9
Treatment by nanofiltration and reverse osmosis of high salinity drilling water for seafood washing and processing Khaled Walha a* , Raja Ben Amar a , Francis Quemeneur b , Pascal Jaouen b a Laboratoire des Sciences de Matériaux et Environnement, Faculté des Sciences de Sfax, BP 759, Sfax, Tunisie Tel. +216 (74) 276 400; Fax: +216 (74) 274 473; email: [email protected] b Laboratoire de Génie des Procédés Environnement-Agroalimentaire (Université de Nantes), GEPEA-UMR-CNRS 6144, Nantes Atlantique Universités, BP 406-44602, Saint Nazaire, France Received 18 November 2006; Accepted 11 May 2007 Abstract The Calembo Company (seafood conditioning, Sfax, Tunisia) uses 50 m 3 /d of drilling water per ton of fish to condition cuttlefish before freezing. The water used is characterised by its high degree of hardness, high sulphate concentration and salinity (67 g.l !1 ). Consequently, generated effluents contain a large amount of salts and a high organic load (COD of 10–20 g.l !1 ). The use of water with high salinity entails difficulties regarding the biological treatment effluents. A membrane bioreactor (MBR) is effective to treat cuttlefish effluent when salinity is lower than 35 g.l !1 . In our case, to be in the application range of such MBR process, it is necessary to decrease the salinity of the drilling water. Recently, nanofiltration (NF) membranes have been studied as pre-treatment unit operations both in thermal and membrane seawater desalination processes. In this work, drilling water will be handled using: (1) ultrafiltration and nanofiltration, as a pre-treatment step before desalination process. Three different commercial UF–NF membranes MT03, MT44 and XP117 (cut-offs ranging between 200 to 4000 Dalton) are tested using a tangential flow filtration cell, and (2) high pressure reverse osmosis (the osmotic pressure of drilling water is about 55 bars) to desalt partially the drilling water. Results showed that the MT03 membrane was effective for the removal of natural organic matter and reducing the sulphate concentration in drilling water. Nanofiltration makes possible to have a standardization of the water quality (raw water can vary according to the season and the drilling place). The treatment by RO at high pressure ($70 bars) allows to reduce significantly the salinity of drilling water (permeate salinity about 2 g.l !1 ). Mixing RO permeate and cuttlefish conditioning effluent together could lead to reduce significantly the salinity of the effluent in view of a biological treatment with a membrane bioreactor (35 g/l or less is required). *Corresponding author. Desalination 219 (2008) 231–239 Keywords: Ultrafiltration; Nanofiltration; High pressure reverse osmosis; Drilling water; Pre-treatment doi:10.1016/j.desal.2007.05.016 0011-9164/08/$– See front matter © 2008 Published by Elsevier B.V.

Treatment by nanofiltration and reverse osmosis of high salinity drilling water for seafood washing and processing Abstract

Embed Size (px)

Citation preview

Treatment by nanofiltration and reverse osmosis of high salinitydrilling water for seafood washing and processing

Khaled Walhaa*, Raja Ben Amara, Francis Quemeneurb, Pascal Jaouenb

aLaboratoire des Sciences de Matériaux et Environnement, Faculté des Sciences de Sfax, BP 759, Sfax, TunisieTel. +216 (74) 276 400; Fax: +216 (74) 274 473; email: [email protected]

bLaboratoire de Génie des Procédés Environnement-Agroalimentaire (Université de Nantes),GEPEA-UMR-CNRS 6144, Nantes Atlantique Universités, BP 406-44602, Saint Nazaire, France

