Sie sind auf Seite 1von 9


Water Research 37 (2003) 4018–4026

Hybrid flotation—membrane filtration process for

the removal of heavy metal ions from wastewater
C. Blocher a,
*, J. Dordaa, V. Mavrovb, H. Chmiela, N.K. Lazaridisc, K.A. Matisc
Institute for Environmentally Compatible Process Technology (upt) Ltd., Im Stadtwald Geb. 47, D-66123 Saarbrucken, Germany
Department of Process Technology, Saarland University, Im Stadtwald Geb. 47, D-66123 Saarbrucken, Germany
Section of Chemical Technology & Industrial Chemistry, School of Chemistry, Aristotle University (Box 116),
GR-54124 Thessaloniki, Greece
Received 21 February 2003; received in revised form 8 May 2003; accepted 14 May 2003


A promising process for the removal of heavy metal ions from aqueous solutions involves bonding the metals firstly
to a special bonding agent and then separating the loaded bonding agents from the wastewater stream by separation
processes. For the separation stage, a new hybrid process of flotation and membrane separation has been developed in
this work by integrating specially designed submerged microfiltration modules directly into a flotation reactor. This
made it possible to combine the advantages of both flotation and membrane separation while overcoming the
limitations. The feasibility of this hybrid process was proven using powdered synthetic zeolites as bonding agents.
Stable fluxes of up to 80 l m2 h1 were achieved with the ceramic flat-sheet multi-channel membranes applied at low
transmembrane pressure (o100 mbar). The process was applied in lab-scale to treat wastewater from the electronics
industry. All toxic metals in question, namely copper, nickel and zinc, were reduced from initial concentrations of 474,
3.3 and 167 mg l1, respectively, to below 0.05 mg l1, consistently meeting the discharge limits.
r 2003 Elsevier Ltd. All rights reserved.

Keywords: Flotation; Microfiltration; Membrane separation; Heavy metals removal; Industrial water reuse

1. Introduction exchange), (iii) membrane processes, (iv) electrolytic

recovery, (vi) liquid–liquid extraction and (vii) flotation
Many industrial wastewater streams (e.g. the metal [3,4]. Even though precipitation and sorption remain the
working, semiconductor and copper industries, mine most popular methods for the removal of heavy metal
water, etc.) contain heavy metals, which are of great ions, membrane filtration and flotation are gaining
environmental concern and must be removed prior to increasing importance nowadays.
water discharge or water recycling [1]. Therefore, the A new and alternative approach combines several
development and application of effective treatment of these processes and involves binding the metals
processes is obligatory and several methods have firstly to a special bonding agent (which can be
been reported in use for the control of metal toxicants selective for certain ions, e.g. lead, zinc, cadmium,
[2]. The most commonly applied physico-chemical copper), separating the loaded bonding agents from
treatment methods are: (i) precipitation as hydroxides, the wastewater stream by separation processes and
carbonates or sulphides, (ii) sorption (adsorption, ion eventually regenerating the bonding agent, making
water as well as metal reuse possible [5]. In this
*Corresponding author. Tel.: +49-681-9345-212; fax: +49- work, a new hybrid process of flotation and membrane
681-9345-380. separation, which combines the advantages of both
E-mail address: (C. Blocher). these processes while overcoming the limitations, has

0043-1354/03/$ - see front matter r 2003 Elsevier Ltd. All rights reserved.
C. Blocher et al. / Water Research 37 (2003) 4018–4026 4019

Nomenclature TMP transmembrane pressure (mbar)

