You are on page 1of 9

Chemical Engineering and Processing 41 (2002) 601– 609

www.elsevier.com/locate/cep

Treatment of textile dye effluent using a polyamide-based


nanofiltration membrane
A. Akbari a,*, J.C. Remigy b, P. Aptel b
a
Uni6ersity of Kashan, Kashan, Iran
b
Laboratoire de Génie Chimique (CNRS UMR 5503), Uni6ersité Paul Sabatier, 118 route de Narbonne, 31062 Toulouse, Cedex, France

Received 6 August 2001; received in revised form 10 October 2001; accepted 12 October 2001

Abstract

Experiments were run with seven dyes and a Desal 5DK membrane (cut-off, 150– 300 g/mole). The effects of concentration, pH
and salt on flux and retention were studied. The cut-off of the membrane explains that the retention of the relative high molecular
weight dyes (as direct red 80 or direct yellow 8) is always almost 100%. For anionic dyes as acid orange 10 or acid red 4, the
amphoteric nature of polyamide explains the lower retention at pH 3 than 6. This effect is more pronounced and reversed for basic
blue 3, a cationic dye. The membrane is sensitive to fouling since most of the dyes are used for polyamide textile dying. Moreover,
the presence of salt leads to a further decrease in flux. © 2002 Elsevier Science B.V. All rights reserved.

Keywords: Nanofiltration; Textile dyes; Fouling; Polyamide

1. Introduction waste constituents (specifically dyes). In the last few


years, technical and economical improvements [2–7]
It is well-established fact that dye effluents dis- have made the treatment of industrial wastewater by
charged without appropriate treatment bring about a membrane systems even more advantageous.
number of undesirable changes in the recipient stream. Ultrafiltration has been successfully applied for recy-
Increased colour concentrations not only make the cling high molecular weight and insoluble dyes (e.g.
water unfit for domestic or industrial uses, but they also indigo, disperse), auxiliary chemicals (polyvinyl alco-
reduce light transmittance, thus limiting aquatic plant hol) and water [8,9]. However, ultrafiltration does not
growth and self-purification processes. In addition, dyes remove low molecular weight and soluble dyes (acid,
exert an adverse effect on fish life. direct, reactive, basic, etc.) [10] but efficient colour
Colour removal by conventional treatment methods removal has been achieved by nanofiltration and re-
(e.g. ozonation, bleaching, hydrogen peroxide/UV, elec- verse osmosis [11 –18].
trochemical techniques) was found to be inadequate [1] As for all membrane processes, the major problem is
because most textile dyes have complex aromatic the decline of permeate flux due to the accumulation of
molecular structures that resist degradation. They are molecules on the membrane surface. This accumula-
stable to light, oxidising agents and aerobic digestion. tion, known as concentration polarisation, leads to an
The need for more efficient treatment processes has increase of membrane fouling. Several approaches have
attracted the attention of environmental scientists and been proposed to limit fouling for the treatment of
engineers towards pressure-driven membrane tech- textile wastewater. Buckley has proposed the use of
niques. The application of membrane filtration pro- microfiltration to remove colloidal species from the
cesses not only enables high removal efficiencies, but exhausted dye bath before nanofiltration [11,12].
also allows reuse of water and some of the valuable Noël et al. have evaluated the performance of two
types of membrane (NF45 and BQ01) for a direct dye
* Corresponding author. Tel.: +33-561-55-7615; fax: + 33-561-55-
solution [13]. They applied an electric field to eliminate
6139. fouling problems. Electro-nanofiltration performance
E-mail address: akbari@chimie.ups-tlse.fr (A. Akbari). was quantified for different potentials and concentra-

