Water Research 38 (2004) 593–600

Metal ions removal from wastewater or washing water from

contaminated soil by ultrafiltration–complexation

Raffaele Molinari*, Saverio Gallo, Pietro Argurio

Department of Chemical and Materials Engineering, University of Calabria, Via P. Bucci, Cubo 45/A, Rende (CS) I-87030, Italy

Received 18 September 2002; received in revised form 10 September 2003; accepted 14 October 2003

Abstract

In the present paper a process for removal of ions from wastewater or from washing water of contaminated soil by

using the weakly basic water-soluble polymer polyethylenimine (PEI) as chelating agent and the Cu

2+

ion as model in

combination with an ultrafiltration process was investigated. The complexing agent was preliminarily tested to establish
the best operative conditions of the process. Next, ultrafiltration tests by using five different membranes were realised to
check membrane performance like flux and rejection. Finally, the possibility for recovering and recycling the polymer
was tested in order to obtain an economically sustainable process. Obtained results showed that complexation
conditions depends on pH: indeed, at a pH>6 PEI–Cu

2+

complexes are formed, while at pH

o3 the decomplexation

reaction takes place. Saturation condition is 0.333 mg Cu

2+

/mg PEI, meaning a ratio PEI/Cu

2+

=3 (w/w). UF tests

showed good results using the PAN 40 kDa membrane reaching an average copper concentration in the permeate of
2 mg/l and a flux of 135.4 and 156.5 l/h.m

2

at 2 and 4 bar, respectively. Metal rejection, permeate flow rate, and

possibility to regenerating and recycling the polymer makes the polymer-assisted ultrafiltration process (PAUF) very
interesting for metal ion removal from waters.
r

2003 Elsevier Ltd. All rights reserved.

Keywords: Ultrafiltration–complexation; Polymer-assisted ultrafiltration; Cu (II) ion removal from water; Wastewater treatment; Soil
remediation

1. Introduction

Water treatment plays an important role in the wide

subject of pollution problems solving and represents
today one of the most important fields of study. In fact,
a ‘‘rational hydrologic resource management’’ is neces-
sary because of increased world’s demand of water,
particularly in these last years owing to lacking of this
resource. Other than in the Mediterranean Middle-East,
it is known that in some regions of Southern Italy the
service of water distribution is not continuous, and this
phenomenon is happening also in Northern Italy
because of anomalous climatic changes, especially in

winter. In this situation some textile industries of
Northern Italy, which request big water consumption,
have taken into account to review their operational
system, treating and recycling wastewater, in order to
find a remedy for unoptimistic prevision for the future.
So, the approach of a pondered water consumption and
its purification and recycling has become very important
as stated also by the European Commission (Council
Directive 96/61/EC) for achieving integrated prevention
and control of pollution through the application of Best
Available Techniques (BAT) to obtain a high level of
protection of the environment as a whole.

The engineered systems associated with wastewater

reclamation, recycling and reuse can play an important
role in the natural hydrologic cycle. An overview of
the cycling of water from surface and ground-
water resources to water treatment facilities, irrigation,

ARTICLE IN PRESS

*Corresponding author. Tel.: [phone]; fax: +39-

0984-496-655.

E-mail address:

[e-mail] (R. Molinari).

0043-1354/$ - see front matter r 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.watres.2003.10.024

municipal, and industrial applications, and to waste-
water reclamation and reuse facilities is described
elsewhere

[1]

. This scheme takes into account that in

most cases the reuse of wastewater can produce water
for non-potable applications.

Metal contamination is a dangerous cause of water

pollution: indeed, e.g., Cu

2+

ions are essential nutrients,

but when people are exposed to Cu levels of above
1.3 mg/l for short period of time, stomach and intestinal
problems occur. Long-term exposure to Cu

2+

leads to

kidney and liver damage.

Separation processes for metal ions removal from

aqueous solutions are a major industrial activity cover-
ing processes ranging from production of potable water,
to leaching and recovery of metals from contaminated
soil or from ores, to detoxification of process water and
wastewater, also for water recycling and reuse. A variety
of separation processes for metal ions have been
developed up-to-date for industrial need: e.g., in water
softening (Mg

2+

, Ca

2+

removal) the first studied

traditional process is the lime-soda method, which
causes a precipitation of hardness

[2]

. This technique,

like the others, consisting practically in an induced
sedimentation,

produce water within international

health standards, but has two important drawbacks:
they produce big amount of sludge

[3–5]

containing

residual of reagents used, which result in a pollution
problem, and treated water may contain residual
coagulants if the process is not correctly controlled or
operated

[6]

. Another important method for metal

removal from polluted water is the ion exchange process,
which produces water within international quality
standards, but it is not a continuous process, because
of regeneration necessity

[2]

.