RM membrane resistance (m1)
ug air superficial velocity (cm s1) RC cake layer (fouling) resistance (m1)
J specific permeate flux (l m2 h1) m viscosity (Pa s)
A membrane surface area (m2)

been developed as the separation stage in the overall variety of applications, especially biological wastewater
process. treatment [10] and drinking water production [11].
In flotation, the various polluting substances are In the hybrid process developed, specially designed
forced to collide and finally get trapped by rising air submerged microfiltration modules (with multi-channel
bubbles, resulting in a highly concentrated (mass flat-sheet membranes) are integrated directly into the
fraction) foam layer at the surface that can be removed flotation reactor. Since most of the bonding agents have
[4]. It is a cost-effective process for high concentrations already been removed by flotation from the bulk
of polluting substances that allows high loading rates. solution and passed to the froth layer, the membranes
However, the removal efficiency could be limited, if are subjected to only a part of the initial concentration
applied in one stage, because some substances may not of the bonding agents. As a consequence, the flux can be
be retained in the froth or fail to become permanently established at a high level even at a very low level of
attached to the bubbles and remain in the bulk solution turbulence created in the hybrid flotation module. Thus
and thus, in the effluent. In contrast, microfiltration energy requirements for membrane operation are mini-
constitutes a physical barrier for the bonding agents but mised. Furthermore, the bonding agents to be regener-
has limitations because of increased fouling at high ated or discharged are concentrated in the froth,
concentrations. Conventionally cross-flow processes are whereas in a conventional process with immersed
applied, where high shear rates on the membrane surface membranes, the concentration in the bleed is the same
are created to prevent fouling of the membranes. The as in the reactor and thus limited. This is due to the fact
corresponding high energy consumption led to the that these conventional systems (e.g. with immersed
development of so-called low-energy systems [6,7], that hollow fibre membranes, [9]) consist of fully mixed
are characterised by immersed membranes, operated by reactors with a uniform concentration in the whole
vacuum or hydrostatic pressure. Fouling control often is reactor volume. The basic idea of the new hybrid process
conducted by air blowing from below the membranes, is illustrated in Fig. 1.
creating shear forces near the membrane surface. Energy The paper will present the first results with the process
consumption of these systems is considerably lower than in question, applied to treat wastewater from the copper
that of the cross-flow systems, yet still significant [8,9]. foil industry. Synthetic zeolites were used as bonding
Furthermore, flux is also low, especially at higher agents as these proved to be highly efficient for heavy
concentrations of suspended solids in the reactor. Up metal removal and at the same time, they can be
to now, immersed membranes have been used for a removed by flotation [12].

Fig. 1. Basic concept of the hybrid system of membrane flotation as compared to a conventional reactor with immersed membranes.
4020 .
C. Blocher et al. / Water Research 37 (2003) 4018–4026

2. Materials and methods Table 2

Characteristics of the ceramic membranes used in the experi-
2.1. Materials ments

Type Multi-channel flat-sheet

The properties of the powdered synthetic zeolite
(produced by Ineos Silicas Ltd.) which was used to bind Geometry 100 mm  100 mm  6 mm
the heavy metals, are presented in Table 1. Zeolite Mean pore size 0.28 mm
Area per membrane 0.018 m2
concentrations were measured gravimetrically in the
No.of membranes operated in 1–3
high-concentration range (froth) and via turbidity hybrid reactor
measurement in the low-concentration range (bulk Permeability (pure water) 3.000 1 m2 h1 bar1
solution). In the latter case, the correlation between Membrane resistance RM 1.2  1011 m1
zeolite concentration and turbidity followed a linear
function up to 1.0 g l1 solid concentration. Heavy metal
concentrations were analysed using atomic adsorption
The cationic surfactant hexadecyl-trimethyl-ammo-
nium bromide (CTABr, Merck, analytical grade) was
used as a collector during flotation experiments. At a pH sealed collector
of 8–10, which was used in the experiments (see Section permeate
‘‘results and discussion’’) the zeolite has a negative channels
surface charge, therefore the collector binds with its
cationic ammonia group to the zeolite, whereas
the hydrophobic end of the molecule is able to be
trapped by the bubbles, making the zeolite rise to the
surface. CTABr was fed as an ethanolic solution with a
final ethanol concentration in the hybrid cell of
0.17 vol%. In most runs, 20 mg CTABr per 1 g zeolite membrane
(c(CTABr)=20 mg g1) were used.
Ceramic multi-channel flat-sheet membranes were
used in the process, supplied by the Hermsdorfer Institut
. Technische Keramik, HITK (Thuringia, Germany).
The characteristics of the membrane type used in the
investigations are presented in Table 2. Modules were
prepared by sealing these at one end and placing a
permeate collector at the other end. A schematic
representation of this module type is shown in Fig. 2.
Fig. 2. Schematic representation of a module of the ceramic
multi-channel flat-sheet membrane used in the tests.
2.2. Lab-scale plant and procedure