0255-2701/02/$ - see front matter © 2002 Elsevier Science B.V. All rights reserved.
PII: S0255-2701(01)00181-7
602 A. Akbari et al. / Chemical Engineering and Processing 41 (2002) 601–609

tions. An electric field was found to be efficient in groups include the carboxyl, sulphonic, hydroxyl,
reducing fouling for both membranes studied. They amino groups, etc.
also noticed that direct dye solution-fouling changes the The dye is bound to the fibre through covalent and
NF45 membrane into a reverse osmosis membrane, ionic bonding, Van der Waals forces and impregnation
while the permeability of the second membrane was of colloidal dye particles into the fibre. Practically,
enhanced. textile dyes are classified into acid, basic, direct, dis-
Ratana et al. have studied three nanofiltration mem- perse, reactive, sulphur or vat (Table 1) [21].
branes for the treatment of effluents containing salt and The acid dyes were so named because of the presence
reactive dye [14]. The membranes were ES20, NTR- in their molecules of one or more sulphonic acid or
729HF and LES90. ES20 and LES90 are negatively other acidic groups. Acid dyes are water-soluble anionic
charged membranes and NTR-729HF, a substituted dyes that are applied to nitrogenous fibres such as
polyvinyl alcohol membrane, is neutral. It was observed wool, silk, polyamide (nylon) and modified acrylic
that the ES20 and LES90 membranes can effectively fibres from acid or neutral baths. Attachment to the
retain colour, producing a permeate suitable for reuse. fibre is attributed at least partly to salt formation
They also noted that for feed containing salt and dye, between anionic groups in the dyes and cationic groups
the transmembrane osmotic pressure difference is en- in the fibre.
hanced, resulting in a substantial flux decline. The Basic dyes, which are water soluble in aqueous solu-
presence of dye in the effluent significantly affects the tion, yield coloured cations. In the late 1930s a few
salt removal effectiveness of ES20 and LES90. basic dyes were developed for acetate and to a lesser
The dye molecule can be considered to be structured extent for nylon. A number of these dyes were later to
from two components the dye chromophore and the become more important for application to acrylic
dye auxochrome [20]. The chromophore includes dou- fibres.
ble bonds (CC) and is responsible for the colour of the The conventional definition [22] of direct dyes is,
dye. When the molecule is exposed to light, the struc- ‘anionic dyes substantive to cellulose when applied
ture of chromophore oscillates, light is absorbed and from an aqueous bath containing an electrolyte’. They
colour is visible. There are several chemical classes of provide the simplest means of colouring cellulose mate-
chromophore groups, which include the azo, an- rials since they are normally applied from a neutral or
thraquinone, triphenylmethane, indigo, sulphur groups, slightly alkaline bath, at or near the boil, to which
etc. The second portion of the molecule is responsible sodium chloride or sulphate is added in such quantities
for the bond between the fibre and the dye. These and at such intervals of time appropriate to the dyeing

Table 1
Dye classification

Dye class Characteristics Application Dye-fibre Dyeing method


attachment
mechanism

Acid Anionic, water soluble Nylon, wool Ionic bond, Van Fibre placed in acidified aqueous
der Waals media pH 2.5–7 dye added and
temperature 100–110 °C
Basic Cationic, water soluble Acrylic, nylon, silk cotton, Ionic bond Fibre placed in acidified aqueous
wool dye bath at pH 5–6.5 and
temperature 105 °C
Direct Anionic, water soluble Cotton H-bond Fibre placed in dye bath slightly
alkaline and electrolyte at
temperature 100 °C
Disperse Colloidal dispersion, very low water Polyester, nylon, acetate, Solid solution Fibre placed in acidified dye bath
solubility cellulose, acrylic formation pH 5.5 temperature 130–210 °C
Sulphur Colloidal, insoluble Cotton Dye precipitated Fibre placed in dye bath dye
fibre dissolved in alkaline sodium
sulphur, dye displaced to fibre with
electrolyte, dye precipitate in situ
with air or peroxide
Reactive Anionic, water soluble Cotton, silk, nylon, wool Covalent bond Fibre placed in dye bath add salt
to displace dye to fibre add alkali
to cause reaction between dye and
fibre
Vat As sulphur dye Cotton As sulphur dye As sulphur dye
A. Akbari et al. / Chemical Engineering and Processing 41 (2002) 601–609 603

Table 2
Characteristics of the dyes tested [21]