Membrane processes play today an important role in

the field of wastewater purification and reuse. This well
consolidated technology is very interesting because of
low operative costs, conceptual simplicity, modularity,
and optimal quality of treated water. Furthermore, the
use of new materials permits to obtain very resistant
membranes, both on chemical and mechanical point of
view, for various applications

[6–8]

.

For separating species with ionic dimensions, reverse

osmosis membranes are required but they will result in
high operative costs, low permeate flow rate and low
ions selectivity. In order to overcome these problems,
the ultrafiltration–complexation, also named polymer-
assisted ultrafiltration (PAUF), was introduced. Ultra-
filtration can be used for removal of trace metals from
aqueous streams, provided that these metals are
primarily bound to water-soluble polymers

[9]

. The

unbound metals pass through the membrane, whereas
the polymers and their complexes are retained. This
PAUF process can be applied for various purposes
such as the treatment of waste effluents, groundwater
and seawater. The advantages of this method are the

low-energy requirements involved in the ultrafiltration
and the high removal efficiency because of effective
binding

[10]

. Several research efforts have been carried

out to study the applicability of PAUF in metal removal
from water of various origins. Pivot of the study are the
consideration on technical and economical feasibility, to
respect the limits fixed by pollution laws: Juang and
Shiau

[10]

studied the metal removal from aqueous

solutions using chitosan-enhanced membrane filtration,
and in other two works

[11,12]

the authors considered

the problem of technical feasibility on the use of PAUF
for brackish water softening, or wastewater treatment by
using three weakly basic, water-soluble polymers like
chitosan, polyethylenimine (PEI), poly(diallyl dimethy-
lammonium chloride) to remove ions like Ca

2+

, Mg

2+

,

Na

+

, K

+

, Cu

2+

and Zn

2+

. Tabatabai et al.

[13]

studied

the feasibility of PAUF for water softening in the
removal of Ca

2+

and Mg

2+

ions from hard water by

using the polymer sodium polystyrene sulfonate (PSS).
They demonstrated (with some economical considera-
tions) that the PSS needs to be recovered from the
retentate and regenerated appropriately, to be reused.
Steenkamp et al.

[14]

considered the Copper (II) removal

from polluted water with alumina/chitosan composite
membrane, giving attention prevalently to the problems
related to the synthesis of their composite support
and to the factors which influence metal removal
efficiency, like pore radii variation with temperature
and powder mixtures used, and chitosan coating
thickness. Vieira et al.

[15]

studied an application of

PAUF in metal removal from pulp and paper industry
wastewater.

All the cited works consider the PAUF and its

applicability in metal removal from water, but they use
polymers that in their complexing action to complex
ions of interest, release other ions, like Na

+

or H

+

,

which results in a potentially modification of water
characteristic, or they do not consider the use of the
polymers at their maximum complexation ability,
saturation conditions and chemistry of polymer–metal
complexation.

In the present work, some results of a study on

metallic ions removal from wastewater and from
washing water of contaminated soils by means of PAUF
are reported.

The determination of complexation, de-complexation

and chemical conditions of saturation are discussed; the
results of ultrafiltration tests on five different mem-
branes realised by using the weakly basic poly(ethyleni-
mine) as the selective polymer and the copper as model
ion are reported. This polymer has the advantage of no
release of other ions in treated water because it does not
work by means of ionic exchange reaction, but it
complexes also counter-ions to form neutral complexes.
A criterion to find the membrane with the best
performance is also reported.

ARTICLE IN PRESS

R. Molinari et al. / Water Research 38 (2004) 593–600

594

2. Materials and methods

Copper sulphate penta-hydrate (CuSO

4

5H

2

O) from

Fluka Chemika (purity>99%) was used for preparing
Cu

2+

solutions. Poly(ethylenimine) 50% wt solution in

water from Sigma-Aldrich (MW 10,000 and 60,000) was
the polymer utilised. Other chemicals used were H

2

SO

4

(purity 95–97%) purchased by Riedel de Haen and
NaOH (purity>99%) from Merck.