A simplified flow scheme of the plant used for the flotation takes place and in which the membrane
investigations is presented in Fig. 3. The main part of the modules are installed. The whole bottom of the reactor
plant is the hybrid reactor (net volume 2.2 l) where is made of a porous ceramic plate (mean pore size 1 mm),
through which the air was introduced. This gave a
uniform bubble distribution across the whole cross-
Table 1 section. Resulting from preliminary investigations [13]
Characteristics of the synthetic zeolites used [5] the air superficial velocity (ug) was kept at 0.33 cm s1.
The solution to be treated was prepared in a feed tank
Type Type P
(150 l) at least 3 h before starting the experiments to give
Chemical composition (Na2O)(Al2O3)(SiO2)2 sufficient wetting time. From the feed tank the solution
Molar mass (M) 284 g mol1 entered the hybrid reactor via a small feed regulator
Mean particle diameter 2 mm chamber. Collector was dosed in-line into the feed pipe
Ion exchange/sorption and mixed with the zeolite solution, guaranteeing a
capacity conditioning time greater than 30 min. Permeate was
for zinc (Zn2+) 250 mg g1
withdrawn from the membranes via a suction pump, the
for nickel (Ni2+) 90 mg g1
respective suction pressure was read via electronic
for copper (Cu2+) 200 mg g1
pressure transducer. The permeate flow rate was
C. Blocher et al. / Water Research 37 (2003) 4018–4026 4021

hybrid flotation -
membrane filtration suction pump


feed tank collector feed level

dosing regulator

froth collector
air balance

Fig. 3. Flow scheme of the hybrid process lab-scale plant used for the investigations.

calculated from permeate weight that was measured via copper, nickel ions [5]. Therefore, if capacity is not
electronic balance. When backflushing of the mem- exceeded, almost 100% of the heavy metals are
branes was applied for fouling control, time intervals transferred from the water into the zeolites. This means
were controlled via automated valves. Froth was with- that in terms of process efficiency there is virtually no
drawn from the top via overflow or suction pump. All difference between the removal of zeolites and that of
relevant process data were read continuously by PC. the heavy metals. Considering the concentrations of
Since in industrial applications, the wastewater flow to heavy metals observed in industrial wastewater streams
be treated is a pre-set value, the plant was operated at (up to 500 mg l1 for several ions) and the capacity of the
constant flux. zeolites, the concentrations of practical significance that
had to be investigated were around 10 g l1.
For initial zeolite concentrations of 8 and 12 g l1, the
3. Results and discussion removal efficiency for zeolites and heavy metals during
two selected continuous experiments is presented in
In preliminary experimental investigations, it was Fig. 4. As can be seen, the efficiency of flotation alone is
found that flotation of the zeolite type used is best 90% on average, which means that 10% of the heavy
carried out at alkaline pH values [13]. Furthermore, the metals remain in the bulk solution (although they are
zeolite type used in this research causes a pH shift to sorbed to the zeolites), whereas the zeolite concentration
higher values if dispersed in water. Therefore, all in the membrane permeate is below the detection limit,
experiments were performed at the ‘‘natural’’ pH that leading to an efficiency >99.9%.
resulted from zeolite wetting (pH 8–10). Due to the For a variety of 1–3 day experiments, aimed at the
negative surface charge of the zeolites at that pH, the investigation of different membrane operating modes,
cationic collector CTABr was found to be appropriate the results regarding zeolite/heavy metal removal were
(see Section ‘‘materials and methods’’). In the case of similar. Usually, a slight decrease of flotation efficiency
industrial wastewater, pH has to be adjusted accordingly was observed in the beginning that stabilised after 24 h
before treatment as most of the wastewater streams of operation. Removal efficiency of flotation did not
containing heavy metals are characterised by low pH significantly depend upon initial zeolite concentration,
(e.g. pH 2–3). on average it was 90.7% for 8 g l1 and 86.1% for
12 g l1 initial concentration, respectively. However, a
3.1. Performance of the hybrid process for zeolite/heavy similar removal efficiency for different initial zeolite
metal removal concentrations means that the concentration of zeolites
in the bulk solution is higher, if higher concentrated
The zeolite used is characterized by high capacity and zeolite solutions are used for binding the heavy metals.
fast kinetics for binding heavy metal ions such as zinc, Generally, flotation efficiency could have been increased
4022 .
C. Blocher et al. / Water Research 37 (2003) 4018–4026