Dye (C.I…) M.W (g/mole) Structure Compact formula Charge

Acid red 4 (Aldrich) 402.36 Monoazo C17H13N2NaO5S −1


Acid orange 10 (Aldrich) 496.38 Monoazo C16H10N2Na2O7S2 −2
Basic blue 3 359.90 Oxazine C20H26N3OCl +1
Direct yellow 8 (Aldrich) 540.55 Monoazo C23H19N4NaO5S2 −1
Direct red 80 (Ciba) 1372.55 Polyazo C45H26N10Na6O21S6 −6
Disperse blue 56 (Ciba 305.7 Antraquinone C14H10N2O4Cl 0
Reactive orange 16 (Aldrich) 661.54 Monoazo C20H17N3Na2O11S3 −1

properties of individual dyes. The majority of direct highly probable that this fouling will be irreversible.
dyes belong to the dis-, tris-, and polyazo classes, the The following equations show the reaction schemes
remainder being monoazo, stilbene, oxazine, thiazole between the polyamide fibres and the textile dyes.
and phthalocyanine compounds. Some direct dyes have Below the isoelectric point (4.2) polyamides have a
extensive uses other than on cellulose fibres, many positive charge and can react with anionic dyes (acid,
being of outstanding importance for use on paper, direct, etc.).
leather, wool, silk and nylon.
H3N+ − Polyamide− COOH + DyeSO−
3
The disperse dyes are defined as ‘substantially water
insoluble dyes, behaving substantively for one or more “ DyeSO− +
3 H3N − Polyamide−COOH
Ionic bond
hydrophobic fibre, for example cellulose acetate, and
Above the isoelectric point, polyamides have a nega-
usually applied from fine aqueous dispersion’. Although
tive charge and can react with cationic dyes.
substantially insoluble, these dyes are slightly soluble in
water and uptake by the fibres is believed to take place H2N− Polyamide− COO− + Dye+
from this aqueous phase. The solubility of the dye and
“ H2N− Polyamide− COO−Dye+
hence its uptake equilibrium and dyeing rate may be Ionic bond
modified by the dispersing agent employed. In all cases, the exhaustion is dependent on pH,
The reactive dyes include molecules with active concentration, temperature, class of dye and electrolyte
groups such as mono-, dichlorotriazinyl, di-, as NaCl or Na2SO4.
trichlorophrimidinyl, chloroethylsulphonyl. They form There are two types of reactive dye, which can form
a bond with fibre, and it has been suggested that a covalent bond with polyamide:
self-condensation or the elimination of soluble groups
H2N− Polyamide− COOH +Dye −Cl
fix some of them. A range of reactive procinyl dyes for
nylon was introduced by ICI in 1959. They are applied “ Dye − HN − Polyamide− COOH
Covalent bond
in weakly acid conditions, under which reaction with
H2N− Polyamide− COOH + Dye − x−CH2CH2 −y
the fibre does not occur. Dyeing is then readily ob-
tained on treatment with alkali, a reaction with the “ Dye − x− CH2CH2 − HN − Polyamide −COOH
Covalent bond
fibre takes place and the resulting dyeing is very fast to
(x=O, SO2, … ,y= Cl, SO3H, …)
wet treatments. Free amino groups in the nylon fibre
provide reactive sites, but there is some evidence sug- This study primarily concerned the influence of the
gesting that reactions may also occur with amide interactions between a polyamide membrane (Desal
groups. 5DK) and seven dyes, on the transport properties in
Thus, the textile dyes are reactive towards the chemi- nanofiltration. The effects of such significant parame-
cal functions present at the fibre surface. These same ters as pH, solution concentration and salt effects on
chemical functions are found in the polymer surface two fundamental characteristics of nanofiltration (flux
used for the preparation of membranes. For example, and retention) will be presented.
most nonofiltration membranes are composite mem-
branes, the active surface of which is polyamide.
Polyamide fibres can be dyed by most classes of dyes. 2. Material and methods
Acid, disperse, direct and reactive dyes are classic ex-
amples but basic dye can also be used by changing pH. The Desal 5DK membrane, manufactured by Os-
Thus, it is reasonable to think that nanofiltration of monics, was selected for study. It is a hydrophilic
textile dyes with polyamide membranes will be prob- membrane with a nominal molecular weight cut-off of
lematic, heavy fouling is very probable because the dyes 150–300 g/mole. This membrane is a thin-film com-
will be fixed to the fibres by chemical bonds; and it is posite with an active layer made of polyamide [19].
604 A. Akbari et al. / Chemical Engineering and Processing 41 (2002) 601–609

