what f XBGP value
to use remains uncertain and is a problem of BNR models not MBRs.

3.6    Sludge production - gVSS or TSS/d per gCOD load/d
The MBR system had a higher sludge production (0.41 gTSS/d per gCOD/d at 0.79 VSS/TSS ratio)
than the SST system (0.25 gTSS/d per gCOD/d at 0.81 VSS/TSS ratio) at the same sludge age. This
can be explained in part by the retention of (i) the suspended solids and (ii) unbiodegradable
“soluble” organics that would normally be lost in the effluent from a SST system. However, this
retention of COD accounts for only 25% of the difference in sludge production. In the steady-state
design procedure, this increased sludge production rate in the MBR system can be accommodated
by increasing the influent unbiodegradable particulate COD fraction (fS‟up), determined as 0.224
compared with 0.067 for the SST system. Although the same fS‟up values are expected for the two
systems because they were fed the same wastewater, high fS‟up value have been observed in SST
BNR systems in the past (Ekama and Wentzel, 1999) for the Mitchells Plain wastewater (Musvoto
et al., 1994). More data at different sludge ages are required to determine if this increase remains
consistent and to identify the underlying cause.

4.     Closure
This study so far has clearly demonstrated that BNRAS with membrane solid liquid separation is
entirely feasible, and offers considerable advantages over SST BNR activated sludge systems, such
as improved effluent quality, which is independent of sludge flocculation or settling characteristics
and adjustable sludge mass fractions with recycle flows. To aid the implementation and operation
of membrane BNR systems, design guides and procedures have been developed (Ramphao et al., In
press) but the examples quoted are illustrative only and not for direct application to design. Further
experimental work to determine the kinetic rates of the various biological processes needs to be
undertaken to see if these are similar to the rates in SST BNR systems. The objective is to establish
the applicability of the kinetic models developed for SST BNR systems to MBR BNR systems so
that once a MBR system has been designed (reactor volumes, sludge age, recycle ratios), the
dynamic performance of the system can be evaluated with the standard activated sludge simulation
models (e.g. UCTPHO, Wentzel et al., 1992 with possibly modified kinetic rates), by using a point
settler with the underflow recycled to the last aerobic reactor.

                                                                                      WISA-MTD 2005

5.     Acknowledgments
This research was supported by the Water Research Commission (WRC), Aquator South Africa,
CopaMBR Technology, UK and the University of Cape Town and is published with their

6.     References
Churchouse S and Brindle K (2003) Long term experience with membrane bioreactors. 4th
International meeting on membrane bioreactors for wastewater treatment.        School of Water
Sciences/ Water Biotreatment Club, Cranfield University, Cranfield, UK, 9th Apr 2003.
Ekama GA and Wentzel MC (1999) Denitrification kinetics in biological N & P removal activated
sludge systems treating municipal wastewaters. Wat. Sci Tech. 39(6) 69-77
Hu Z-R, Wentzel MC and Ekama GA (2002) The significance of denitrifying polyphosphate
accumulating organisms in biological nutrient removal activated sludge systems. Wat. Sci. Tech.
46(1/2) 129-138.
Musvoto EV, Casey TG, Wentzel MC, Ekama GA and Marais GvR (1994)                   The effect of
incomplete denitrification on anoxic-aerobic (low F/M) filament bulking in nutrient removal
activated sludge systems. Wat Sci Tech. 29(7) 295-299.
Ramphao M, Wentzel MC, Merritt R, Ekama GA, Young T and Buckley CA (In press) The impact
of membrane solid liquid separation on the design of biological nutrient removal activated sludge
systems. Biotech & Bioeng.
Ramphao M, Wentzel MC, Merritt R, Ekama GA, Young T and Buckley CA (2004) Performance
and kinetics of biological nitrogen and phosphorus removal with ultrafiltration membranes for
solid-liquid separation. UCT Research Report W120, Dept of Civil Eng., Univ of Cape Town,
Rondebosch, 7701, Cape South Africa.
Wentzel MC, Ekama GA and Marais GvR (1992) Processes and modelling of nitrification
denitrification biological excess phosphorus removal systems - a review. Wat Sci Tech. 25(6) 59-

                                                                                  WISA-MTD 2005


                                    SKO Ntwampe and MS Sheldon
  Department of Chemical Engineering , Cape Peninsula University of Technology, Cape Town, 8000, South Africa

There are an increasing number of areas where membrane bioreactors are used to immobilise
microorganisms and in turn these organisms produce enzymes that can be used in various
applications. Growing interest was focused on fungi, which have potentialities in solving
environmental problems. This is the case of P. Chrysosporium, which produces extra cellular
enzyme (Manganese Peroxidase (MnP) and Lignin Peroxidase (LiP)) that can degrade aromatic
compounds and remove coloured compounds from effluents [1-3].

Growth kinetics are the most important and critical parameters of biofilms as they can be used to
model mass transport and reactions [4]. To our best knowledge, there are not any available growth
kinetic models developed for continuous membrane bioreactor system immobilised with P.
Chrysosporium biofilm and nutrient media supplied through the substratum while air is supplied on
the shell side. The aim of the present study was to quantify growth kinetics of the biofilm
immobilised on a vertically placed membrane bioreactor for a specified period of time.

P. Chrysosporium strain BKMF 1767 was grown in single fibre capillary membrane bioreactors
(SFCMBR). 15 SFCMBR with working volumes of 20.4 ml and active membrane length of 160
mm were used. The bioreactor was operated at a dead-end filtration mode for 11 days. Nutrients
were pumped at 1.67 ml/hr through the lumen of the polysulphone membrane while airflow rate of
1-3 L/hr was used. Permeate samples were taken every 24hrs.
On days 3, 5, 7, 9 and 11; 3 SFCMBR were taken off per day to measure the dry biofilm density
developed on the polysulphone membranes after drying at 60 0C for 12 hours in an incubator to
remove excess moisture, by using a helium pycnometer. Concentration of glucose content in the
permeate samples was determined by using a D-glucose test kit. A Clark-type oxygen microsensor
was used to take oxygen measurements across the biofilm at intervals of 50m during the stationary
(steady state) phase. PROFILE 1.0 software [5] was used to quantify the flux of oxygen.

                                                                                              WISA-MTD 2005

A reaction rate constant value of 0.035hr-1 was used to model the dry density logistic curve. The
primary growth phase reached structural steady state at an average density of 9.7E5g/m3 after
192hrs with the secondary growth phase occurring after 216hrs. The glucose based Monod
saturation constant, K g , was calculated as 6.59E03g/m3, with the maximum substrate utilization

rate, rm , as 138.89 g/m3.hr. The growth yield coefficient, Y X / g , was calculated as 0.195, with 0.14g

as average dry biomass after 192hrs. A glucose maintenance coefficient, m g , of 0.0428hr-1 was

obtained during the biofilms stationary phase. The maximum specific growth rate,  m ax , was 2.8E-

5hr-1. During the biofilms stationary phase, the oxygen based Monod saturation constant, K O2 , was

calculated as 6.25g/m3 with biofilm diffusion, D f , as 4.04E-5m2/hr. The reaction rate constant, k O2 ,

for the oxygen profile was calculated as 2.975E-6hr-1.

