Desalination 157 (2003) 73–80

Combination of physico-chemical treatment and nanofiltration
to reuse wastewater of a printing, dyeing and finishing textile
industry
A. Bes-Piá*, J.A. Mendoza-Roca, M.I. Alcaina-Miranda, A. Iborra-Clar,
M.I. Iborra-Clar
Department of Chemical and Nuclear Engineering, Universidad Politécnica of Valencia,
Camino de Vera s/n, 46071 Valencia, Spain
Tel. +34 (96) 3879633; Fax +34 (96) 3877639; email: mbespia@iqn.upv.es

Received 23 December 2002; accepted 30 December 2002

Abstract
The main goal of this work was to study the feasibility of the combination of physico-chemical treatment with
nanofiltration to reuse wastewater of a printing, dyeing and finishing textile industry. For the physico-chemical
treatment two coagulants (one containing Al3+ and another containing Fe2+) were compared by carrying out jar-tests
using different chemical concentrations and pH values. After that, nanofiltration experiments with physico-chemically
treated wastewater were performed at different operating pressures and cross-flow velocities. The results showed
that the COD and conductivity of the nanofiltration permeates were lower than 100 mg/L and 1000 µS/cm respectively.
Keywords: Phisico-chemical treatment; Nanofiltration; Wastewater reuse

1. Introduction
Due to high water consumption in the textile
industry it is essential to study its reuse. Previous
experiments of the research group with wastewater of a textile plant that mainly manufactures
*Corresponding author.

socks, stockings and panties proved that the
combination of physico-chemical treatment and
membrane technologies could produce water for
reuse in the factory [1]. In this case, an attempt is
made to apply these technologies to the effluent
of another textile plant, optimizing the operating
conditions of the membrane process.

Presented at the European Conference on Desalination and the Environment: Fresh Water for All, Malta, 4–8 May 2003.
European Desalination Society, International Water Association.

0011-9164/03/$– See front matter © 2003 Elsevier Science B.V. All rights reserved

3.A. COD was determined with Spectroquant Nova 60 from Merck and BOD5 with the Oxitop system from WTW. having only a slight effect on colour [7. Jar-tests Physico-chemical experiments were carried out in a multiple stirrer Jar-Test apparatus from Selecta. electrical repulsion forces prevent from colloids aggregation. dyeing and finishing textile plant. Objectives The objectives of this work were the following: • Evaluation of the efficiency of two coagulants in the physico-chemical treatment for wastewater of a printing. these techniques should be applied in combination with conventional treatments [3. Additional treatments like membrane techniques could be necessary according to the environmental laws for the disposal of treated water or for its reuse in a textile plant. and conductivity. Jar-tests are a valuable tool in wastewater treatment to evaluate the efficiency of the physicochemical treatment [5]. Biological treatment by activated sludge offers high efficiencies in COD removal but it does not eliminate completely the colour from the water [2]. 3.A. Bes-Piá et al. dyeing and finishing textile industry is proposed. The first step consisted in the characterization of the wastewater samples. 3. Material and methods This work was carried out in three steps. The chemicals used in the jar-tests were the commercial products DK-FER 505-1 from Acideka S. the clarified water was treated with nanofiltration membranes in a laboratory plant to improve the quality of the physico-chemically treated wastewater. The next step was a physicochemical treatment by means of jar-tests to reduce COD. In this work.74 A. Chemical oxidation by ozone.8]. the coagulant was added and rapidly mixed (180 rpm) during 3 min. and UPAX-33 from Kemira S. • Optimization of the best operating conditions for nanofiltration process (feed pressure and cross flow velocity). The general procedure consisted in introducing 900 mL of the sample in the jars. In order to achieve an effective agglomeration. The possible approaches for the membrane purification treatment are nanofiltration or reverse osmosis. or a combination of UV-radiation and ozone and H2O2. • Selection of the membrane according to the salts and COD retention and permeate flow rate. the paddles were withdrawn so . The optimum operating conditions (pH. In this way the colloids can be settled. Wastewater characterization The parameters analysed were COD. since ultrafiltration membranes can hardly remove COD and conductivity. pH. the compression of the thickness of the electrical double layer or a charge reduction of the particles have to be carried out. but the permeate quality is good enough for its reuse in rinse processes (COD <100 mg/L and conductivity <1000 µS/cm). This implies the zeta potential reduction [6]. Thus. According to the double layer theory. chemical concentrations) are determined by means of these experiments.4]. are of great interest but their costs are very high to treat raw textile wastewater. the combination of the physicochemical treatment with nanofiltration to reuse wastewater of a printing. Once the best efficiencies were obtained.1.2. After that. • Study of the water quality after treating the wastewater with a combination of physicochemical treatment and nanofiltration membranes. BOD5. 2. / Desalination 157 (2003) 73–80 Many processes have been studied to treat textile wastewaters. Nanofiltration does not reach the retentions of reverse osmosis.