The supplier, the chemical structure, the molecular the feed solution. After the measurements, the mem-
weight and the number of charges of the dyes tested brane was rinsed with distilled water and the fouled
are given in Table 2 and Fig. 1. Model dye solutions membrane permeability was measured with distilled
were prepared by dissolving dyes (as supplied) in dis- water. A new membrane was used for each experi-
tilled water at a concentration of 1 g/l without adding ment.
any auxiliary compounds. Addition of NaOH and The concentrations are determined either by visible-
H2SO4 to the solution was used to set the pH. UV spectrophotometry for the dye solution at the
All the nanofiltration experiments were performed in maximum absorption wavelength or by conductivity
a cylindrical stirred cell (300 cm3) with an active mem- for the Na2SO4 solution.
brane area of 36.3 cm2. Tests were first run using Adsorption tests were performed in a small cylindri-
distilled water under 20 bar in order to compact the cal filtration cell (active membrane surface area of 13.2
membrane. The filtration was then carried out at a cm2). A clean wet sample of membrane was installed
pressure of 10 bar with or without stirring. The stir- and 100 cm3 of the dye solution was poured into the
ring rate was measured by a stroboscope. After 10 min cell. Adsorption was followed versus time by measur-
filtration, the permeate sample was collected for 10 ing the decrease in absorbance of the solution. The
min and weighted. During the nanofiltration operation increase in dry weight of the exposed membrane was
with dye solutions, the permeate samples were sent estimated by weighting another membrane sample of
back to the cell to ensure a constant concentration of the same surface area.

Fig. 1. Chemical structure of tested dyes.


A. Akbari et al. / Chemical Engineering and Processing 41 (2002) 601–609 605

3. Definitions Table 3
Flux and retention of dye solutions with or without stirring (DP=10
bar, temperature 25 °C, concentration 1 g/l and pH 6)
The distilled water flux is proportional to the applied
pressure and distilled water permeability, Lp (l/h m2 Dye (C.I…) Flux (l/h m2) Retention (%)
bar) is defined as [19]:
Re=0 Re= 4100 Re=0 Re=4100
Jw
Lp = (1)
DP Direct red 80 27.1 41.1 99 100
Disperse blue 56 28.3 41.3 99.4 100
where, Jw is the permeate flux (1/h m2) and DP the Acid red 4 13.3 21.7 83 96.5
Basic blue 3 13.1 18.6 21.2 41
applied pressure (bar or 105Pa). Taking account the
osmotic pressure, Noel et al. applied Eq. (2) for salt
solution permeability Ls (1/h m2 bar):
Js
Ls = with DPe =DP −D^ and 4.2. Effect of hydraulic conditions on membrane
DPe
properties
D^ = ^(Cm)− ^(Cp) (2)

where, Js is the permeate flux (1/h m2), DPe the effective Table 3 shows the effect of stirring on the flux and
pressure (bar),  the osmotic pressure (bar). Cm and Cp the retention of four dye solutions.
are the concentration of the concentrated boundary and The measurements were performed after 20 min of
of the permeate, respectively. filtration. As shown by these data, stirring of the bulk
The osmotic pressure can be calculated according to solution at a velocity corresponding to a Reynolds
Van’t Hoff’s equation [13,18]: number of 4100 (turbulent regime) improved the dye
solution permeability and the retention. It should be
pointed out that for higher velocities, the fluxes de-
^=% i · R · T · C (3)
creased while the retention remained constant. This is
due to the formation of a vortex, which limits the
where, i is the number of ions per molecule of solute, R membrane surface area. For maximum stirring velocity,
the perfect gas constant, T the temperature (°K) and C the unfavourable concentration polarisation effect was
the concentration (mole/1). practically eliminated. All the following experiments
When concentration polarisation is limited by stir- were conducted in these conditions.
ring, (Cm) is close to (Cb) thus D^ = ^(Cb)−
^(Cp) ((Cb): concentration of bulk). 4.3. Fouling
The retention of dye solution is defined as:
Ab −Ap Fig. 2 shows the evolution of the permeability ratio
R% = 100 (4) and the retention as a function of time. The initial and
Ab
final distilled water permeability are also shown at
Ab and Ap are the absorbance of the bulk and time= 0 and 140 min, respectively.
permeate solutions, respectively.
The experiments were carried out at ambient temper-
ature (q). Fluxes are given at 25 °C after correction of
the influence of the water viscosity [23]:

J25 = Jq exp(0.0239(25−q)) (5)

The viscosity of the solutions is assumed to be equal


to the water viscosity at the same temperature.

4. Results

4.1. Initial water permeability

The membrane samples did not all have exactly the


same initial permeability, Lp varied from 3.6 to 4.5 l/h. Fig. 2. Permeability ratio (a) and retention (b) vs. time (DP=10 bar,
m2. bar. temperature 25 °C, concentration 1 g/l, Re= 4100 and pH 6).
606 A. Akbari et al. / Chemical Engineering and Processing 41 (2002) 601–609

Fig. 3. Dye exhaustion by the membrane (concentration 1 g/l, temper-


ature 25 °C).

A flux decrease is observed for all solutions (Fig. 2a)


which was not stabilised after 2 h of operation. This
decrease is attributed to the fouling of the membrane.
After rinsing with water, it is surprising to observe that Fig. 4. Influence of concentration on the permeability ratio (a) and
the water permeability was higher than the initial value retention (b) of dye solutions (DP= 10 bar, temperature 25 °C and
for the anionic dyes (direct red 80 and acid red 4). Re=4100).
On the other hand, the retention was stable, except
for the basic dye (Fig. 2b). The sharp drop in retention As previously observed, for all except the cationic
for the basic blue 3 can be correlated to the adsorption dye, there was a net decrease in flux versus time, while
of this dye onto the active polyamide surface of the retention remained constant.
membrane as shown in Fig. 3. Adsorption equilibrium
was reached in 2 h. 4.5. Influence of the concentration

4.4. Effect of pH on the permeability and retention of Since the purpose of nanofiltration is to concentrate
dye solutions the effluent, it was important to have an idea of the
influence of concentration on the performance. In
Table 4 shows the effect of pH on the retention and nanofiltration membrane separation, concentration
permeability of 6 dye solutions. It can be seen that the plays a significant role. In general, the higher the con-
Desal 5DK membrane retained direct red 80 and direct centration, the higher the osmotic pressure and conse-
yellow 8 completely at pH values of 3 and 6. For these quently the lower the permeate flux. For the solutions
dyes, the pH also had no influence on the permeability of direct red 80, acid red 4 and disperse blue 56, it can
ratio. The results were slightly different for the 2 other be seen in Fig. 4 that the permeate flux decreased with
anionic dyes (acid red 4 and acid orange 10), the the concentration increase and the flux was always
permeability ratio and; to a lesser degree, the retention higher at lower solution concentration.
improved when the pH increased from pH 3 to 6. The
last anionic dye was only studied at pH 6, and as for 4.6. Influence of the addition of sodium sulphate
the other anionic dyes, the retention was high.
The behaviour was quite the opposite for the cationic Preliminary studies were done to determine the per-
dye (basic blue), both permeability ratio and retention formance of the membrane in the treatment of solu-
decreased as the pH was increased from pH 3 to 6. tions composed of sodium sulphate and dyes. A direct

Table 4
Influence of pH on the permeability ratio and retention of dye solutions (DP= 10 bar, temperature 25 °C, Re=4100, concentration 1 g/l)