1. Tien, M. and Kirk, T.K. 1984. Lignin degrading enzyme from Phanerochaete Chrysosporium:
Purification, characterization, and catalytic properties of a unique H2O2-requiring oxygenase. Proc.
Natl. Acad. Sci. USA 81: 2280-2284
2. Walsh, G. 1998. Comparison of pollutant degrading capacities of strain of Phanerochaete
chrysosporium in relation to physiological differences. Honours Thesis. Rhodes University.
3. Bumpus, J.A. and Aust S.D. 1987. Biodegradation of environmental pollutants by the white rot
fungus Phanerochaete chrysosporium. Involvement of the lignin degrading system. Bioassays. 6:
4. Lewandowski, Z. and Beyenal, H. 2001. Limiting-current microelectrodes for quantifying
mass transport dynamics in biofilms. Methods in enzymology. Vol. 337: 339-359
5. Berg, P., Risgaard-Petersen, N. and Rysgaard, S.                1998. Interpretation of measured
concentration profiles in sediment pore water. Limnology and Oceanography. Vol. 43 (7). 1500-

                                                                                        WISA-MTD 2005

                                      PLENARY LECTURE 2

                                  NATURAL BROWN WATER

     RD Sanderson, SK Sikder, MB Mbanjwa, DA Keuler, DS McLachlan, FJ Reineceke

     University of Stellenbosch, Department of Chemistry & Polymer Science, Stellenbosch, 7600, South Africa

Brown-coloured 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 [1]. However, this
coloured water is unacceptable for potable use.           The colour in natural brown water is produced
mainly by the presence of natural organic matter (NOM), 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, and fulvic acid. These contain stable ring structures and carboxylic acid and
phenolic compounds. Among various methods for treating coloured water, membrane filtration is
one of the most recent. However, membrane fouling causes permeability reduction and the
subsequent decline in flux through the membrane and the process rapidly becomes uneconomical,
unless the membranes are periodically cleaned. Over the last few years, there have been a number
of fouling studies using NOM and humic acid [2-5]. Ultrasonic time-domain reflectometry (UTDR)
is a more recent and versatile in-situ non-invasive measurement technique in real-time.

In this study, a broader picture of membrane fouling and defouling during microfiltration (MF) of
brown water is presented. The ultrasonic frequency spectra have been processed by “Wavelet
Transform” software that is capable of showing the change between consecutive measurements with
a higher resolution.

Non-destructive (NDT), non-invasive ultrasonic techniques have long been used to evaluate the
properties of thin layers. However, the interference of ultrasonic signals reflected from multiple
layers limits the development and the more extensive use of the ultrasonic technique. Here we
describe a technique to suppress the effects of the interference of ultrasonic signals. This new
                                                                                                WISA-MTD 2005

ultrasonic technique offers an unexpected sensitivity in the detection of membrane fouling during
liquid separation by membranes. In situ ultrasonic reflection data indicate fouling deposition in a
very short time, i.e. 1 s. By producing differential signals, obtained by comparing reference and test
waveforms, the fouling process can be detected and monitored. A linear relationship between
fouling resistance and the amplitude of differential signals exists. In the case of fouling layer
thickness, the resolution exceeds the theoretical limit h            0.25, where h is the layer thickness and

were obtained, down to (h                                                     sforms support the findings and
add information on other physical properties such as density and porosity of fouling layers and the
fouling process. Data at 5-seconds fouling is already possible, i.e. changes to the membrane are
observed before significant flux change.
The first picture illustrates Wavelet transform and superimposed waveform (red) for an ultrasonic
signal reflected off a nylon membrane (T – top of membrane, S – nylon support).

The second picture illustrates Wavelet transform and superimposed waveform (red) for an
ultrasonic signal reflected off a fouled nylon membrane after 60 minutes (T – top of membrane, S –
nylon support, F – top of fouling layer).

                        1                                                         2


[1] C.D. Swartz and H.A. de Villiers, Guidelines for the Treatment of Cape Coloured Water, Report to Water
    Research Commission of South Africa, Report No 534/1/98, 1998, pp. 2.1-3.4.
[2] A. Maartens, P. Swart and E.P. Jacobs, Humic membrane foulants in natural brown water: characterization and
    removal, Desalination, 115 (1998) 215-22.
[3] G. Amy and J. Cho, Interactions between natural organic matter (NOM) and membranes: rejection and fouling,
    Water Science and Technology, 40 (1999) 131-139.
[4] T. Carroll, S. King, S. R. Gray, B.A. Bolto, and N.A. Booker, The fouling of microfiltration membranes by NOM
    after coagulation treatment, Water research, 34 (2000) 2861-2868.

                                                                                               WISA-MTD 2005

[5] J. Cho, G. Amy and J. Pellegrino, Membrane filtration of natural organic matter: comparison of flux decline,
    NOM rejection, and foulants during filtration with three UF membranes, Desalination, 127 (2000) 283-298.
[6]. A.P. Mairal, A.R. Greenberg, W. B. Krantz, Leonard J. Bond, J. Membr. Sci. 159, 185 (1999), A.P. Mairal, A.R.
    Greenberg, William B. Krantz, Desalination 130, 45 (2000).
[7]. G.Y. Chai, W.B. Krantz, A.R. Greenberg, paper presented at proceeding of 6 th World Congress of Chemical
    Engineering, Melbourne, Australia, 23-27 Sep. 2001.
[8]. J. Li, R.D. Sanderson, E.P. Jacobs, J. Membr. Sci. 201, 17 (2002).
[9]. R.D. Sanderson, J. Li, L.J. Koen, L. Lorenzen, J. Membr. Sci., 207, 105 (2002).
[10] J. Li, R.D. Sanderson, Desalination 146, 169 (2002).
[11] J. Li, V.Y. Hallbauer-Zadorozhnaya, D.K. Hallbauer, R.D. Sanderson, Desalination 146, 177 (2002).
[12] R.D. Sanderson, D.K. Hallbauer, J. Li, V Yu Hallbauer-Zadorozhnaya, S. Marke, J. Schiller, SA provisional patent
    application: 2002/4753
[13] J. Li, V Yu Hallbauer-Zadorozhnaya, D.K. Hallbauer, R.D. Sanderson, Ind. & Eng. Chem. Res. 41, 4106 (2002).

                                                                                                   WISA-MTD 2005

                               DEMINIRALISATION PLANT

                                      J. D. Janse van Noordwyk
                           Senior Consultant Eskom Holdings Tutuka Power Station

Eskom Holdings generates 96% of the country‟s electricity requirement. The power stations are a
combination of fossil, hydro, pump storage and nuclear. The fossil-fired power stations are of
different designs to utilize primary energy at the source.

Direct cooled stations were built in areas where water resources are limited. Traditional wet cooled
stations were constructed during the late 70‟s and early 80‟s. Tutuka is one of Eskom‟s large six
pack wet cooled power stations. The raw water supply to Tutuka is from the Usuthu Vaal system.
The salinity of this water is considerably higher than the water supply from the Komati system.

The tied coal mine for Tutuka is a very deep underground mine. The coal seam height varies from
1.6 to 1.9 meters. Total coal extraction mining method is primarily used. Due to the depth of the
mine and the applied mining method, the aquifers above the mine roof have broken and the ancient
water is being released into the mining area. This mine effluent water is highly saline and may not
be released into the environment.