The used ultrafiltration plant (

Fig. 1

) permitted to test

simultaneously five different UF membranes. It was
constituted by three sections: electric panel, to control
plant working; alimentation section, constituted by
a feed reservoir of 25 l, in which a cooling coil
was immersed with a thermostat and a level control
sound; ultrafiltration section, with a centrifugal pump
LOWARA CKM 70/34 that generates the flow (max
flow rate=1.5 m

3

/h; max pressure=4.5 bar), five steel

plane cells (useful membrane surface area 14.18 cm

2

)

equipped with manometers and flow meters to control
operative transmembrane pressures and tangential flow
rates of the concentrate (retentate).

The thermostatic system controlled the operative

temperature, resulting isothermal ultrafiltration runs
(t ¼ 25



C). The coolant was simply tap water.

Before each ultrafiltration run the membranes were

characterised with demineralised water in the same
plant, in order to evaluate the relative membrane
permeability and the membrane fouling after the
ultrafiltration runs. During this characterisation, opera-
tive transmembrane pressure was fixed first at 4 bar to
stabilise the compaction, and later it was decreased
to 2 bar. Ultrafiltration tests were carried out setting
the operative pressure first at 2 bar and after increased
at 4 bar, because it is known that fouling tendency

increases with transmembrane pressure. The plant
was operated in batch mode recycling the five permeates
to feed reservoir. Every 30 min (starting by the
time t ¼ 0 min) permeates were collected for 2 min
and their volumes measured in order to calculate
instantaneous flux; they were also analysed to determine
copper concentration. Each ultrafiltration run was
stopped (or pressure was changed) when steady state
was reached, that means permeate flux and copper
concentration were practically constant. The aver-
age time to reach the steady state was between 2.5
and 3 h.

Permeates collected in the steady-state condition were

also submitted to total organic carbon (TOC) measure-
ments, in order to verify if the polymer passed through
the membranes. The thermostat LT 100-1, the photo-
meter LASA 100 and the analytical kits LCK 380 and
LCK 381 (depending on the concentration of the total
carbon estimated in the sample) from Dr. Lange, were
used for carrying out TOC measurements.

Five different UF membranes were tested measuring

retention and water permeate flux. Some of their
characteristics are reported in

Table 1

where fluxes

measured with distilled water at 2 and 4 bar are also
reported. The retention R

TC

; for the target component

(TC), was measured by using its definition:

R

TC

¼ 1  ½C

TC;P

=C

TC;F

;

where C

TC;P

and C

TC;F

are the concentrations of the TC

(Cu

2+

) in the permeate and in the feed solution,

respectively.

The other important parameter, the volume permeate

flux (J), generally expressed as the volume obtained per
unit time (t) and per unit of membrane surface (S), was

ARTICLE IN PRESS

Feed

reservoir

Pump

Flow meters

Plane cells

Feed

header

Permeates

Permeate

header

Retentate

header

Fig. 1. Flowsheet of the ultrafiltration laboratory plant able to tests simultaneously five UF membranes.

R. Molinari et al. / Water Research 38 (2004) 593–600

595

measured by the following equation:

J ¼ V =ðtSÞ:

Determination of copper concentration was carried

out by using an analytical kit (Carlo Erba Reagenti),
based on a colorimetric reaction and absorbance reading
at a 600 nm wavelength. Absorbance reading was
performed by using a Recording Spectrophotometer
UV–Visible 160 A (Shimadzu Corporation-Analytical
instruments division).

A pH meter (Orion Research Incorporated–Expand-

able ion Analyzer EA 920) with a combined glass
electrode was used for pH measurements.

3. Results and discussion

The water-soluble polymer PEI was considered in this

preliminary phase of our work. It shows a good affinity
for the model ion Cu

2+

and has the advantage of no

release of counter ions in the treated water with respect
to polymers that work with an ion-exchange mechanism.
In

Fig. 2

, the mechanism of PEI–copper interaction is

reported

[16]

, where the lone-pair of nitrogen binds the

copper according to the acid and base Lewis theory.