The operating time of the plant was established according to the steady state conditions. 1. speed control. feed pump. 0. L/m2h bar* >97 98 0. 6.A.10. stirring. 11. Figs. 2. and three different feed flow rates (0.2. % % Dow NF-90 85–95 Osmonics Desal — DL-5 Osmonics Desal — DK-5 Permeability. Results Table 2 shows the average values of the measured parameters of the textile wastewater.481 3.0. pH and conductivity of the clarified water and the sludge volume after 30 min sedimentation (V30) were measured. 8. security valve.1 N and NaOH 0. at the end of each experiment. respectively. 1. 3. rejection stream. These values are typical for textile effluents. The permeate fluxes JP (L/m2h) and salt retentions RSALT (%) were determined. The coagulant concentrations are referred to Fe2+ and Al2O3 for DK-FER 505-1 and UPAX33.22 m/s respectively. Scheme of NF laboratory plant.3. since in this medium the coagulants drive to the formation of positively charged metal hidroxy complexes that specifically adsorb onto colloids. In all tests COD. manometer. regulation valve. heat exchanger. turbidity. / Desalination 157 (2003) 73–80 that the particles could settle. In Fig.20 MPa). permeate stream. In addition. COD was analysed.66 and 2. 2 and 3 show the COD obtained after jar-tests on varying the DK-FER 505-1 and UPAX-33 concentrations. thermometer. °C pH Conductivity. 10. retention. Conductivity and COD are quite important and they have to be reduced to produce water with enough quality to be reused. feed tank.11. 4. regulation valve. 14. Table 2 Wastewater characterization Parameter T. 4.53 490 1630 . 9–9’. 1. 12. 1. The pH of the samples was changed by addition of HCl 0. For each membrane. The tested NF membranes are described in Table 1.15 and 0. mg/L Feedwater 20 12.1 and 0. experiments with three different transmembrane pressures (0. 5. The cross flow velocities related to these flow rates 2 14 3 4 13 1 8 5 6 7 75 Table 1 Tested NF membranes Membrane NaCl MgSO4 retention.4 m3/h) at 25°C were performed. mg/L COD.3 and 0.863 96 3. The series of experiments were carried out using an experimental design obtained from Statgraphics Plus 4. filtration system. mS/cm BOD5. 0. The configuration of the plant can be observed in Fig. 13. NF module. NF module is plane and its effective membrane area is 0. Bes-Piá et al. The influence of coagulants concentration and pH values were studied. 2 it can be 12 9 10 9’ 11 Fig. Experiments with membranes Experiments were carried out using a laboratory nanofiltration (NF) plant.5 N.012 m2. pH values were selected in the alkaline range. 7. 3.562 *Experimental values are 1. It was about 8 h in all cases.0 4.

It is important to emphasize that a total colour removal was accomplished. In Fig. Fig. 5 show the COD obtained after jar-test for different wastewater pH values. Influence of UPAX-33 concentration on COD of treated water. Influence of wastewater pH on COD of treated water using 700 mg/L of DK-FER 505-1. The other tested pH did not improve the COD efficiency. In Fig. The optimum operating conditions were: pH = 12.e.0 11. This occurred because of the insufficient reduction in the zeta potential of the wastewater. Influence of DK-FER 505-1 concentration on COD of treated water. .0 and CFe2+ = 700 mg/L. 3. 4 and Fig. Higher coagulant concentrations did not improve the efficiency.5% at pH 12 (the raw wastewater pH). Bes-Piá et al. / Desalination 157 (2003) 73–80 1600 1600 1200 1200 COD (mg/L) COD (mg/L) 76 800 400 800 400 0 0 500 600 700 200 8 00 DK-FER 505-1 coagulant (mg/L) 300 400 500 UPAX-33 coagulant (mg/L) Fig. 4. 4 it can be observed that the COD removal slightly increased with wastewater pH.5%) was reached with a concentration of 700 mg/L.0 (COD removal of 66.0% removal). Influence of wastewater pH on COD of treated water using 300 mg/L of UPAX-33.5 11. In Table 3 the characteri-zation of the clarified water after the test at these conditions can be observed. observed that DK-FER 505-1 concentrations of 500 mg/L and 600 mg/L produced only a slight COD reduction. 600 400 200 600 400 200 0 9 10 11 12 pH Fig. In Fig.A. 5 it is shown that the best yield was reached at pH 11. Thus.4 pH Fig. 2. 0 10. 5. 1000 1000 800 800 COD (mg/L) COD (mg/L) Fig. 58.0%). reaching a COD removal of 70.0 10. For this value the COD of the clarified water was 680 mg/L (i. experiments showed that DK-FER 505-1 coagulant provided the best results for the physicochemical treatment of this textile wastewater. The maximum COD removal yield (72. 3. for UPAX-33 the selected coagulant concentration was 300 mg/L.