Brief description of project

During 1997 the influx of water into the mine increased drastically and the mining operation was at
risk. The tied Colliery and Eskom Holdings embarked on a joint venture to ensure continued mining
operation while protecting the environment. The aim was to recover the mine water into the cooling
water system at the power station and use desalination technology to reduce the effluent for disposal
on the station‟s ash system.
                                                                                     WISA-MTD 2005

Studying the latest technologies and obtaining the advice from membrane manufacturers/suppliers,
the option was to install a Spiral Wound Reverse Osmosis (SRO) plant.

The turnkey project was motivated and construction of a 12,5 Ml/d plant commenced July 1998 and
was completed by May 1999. The plant was designed to desalinate a combination of cooling water
and mine effluent water.

Tutuka uses approximately 100 Ml of raw water per day at the present production. This water
contains some 9 tons of dissolved salt. To prevent environmental pollution, the mine effluent water
has to be treated and used as cooling water make up at the power station. Presently 7 Ml/d of mine
effluent is recovered to the power station. This volume contains about 30 tons of dissolved salt. The
total quantity of salt has to be sidelined and disposed per day to maintain the cooling water

Cold lime softening is done on a mixture of 4-5 Ml of concentrated cooling water and 7-8 Ml of
mine effluent water. The alkalinity and suspended solids are reduced with this process. The clarifier
overflow is then neutralized to pH 7.5, sterilized with chlorine gas before sand and cartridge
filtration and is fed into the SRO plant.

The SRO plant comprises three parallel trains. Each train has three stages with 20, 10 and 5
pressure vessel arrangements and 7 membranes per pressure vessel. An inter stage booster pump is
installed upstream of the third stage. The design water recovery is 87% with a 80% salt rejection.
Design flux is 22 LMH. As the feed water organic loading is high and unpredictable, Trisep X 20
membranes were selected.

Results and Discussion

The water recovery is automatically controlled at 87 %. The SRO plant has maintained this water
recovery throughout the life of the originally installed membranes. Actual salt rejection of 98-95%
is obtained. The salt rejection was still well above the design value when the membranes were
replaced after 5 years of operation. The requirement for the replacement was basically due to

                                                                                     WISA-MTD 2005

increased pressure drop across the membranes. The contributing factor was passing cartridge filters
and excursions on the pretreatment.
During the commissioning phase it was soon realized by the power station project team that the first
stage permeate quality was much better than the raw water quality. The TDS of this permeate was
about 20% that of the filtered water supply to the demineralization plant used for boiler feed water
production. Furthermore, the permeate contained virtually no calcium, magnesium, sulfate, organic
matter or silica. The benefit of using the first stage permeate as feed to the demineralization plant
was realized and a project was successfully motivated to effect the supply of first stage permeate to
the demineralization plant.

Extended runs between regenerations of ion exchange resins on the demineralisation plant were
obtained. Reduced frequency of brine washes was possible. The reduction in total organic carbon
(TOC) in the demineralized water was a significant benefit for Eskom. Organic carbon that slips
through the demineralization process is oxidized to carbon dioxide in the steam production process
and increases impurities in the steam supply to the turbines. The normal TOC in the outlet of the
mixed bed demineralizers use to be 100-150 ppb. After changing the feed supply to first stage
permeate, the TOC reduced to below 25 ppb. This was in fact a major benefit in using first stage
permeate as feed to the demineralization process.

During 2004 the cost of demineralized water production decreased from R 712/Ml to R 426/Ml.
Regeneration chemical cost reduction of R277 375 was achieved. This was due to a reduction of
90 232 kg sulfuric acid and of 76 473 kg sodium hydroxide.

The use of membrane technology to reduce the TDS of water supply to the demineralization process
surely paid off for Tutuka. Subsequently another power station has implemented this technology
with remarkable reduction in cost and improvement in quality.

Tutuka Power Station SRO experience

                                                                                     WISA-MTD 2005


                                          V L Pillay1 and E P Jacobs2

   Durban Institute of Technology, Depart of Chemical Engineering, S4 L1 Steve Biko Campus, Durban Institute of
                         Technology, Mansfield Road, Durban, 4001; email : pillayvl@dit.ac.za

              University of Stellenbosch, Department of Chemistry and Polymer Science, Stellenbosch, 7600

The provision of adequate drinking water is a national developmental priority in most developing
economies. Membrane technology is the fastest growing technology internationally for drinking
water production. Ultrafiltration (UF) and microfiltration (MF) are membrane processes that are
ideal for producing a high quality of drinking water from non-saline raw surface waters, without
requiring any chemicals to be added to the water. These processes offer various advantages over
conventional chemical treatment methods that make them ideal for applications in rural and peri-
urban areas.

In South Africa, however, the use of UF and MF in potable water production is virtually unknown.
Possible reasons for this include the perceptions that membrane systems are very “high tech” and
require very highly skilled operators; that membrane systems are extremely expensive to set up and
operate; that there is insufficient local expertise to service the technology; and the general
perception that membrane technology would not be sustainable in South Africa‟s rural and peri-
urban areas.

Six years ago a project was initiated to develop a UF system for potable water production in South
African conditions. The project involved the University of Stellenbosch and the Durban Institute of
Technology, and was sponsored by the Water Research Commission. A major aim of the project
was that the final system should meet all the requirements for sustainability, i.e. guaranteed high
quality of water, reliable and robust, easily operation and maintainence, and cost effectiveness.
After six years of intensive development and field evaluations, a truly South African water
treatment system has been developed. This paper will describe the development of the local CUF
system. Important criteria for sustainability in developing economies will be discussed, and it will
be shown that the local system fits these sustainability criteria. The advantages and disadvantages
of employing this system to solve urgent water needs will be highlighted.

                                                                                                   WISA-MTD 2005


                             Assoc.Prof. T Leiknes1, Prof. HØdegaard1
 NTNU-Norwegian University of Science and Technology, Department of Hydraulic and Environmental Engineering,
            S.P. Andersensvei 5, N-7491 Trondheim, Norway, torove.leiknes@ntnu.no, +47 7359 4758

Surface water in northern climates commonly has a high content of natural organic matter (NOM),
resulting in high colour and total organic carbon (TOC) content, which can be of concern when
used as a raw water source for potable water. Removal of NOM is required since coloured water is
unattractive to consumers, results in colouring of clothes during washing, can cause odour and taste,
increases corrosion and biofilm growth in the distribution network, and is a precursor to the
formation of disinfection by-products (DBP) when water is disinfected. Ozonation/biofiltration is a
novel alternative treatment process for waters with a high NOM concentration [1,3]. A hybrid
process combining ozonation / biofiltration / rotating membrane-disc system has been investigated
in this study where the rotating membrane-disc filtration module downstream the ozonation process
was successfully used both as the biofilm reactor as well the separation step.

Objective: to develop a hybrid process combining ozonation and bio-membrane filtration. The
effect of biofilm formation on the membrane disks, biodegradation, control and management of
biofouling has been investigated in this study. An underlying process requirement in the treatment
scheme is to have sufficient biomass to obtain the necessary biodegradation of the biodegradable
NOM while preventing excessive biofouling which prohibits the membrane performance.