Received 18 November 2006; Accepted 11 May 2007

Abstract

The Calembo Company (seafood conditioning, Sfax, Tunisia) uses 50 m3/d of drilling water per ton of fish tocondition cuttlefish before freezing. The water used is characterised by its high degree of hardness, high sulphateconcentration and salinity (67 g.l!1). Consequently, generated effluents contain a large amount of salts and a highorganic load (COD of 10–20 g.l!1). The use of water with high salinity entails difficulties regarding the biologicaltreatment effluents. A membrane bioreactor (MBR) is effective to treat cuttlefish effluent when salinity is lower than35 g.l!1. In our case, to be in the application range of such MBR process, it is necessary to decrease the salinity ofthe drilling water. Recently, nanofiltration (NF) membranes have been studied as pre-treatment unit operations bothin thermal and membrane seawater desalination processes. In this work, drilling water will be handled using:(1) ultrafiltration and nanofiltration, as a pre-treatment step before desalination process. Three different commercialUF–NF membranes MT03, MT44 and XP117 (cut-offs ranging between 200 to 4000 Dalton) are tested using atangential flow filtration cell, and (2) high pressure reverse osmosis (the osmotic pressure of drilling water is about55 bars) to desalt partially the drilling water. Results showed that the MT03 membrane was effective for the removalof natural organic matter and reducing the sulphate concentration in drilling water. Nanofiltration makes possibleto have a standardization of the water quality (raw water can vary according to the season and the drilling place). Thetreatment by RO at high pressure ($70 bars) allows to reduce significantly the salinity of drilling water (permeatesalinity about 2 g.l!1). Mixing RO permeate and cuttlefish conditioning effluent together could lead to reducesignificantly the salinity of the effluent in view of a biological treatment with a membrane bioreactor (35 g/l or lessis required).

*Corresponding author.

Desalination 219 (2008) 231–239

Keywords: Ultrafiltration; Nanofiltration; High pressure reverse osmosis; Drilling water; Pre-treatment

doi:10.1016/j.desal.2007.05.016

0011-9164/08/$– See front matter © 2008 Published by Elsevier B.V.

K. Walha et al. / Desalination 219 (2008) 231–239232

1. IntroductionDuring the last decades, seawater has became

an important source of fresh water [1] because ofthe changing weather patterns, increased indus-trialisation and tendency in recent years for theworld swelling population to dwell in areas wherelocal supplies of high quality fresh water are lessthan adequate. Drinking water is prepared fromseawater using reverse osmosis (RO) or multi-stage distillation processes [2]. Seawater is char-acterised by having high degree of hardness,microrganisms contents and high total dissolvedsolids (TDS). These properties lead to majorproblems such as salt precipitation, membranefouling, high-energy requirements and the neces-sity to use high quality construction materials.

To solve seawater desalination problems andto minimise their effects on productivity andwater costs of conventional plants, ultrafiltration(UF) and nanofiltration (NF) membranes havebeen evaluated and employed in pre-treatmentfacilities for both RO and thermal processes[3–7]. NF is a type of pressure-driven membranethat has properties between those of UF and RO.NF membranes have the advantages of providinga high water flux at low operating pressure (10–25 bars) while maintaining a high organic matterrejection [8]. The NF process benefits from easeof operation, reliability and comparatively lowenergy consumption as well as high efficiency ofpollutant removal [9]. Several research projectsinvestigated natural organic matter (NOM)removal by nanofiltration [10]. Almost completeNOM rejection (>90%) could be achieved withNF membranes with molecular weight cut off<300 g/mol [11–13]. NF may constitute a preli-minary treatment for seawater. In addition, thecorrect choice on NF membrane is of majorimportance for the pre-treatment of the drillingwater, which will make or break the economicalfeasibility of the whole process.

The processing industry of seafood products isundergoing a great development in seaside

countries nowadays. In Tunisia, the main activityin this domain consists in freezing products withhigh trading values destined for export. Theregion of Sfax in the south of Tunisia has actuallyabout 30 units where processed seafoods mainlyconsist of the kingly prawn (Penaeus kerathurus),cuttle-fish (Sepia officinalis), octopus (Octopusvulgaris) and calamary (Loligo vulgaris). Theaverage daily water consumption for this industryis estimated in the region of Sfax, Tunisia, toabout 1000 m3/d.

The Calembo Company (our industrial partnerin this study) is situated in the side of the sea anduses 50 m3/d per ton to condition cuttlefish beforefreezing [14,15]. The transformation processingconsists in washing, draining and peeling thecuttlefish which is then deep-freezed beforeexportation. The drilling water used for this pur-pose is characterised by its high degree of hard-ness, 1534EF (1EF = 10 mg.l!1 of CaCO3), highsulphate concentration and salinity (67 g.l!1).Consequently, generated effluents contain a largeamount of salts as well as a high organic load(COD about 10–20 g.l!1).