The mechanism of PEI–copper interaction can be

described by the following equilibrium reactions:

PEI þ nH

2

O

"PEIH

nþ
n

þ nOH



;

ð1Þ

PEI þ aCu

2þ

"PEICu

2aþ
a

;

ð2Þ

where 0

pnp %n and 0pap%a with %n equal to the number

of monomers contained in a single polymeric chain and

%a representing the maximum complexation ratio of the

polymers with copper ions (

%a ¼ %n=4 as showed in the

idealised structure previously reported). In particular,
considering the longer polymeric chain (MW 60 kDa)
and considering the monomeric unit –CH

2

–CH

2

–NH–

(MW 43.062 Da) we obtain that

%n ¼ 1393:

Reactions (1) and (2) are competitive for the polymer

because, depending on pH conditions, it is able to
complex copper ions by means of Eq. (2) or stays
in aqueous solution like PEI H

nþ
n

at low pH incapable to

interact with copper. Measuring the pH of PEI in
aqueous solutions, the K

eq

of Eq. (1) was practically

equal to zero, meaning that this equilibrium reaction is
hardly shifted at left. So, the polymer at its natural pH in
water stays prevalently as PEI, and not in the form
PEI H

nþ
n

:

First step of this research consisted in the determina-

tion of optimal chemical conditions (pH) for copper
complexation (bound) and de-complexation (release).
The determination of release condition is fundamental
for recovering and recycling the binding polymeric
agent. Indeed, it should be taken into account that the
PAUF process appears to be economically more feasible
if the polymer could be regenerated and reused, so that
the process should be represented as reported in

Fig. 3

.

Complexation and de-complexation conditions were

determined by means of some tests conceptually very
similar to L–L extraction. They were carried out in
isothermal conditions at a temperature of 25

71



C b y

preparing 20 ml of aqueous solutions containing poly-
mer and copper at concentrations respectively of 150
and 50 ppm, and changing the pH. In order to evaluate
the influence of polymer molecular weight on copper
complexation, these experiments were realised both with
PEI 10,000 and PEI 60,000. To quantify the copper–
polyethylenimine (Cu–PEI) complex formation, the
spectrophotometric technique was used. By previous
scanning of various samples, the peaks of absorbance vs.
wavelength showed a maximum absorbance for Cu–PEI
complex in aqueous media at a wavelength of 620 nm.
By taking advantage of complex reading at this
wavelength whilst no reading was observed for Cu

2+

alone, it was possible to establish if the complex was or
not formed by simple experiments in test tubes without
using the membrane separation process. The complexa-
tion–decomplexation process was quantified by plotting

ARTICLE IN PRESS

Table 1
Some characteristics of the tested UF membranes

Membrane type

Material

Cut-off (kDa)

Producer

Water flux (l=h m

2

) (2–4 bar)

Iris 10

Polyether sulphone (PES)

10

Tech-Sep

33.85–55.00

FS 40 PP

Fluoride-polypropylene

40

Dow

220.0–397.7

GR 40 PP

Polysulphone-polypropylene

40

Dow

220.0–444.3

Iris 30

Polyether sulphone (PES)

30

Tech-Sep

114.2–207.3

PAN 40

Polyacrylonitrile

40

Tech-Sep

291.1–528.9

Fig. 2. Idealised structure of the polymeric complex PEI–
copper (II) ions.

R. Molinari et al. / Water Research 38 (2004) 593–600

596

the complexation percentage C

%¼ ðABS=ABS

max

Þ

100; where ABS

max

is the maximum value of the

absorbance obtained from the experiments which
corresponds to the maximum amount of complex
(100%). Obtained results for PEI 10,000, reported in

Fig. 4

, show that this polymer is able to complex copper

ion at pH 6 or higher, while the decomplexation happens
at pH

o3 ðC%o10%Þ: The results of complexation tests

with PEI 60,000 are not reported because the obtained
data were practically equal to that ones obtained with
the smaller chain-polymer. Indeed, the number of single
complexes in

Fig. 2

, [Cu

2+

–(NH)

4

], depends only on

total monomeric units present in the overall polymer
chains and not on polymer molecular weight. Similar
results were obtained by working at different polymer
concentrations (300 and 50 ppm), showing that pH of
maximum bonding does not depend also on polymer
concentration.

These results agree with the chemical mechanism of

polymer–copper interactions. Indeed, at high pH the
complexation reaction (2) takes place.