The study was conducted with a representative source water, colour of 50 mg Pt/l, (UV 254-abs of 25
m-1, TOC of 6 mg/l). To maintain consistent operating conditions throughout the experimental
period, a constructed feed water was used consisting of a mixture of formic, acetic and oxalic
(dihydrate) acids, glioxylic and pyruvic ketoacids, formaldehyde and glyoxal, which were added to
tap water in such a way that the ratios and concentrations represented ozonated water with a colour
reduction from 50 to 10 mg Pt/l [2,3]. The water treatment efficiency and biodegradation of
ozonation by products was measured by analysing the removal of TOC.

                                                                                            WISA-MTD 2005

The membrane system used in this study was a Rotary Multi-Disk Type Membrane Separator
provided by Hitachi Plant Engineering & Construction Co., Ltd., Japan. The system is based on flat
sheet membranes attached to rotary disks positioned on two shafts. Permeate is extracted through
the shaft by a vacuum pump where the treated water first passes through the membrane and into the
disk before entering the shaft. Six disks, diameter: 210 mm (polysulfone UF membranes; MWCO
750 kDaltons) are mounted in a 10 L reactor (8.44 L effective volume) giving a total surface area of
0.3 m2. Rotational speeds of the disks can be varied between 0-500 rpm. The membrane disks were
fully submerged in the membrane reactor and operated at a constant flux of 20 L/m2.h with an
operating pressure range of 0.1-0.5 bar (Figure 1).
 Raw water        Ozonated water


O2 generator
                                    Bioreactor and
                                    membrane unit


Figure 1: Schematic of experimental configuration and picture of pilot unit.

Results and discussion
Biofouling due to the growth of the biofilm on the membrane surface was measured by monitoring
the development of the TMP as a function of operating conditions. The operating conditions applied
in this study are based on experiences from initial investigations of the process where two modes of
operation were chosen [4,5]. The effect of rotation and hydraulic shear on the development and
control of the biofilm was first investigated. The system was operated with a constant rotation of 10
rpm. After approx. 1280 hours TMP increased to the maximum level of operation set and a periodic
shear induced cleaning was performed. The membrane disks were spun at 800 rpm for 15 minutes
once every 24 hours. Results show that a sustainable production of drinking water was obtained,
however, a complete cleaning and recovery of the membrane was not achieved (Figure 2 A).
Attempts at removing more of the biofilm was done by extending the cleaning time of 15 minutes
and starting and stopping the rotation during the cleaning cycle without being able to register any
noticeable effects or reduction on the TMP values. A mechanical cleaning was then investigated
consisting of adding small polyurethane sponge cubes (approx. 3x3mm) to the membrane chamber.
No rotation was applied during operation and when the TMP reached a value of 0.5 bar a 2 minute
rotation speed of 425 rpm was applied. Rotation rate and time was controlled to ensure enough
biomass was kept within the system after the cleaning cycle to maintain necessary biodegradation.
With the addition of sponges the removal of excess biofilm was achieved. The fouling rate between
                                                                                     WISA-MTD 2005

each cleaning cycle is also similar indicating the growth of the biofilm and biodegradation is
maintained throughout the operation (Figure 2 B). A cleaning frequency of approx. once a week
was necessary with this mode of operation giving essentially a 100% water recovery.
                        0,7                                                               100 %
                                                                                                                                            0,6                                                   100 %
                                    TMP                                                   90 %

                                                                                                  % removed of total TOC
                        0,6         % TOC removed
                                                                                          80 %                                              0,5

                                                                                                                                                                                                              % removal of TOC
                        0,5                                                               70 %
          TMP, kg/cm2

                                 No shear cleaning                                                                                          0,4

                                                                                                                               TMP, kg/cm
                                                                                          60 %
                                                                                          50 %                                                                                                     50 %
                                                             Periodic shear cleaning                                                        0,3
                                                             (daily: 15 min, 800 rpm)     40 %

                        0,2                                                               30 %                                              0,2
                                                                                                                                                                                                  (25-30 %)

                                                                                          20 %
                        0,1                                                                                                                 0,1
                                                                                          10 %
                        0,0                                                                0%                                               0,0                                                        0%
                          1000     1100       1200    1300          1400        1500    1600                                                  1500   1600   1700   1800   1900   2000   2100    2200
                                                     Time, h                                                                                                         Time, h
      A                                                                                                                    B
Figure 2: System performance; A - shear induced cleaning, B - mechanical cleaning.

Reduction in total TOC was approximately 25% during all modes of operation, indicating a total
degradation of the available biodegradable carbon added to the constructed water was achieved.
Hydrodynamic shear forces created by rotating the membrane disks were not sufficient to control
and manage biofouling of the membranes. Biofouling was controlled by combining shear induced
cleaning and adding sponges for mechanical cleaning of the membrane surface. The results show
that combining ozonation with biofilm-membrane filtration is a potential process alternative for
producing potable water from water sources with high concentrations of NOM and that biofouling
can be controlled and managed by optimizing the mode of operation.

[1]                     H. Fløgstad, H. Ødegaard, “Treatment of humus waters by ozone”, Ozone Science and
                        Engineering, 1985, Vol. 7, pp.121-136
[2]                     K. Carlson, G. Amy, “The Formation of Filter-Removable Biodegradable Organic Matter
                        during Ozonation”, Ozone Sci. Eng., 1997, 19, 179-199.
[3]                     E. Melin, B. Eikebrokk, M. Brugger, H. Ødegaard, “Treatment of Humic Surface Water at
                        Cold Temperatures by Ozonation and Biofiltration”, Proc. IWA 3‟rd World Water Congress,
                        Melbourne, Australia, 7-12 April 2002
[4]                     T. Leiknes, H. Ødegaard, H. Ohme, M. Lazarova, “Ozonation/biofiltration for NOM-
                        removal using rotating disk membranes”, Proc. IMSTEC‟03 – 5th International Membrane
                        Science and technology Conference, Sydney, Australia, 10-14 November 2003.
[5]                     T. Leiknes, M. Lazarova, H. Ødegaard, “Development of a hybrid ozonation biofilm-
                        membrane filtration process for the production of drinking water”, Water Environment-

                                                                                                                                                                                               WISA-MTD 2005

Membrane Technology (WEMT) 2004 Conference, June 7-10, 2004, Seoul, Korea.
Proceedings / Water Science and Technology

                                                             WISA-MTD 2005

                                              NATURAL HEADS

                                      EP Jacobs1, VL Pillay2 and S Victor3

     Department of Chemistry and Polymer Science, University of Stellenbosch, Stellenbosch; 2 Department of Chemical
                 Engineering, Durban Institute of Technology, Durban; 3 Hydrophil (Pty) Ltd, Stellenbosch

Membranes are traditionally operated in cross-flow mode to reduce the negative effects that
concentration polarization have on membrane performance.                       However, cross-flow filtration is
energy intensive, which has an impact on operating cost. Four times less power is consumed in
dead-end filtration as opposed to cross-flow filtration and dead-end filtration has therefore become
the standard technology in drinking water production using capillary membranes.                             Fouling is
controlled by means of frequent backwash with permeate. In recent years, air is introduced into the
membrane lumen during backwash, effectively creating two-phase flow, to enhance flux restoration
during backwash.           However, compressors add to the operating cost of such drinking water

We have designed and installed a drinking water treatment facility that operates without pumps and
therefore with virtual no electrical power consumption. The answer lies in the effective use of
hydrostatic head.