Dye (C.I…) Ls/Lp Retention (%)

pH 3 pH 6 pH 3 pH=6

t= 1 h t= 2 h t= 1 h t= 2 h t= 1 h t =2 h t =1 h t=2 h

Direct red 80 0.73 0.66 0.72 0.64 100 100 100 100
Direct yellow 8 0.51 0.44 0.66 0.49 99.3 100 100 100
Basic blue 3 0.52 0.38 0.21 0.18 97.3 96 91 75
Acid red 4 0.47 0.40 0.73 0.65 96.3 95.5 96.1 95.9
Acid orange 10 0.74 0.64 0.90 0.83 95.2 94.3 98.1 97.9
Reactive orange 16 – – 0.94 0.86 – – 95.6 95.8
A. Akbari et al. / Chemical Engineering and Processing 41 (2002) 601–609 607

dye and a reactive dye were chosen since they are the membrane; retention is controlled by a sieving
widely used to colour cellulose fibres and the process mechanism. However, strong interactions (ionic and
requires large amount of salt and causes serious envi- Hbonding) take place at the surface of the mem-
ronmental problems. The dye concentration was 1 g/l brane at both values of pH, this explains the de-
and the sodium sulphate concentration was 10 g/l. The crease in permeability versus time (Fig. 2a and
applied pressure was 10 bars for the all experiments and Table 4).
the Reynolds number was 4100. The results are pre- 2. For the three other anionic dyes (acid orange 10,
sented in Table 5. Dye retention was 100% in all cases acid red 4 and direct yellow 8) their MW is lower
for the direct dye while the retention of reactive dye and the sieving mechanism was slightly less impor-
was lower (96% or less). The retention of salt was tant than for the direct red 80. The pH effect begins
noticeably affected by the presence of the dye. The to be noticeable (Table 4), when the membrane
more the dye was retained, the less the salt was carries the same charge as the dye (pH 6) repulsion
retained. between the membrane and the solute explains the
slightly better performance in terms of permeability
and retention.
5. Discussion 3. For the cationic dye (basic blue 3) ionic interactions
play a major role in the transport properties as
In nanofiltration, there are two principal mechanisms shown in Fig. 2 and Table 4. At pH 6 there is a
of separation: strong ionic interaction between the membrane and
transport of uncharged solutes takes place by convec- the solute (Fig. 3). The initial retention of 100% can
tion due to a pressure difference and by diffusion due be explained by the binding of the dye to the
to a concentration gradient across the membrane. A membrane. With a progressive saturation of COO−
sieving mechanism is responsible for the retention. groups, retention began to decrease to reach 75%
For charged components electrostatic interactions after 2 h, which corresponds to the time needed to
take place between the components and the mem- reach adsorption equilibrium (Fig. 3). On the con-
brane charges. The transport of ions across the mem- trary, at pH 3, there is a repulsion between the
brane can be described in terms of diffusion and membrane charges and the cationic dye, retention
migration, as a result of concentration and electrical increased from 75% at pH 6 to 96% at pH 3, while
potential gradients, as well as convection caused by the flux doubled (Table 4).
transmembrane pressure. 4. In the presence of salt an interesting result was the
As mentioned in the introduction, the polyamide decrease in rejection (Table 5) for the salt and the
membrane has a cut-off around 150– 300 and carries dye when they were mixed compared with the rejec-
negative charges (COO−) at pH 6 and positive charges tion observed when they were alone (excepted for
at pH 3 (NH+ 3 ). Based on that and on the schematic the high molecular weight direct red 80, always
dyeing mechanisms briefly recalled in the introduction, totally retained by a sieving mechanism). The reason
most of the data obtained in this paper can be qualita- for this decrease is that, in the presence of a high
tively explained. salt concentration, Donnan exclusion becomes less
1. For high molecular weight (1373 g/mole) direct red effective and the anionic species are repelled less.
80, retention was 100% in all cases whatever the 5. A notable result concerns the apparent increase in
conditions, the size of the molecule is larger than the permeability after nanofiltration of anionic dyes
pore size of the membrane and it cannot enter into (Fig. 2a, values at 140 min and Table 5, last line).