In order to determine the bonding capacity (satura-

tion condition) of PEI (maximum copper amount, e.g.
grams, that can be complexed by a fixed amount, e.g.
1 g, of polymer), some complexation tests were carried
out with a polymer concentration of 150 mg/l (volu-
me=20 ml) and changing copper concentration at a

fixed pH value (equal to 6). It should be taken into
account that one important cost of PAUF process is
represented by polymer consumption and its bonding
efficiency: this justify the convenience to use the polymer
at its maximum complexation capacity. Obtained
results, reported in

Fig. 5a

for PEI 10,000 and 60,000,

show that maximum copper that can be complexed by
150 mg/l PEI is equal to 50 mg/l, that is 0.333 mg Cu

2+

/

mg PEI. This value is also confirmed in

Fig. 5b

for

various concentration values of Cu

2+

and polymer. The

excess of copper remains in solution like hydroxide. In
fact, taking into account that the solubility product
constant K

sp CuðOHÞ

2

¼ ½Cu

2þ

½OH





2

¼ 10

19:9

; at 25



C,

maximum feed pH to avoid copper hydroxide formation

ARTICLE IN PRESS

Recovered

Metal solution

Polymer Recycle

Metal

containing

Feed

Permeate (water to
reuse or discharge)

Retentate

Complexation

Polymer

make-up

MEMBRANE

STEP

Polymer

Regeneration

Fig. 3. Schematic principle of the PAUF separation process.

0

20

40

60

80

100

2

4

6

8

10

pH

C %

300/50

150/50

50/50

Fig. 4. Cu–PEI complex formation C

%ð¼ ðABS=ABS

max

Þ

100Þ vs. pH in complexation tests of PEI 10,000 (300, 150 and
50 mg/l) with copper (50 mg/l).

0

20

40

60

80

100

0

30

60

90

120

150

Copper concentration [mg / l]

C %

PEI 60000

PEI 10000

R

2

= 0.9857

0

40

80

120

0

100

200

300

400

Polymer concentration [mg / l]

[C

u

2+

]

max

(a)

(b)

Fig. 5. (a) Cu–PEI complex formation C

%ð¼ ðABS=ABS

max

Þ

100Þ vs. copper concentration (initial PEI concentration=
150 mg/l, pH=6) and (b) determination of the binding capacity
of PEI 60,000 (pH=6).

R. Molinari et al. / Water Research 38 (2004) 593–600

597

and precipitation is approximately 5.6 for an aqueous
concentration of copper equal to 50 mg/l, while in these
complexation tests operative pH was 6.

Important result of the complexation tests is the

identical binding capacity obtained operating with PEI
10,000 and 60,000. This is a consequence of the
concentration unit used: in fact, PEI binding capacity
depends only on the number of complexation sites that
are the same for a same weight amount of polymer.
Obviously, this is not applicable when referring to molar
concentration (polymer moles/litres of solution).

Adding the previous consideration on identical

operative pHs and binding capacity for the two size of
PEI used and taking into account that in PAUF the
complexation step is followed by a membrane filtration,
the PEI 60,000 is more interesting than the smaller one,
permitting the use of membranes with bigger pores for
obtaining low operating costs.

From the above results the operative conditions for

ultrafiltration tests were ratio PEI/Cu

2+

=3(w/w) and

pH=6.

Ultrafiltration tests were carried out using five

different membranes (see

Table 1

), two operative

trans-membrane pressures (2 and 4 bar), pH approxi-
mately equal to 6 and three different weight concentra-
tions of PEI and Cu

2+

(150/50, 375/125, 600/200).

Working at increasing PEI/Cu

2+

concentrations per-

mitted to simulate the increase of retentate concentra-
tion in a hypothetical industrial plant where the
permeate free of metals is withdrawn using the PAUF
technique. We used this approach because of the small
volume of permeate collected during an ultrafiltration
run (as a consequence of the small useful membrane
surface area); indeed, the increase of concentration in
the retentate by a direct withdrawn of permeate would
require a lot of working hours of the laboratory plant.

Each test, at different copper and polymer concentra-

tions, was carried out by starting with a new set of the
five membranes: in fact, to realise a process with an
industrial applicability it is certainly important to
evaluate membrane fouling, regeneration and reuse,
but goal of our preliminary tests was to evaluate and
compare membrane performances at increasing reten-
tate concentrations, which practically results in an
increase of membrane fouling and/or concentration
polarisation.