In this instance, the source water gravitates naturally into a tank, on the slope of a mountain. The
ultrafiltration plant is located 25m lower down the mountainside.                          The permeate tank is
interpositioned between the feed tank and the membrane plant. This gives a net driving pressure of
1 bar and an effective operating pressure of 50kPa, because of losses generated by a constant flow
valve positioned in the permeate line. Because the permeate tank is 10m above the ultrafiltation
plant, backwash is also conduced under gravity.

A PLC and three actuated valves control the complete filtration process. None of the valves are
energized during the filtration sequence, whereas all the valves are energized during the backwash

                                                                                                     WISA-MTD 2005


                                                  M Aziz

Department of Chemical Engineering, Cape Peninsula University of Technology , Zonneblom Campus, P. O. Box 652,
                           Cape Town, 8000, South Africa; email: azizm@ctech.ac.za

A wieldy supported liquid membrane (SLM) system for the permeation and efficient transport of
copper ion is introduced.       Supported liquid membranes seem to be an imminent separation
technique, because of its relatively low cost and energy saving in analogy to standard solvent
extraction processes.    The SLM used is a thin tubular porous polyvinyldienefluoride (PVDF)
membrane impregnated with LIX 984N (oxime derivative), dissolved in an industrial diluent. A
model is proposed describing the transport mechanism, which consists of diffusion process through
the source aqueous diffusion layer, fast interfacial chemical reaction and diffusion through the
membrane. The experimental results are explained by equations describing the rate transport.
Various rate-controlling processes take place as the metal transport occurs.

Keywords: supported liquid membrane, tubular PVDF membrane, copper and permeation

                                                                                             WISA-MTD 2005


         MB Mbanjwa1, 3, SK Sikder1, DA Keuler1, FJ Reineke1, DS Mclachlan1, J Schiller2, W
                                                Scharff2, RD Sanderson1

         UNESCO Associated Centre for Macromolecules & Materials, Department of Chemistry and Polymer Science,
                                University of Stellenbosch, P. Bag X1, Matieland 7602, S. Africa
                             IfU GmbH, Gottfried-Schenker-Str. 18, 09244 Lichtenau, Germany
    Cape Peninsula University of Technology, Department of Chemical Engineering, Cape Town Campus P. O. Box 652
                                                  Cape Town 8000, S. Africa

Fouling of membranes is universally accepted as one of the major setbacks in operation of pressure-
driven membrane systems for water and wastewater purification. The occurrence of membrane
fouling can be very complex and depends on several different operating parameters such as the type
of membrane filtration, nature of the effluent and the solute thereof, system flow conditions and
system hydrodynamics. During fouling, there is decrease in permeate flux and consequent loss of
productivity. To overcome this constraint, more energy is needed and therefore increased operation
costs [1]. Therefore, finding ways to minimize fouling using cost effective methods that are
environmental friendly should be of paramount priority.

However, progress in developing effective control and prevention of fouling is impeded by the
absence of suitable fouling measurement and characterization techniques [2-4]. Available methods
are largely direct-experimental and use destructive techniques (e.g. autopsies) to infer the extent of
fouling. Recent studies have shown the ultrasonic measurement technique to be a suitable non-
destructive method for measuring fouling in water treatment membrane systems in operation [4-5].

An ultrasonic fouling index meter was developed after several fouling tests using ultrasonic
measurement technique during membrane filtration. Key tests have been performed using different
effluents namely, dissolved air flotation wastewater from Mondi Kraft mill, beer-brewing
wastewater, synthetic effluents (bovine serum albumen, polyethylene glycol and humic acid) and
natural brown water [6]. The core successes have been in reverse osmosis (RO), ultrafiltration (UF)
and microfiltration (MF). Investigations have been extended to nanofiltration (NF) as well.

                                                                                                   WISA-MTD 2005

Although groundwork has been done using capillary and spiral wrap modules, the index meter is
based on the experimental work done on flat-bed modules. A prototype instrument is currently
under improvement by the Department of Chemistry, University of Stellenbosch, in collaboration
with IfU GmbH (Lichtenau, Germany). The test will then continue for industrial trials.

Technique/ Experimental
A MF membrane (Pall Corporation) with a nominal pore size of 0.2 m was used. The brown water
from the Buffelsrivier municipal reservoir, South Africa, was fed into the filtration cell at
100ml/min. The water used had a pH = 7.6, colour = 205 Pt, turbidity = 1.1 NTU and DOC = 8.2
mg/L. The crossflow velocity across the 0.0032 m2 effective membrane surface was 2.08cm/s. The
pressure in the system was kept below 10 KPa (g) and the initial permeate flux was 31 L/m 2/h. The
filtration was stopped after 13 hours when the permeate flux was very low.

An ultrasonic transducer, coupled with ultrasonic gel to the top of a perspex cell, was used as a
measuring device in conjunction with an electrical device for pulsing and receiving electrical
voltage. An electrical voltage signal was supplied to 7.5 MHz an ultrasonic transducer to convert
the electrical energy to mechanical energy and vice versa. The transducer picked up the reflected
waves, which were displayed as time-amplitude changes on an oscilloscope. The reflected wave is
detected and analysed to define the presence of any discontinuities between the solution-membrane
interface. A new solution-fouling interface was then measured during fouling formation. The first
signal was measured. This data was stored in digital format for analysis. Gabor wavelets were then
applied for time-frequency representation for analysis of frequency spectrum of the signals as
function of time. This was achieved using AGU-Vallen Wavelet mathematical software.

Results and Discussion
The permeate flux reduced to 14 L/m2/h after 13 hours In figure 1(b), a peak (F) on the waveform
resulted from an interface between the solution and the membrane. Colour changes signify density
changes as less energy is reflected due to an additional surface with less density. The fouling layer
had less density as it was not compacted since the system was operated under low pressure.
Presence of fouling was confirmed with scanning electron and optical microscopes and with eye

                                                                                     WISA-MTD 2005

              (a)                                   (b)
Figure 1. Ultrasonic signal waveforms taken (a) at the beginning of test and (b) after 13 hrs. Below
the waveforms are corresponding Gabor wavelets. The colour legend for z axis is on the right.

[1] L. Song, Flux Decline in Crossflow Microfiltration and Ultrafiltration: Mechanisms and
    Modeling of Membrane Fouling, J. Membrane Sci., 139 (1998) 183-200.
[2] G.R. Shetty and S. Chellam, Predicting membrane fouling during municipal drinking water
    nanofiltration using artificial neural networks, J. Membrane Sci., 217 (2003) 69-86.
[3] J. Li, R.D. Sanderson and E.P. Jacobs, Non-invasive visualization of the fouling of
    microfiltration membranes by ultrasonic time-domain reflectometry, J. Membrane Sci., 201
    (2002) 17-29.
[4] J. Li, D.K. Hallbauer and R.D. Sanderson, Direct monitoring of membrane fouling and
    cleaning during ultrafiltration using a non-invasive ultrasonic technique J. Membrane Sci., 215
    (2003) 33-52.
[5] R.A. Peterson, A.R. Greenberg, L.J. Bond and W.B. Krantz, Use of ultrasonic TDR for real-
    time non-invasive measurement of compressive strain during membrane compaction.
    Desalination, 116 (1998) 115-122.