Salinity has a beneficial effect on the seafoodconservation (the seawater inflates the cuttlefishby osmotic chock), but has difficulties regardingthe further biological treatment of effluents [16].The regulation on wastewaters treatment isbecoming more and more restrictive since itinvolves an urgent demand for appropriate andefficient treatment for such effluents. Grelier etal. [17] showed that the treatment of cuttlefisheffluent (COD .3 g.l!1, salinity .35 g.l!1) usinga bioreactor membrane, Biosep®, was effective. Inour case, to fit in the applicability range of sucha MBR process, it is necessary to reduce thesalinity of drilling water. This can be carried outby using high pressure reverse osmosis (theosmotic pressure of drilling water being close to55 bars). The drilling water being rather rich insulphates, a phenomenon of CaSO4 precipitationcan take place during the desalination step; thus

K. Walha et al. / Desalination 219 (2008) 231–239 233

it is very necessary to reduce the sulphateconcentration.

In this paper, in the framework of theexperimental work, drilling water will be handledusing:C UF and NF as a pre-treatment step of the

desalination process, for reducing, on onehand, NOM, which can be presented in thedrilling water; and on the other hand, thesulphate concentration. Three different com-mercial membranes — MT03, MT44 andXP117 — (cut-offs ranging between 200 to4000 Dalton) were tested using tangentialfiltration cell.

C Reverse osmosis (constant concentration) todesalt partially the drilling water.

2. Materials and methods

2.1. Saline water

The experimental work consisted in evaluatingperformances of the UF and NF processes (com-

Table 1Mineral composition of the raw water

Element

pH 7.2Salinity (g.l!1) 67TOC (mg.l!1) 3.86Hardness (EF) 1534

Concentration mg.l!1 éq g.m!3

Ca2+ 1,664 83.2Mg2+ 2,718 223.7Na+ 22,245 967.17K+ 200 5.12HCO3

! 240 3.93Cl! 40,046 1128.05SO4

2! 5,510 114.8Σ anions 26,327 1246.7Σ cations 45,796 1279.7

parison of three membranes) with an applicationto a prefiltered (0.2 µm) drilling water providedby Calembo (Sfax, Tunisia). The composition ofthe aforementioned water is given in Table 1. Theconcerned water is characterised by having highdegree of hardness and salinity. Analyses resultsshow that the ionic balance is validated (differ-ence 2.6%).

2.2. Apparatus

UF and NF experiments were performed witha Microlab 40 plant (Fig. 1) equipped with atubular membrane and was capable of operatingunder pressure from 10 to 40 bars. In RO experi-ments, high pressure ($70 bars) is required toovercome the high osmotic pressure of the feedwater (~55 bars).

2.3. Membranes

Three organic tubular membranes (MT03,MT44 and XP117) provided by PCI were testedfor UF and NF experiments. The effective mem-brane area was 0.033 m2 (length 0.82 m and

Fig. 1. NF pilot plant for the experiments.

K. Walha et al. / Desalination 219 (2008) 231–239234

Table 2Characteristics of the tested membranes

UF NF RO

XP117 MT44 MT03

Cut-off (Dalton)Pure water permeability (l.h!1.m!2); P = 20 bars, T = 25EC)Materiala

Recommended operating pressure (bar)Recommended pH rangeMaximum temperature (EC)

4000320PES302–1180

1000–4000185PES202–1260

200115PA/PES303–960

<20040PA/PES1202–1040

aPES, polyethersulfone; PA, polyamide.

0.0127 m diameter). Furthermore, the RO mem-brane surface was 1.2 m2. The conditions underwhich the membranes can operate and also someproperties of the studied membranes (data frommanufacturer) are presented in Table 2.

2.4. Methods for membrane testing

Before being used in any experimental work,each membrane was cleaned by means of stan-dard procedures to remove preservatives andrinsed with deionised water MilliQ (Millipore,France) until conductivity of the permeateremained below 1 µS.cm!1.