Obtained results, summarised in

Figs. 6–8

, show that

increasing copper and then polymer concentration (ratio
PEI/Cu

2+

=3 fixed) in the retentate, the separation

efficiency (R

%) decreases, that will results in a copper

and polymer concentrations increase in the permeate
and a little decrease of permeate flux. By increasing
polymer concentration in the retentate (

Fig. 8

), rejection

first decreases, because of fouling, but increasing
retentate concentration this tendency changes, because
of the formation of a selective dynamic layer (by

concentration polarisation) that improves the separa-
tion. This causes a little permeate flux decrease too,
because of mass transfer resistance increase. The
inspection of the membranes at the end of each run
permitted to confirm cake formation: in fact, each used
membrane presented a thin layer on its filtering surface.
This cake was cerulean, practically the colour of the
polymer–copper complex, but it not gave too much
intensity of the fouling after simple washings with tap

ARTICLE IN PRESS

0

100

200

300

150

300

450

600

Polymer concentration in the retentate [mg / l]

Flux [l/h*

m

2

]

Iris 10 kDa

PAN 40 kDa
DOW GR 40

DOW FS 40 PP
Iris 30 kDa

Fig. 6. Comparison of flux through the five membranes by
increasing polymer concentration (values at steady state,
P ¼ 2 bar, PEI=Cu

2þ

¼ 3).

0

2

4

6

8

10

150

300

450

600

Polymer concentration in the retentate [mg / l]

Copper Cp [mg

/l

]

Iris 10 kDa
PAN 40 kDa
DOW GR 40
DOW FS 40 PP
Iris 30 kDa

Fig. 7. Comparison of copper concentration in the permeate
(C

p

) for the five membranes by increasing polymer concentra-

tion (values at steady state, P ¼ 2 bar, PEI=Cu

2þ

¼ 3).

96.0

98.0

100.0

150

300

450

600

Polymer concentration in the retentate [mg / l ]

R%

Iris 10 kDa

Iris 30 kDa

PAN 40 kDa

DOW GR 40

DOW FS 40
PP

Fig. 8. Comparison of copper rejections (R

%) for the five

membranes by increasing polymer concentration (values at
steady state, P ¼ 2 bar, PEI=Cu

2þ

¼ 3).

R. Molinari et al. / Water Research 38 (2004) 593–600

598

water. Indeed, the initial water permeate flux obtained
with demineralised water (see

Table 1

, column at 2 bar)

was about the same of that one obtained in operating
conditions (see

Fig. 6

).

It should be taken into account that an optimal

PAUF processes should generate high permeate flux (J

P

)

with low copper concentration (C

p

). So, in order to

compare membrane performances, an appropriate para-
meter J

p

=C

p

was introduced. This parameter has no

dimensional significance, but it answers to the previous
requirements to optimise PAUF processes.

Data of our optimisation parameter J

p

=C

p

; reported

in

Figs. 9 and 10

at 2 and 4 bar, respectively, show that

the PAN 40 kDa membrane gives the best combination
of the two parameters. Furthermore, it is better
operating at P ¼ 2 bar rather than at 4 bar as the higher
J

p

=C

p

value shows.

The economical feasibility of PAUF process depends

also on polymer regeneration, so some UF tests were
carried out in the laboratory plant, with a set of five
membrane previously used and with the following
operative

conditions:

PEI=150 ppm;

Cu=50 ppm;

pH=3 (de-complexation conditions). The permeates
withdrawn at established time were analysed to deter-
mine copper and TOC concentrations: obtained data

showed that all the copper passed through the mem-
brane, while the polymer remained in the retentate
(rejection of 95% with PAN 40 kDa membrane), that
means a good possibility of polymer regeneration,
recovery and reuse.

4. Conclusions

The described study on polymer assisted ultrafiltra-

tion shows: (i) the importance to use optimal chemical
conditions to obtain a maximum binding capacity of the
polymer; (ii) the role of different types of UF
membranes in order to employ this process in metal
ions removal from various types of wastewaters. In the
case of Cu

2+

removal by its complexation with PEI best

results were obtained for the membrane PAN 40 kDa,
reaching an average copper concentration in the
permeate of 2 mg/l and a flux of 135.4 and 156.5 l/h.m

2

at 2 and 4 bar, respectively. The obtained results show
that use of PAUF process with PEI does not reach a
complete removal of the metal, but can reach the
objective of the purification process that is to decrease
metal concentration down a certain value required by
reuse or fixed by water laws for discharge.