R.D. Sanderson, D.K. Hallbauer, J. Li, V.Y. Hallbauer-Zadorozhnaya, S. Marke and J. Schiller,
Flat bed type slave unit for the detection and monitoring of fouling of membranes used in liquid
separation processes, SA Provisional Patent: (SA 2002/4753), 2002.

                                                                                    WISA-MTD 2005

    SK Sikder, MB Mbanjwa, DA Keuler, DS Mclachlan, FJ Reineke And RD Sanderson

  UNESCO Associated Center For Macromolecules & Materials, Department Of Chemistry And Polymer Science,
                   University Of Stellenbosch, Private Bag X1, Matieland 7602, South Africa

A beer brewing industry produces a large quantity of wastewater that contains high concentrations
of organic and inorganic compounds. Organic compounds are generally easily biodegradable and
arise mainly from dissolved carbohydrates, alcohol, volatile fatty acids and suspended solids. The
solids mainly consist of yeast, spent grains, trub, etc. The ratio of biological oxygen demand (BOD)
to chemical oxygen demand (COD) of this wastewater ranges between 0.6-0.7 [1-2]. Though there
may be large variations in flow and strength, this wastewater is highly polluting and has the
potential to cause considerable environmental problems. It can reduce the efficiency of the
municipal treatment plants and affect the water quality in many ways, including an increase in
organic matter, BOD and COD. In order to control pollution and protect the environment, this
wastewater cannot be discharged to sewers and watercourses without any treatment [1, 3].

The implementation of low-cost, efficient and simple mitigation measures are required for beer-
brewing wastewater [3]. Though a number of traditional systems are being used to treat this
wastewater, the application of membranes is the most recent and most promising. It can be used to
reduce the volume of wastewater or to produce recycle water for technical purposes. However,
fouling of membranes during filtration of this wastewater causes permeability reduction and the
subsequent decline in flux through the membrane. The membrane needs periodic cleaning or
replacement [4-6]. An understanding of the nature of fouling is of significant importance for
developing an efficient membrane filtration system.

For large-scale applications of membranes, on-line monitoring of the growth of the fouling layer
would be most useful. Ultrasonic time-domain reflectometry (UTDR) is a more recent and versatile
in-situ non-invasive measurement technique in real-time, which offers this possibility. Previous
studies with this technique confirm that a direct correlation exits between the change in ultrasonic
signal amplitude and fouling layer initiation [5-6]. In the present study, a broader picture of
membrane fouling during microfiltration (MF) of beer-brewing wastewater is presented. The
                                                                                              WISA-MTD 2005

ultrasonic frequency spectra have been processed by “Wavelet Transform” software that is capable
of showing the change between consecutive measurements with a higher resolution.

An amphoteric nylon 6,6 membrane (Biodyne A, Pall Corporation) with a nominal pore size of 0.45
m was used into a rectangular flat-bed module made of polymethyl methacrylate (perspex). The
effective membrane area was 0.0032 m2. The beer wastewater from the Heldebrau Breweries
(Somerset West, SA) was used as the fouling effluent. The ultrasonic measurement system
consisted of a 7.5 MHz ultrasonic transducer, a pulser-receiver and a digital oscilloscope. The
pulser-receiver generated the required voltage signal to trigger the transducer to send out an
ultrasonic wave. The oscilloscope captured and displayed the signal. The recorded data were
analysed and presented as different graphs and contour diagrams using MS Excel and AGU-Vallen
Wavelet (Vallen-Systeme GmbH, Germany) softwares. A pump was used to feed the cell that kept
the feed flow rate constant at 50 ml/min (Reynolds‟s number 980). The experiments were carried
out at room temperature (~25C) and the pressure across the membrane was 100 kPa. The run was
continued until the permeate flux dropped down to a very low value (1-3 hrs). The interval of data
recording was 5 seconds at the start, but ever-increasing intervals were used at the later stages.

Results and Discussion
The time-domain ultrasonic signals or waveforms (0s, 5s, 10s, 15s, 30s and 60s) are superimposed
in Fig. 1a. A change in signals was observed from the very beginning (5 s) of the fouling operation.
Because of the foulant deposition on the membrane surface, a change in membrane topography
occurred that changed the acoustic impedance on the membrane surface and caused a change in
magnitude of the membrane echo. A new peak (peak F) continued to develop at an earlier arrival
time than the membrane peak. This peak was referred as the „fouling peak‟ since it resulted from the
fouling layer. As the fouling progressed, further growth and consolidation of the fouling layer
occurred, due to further deposits on the previous layer.
                  2         5s                               F
                1.5         10s
Amplitude (V)

                0.5         1min






                 6.10E-06           6.30E-06      6.50E-06       6.70E-06
                                               Time (s)

                                   (a)                                      (b)   0s   (c)   15s   (d)     30s
Fig. 1. (a): waveforms, (b) - (d): wavelets of 0s, 15s and 30s waveforms respectively.
                                                                                                         WISA-MTD 2005

The waveforms were transformed into time-frequency contour maps known as „Gabor wavelets‟,
where the wavelet magnitude is represented by colour coding. (Figs. 1b, 1c and 1d). The initiation
and growth of fouling could be seen more clearly from the waveforms and respective wavelets
within very short intervals. The fouling growth was also confirmed by the physical observation of a
change in membrane colour and a reduction in permeate flux. These changes also continued as the
fouling proceeded with time.

      It has been shown that ultrasonic signals from the in-situ detection of membrane fouling when
 presented as wavelet transforms (a completely new approach) give a higher sensitivity and better
  resolution than has previously. Thus, the fouling growth at 5s intervals could be distinguished in
                                               this study.

We gratefully acknowledge the financial support of the Water Research Commission of South
Africa. We are also thankful to Heldebrau Breweries (Stellenbosch) for supplying the beer effluent.

[1]       C. Cronin and K.V.Lo, Anaerobic Treatment of Brewery Wastewater using UASB eactors
          seeded with Activated Sludge, Bioresource Technology, 64 (1998) 33-38.
[2]       W. Driessen and T. Vereijken, Recent Developments in Biological Treatment of Brewery
          Effluent, Institute and Guild of Brewing Convention, Livingstone, Zambia, March2-7, 2003.
[3]       W. Parawira, I. Kudita, M.G. Nyandoroh and R. Zvauya, A Study of Industrial Anaerobic
          Treatment of Opaque Beer Brewery Wastewater in a Tropical Climate using a full-scale
          UASB Reactor seeded with Activated Sludge, Process Biochemistry, 40 (2005) 593-599.
[4]       L. Song, Flux Decline in Crossflow Microfiltration and Ultrafiltration: Mechanisms and
          Modeling of Membrane Fouling, J. Membrane Sci., 139 (1998) 183-200.
[5]       J. Li, R.D. Sanderson and E.P. Jacobs, Non-invasive visualization of the fouling of
          microfiltration membranes by ultrasonic time-domain reflectometry, J. Membrane Sci., 201
          (2002) 17-29.
[6]       J. Li and R.D. Sanderson, In situ measurement of particle deposition and its removal in
          microfiltration by ultrasonic time-domain reflectometry, Desalination, 146 (2002) 169-175.