Flotation efficiency (bulk solution)
----- Membrane flotation efficiency (hybrid system)
Zeolite removal [%]

-1 -1
Initial zeolite concentration: 12 g l Initial zeolite concentration: 8 g l
-1 -1
c(CTABr): 67 mg g c(CTABr): 20 mg g
20 Membrane area: 0.054 m² Membrane area: 0.018 m²
-2 -1 -2 -1
Permeate flux: 20 l m h Permeate flux: 40 l m h
-1 -1
Air superficial velocity: 0.33 cm s Air superficial velocity: 0.33 cm s
0 10 20 30 40 50 60
Operating time [hours]

Fig. 4. Removal efficiency of flotation and the hybrid system (membrane flotation) as a function of operating time for two selected
experiments, heavy metal removal equals zeolite removal.

by using higher collector concentrations. However, this flushing proved to be necessary. A cycle of 10 s back-
led to stronger foaming, making foam collection difficult flushing every 30 min was found to be sufficient,
and at the same time increasing the water content in the resulting in TMP of 45 and 90 mbar (mean of two
foam. cycles between backwashing) for imposed fluxes of 40
Besides the quality of the treated water, the second and 80 l m2 h1, respectively (Fig. 5). The relatively
important goal for nearly any separation process is to long time between backflushes resulted in a permeate
achieve high concentrations in the concentrate fraction. loss of less than 5% by the backwashing.
This is also true for our case where the bonding agents In Fig. 5, it can also be seen that the increase in
(e.g. zeolites) will either be regenerated or be disposed transmembrane pressure when operating the membranes
of. Therefore, a high zeolite concentration in the froth is at 40 l m2 h1 without backflushing is fully reversible.
aimed at. Although the system was not optimised yet Switching the operating mode of a membrane, that had
with respect to froth structure, average zeolite mass been operated with continuous permeate withdrawal for
fractions of 61.5% and 83.0% were achieved in 24 h, to backflushing resulted after less than 1 h in the
experiments with 8 and 12 g l1 initial zeolite concentra- same transmembrane pressure as for a membrane that
tion (approximately 0.8–1.2%), respectively. Thus the had been operated with backflushing from the begin-
concentrations in the froth were of nearly two magni- ning. This proves that the loss in membrane perfor-
tudes higher than those of the initial zeolite solution. mance is attributed to an easily removable cake layer,
This is considered a major advantage of this hybrid which was also visually observed. Using the resistance-
system investigated as similar fractions of bonding in-series-model [14] which describes filtration resistance
agents in the retentate cannot be achieved with a as the fraction of TMP to flux, the resistance of the cake
conventional membrane filtration stage, be it in cross- layer can be evaluated according to
flow mode or in an immersed membrane system. It TMP
would mean operating the conventional membrane J¼ 3:6  108 ; ð1Þ
separation unit at concentrations of >500 g l1 which
is neither practical nor economical. where, J is permeate flux (l m2 h1), RM is membrane
resistance (m1), RC is the resistance of the cake layer
3.2. Membrane performance (or in general fouling resistance) (m1), TMP is
transmembrane pressure (mbar) and Z is viscosity of
Membranes were operated at constant flux in all water to be filtered (103 Pa s).
experiments. For an initial zeolite concentration of For the experiments with 40 and 80 l m2 h1 (result-
8 g l1, the membranes could be operated at 20 l m2 h1 ing in TMP of 45 and 90 mbar, respectively) the same
without backflushing. Resulting transmembrane pres- value for RC of 2.9  1011 m1 is obtained which means
sure (TMP) was as low as 15 mbar. However, to achieve that the higher imposed flux did not create a higher cake
stable membrane performance with higher fluxes, back- resistance. Cake resistance, in general, is of the same
C. Blocher et al. / Water Research 37 (2003) 4018–4026 4023