In UF and NF experiments the tangentialvelocity (U) of the solution over the membranesurface was 2.5 m.s!1. The temperature was main-tained at 25EC by an automatic temperature con-troller and a heat exchanger in the recirculationloop. The filtration experiments were carried outin total recirculation mode, i.e. both the concen-trate and the permeate streams were recirculatedinto the feed tank, so that the feed concentrationis kept approximately constant. When the equili-brium was reached (after 120 min), experimentswere conducted in a concentration mode: theconcentrate of the initial step was circulated backto the feed tank, thereby increasing salt concen-tration while permeate was collected in anotherreservoir. The experiments were stopped when

the VRF was equal to 3. The RO experiments(constant concentration) were carried on drillingwater, 25°C feed water temperature.

2.5. Calculation

The volumetric balance can be written as:

Vf = Vp + Vr (1)

where Vf is the feed volume at the beginning (inl), and Vr and Vp are respectively the concentrate(feed) and the permeate volumes (in l) during aconcentration operation.

In concentration mode, the molar (or mass)balance on a solute can be expressed as:

Vf Cf = Vp Cp + Vr Cr (2)

The water permeability, WP (l.h!1. m!2. bar!1) wascalculated as:

WP = Vp /(t.A.P) (3)

where A is the effective membrane area (m2), t istime (h), and P is the applied pressure (bar).

The volume reduction factor (VRF) was cal-culated using the following equation:

VRF = Vf /Vr (4)

K. Walha et al. / Desalination 219 (2008) 231–239 235

The instantaneous retention R of ions and pollu-tants was calculated from the analysis of samplestaken at specific VRF values according to thefollowing equation:

Rinstantaneous (%) = 100 × (1!Cp/Cr) (5)

where Cp and Cr are the instantaneous concen-trations (mg.l!1) at specific VRF in the permeateand the bulk solution (concentrate) respectively.

2.6. Analytical procedures

The analytical procedures utilized in this studywere those recommended by Afnor [18]. Con-centrations of Na+, Ca2+, Mg2+, Cl! and SO4

2! weredetermined by ionic chromatography with DionexDX 120 equipped with a CS 15 column forcations and AS 12 A column for anions. The totalorganic carbon (TOC) was measured by a TOCmeter (Shimadzu TOC-5000 A).

3. Results and discussion

3.1. Results obtained by UF and NF

3.1.1. Variation of permeate flux with pressureat constant concentration

Fig. 2 shows that for all membranes the per-meate flux was found to increase linearly with thetransmembrane pressure (verification of Darcy’slaw) with all curves passing through the origin inaccordance with the null value of the pressure.The water flux for XP117 membranes was thehighest among the studied membranes under allpressures while MT03 membrane bears the lowestflux. The water permeabilities as calculated fromEq. (3) were 11.18, 7.27 and 5.59 l. h!1.m!2.bar!1

for the XP117, MT44 and MT03 membranes,respectively. These results can be related to themembranes characteristics. For example, XP 117has relatively the highest molecular weight cut-off and was thus expected to have high water

Fig. 2. Variation of permeate flux vs. transmembranepressure (constant concentration, i.e., total recycling).

permeability. On the other hand, MT03 has thelowest permeability for its lower molecularweight cut-off.

3.1.2. Variation of permeate flux vs. time atconcentration mode

Experiments were stopped (final concen-tration) when the VRF was equal to 3; from 2.4 Lof feed (drilling-water), we obtained 1.6 L of per-meate and 0.8 L of concentrate. The operatingtime was dictated by the dead volume of theequipment (0.8 L). Fig. 3 shows the variation ofthe concentration flux which decreases veryweakly during the filtration test of drilling waterand using the three UF and NF membranes.Experiment duration was 13, 20 and 26 min withthe XP117, MT44 and MT03 membranes respec-tively. This is in agreement with the water per-meabilities of the membranes. After treatment byfiltration process, all the membranes were washedwith demineralized water. Results showed thatthe rinsed membrane had approximatively thesame initial permeability. Thus, the irreversiblefouling phenomenon can be neglected.