                                                                                      WISA-MTD 2005


            S.P. Roux, E.P. Jacobs*, A.J. van Reenen, C. Morkel and M. Meincken
 Department of Chemistry and Polymer Science, University of Stellenbosch, Private Bag X1, Matieland 7602, South
                            Africa, Fax: +27218084967, email: epj@maties.sun.ac.za

Polysulphone (PSU) is one of the most popular thermoplastic materials used for manufacturing
ultrafiltration (UF) and microfiltration semi-permeable membranes [1,2].                      However, PSU
membranes remain susceptible to adsorptive fouling due to the hydrophobic character of the
material [3].    Synthesizing an amphiphilic PEO-containing copolymer for incorporation into
membrane formulations allows the preparation of modified (hydrophilic) PSU UF membranes
[4,5,6]. This method, uniquely suited to the phase-inversion membrane manufacturing process, also
solves many of the problems associated with conventional membrane surface modification
techniques [6]. However, cost-effective, large-scale manufacturing of such membranes will be
severely hampered by high-temperature annealing as prescribed by Hancock et al [4]. Annealing
was therefore compared with the amount of copolymer added to membrane formulations in order to
ascertain the influence of both annealing, as well as the amount of copolymer added, on the degree
of membrane surface modification achieved.

Brief description of project
A branched PSU-block-PEO copolymer was synthesized as described by Hancock [6] and this
copolymer was used in preparing membrane test strips. Modified and unmodified PSU membrane
test strips (control) were annealed at 85 °C for up to 25 h. Samples were taken at 5 h intervals for
analysis by dynamic contact angle, energy dispersive spectrometry, protein adsorption and atomic
force microscopy techniques. Membrane samples containing different amounts of the branched,
block copolymer were also prepared and similarly analyzed for the degree of surface modification

The main aim of this project includes the large-scale synthesis of this membrane surface-modifying
branched PSU-block-PEO copolymer and the incorporation of this copolymer into membrane
formulations for preparing ultrafiltration capillary membranes via the phase inversion method. If

                                                                                               WISA-MTD 2005

successful, this technique will lead to the development of high-flux ultrafiltration capillary
membranes with a reduced vulnerability to adsorptive fouling.

Results and discussion
Our results strongly indicate that optimal membrane surface hydrophilisation depends on surface
segregation of the hydrophilic PEO segments of the amphiphilic copolymer during phase-inversion,
rather than annealing.   These results also show that increasing the amount of the membrane
modifying copolymer added to the membrane formulation can increase the degree of surface
modification achieved significantly.


   1. Koehler, J.A., Ulbricht, M. and Belfort, G., (2000) Intermolecular forces between a protein
       and a hydrophylic modified polysulfone film with relevance to filtration, Langmuir 16,

   2. Nunes,S.P., Sforça, M.L. and Peinemann, K-V., (1995) Dense hydrophilic composite
       membranes for ultrafiltration, Journal of Membrane Science 106, 49-56.

   3. Steen, M.L., Hymas, L., Havey, E.D., Capps, N.E., Castner, D.G. and Fischer, E.R., (2001)
       Low temperature plasma treatment of asymmetric polysulfone membranes for permanent
       hydrophilic surface modification, Journal of Membrane Science 188 97-114.

   4. 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.

   5. Hancock, L.F., (1997) Phase inversion membranes with an organized surface structure from
       mixtures of polysulfone and polysulfone-poly(ethylene oxide) block copolymers, Journal of
       Applied Polymer Science 66, 1353-1358.

   6. Hancock, L.F., Fagan, S.M. et al (2001) Highly branched block copolymers, US Patent
       6,172,180 B1.

                                                                                   WISA-MTD 2005


                 WISA-MTD 2005

                                   CERAMIC MEMBRANES

                                     MS Sheldon and K Mohamed

Department of Chemical Engineering, Cape Peninsula University of Technology, Zonnebloem Campus, P.O. Box 652,
                                       Cape Town, 8000, South Africa

Membrane bioreactors (MBR) can be applied for the production of bioactive molecules such as
enzymes. In the design of a MBR, the membrane morphology is the most important aspect, as the
membrane should be developed or chosen to provide maximum surface area and an ideal
environment for biofilm development. Ceramic membranes are becoming an attractive alternative
to polymeric membranes due to their excellent chemical, thermal and mechanical properties. The
objectives of this study were to (i) test the suitability of a ceramic membrane for biofilm
development and enzyme production and (ii) compare this with a polysulphone (PS) membrane.

Single capillary membrane bioreactors (SCMBR) were used to immobilize and continuously culture
a biofilm of Phanerochaete chrysosporium (strain BKMF 1767) for the continuous production of
extracellular enzymes, Lignin Peroxidase (LiP) and Manganese Peroxidase (MnP). Horizontal and
vertical set-ups were tested with both the polysulphone and ceramic membranes.

Average biofilm thicknesses developed on the PS membrane in the horizontal position were 555m,
420m and 253m at the inlet, centre and end respectively. Much thicker biofilms were developed
on the PS membranes in the vertical positions. On the ceramic membrane in the horizontal position
average biofilm thicknesses developed were 970m, 940m and 470m at the inlet, centre and
outlet respectively.   Again thicker and more even biofilms were produced when the ceramic
membranes were in the vertical position. LiP and MnP production above 1900 and 2000 U/L were
achieved respectively with both membranes in both the horizontal and vertical positions. The
polysulphone membranes lacked mechanical and thermal stability of the ceramics. The ceramic
membranes could be chemically cleaned and re-used. In conclusion the ceramic membrane was
found to be chemically inert in physiological growth media, rigid with good strength, porous to
facilitate mycelial attachment and it is also reusable after the removal of the fungus.

                                                                                            WISA-MTD 2005


                                  B Godongwana. and MS Sheldon*
Department Of Chemical Engineering, Cape Peninsula University of Technology, P.O. Box 652, Zonnebloem Campus,
                                       Cape Town, 8000, South Africa

Membrane modules have a wide range of application in industry (more prominently in wastewater
treatment), and these include use as membrane bioreactors. To effectively design and to
commercially implement a membrane bioreactor a clear understanding and modelling of the
parameters such as momentum and mass transfer inside the reactor is required.

A number of experimental and theoretical investigations have been conducted with the aim of
modelling flow behaviour in hollow fiber and capillary membranes; and many of these have
demonstrated the significance of recirculation on the overall performance of the membranes[1,2,3].
These investigations however, were not successful in accurately describing the detailed convective
recirculation nor the velocity profiles in capillary fiber membrane bioreactors in both the vertical
and horizontal configuration.

The aim of this study is to derive mathematical models to predict flow behaviour, pressure and
velocity profiles, of nutrient through a vertical orientated single fiber capillary membrane bioreactor
(SFCMBR). The nutrient flows through the lumen of the membrane with a fungus immobilised on
the external skin of the membrane. The models was tested at different growth rates and biofilm
thicknesses of the fungus.