malfunction of collector dosing

Transmembrane pressure [mbar]




-2 -1
(3) 40 (0-24 h) / 80 l m h (25-68 h), with backflush
(1) -2 -1
(2) 40 l m h , backflush starting after 26 hours
-2 -1
(1) 20 l m h , no backflush
0 12 24 36 48 60
Operating time [h]

Fig. 5. Transmembrane pressure as a function of time for the hybrid process at various operating modes, under the following
conditions: c(zeolites)=8 g l1, c(CTABr)=20 mg g1, ug=0.33 cm s1, backflushing (if applied) every 30 min for 10 s.

-1 -2 -1
(4) 12 g l (40 l m h , with backflush)
Transmembrane pressure [mbar]

-1 -2 -1
(3) 8 g l (40 l m h , with backflush)
-1 -2 -1
(2) 12 g l (20 l m h , no backflush)
-1 -2 -1
100 (1) 8 g l (20 l m h , no backflush)


0 12 24 36 48
Operating time [h]

Fig. 6. Transmembrane pressure as a function of time for the hybrid process at 8 and 12 g l 1 zeolite concentration, under the
following conditions: c(CTABr)=20 mg g1, ug=0.33 cm s1, backflushing (if applied) every 30 min for 10 s.

magnitude as membrane resistance and thus quite low, concentrations led to higher TMP, yet stable hydraulic
which can be attributed to a highly permeable cake layer performance was achieved for both concentrations and
which the zeolites form. The results can also be imposed fluxes (with the respective operating mode).
considered in terms of the concept of ‘‘critical flux’’ In all investigations, the air supply was detemined by
(the flux up to which the relation between flux and TMP that which was required for the flotation process. No
is linear), which is currently often applied for micro- additional aeration of the membranes was carried out.
filtration processes [9,15]. Although this concept does This has to be considered when relating the membrane
basically not imply backflushes, from a process point of performance to conventional immersed membrane
view it can also be concluded that the membranes in our systems, as the air superficial velocity used in the hybrid
process were operated in the subcritical flux region. Thus system was about 10 times lower than that usually
the flux of 80 l m2 h1 seems not to be the limiting flux applied in these systems [16,9]. This results in an equally
for the process investigated. lower energy consumption which, generally, still makes
The influence of initial zeolite concentrations on up a great portion of operating costs in microfiltration
hydraulic membrane performance is shown in Fig. 6, systems. In the hybrid system, these energy costs for
where the resulting TMP for two different fluxes and membrane operation were eliminated.
operating modes is shown for 8 and 12 g l1, respec- From these results, it can be concluded that the basic
tively. Generally, it can be said that higher initial zeolite idea of operating microfiltration membranes integrated
4024 .
C. Blocher et al. / Water Research 37 (2003) 4018–4026

(3) flotation efficiency 83 % (cbulk = 1.3 g l )

Transmembrane pressure [mbar]