3.1.3. Removal of NOM and pollutants (con-centration mode)

Analyses of TOC performed on the initialfeed, permeate and concentrate at the end ofconcentration operation (VRF = 3) are put

K. Walha et al. / Desalination 219 (2008) 231–239236

together in Table 3. Organic materials rejectionwith MT03 membrane is much higher (about89%) than with MT44 (72 %) and XP117 ones(65%). From a strict de-pollution point of view,considering values of permeate, UF and NF aresuitable for the removal of NOM from drillingwater.

Fig. 3. Variation of permeate flux vs. time (P = 20 bars,T = 25±C, VRF: 1–3).

Table 3Concentrations and rejection rates, R (%) of TOC (P =20 bars, T = 25EC, VRF = 3)

Concentration TOC (mg.l!1) R (%)

Membrane Feedwater

Permeate Concentrate

MT03 3.86 1 9.1 89MT44 3.86 2.07 7.6 72XP117 3.86 2.31 6.6 65

Table 4Rejection rates, R (%) of cations and anions (P = 20 bars,T =25EC, drilling water)

Ions XP117 MT44 MT03

Ca2+ 6.1 6.3 14Mg2+ 6.5 11.9 28Na+ 4.3 5.3 10.8Cl! 3.2 5.3 9.8SO4

2! 23.7 30 61.4

Table 4 shows that the higher rejection of ionsis obtained with the MT03 membrane, whileXP117 gives the lower rejection. If we comparethe retentions observed for Na+, Ca2+ and Mg2+

using different membranes, we can observe thatHomeister series order was respected: Mg2+ >Ca2+ > Na+. This observation is not surprising ifwe compare monovalent and divalent cationsconcentrations. The initial concentrations in cal-cium and magnesium in feed water are different(see Table 1) and in terms of rejection magnesiumis always better retained than calcium due to itshigher hydration energy. The hydration energieswere 1584 and 1921 kJ.mol!1 for Ca2+ and Mg2+

respectively [19]. In addition, ionic strengtheffects could also play a role by screening theelectrokinetic effects occurring under dilute con-ditions. In the case of anions, the difference indiffusion coefficient between ions determinedtheir different retention. Ions with the lowerdiffusion coefficient (1.06×10!9 m2.s!1 for SO4

2!)characterised the higher retention, whereas ions

Table 5Verification of mass balance for TOC (P = 20 bars,T =25EC, drilling water)

Membrane CfVf (mg)

CrVr + CpVp(mg)

Error(%)

MT03 9.26 = 3.86×Vf 8.88 !4.2MT44 9.26 9.39 1.4XP117 9.26 8.98 !3

Table 6Evolution of rejection rate, R (%) vs. transmembranepressure (drilling water, MT03 membrane, VRF = 3)

Ions 20 bars 30 bars

Ca2+ 14 18.1Mg2+ 28 35.4Na+ 10.8 17.4Cl! 9.8 17.7SO4

2! 61.4 70

K. Walha et al. / Desalination 219 (2008) 231–239 237

with the higher diffusion coefficient (2.03×10!9

m2.s!1 for Cl!) had the lowest retention [20]. Sul-phate ions are better retained than the other ions,which was one of our objectives.

3.1.4. Verification of mass balanceThe calculation of the CfVf values and the sum

of CpVp and CrVr allows the verification of themass balance. Table 5 shows that the mass bal-ance for TOC was validated for all membrane(the maximum error is 4.2%). The same resultswere obtained for all ions.

3.1.5. Effect of pressure on salt rejectionTable 6 shows that the rejection of both

cations and anions always increases withincreasing pressure for MT03 membrane, therebyindicating that the transfer mechanism is mostlydiffusional [21,22]. It is also noticed that reten-tion rates of sodium and chlorides are weak, butincrease regularly with increasing pressure.

Results showed that the MT03 membrane wasefficient for the removal of NOM and reducingthe sulphate concentration in drilling water. Thetreatment by nanofiltration makes possible to geta standardization of the water quality (feed watercan vary according to the season and the drillingplace).