Table 5
Permeability ratio and retention of the membrane with the solution containing dye and Na2SO4 (DP= 10 bar, temperature 25 °C, Re=4100 and
pH 6)

Solution Ls/Lp %R (Na2SO4) %R (Dye) D^ (bar)

t= 1 h t =2 h t =1 h t= 2 h t= 1 h t=2 h t=1 h

Distilled water before NF 1.00 – – – – – –


10 g/l Na2SO4 0.88 0.70 96.8 96.1 – – 4.67
1 g/l Direct red 80 0.85 0.76 – – 100 100 0.13
1 g/l Direct red 80+10 g/l Na2SO4 0.79 0.65 78.6 72.8 100 100 2.91
1 g/l Reactive orange 16 0.86 0.78 – – 95.6 95.8 0.11
1 g/l Reactive orange 16+10 g/l Na2SO4 0.79 0.69 88.2 87.2 86.6 83.0 3.47
Distilled water after NF 1.10 – – – – – –
608 A. Akbari et al. / Chemical Engineering and Processing 41 (2002) 601–609

We may assume that this is due to the charges of the toral grant from the government of Iran and Société
dyes adsorbed at the membrane surface, this proba- Française d’Exporatation des Ressources Educatives
bly increases the hydrophilicity of the skin layer of (SFERE).
the membrane. However, this effect is not perma-
nent and after a longer cleaning in water, the initial
permeability is restored. Appendix A. Nomenclature
6. A more general comment concerns the fouling prob-
lem and the performance in terms of flux. Although Ab absorbance of the bulk solution
the experiments were run in a turbulent flow regime Ap absorbance of the permeate
(low concentration polarisation), fouling was heavy C concentration (mole/l)
and the values of the permeability ratio, after 2 h of Cb concentration of bulk (mole/l)
operation, were often between 0.6 and 0.8 which Cm concentration at the membrane surface (mole/l)
corresponds to a permeate flux of only 25– 35 l/h Cp concentration of the permeate (mole/l)
per m2 under a pressure of 10 bar. These values will i number of ions per molecule of solute
certainly be lower in the long term and full-scale Js permeate flux of salt and dye solutions (l/h m2)
operation with conventionally spiral wound mod- Jw permeate flux of distilled water (l/h m2)
ules. The fouling is more or less reversible as a Lp distilled water permeability (l/h m2 bar)
function of the class of the dye and of the molecular Ls permeability of salt and dye solutions (l/h m2
weight. Qualitatively, it could be expected that the bar)
reversibility decreases in the following order, direct Pe effective pressure (bar or 105 Pa)
rouge 80Bdirect yellow 8Bacid red 4B acid or- R perfect gas constant
ange 10Bdisperse blue 56B reactive orange 16 T temperature (°K)
basic blue 3. Since the fouling is mainly due to the
amphoteric nature of the polyamide skin layer, the Greek letters
development of less fouling-sensitive nanofiltration D Difference
membranes will be necessary to improve the eco-  osmotic pressure
nomics of the process. q temperature (°C)