(2) flotation efficiency 93 % (c bulk = 0.6 g l )
(1) flotation efficiency 94 % (cbulk = 0.5 g l ) (3)


0 1 2 24
Operating time [h]

Fig. 7. Transmembrane pressure as a function of time for the hybrid process for various flotation efficiencies, under the following
conditions: c(zeolites)=8 g l1, c(CTABr)=20 mg g1, J=40 l m2 h1, no backflushing, ug=0.33 cm s1.

c(CTABr) = 50 mg g
c(CTABr) = 100 mg g
Removal efficiency of flotation [%]





0 5 10 15 20 25
Operating time [h]

Fig. 8. Removal efficiency of flotation as a function of time for the hybrid process in wastewater treatment from copper foil
production, for two collector concentrations, under the following conditions: c(zeolites)=10 g l1, A=0.054 m2, =40 l m2 h1,
ug=0.33 cm s1.

with flotation was proven to have beneficial effects in 40 l m2 h1 without backflushing was higher, as bulk
terms of membrane performance and energy efficiency. zeolite concentration increased.
However, it was also found out that the membranes are
sensitive to small changes in flotation efficiency. In 3.3. Application of the process to treat real wastewater
Fig. 7, the transmembrane pressure over a 1-day period
of operation is shown for three experiments that were The developed process was used to treat wastewater
performed basically under the same conditions. Yet from a company producing copper foils that are
because of insufficient homogeneous collector dosing in polluted with copper, zinc and nickel ions in concentra-
the beginning, flotation efficiency differed, resulting in tions of 474, 167 and 3.3 mg l1, respectively. To achieve
different zeolite concentrations in the bulk solution satisfying flotation efficiency, higher collector concen-
(cbulk). Increase of TMP for an imposed flux of trations were needed for this wastewater than for the
C. Blocher et al. / Water Research 37 (2003) 4018–4026 4025

Table 3 * Extremely high concentrations of bonding agents in

Concentrations of heavy metal ions in wastewater from the the froth were achieved, making regeneration or
copper foil producing industry and in the treated water disposal easier.
in relation to the limit required by the company for water reuse * It was shown that the process is capable of treating
(in mg l1)
wastewater from copper foil production to obtain a
Ion Waste water Treated water Water reuse quality that allows water reuse.
limit (required
by the Further research will focus on optimising operating
company) conditions with different industrial wastewater (includ-
Cu(II) 474 o0.05 o0.1 ing the investigation of extreme flotation conditions with
Zn(II) 167 o0.05 o0.1 increased air flow rates to further limit membranes
Ni(II) 3.3 o0.05 o0.1 fouling) as well as froth collection and regeneration of
the bonding agents.