These Pareto charts display a frequency histogram where the length of each bar is proportional to the estimated effect and interactions of the feed flow rate (B) and feed pressure (A) on permeate flux.8 35.0 43. Bes-Piá et al.3 5.A.5 4. 7.0 5. L/m2h RSALT.1 42.5 28. L/m2h 10 10 10 15 15 15 20 20 20 200 300 400 200 300 400 200 300 400 83.7 21.0 36.7 55.0 72.8 52.7 30. % JP.8 40.3 33.7 7.3 45.6 83. L/m2h RSALT. bar Feed flow rate.2 28.8 8.1 35.27 25. it can be seen that only feed pressure 0 2 4 6 8 10 Standardized effect Fig.0 37. Standardized Pareto chart for permeate flux of NF-90.2 55.7 37.3 67. mg/L 77 A:P B:Q 6.6 25.9 44.2 78.0 60.0 77. Standardized Pareto chart for permeate flux of DK-5.8 87.0 42. Table 4 Salt rejections and permeate fluxes at the steady state conditions in the different experiments Operating conditions NF-90 DL-5 DK-5 Feed pressure.1 38.cm–1 Turbidity.6 7. The cross line indicates the significance of each parameter.3 85.6 18. For NF-90 it can be observed that feed pressure and feed flow rate were significant variables.2 34.7 51. In Table 4 salts rejections (RSALT) and permeate fluxes (JP) at the steady state conditions for the tested membranes can be observed. A:P B:Q BB AA AB 0 1 2 3 4 5 Standardized effect Fig.5 .14 3.8 86. pH = 12) pH Conductivity.8 2.3 42. mL/L COD.39 40. % JP.5 39.2 85. NTU V30.14 48.9 29.4 16.2 2.5 2.80 4. / Desalination 157 (2003) 73–80 Table 3 Clarified water characterization (Jar-test conditions: CFe2+ = 700 mg/L.7 260 448 AB BB AA Nanofiltration experiments were made using wastewater treated with DK-FER 505-1 as feed water. mS. while the influence of the feed flow rate was substantially lower.8 34. 6–8 show the standardized Pareto charts for permeate flux of the membranes tested. Figs. L/h RSALT. However.1 26. Feed pressure variations produced an important increase on permeate flux. 6. % JP.8 21.

15 Q (L/h) Standardized effect Fig. 11–13 show the standardized Pareto charts for retention salts of the membranes tested. In all cases salts retentions did no depend on the feed flow rate. However. Influence of feed pressure and feed flow rate on permeate flux in NF-90. 20 DL-5 DK-5 60 40 20 0 5 01 51 20 25 P (bar) Fig. Similarly. the higher salts retentions. Influence of feed pressure on permeate flux in DL-5 and DK-5. Standardized Pareto chart for permeate flux of DL-5. By means of DK-5 and NF-90. 11–13 show the standardized Pareto charts for retention salts of the membranes tested. the highest permeate flux was achieved at 20 bar and 200 L/h. Thus. a graph showing the evolution of this variable was performed (Fig. the feed pressure have been represented for DL-5 and DK-5. The low permeate COD values (50 mg/L) can be highlighted. the obtained salts rejections values have been studied using the same types of graphs. Table 5 summarizes the results of the permeate analysis. 10. Similarly. the higher feed pressure. The highest salts retentions were obtained with NF-90. Figs. The permeate flux values correspond with the average values calculated for the tested feed flow rate. the obtained salts rejection values have been studied using the same types of graphs. No influence of the feed flow rate was observed at 10 and 15 bar. 80 JP (L/m 2h) influenced significantly on permeate flux for DL-5 and DK-5. 9 shows the increase of the permeate fluxes with the feed pressure. the highest feed pressure. / Desalination 157 (2003) 73–80 8-10 6-8 4-6 2-4 A:P 10 Flux (l/m 2 h) BB AB B:Q 8 6 4 2 0 200 AA 0 2 4 6 8 300 400 10 P (bar) Fig. the highest permeate flux. 9.78 A. In order to illustrate the obtained results. 10 the permeate fluxes vs. As it can be expected. 8. Fig. DL-5 and DK-5 yielded very similar values for each tested pressure. 10). Bes-Piá et al. 9). As can be seen in Fig. at 20 bar a slight influence was noticed. 14. . In Fig. Figs. a response surface graph was plotted if both feed pressure (P) and feed flow rate (Q) were significant variables (Fig. it was possible to produce water with sufficient quality to be reused in the textile plant. If only the feed pressure was significant. In all cases salts retentions did no depend on the feed flow rate.