3.2. Results obtained by RO

The critical driving pressure — pressure limitcorresponding to the first permeate flux measur-able — is observed at 55 bars (Fig. 4), which is inaccordance with the theoretical value calculatedfrom the osmotic pressure from the Van’t Hofflaw with a value of 54 bars at 25EC. As shown inTable 7, the salinity of permeate depends on thetransmembrane pressure; it decreases with theincreasing applied pressure. This is due to theincrease of retention rates for all ions with thepressure. The treatment by RO at high pressureallows to reduce significantly the salinity of

Fig. 4. Evolution of permeate flux vs. transmembranepressure. (Drilling water at 67 g.l!1, RO membrane madefrom polyamide–polyethersulfone, T = 25EC).

Table 7Evolution of the permeate salinity vs. transmembranepressure (drilling water at 67 g.l!1, RO membrane, 25EC,constant concentration)

Pressure (bars) 60 70 80Salinity of permeate (g.l!1) 8.22 4.11 2.39

drilling water. At 80 bars, the salinity of permeatecloses 2.5 g.l!1 (the retention R of ions is greaterthan 95%). Mixing RO permeate and cuttlefishconditioning effluent together allows the signi-ficant reduction of the effluent salinity, in view ofa biological treatment with a membrane bio-reactor (35 g.l!1 or less is required).

4. Conclusions

The drilling water used in the region of Sfax,Tunisia, to condition cuttlefish before freezing ischaracterised by a high salinity (67g.l!1) and isparticularly rich in sulphates. To determine theirsuitability as a pre-treatment process before de-salination, three commercial UF and NF mem-

K. Walha et al. / Desalination 219 (2008) 231–239238

branes (MT03, MT44, XP117) were evaluated inthe framework of a laboratory-scale study. Thepreliminary study leads to encouraging results.Indeed it has been shown that the MT03 NFmembrane has high rejection and rather lowpermeation flux while the XP117 UF membranehas a lower salt rejection with better permeationflux. The MT03 membrane is thus suitable for theremoval of NOM and sulphates concentrationreduction in the considered drilling water. Fur-thermore, the use of NF as a pre-treatmentprocess allows to standardise the water quality(water composition can vary according to theseason and drilling place).

Treatment by RO at high pressure (70 bars)allows to reduce significantly the salinity of thedrilling water (permeate salinity around 2.5 g.l!1).Mixing RO permeate and cuttlefish conditioningeffluent together can significantly reduce thesalinity of the effluent in view of biological treat-ment with a membrane bioreactor (35 g.l!1or lessis required).

Acknowledgements

This work has been accomplished as part of aFranco-Tunisian research project CMCU 2004(04PRE01) “Gestion des ressouces en eaux nonconventionnelles. Procédés de traitement pourpotabilisation et réutilisation”. We wish to thankMrs. Maryse Chaplain-Derouiniot, Centre ofResearch and Transfer of Technologies (CRTT),Saint Nazaire, France, for assistance in theexperiments.

References

[1] B.V. Brugen and C. Vandecasteele, Distillation vs.membrane filtration: Overview of process evolutionsin seawater desalination, Desalination, 143 (2002)207–218.

[2] M. Al-Sofi, A.M. Hassan, O.A. Hamed, A.G.I.Dalvi, M.N.M. Kither, G. Mustafa and K. Bamar-

douf, Optimization of hybridized seawater desali-nation process, Desalination, 131 (1998) 147–156.

[3] A.M. Hassan, M. Al-Sofi, A. Al-Amoudi, A. Jama-luddin, A. Farooque, A. Rowaili, A.G.I Dalvi, M.N.Kither, G. Mustafa and I. Al-Tisanand, A newapproach to thermal seawater desalination processusing nanofiltration membranes (Part 1), Desali-nation, 118 (1998) 35–51.

[4] M. Al-Sofi, A.M. Hassan, A.G.I. Dalvi, G. Mustafaand M.N. Kither, Nanofiltration as a means ofachieving higher TBT of >120°C in MSF, Desali-nation, 118 (1998) 123–129.