zeolite solutions investigated above. Also, in general,
lower efficiency for flotation was observed. Nevertheless,
It is gratefully acknowledged that this work is
with collector dosing of 100 mg CTABr per 1 g zeolite
supported by the European Union within the 5th
about 75% zeolite removal by flotation was achieved
framework programme as part of the METASEP
(Fig. 8). Again the membranes proved to be a total
project (contract no EVK1-CT-2000-00083).
barrier for the zeolites.
Table 3 contains details of the heavy metal concen-
trations in the permeate, after 7 h continuous operation.
For all three metals, concentration was below the References
detection limit, consistently meeting the water reuse
limit of the company. Corresponding removal efficiency [1] Kim JS, Akeprathumchai S, Wickramasinghe SR. Floccu-
was >99.99% for copper, >99.97% for zinc and lation to enhance microfiltration. J Membr Sci 2001;182:
>98.5% for nickel. 161–72.
Membrane performance in these experiments with [2] Cooney E, Luo B, Patterson JW, Petropoulou C. Toxicity
reduction methodologies. In: Water quality management
real wastewater was similar to the ones shown above.
library, vol. 3, Lancaster, USA, Technomic Publishing,
Even in 1 week of continouous experiments no fouling 1992 p. 109–81.
was observed that could not be eliminated by back- [3] Lazaridis NK, Jekel M, Zouboulis AI. Removal of Cr(VI),
flushing. Mo(VI) and V(V) ions from single metal aqueous solutions
by (i) sorption or (ii) nanofiltration. Separ Sci Technol,
2003, in press.
4. Conclusions [4] Matis KA (editor.), Flotation science and engineering.
New York: Marcel Dekker, 1995. 584pp.
[5] Mavrov V, Erwe T, Chmiel H. Selective separation of
The investigations presented prove that the concept of heavy metals from industrial wastewater streams.
a hybrid system of flotation and immersed microfiltra- In: Kungolos AG, Liakopoulos AB, Korfiatis GP,
tion membranes is promising. Tests were conducted for Koutsospyros AD, Katsifarakis KL, Demetracopoulos
the separation of zeolites within the process of heavy AD, editors. Proceedings of the International Conference
metal removal from industrial wastewater with the on Protection and Restoration of the Environment
following conclusions: VI, Skiathos, Greece, 2002. p. 707–14.
# e PL, Smith BM, Deutschmann AA, Rodrigues CFF,
[6] Cot!
* The beneficial interaction of flotation and immersed Pedersen SC. Frameless array of hollow fibre mem-
membranes was shown: Membranes guarantee an branes and method of maintaining clean fiber surfaces
almost 100% removal of the zeolites while flotation while filtering a substrate to withdraw a permeate.
maintains a low concentration of bonding agents in World patent WO 94/11094, Zenon Environmental Inc,
the bulk solution where the membranes are placed. 1994.
[7] Yamamoto K, Hiasa M, Mahood T, Matsuo T. Direct
This allows high membrane fluxes despite the low
solid–liquid separation using hollow-fibre membrane in an
turbulence created by the flotation air. activated sludge aeration tank. Water Sci Technol
* No energy in addition to that needed for flotation 1989;21(4–5):43–54.
was necessary for fouling control of the membranes, [8] Stephensen T, Judd S, Jeffersen B, Brindle K. Membrane
leading to low operating costs of the process if bioreactors for wastewater treatment. London: IWA
compared to conventional membrane filtration. Publishing; 2000.
4026 .
C. Blocher et al. / Water Research 37 (2003) 4018–4026

[9] Guibert D, Ben Aim R, Rabie H, Cot! # e P. Aeration [13] Matis KA, Zouboulis AI, Gallios GP, Erwe T, Bl.ocher C.
performance of immersed hollow-fibre membranes in a Application of flotation for the separation of metal-loaded
bentonite suspension. Desalination 2002;148:395–400. zeolites. Hydrometallurgy, 2003, submitted for publication.
[10] Blocher .
C, Noronha M, Funfrocken E, Dorda J, Mavrov [14] Cheryan M. Ultrafiltration and microfiltration handbook.
V, Janke HD, Chmiel H. Recycling of spent process water Lancaster, USA: Technomic Publishing; 1998.
in the food industry by an integrated process of biological [15] Cho BD, Fane AG. Fouling transients in nominally sub-
treatment and membrane separation. Desalination 2002; critical flux operation of a membrane bioreactor. J Membr
144:143–50. Sci 2002;209:391–403.
# e PL, Mourato D, Gungerich
[11] Cot! . C, Russell J, Houghton [16] Nagaoka H, Tanaka A, Toriizuka Y. Measurement of
E. Immersed membrane filtration for the production of effective shear stress working on flat-sheet membrane by
drinking water: case studies. Desalination 1998;117:181–8. air-scrabbling. In: Proceedings of the Fifth Conference on
[12] Zouboulis AI, Zamboulis D, Matis KA. Foam flotation of Membranes in Drinking and Industrial Water Production,
zeolites: application on zinc ion removal. Separ Sci .
Mulheim/Ruhr, Berichte aus dem IWW Bd. 37a, 2002.
Technol 1991;26:355–65. p. 235–43.