Acknowledgements

The authors wish to thank the National Interuniver-

sity Consortium ‘‘Chemistry for the Environment’’
(INCA) which partially supported this work within the
Sisifo Project and the National INCA Plane ‘‘Remedia-
tion of contaminated soil’’.

References

[1] Asano T, Levine D. Wastewater reclamation, recycling

and reuse: past, present and future. Water Sci Technol
1996;33(10–11):1–14.

[2] Brisi C, editor. Chimica applicata. Torino: Libreria

Editrice Universitaria Levrotto & Bella; 1990.

[3] Kruithof JC, Kopper HMM. Experiences with ground-

water treatment and disposal of the eliminated substances
in the Netherlands. Aqua 1989;38:207–16.

[4] Drioli E, Molinari R. Operazioni a membrana nel

trattamento di acqua civili ed industriali, 3



Convegno di

Terra d’Otranto: ‘‘L’Acqua fattore di crescita del territor-
io: qualit

"a, trattamento, riuso e gestione’’, Lecce 25–26

Maggio 1995. p. 271.

[5] Molinari R, Grande C. In: Palmisano L, editor. Processi

innovativi a Membrana nel Trattamento delle Acque.
Abbiategrasso, MI: Edizioni spiegel; 2000. p. 159–86
[Chapter 9, ISBN 88-7660-159-7].

[6] Scott K. Handbook of industrial membranes. Oxford:

Elsevier Advanced Technology; 1995.

ARTICLE IN PRESS

0

200

400

600

800

150

300

450

600

Polymer concentration in the retentate [mg / l]

(Jp

/Cp)

Iris 10 kDa

PAN 40 kDa

DOW GR 40

Iris 30 kDa

DOW FS 40 PP

Fig. 9. Comparison of J

p

=C

p

for the five membranes by

increasing polymer concentration (values at steady state,
P ¼ 2 bar, PEI=Cu

2þ

¼ 3).

0

200

400

600

150

300

450

600

Polymer concentration in the retentate [mg / l]

(Jp

/Cp)

Iris 10 kDa

PAN 40 kDa

DOW GR 40

Iris 30 kDa

DOW FS 40 PP

Fig. 10. Comparison of J

p

=C

p

for the five membranes by

increasing polymer concentration (values at steady state, P ¼ 4
bar, PEI=Cu

2þ

¼ 3).

R. Molinari et al. / Water Research 38 (2004) 593–600

599

[7] Aptel P, Vial D. Membranes in water treatment and

potabilization. In: Haber C, Drioli E, editors. Proceedings
of the Membrane Separation Processes. Brazil: Rio de
Janeiro; 1992. p. 291.

[8] Mulder M. Basic principles of membrane technology.

Dordrecht,

Holland:

Kluwer

Academy

Publishers;

1991.

[9] Rumeau M, Persin F, Sciers V, Persin M, Sarrazin J.

Separation by coupling ultrafiltration and complexation of
metallic species with industrial water soluble polymers.
Application for removal or concentration of metallic
cations. J Membr Sci 1992;73:313–22.

[10] Juang RS, Shiau RC. Metal removal from aqueous

solutions using chitosan-enhanced membrane filtration.
J Membr Sci 2000;165:159–67.

[11] Juang RS, Chiou CH. Ultrafiltration rejection of dissolved

ions using various weakly basic water-soluble polymers.
J Membr Sci 2000;177:207–14.

[12] Juang RS, Chiou CH. Feasibility of the use of polymer-

assisted membrane filtration for brackish water softening.
J Membr Sci 2001;187:119–27.

[13] Tabatabai A, Scamehorn JF, Christian SD. Economic

feasibility study of polyelectrolyte-enhanced ultrafiltration
(PEUF) for water softening. J Membr Sci 1995;100:
193–207.

[14] Steenkamp GC, Keizer K, Neomagus HWJP, Krieg H.

Copper (II) removal from polluted water with alumina/
chitosan composite membranes. J Membr Sci 2002;197:
147–56.

[15] Vieira M, Tavares CR, Bergamasco R, Petrus JCC.

Application of ultrafiltration–complexation process for
metal removal from pulp and paper industry wastewater.
J Membr Sci 2001;194:273–6.

[16] Geckeler KE, Volchek K. Removal of hazardous sub-

stances from water using ultrafiltration in conjunction with
soluble polymers. Environ Sci Technol 1996;30(3):725–34.

ARTICLE IN PRESS

R. Molinari et al. / Water Research 38 (2004) 593–600

600