[5] M. Al-Sofi, Seawater desalination — SWCCexperience and vision, Desalination, 135 (2001) 121–139.

[6] A.M. Hassan, A. Farooque, A. Jamaluddin, A. Al-Amoudi, M. Al-Sofi, A. Al-Rubaian, M.N. Kither,I. Al-Tisan and A. Rowaili, A demonstration plantbased on the new NF-SWRO process, Desalination,131 (2000) 157–171.

[7] A. Criscuoli and E. Drioli, Energetic and exergeticanalysis of an integrated membrane desalinationsystem, Desalination, 124 (1999) 243–249.

[8] L.P. Raman, M. Cheryan and N. Rajagopalan, Con-sider nanofiltration for membrane separation, Chem.Eng. Prog., 90 (1994) 68–74.

[9] P. Erikson, Nanofiltration extends the range of mem-brane filtration, Environ. Prog., 7 (1988) 58–61.

[10] A. Gorenflo, D. Velazquez Padron and F.H.Frimmel, Nanofiltration of a German groundwater ofhigh hardness and NOM content: Performance andcosts, Desalination, 151 (2002) 253–265.

[11] P. Fu, H. Ruiz, J. Lozier, K. Thompson and C.Spangenberg, A pilot study on groundwater naturalorganics removal by low-pressure membranes,Desalination, 102 (1995) 47–56.

[12] E. Ivrine, D. Welch, A. Smith and T. Rachwal,Nanofiltration for colour removal — 8 yearsoperational experience in Scotland, Proc. Membranesin Drinking Water and Industrial Water Production,Paris, 1 (2000) 247–255.

[13] T. Thorsen, A. Harz and H. Ødegaard, Influence ofraw water characteristics and membrane pore size onthe performance of ultra-filters for NOM (humicsubstances) removal, Water Supply, 16 (1998) 497–507.

[14] I. Aloulou, K. Walha, R. Ben Amar, P. Jaouen and

K. Walha et al. / Desalination 219 (2008) 231–239 239

F. Quéméneur, Etude préliminaire du traitementd’effluents contenant de l’encre de seiche parcentri-fugation et procédés à membrane, RevueSci. l’Eau, 19(4) (2006) 383–392.

[15] E. Ferjani, E. Ellouze and R. Ben Amar, Treatmentof seafood processing wastewaters by ultrafiltration-nanofiltration cellulose acetate membrane, Desali-nation, 177 (2005) 43–49.

[16] S. Sadok, A. Abdelmouleh and A. El Abed, Com-bined effect of sepia soaking and temperature on theshelf life of peeled shrimp Peneaus kerathurus, FoodChem., 88 (2004) 115–122.

[17] P. Grelier and A. Tazi-Pain, Traitement des eaux delavage de seiches par le procédé Biosep®, 2ème Ecoled’été UNESCO sur les procédés membranaires,“Traitement et réutilisation des eaux, apport destechniques membranaires”, Monastir, Tunisie, 2005.

[18] Afnor (Agence Française de Normalisation), Qualitéde l’eau, Recueil des normes françaises: 2nd ed.,Vol. 3, Paris, 1997.

[19] M. Pontié, A. Lhassani, C.K. Diawara, A. Elana,C. Innocent, D. Aureau, M. Rumeau, J.P. Croué,H. Buisson and P. Hemery, Seawater nanofiltrationfor elaboration of usable salty waters, Desalination,167 (2004) 347–355.

[20] E. Wasik, J. Bohdziewicz and K. Cwiklak, Ionbalance in NF-treated water, Desalination, 186(2005) 81–87.

[21] K. Mehiguene, Y. Garba, S. Taha, N. Gondrexon andG. Dorange, Influence of operating conditions on theretention of copper and cadmium in aqueous solu-tions by nanofiltration: experimental results andmodeling, Sep. Purif. Technol., 15 (1999) 181–189.

[22] E.Sh. Mohamed, G. Papadakis, E. Mathioulakis andV. Belessiotis, An experimental comparative study ofthe technical and economic performance of a smallreverse osmosis desalination system equipped withan hydraulic energy recovery unit, Desalination, 194(2006) 239–250.