mS/cm COD. 13. L/h Flux. mg/L 20 20 20 200 200–400 200–400 8–10 60–80 50–60 85–90 45–50 55–65 0. % Conductivity. Standardized Pareto chart for retention salts of DL-5.71 48 98 50 . Standardized Pareto chart for retention salts of NF-90 membrane. 11. 12.A.09 1. bar Feed flow rate. L/m2h RSALT.33–1. 10 Selection of operating conditions Variables Effluent analysis Feed pressure. Bes-Piá et al. Table 5 Analysis of permeate streams at the best operating conditions for each membrane Membrane NF–90 Desal DL–5 Desal DK–5 25 P (bar) Standardized effect Fig.90–2. Standardized Pareto chart for retention salts of DK-5. Influence of feed pressure on salts rejection in tested membranes. Fig.46 1. 14. / Desalination 157 (2003) 73–80 A:P A:P B:Q B:Q AB AA BB BB AA AB 0 1 2 3 4 0 5 79 1 2 3 4 Standardized effect Standardized effect Fig. 100 A:P RSALT (%) BB B:Q AA DL-5 DK-5 NF-90 80 60 40 20 AB 0 0 1 2 3 4 5 15 20 Fig.

M. [7] M. G. However. 233 (1995) 42–61. Salts rejection and permeate flux rates were dependent basically on feed pressure.A.A. Corrieri and G. I. [4] S Baig and P. Revista de la Industria Textil. Antonelli and M. the management of the retentate stream has to be deeply studied. M.I. design and implementation of wastewater treatment plants in the textile industry. Wat. G. Buckley and B. Groves. D. Marcucci. Though the permeate flux rates of NF-90 were lower than for the other membranes. Botino.5% COD removal efficiency). With UPAX-33 worse efficiencies were achieved. C. since the salts rejections were substantially higher than for the other membranes. Indutria textil: depuración biológica o físicoquímica?. Technol. Conclusions References By means of a physico-chemical treatment using the coagulant DK-FER 505-1 at pH 12 and with a concentration CFe2+ = 700 mg/L. / Desalination 157 (2003) 73–80 5. no influence was found on the studied variables.O. Reuse of wastewater of the textile industry after its treatment with a combination of physico-chemical treatment and membrane technologies.A. Theory and Practice. Nosenzo. Iborra-Clar and M. Nanofiltration of the physico-chemically treated wastewater produced a permeate with a COD lower than 100 mg/L for the three tested membranes. Ozone treatment for biorefractory COD removal.. Desalination. Rozzi.80 A. Part I: Closed loop treatment/recycle system for textile sizing/desizing effluents. for its support in the investigation project.. Acknowledgment We thank Colortex 1967 S.. Coagulants and Flocculants. Wat. Liechti. 1995. Technol. [8] A. Pretoria. Bes-Piá et al. A guide for the planning.L. 1983. 149 (2002) 169–174. AlcainaMiranda. Arcari. G. Ozone treatment of textile wastewater for reuse. Southworth. Sci. for the feed flow rate range that was studied. [1] A. Desalination. Iborra-Clar. this was the selected membrane. Bes-Piá. Prior to an industrial operation. Ciardelli. Mendoza-Roca. Hart. G. Technol. [2] M. Sci. Kim. Treatment and reuse of textile effluents based on new ultrafiltration and other membrane technologies. COD of a textile effluent can be reduced to values lower than 500 mg/L (72. Ciabatti. . Sci. 44(5) (2001) 61–67. A. [6] Y. 40(4–5) (1999) 409–416. 43 (2001) 197–204. Capanelli. [3] G. Wat.H. Huertas López. Ciardelli.A. 138 (2001) 75–82.. Crespí Rosell and J.I. Campanelli and A.R. [5] O. J. Membrane treatment of secondary textile effluents for direct reuse.