6. Conclusions

The present work investigates the performance of a References


polyamide nanofiltration membrane in treating
[1] O. Marmargne, C. Coste, Colour removal from textile plant
coloured textile effluent. For anionic dyes (acid red 4, effluents, American Dye. Rep., (1996) 15 – 21.
acid orange 10, direct red 80, direct yellow 8, and [2] A. Rozzi, F. Malpei, L. Bonomo, R. Bianchi, Textile wastewater
reactive orange 16), the membrane generally showed reuse in northern Italy (COMO), Water Sci. Tech. 39 (1999)
acceptable rejection due mainly to its relatively low 121 – 128.
cut-off and the cationic dyes were more than 95% [3] A. Rozzi, G. Bergna, C. Zaffaroni, Micro and nanofiltration of
secondary textile/domestic effluents for reuse. Proceedings Con-
retained whatever the range of pH and concentration ference New Developments in Membrane and Filtration System
used. In particular, direct red 80 and direct yellow 8 Hilton Head (SC, USA), February 1996, pp. 14 – 16.
were 100% retained and produced a permeate suitable [4] A. Rozzi, M. Antonelli, M. Arcari, Membrane treatment of
for reuse. However, the membrane suffered from flux secondary textile effluents for direct reuse, Water Sci. Tech. 40
decline due to its sensitivity to fouling. In the presence (1999) 409 – 416.
[5] J.J. Porter, Membrane filtration techniques used for recovery of
of salt, the osmotic pressure resulted in a further de- dyes, chemicals and energy, Tex. Chemist Colorist 22 (1999)
crease in flux. 21 – 25.
For the cationic dye tested (basic blue 3), the influ- [6] J. Sojka-Ledakowicz, T. Koprowski, W. Machnowski, H.H.
ence of the Donnan exclusion phenomena was clearly Knusdsen, Membrane filtration of textile dye-house wastewater
observed and the membrane was not suitable for such for technological water reuse, Desalination 119 (1998) 1 –10.
[7] L. Bonomo, R. Bianchi, C. Capra, V. Mezzanotte, A. Rozzi,
an application. Nanofiltration and reverse osmosis treatment of textile dye
With these conclusions in mind, we now intend to effluents, Tech. Innovantes en Epuration des Eaux 6 (1992)
develop a membrane that is less sensitive to fouling. 327 – 336.
[8] N.K. Majewska, T. Winnicki, J. Wisniewski, Effects of flow
conditions on ultrafiltration efficiency of dye solutions and tex-
tile effluents, Desalination 71 (1989) 127 – 135.
Acknowledgements [9] T.H. Liu, K.M. Simms, S.A. Zaidi, Selection of ultrafiltration
nanofiltration membrane for treatment of textile dyeing wastew-
This study has been financially supported by a doc- ater, Water Treat. 9 (1994) 189 – 198.
A. Akbari et al. / Chemical Engineering and Processing 41 (2002) 601–609 609

[10] K.M. Nowak, Synthesis and properties of polysulfone mem- weight dye, Desalination 129 (2000) 237 – 245.
branes, Desalination 71 (1989) 83 –95. [17] G. Chen, X. Chai, P. Yue, Y. Mi, Treatment of textile desiz-
[11] K. Treffry-Goatley, C.A. Buckley, G.R. Grove, Reverse osmo- ing wastewater by pilot scale nanofiltration membrane separa-
sis treatment and reuse of textile dyehouse effuents, Desalina- tion, J. Membr. Sci. 127 (1997) 93 – 99.
tion 47 (1983) 313 – 320. [18] Y. Xu, R.E. Lebrun, Treatment of textile dye plant effluent by
[12] C.A. Buckley, Membrane technology for the treatment of dye- nanofiltration membrane, Sep. Sci. Tech. 34 (1999) 2501 –2519.
house effluents, Water Sci. Tech. 25 (1992) 203 –209. [19] C. Combe, Estimation de la sélectivité en nanofiltration á par-
[13] I.M. Noel, R. Lebrun, C.R. Bouchard, Electro-nanofiltration tir des propriétés du matériau membranaire, Doctoral thesis,
of a textile direct dye solution, Desalination 129 (2000) 125 – Université Paul Sabatier, Toulouse (France) 1996.
136. [20] T. Vickerstaff, The Physical Chemistry of Dyeing, second ed.,
[14] J. Ratana, S. Anawat, L. Piyanoot, Performance evaluation of Oliver & Boyd, London, 1954.
nanofiltration membrane for treatment of effluents containing [21] Colour index international, Society of Dyers and Colourists,
reactive dyes and salt, Desalination 130 (2000) 177 –183. third ed., 1999.
[15] J.C. Watters, E. Biagtan, O. Senler, Ultrafiltration of a textile [22] E.R. Trotman, The Dyeing and Chemical Technology of Tex-
plant effluent, Sep. Sci. Tech. 26 (1991) 1295 –1313. tile Fibres, fourth ed., Griffin, London, 1970.
[16] X. Xu, J.L. Gaddis, H.G. Spencer, Dynamic formation of a [23] R.C. Weast, M.J. Astle, W.H. Beyer, CRS Handbook of
self-rejecting membrane by nanofiltration of a high formula- Chemistry and Physics 78th ed., Boca Raton (USA), 1997.

You might also like