The models developed were solutions to the Navier Stokes for the conditions of operation of the
bioreactor under investigation. The pressure profiles obtained from the model were identical to
those of Bruining[1], however not exactly in agreement with the profiles obtained from
experimental data.      The trends however were found to be the same in all three profiles
(experimental, Bruining and the developed model): The pressure at a point on the lumen side of the
membrane increases with increasing biofilm growth; the pressure decreases along the length of the
membrane; the slope of the pressure curves changes after day 8 of feeding nutrient. The velocity
profiles demonstrated that: the axial velocity decreases along the length of the membrane, and that
the velocity profile approaches that of a pipe with increasing growth.

                                                                                            WISA-MTD 2005


1.    Bruining, W.J. 1989. A general description of flows and pressures in hollow fiber membrane
     modules. Chemical Engineering Science, 44: 1441-1447.
2.    Kelsey, L.J., Pillarella, M.R. & Zydney, A.L. 1990. Theoretical analysis of convective flow
     profiles in a hollow fiber membrane bioreactor. Chemical Engineering Science, 45: 3211–
3.    Thakaran, J.P. & Chau, P.C. 1986. Operation and pressure distribution of immobilised cell
     hollow fiber bioreactors. Biotechnology and Bioengineering, 28: 1064–1071.

                                                                                  WISA-MTD 2005


                                         M Aziz and S Mneno

 Department Chemical Engineering, Cape Peninsula University of Technology, Zonneblom Campus, P. O. Box 652,
                   Cape Town, 8000, South Africa, Tel.: 460 3612, e-mail: azizm@ctech.ac.za

Supported liquid membranes (SLM‟s) extraction is a promising technique for selective removal and
concentration of metal ions from solution. The permeation of metal species through SLM‟s can be
described as the simultaneous extraction and stripping operation combined in a single stage. A thin
layer of organic reagent (extraction) is immobilized in a micro-porous inert support. This support is
interposed between the feed solution (aqueous phase) in which the valuable metal is dissolved and
the second (stripping) phase, in which enrichment of the metal occurs by trans membrane diffusion.

The aim of this study is to determine mass transfer models to describe the permeation of Copper,
through counter-current ultra-filtration using a tubular supported liquid membrane. In view of the
simplicity of operation, economical utilization of the extractant and lack of solvent entrainment,
SLMs are more advantageous than conventional solvent extraction processes.

Although extraction through other membranes has been utilized, tubular membranes are still very
much a novelty.

Keywords:    Supported liquid membrane; Extraction; tubular membrane; Copper

                                                                                              WISA-MTD 2005


                                   XP Mbulawa1, EP Jacobs2 and VL Pillay1

     Department of Chemical Engineering, Durban Institute of Technology; 2Institute of Polymer Science, University of

Aerobic biological processes are the key to wastewater treatment and are deployed in the activated
sludge process, which is the most common treatment technique worldwide. Immersed membrane
bioreactors (IMBRs) combine the biological process with membrane separation, greatly enhancing
the treated water quality. The treatment of wastewaters with a high organic content in conventional
aerobic processes is often limited by insufficient oxygen, due to the low solubility of oxygen in
water. Some aerobic reactors attempt to overcome this limitation by utilising pure oxygen instead of
air. However, using conventional aerators with pure oxygen leads to further problems such as
stripping of volatile organic compounds, and a poor efficiency of oxygen utilisation due to oxygen
bypassing the bed in the form of bubbles. Alternative approaches have focussed on utilising
microporous membranes and membranes coated with a polymer layer as aerators. Here, oxygen is
introduced into the reactor as molecular oxygen, rather than as oxygen bubbles, greatly increasing
the mass transfer of oxygen to the liquid and the oxygen utilisation efficiency. Previous studies
have highlighted silicone membranes as having better permeation characteristics than other
polymeric membranes. The overall objective of this study was to evaluate and characterise the use
of silicone membranes as aerators in IMBRs. The specific aims are to characterise oxygen mass
transfer from silicone membranes under a range of geometric and operating variables that could be
encountered in IMBRs.

The first stage of the project concerned developing a suitable experimental rig. Previous studies on
mass transfer from membranes used some form of closed shell-and-tube experimental rig, using a
dissolved oxygen (DO) probe to infer mass transfer to the liquid. For this study, a rectangular tank
was chosen, as it enabled a far wider range of geometric and operating variables to be investigated.

From the results that came out of the study, mass transfer is strongly influenced by system
hydrodynamics. A range of water crossflow velocities were tested and it showed that higher
crossflow yielded higher mass transfer coefficients. Although oxygen partial pressure had an effect
on mass transfer it was noted that the process is driven by pressure difference (across the interface)

                                                                                                    WISA-MTD 2005

as well as concentration difference (from the surface to the bulk liquid). This made it difficult to
single out the effect that pressure has on mass transfer.
Mass transfer was also evaluated at elevated water temperatures. The results do not show a
significant increase on mass transfer with increasing temperature. From theory one should expect
mass transfer to decrease with increasing temperature due to the reduced oxygen solubility.
However another possibility is that at higher temperature there is reduced viscosity, which reduces
the liquid boundary layer resistance hence one should expect an improved mass transfer rate.

                                                                                    WISA-MTD 2005

                            SODIUM CHLORIDE FROM BRINE STREAMS

                     L. Mariah1, C. A. Buckley1, D. Jaganyi2, E. Drioli3 and E. Curcio3
    Pollution Research Group, Department of Chemical Engineering, University of KwaZulu-Natal, Durban, South Africa
          Department of Chemical and Physical Sciences, University of KwaZulu-Natal, Pietermaritzburg, South Africa
                           Institute on Membrane Technology, ITM-CNR, c/o University of Calabria, Italy


This research stems from the need for Sasol to ensure an adequate supply of water of sufficient
quality and to sustainably manage the inorganic brines from its manufacturing and mining activities.
The investigation is based on the premise that the inorganic concentrates can be separated into high
value chemicals and reusable water. The chlor- alkali salt preparation circuit is being used as an
example of a process that requires a pure inorganic salt for a subsequent synthesis step

Aims and Experimental Procedures

The primary aim is to produce suitably purified and concentrated brine streams, which could
substitute for the imported raw salt used as the feed chemical in the chlor alkali process. Membrane
distillation is used to achieve this by producing fresh water by concentrating aqueous solutions
resulting in highly concentrated mother liquors in which nucleation and crystal growth occurs.
Further, the integration of various membrane operations results in a reduction of pre-treatment costs
and increase in performance of the processes.

Results and Discussion

Nanofiltration was used as pre-treatment to the crystalliser to remove hardness, turbidity and micro-
organisms thereby accelerating the achievement of supersaturation. The concentrate was then

                                                                                                          WISA-MTD 2005

treated by membrane distillation so that crystals (magnesium sulphate or sodium chloride) were
formed. The kinetics of the crystallisation process was investigated.


CURCIO E, CRISCUOLI A and DRIOLI E (2001) Membrane crystallizers. Ind. Eng. Chem. Res. 40

GIANADDA P (2003) The development and application of combined water and materials pinch
analysis to a chlor-alkali plant Doctor of philosophy thesis, University of Natal.

                                                                                     WISA-MTD 2005




                                WISA-MTD 2005