- Email: [email protected]
Selective demineralization of water by nanofiltration Application to the defluorination of brackish water
PII: S0043-1354(01)00020-3
Wat. Res. Vol. 35, No. 13, pp. 3260–3264, 2001 # 2001 Elsevier Science Ltd. All rights reserved Printed in Great Britain 0043-1354/01/$ - see front matter
SELECTIVE DEMINERALIZATION OF WATER BY NANOFILTRATION APPLICATION TO THE DEFLUORINATION OF BRACKISH WATER A. LHASSANI1,2, M. RUMEAU2*, D. BENJELLOUN1 and M. PONTIE3 2
1 De´partement de Chimie, Faculte´ des Sciences Dhar Mahrez. B.P. 1796 Fe`s, Maroc, France; Hydrosciences, Place Eugene Bataillon-Case MSE, UM 2, F-34095 Montpellier Cedex 5, France; and 3 Electrochimie et chimie analytique, ENSCP, F-75005 Paris, France
(First received 1 December 1999; accepted in revised form 28 December 2000) Abstract}Nanofiltration is generally used to separate monovalent ions from divalent ions, but it is also possible to separate ions of the same valency by careful application of the transfer mechanisms involved. Analysis of the retention of halide salts reveals that small ions like fluoride are the best retained, and that this is even more marked under reduced pressure when selectivity is greatest. The selectivity desalination of fluorinated brackish water is hence feasible and drinking water can be produced directly at much lower cost than using reverse osmosis by optimizing the pressure for the type of water treated. # 2001 Elsevier Science Ltd. All rights reserved Key words}nanofiltration, defluorination, selective demineralization, transfer mechanisms
INTRODUCTION
Some brackish waters encountered in nature contain undesirable ions like fluorides. The regular ingestion of water with a fluoride ions concentration higher than 2 mg l1 causes serious ailments like dental and bone fluorosis (Pontie´, 1996). This investigation concerns the use of nanofiltration to remove fluorine from slightly brackish fluorinated water. The objective was to identify optimum condition for the operation of the pilot unit as a first step towards industrial application. Many constraints must be taken into account, including compliance with standards and minimization of operating cost, in order to select the operating parameters. Technologies using membrane processes are being used in many applications. Many processes today have progressed from the research stage to industrial application. Since the 1970 s, these separation processes have also spread considerably to other fields, particularly for brackish water desalination (Pontie´ et al., 1994). Reverse osmosis allows total desalination, which then entails remineralization, while electrodialysis can be used for partial and possibly selective desalination (Diey et al., 1994). Nanofiltration is a new process that is still little used in the water industry, but is beginning to
*Author to whom all correspondence should be addressed. Tel.: +33-467-143605; fax: +33-467-144774; e-mail: [email protected]
compete with the other two membrane techniques for the treatment of brackish water (Rumeau and Pontie´, 1998). Nanofiltration can provide selective desalination and is generally used to remove divalent ions, such as sulfates and calcium ions. But it is also important to determine whether this process can be used to separate ions of the same valency. This technology also offers the great advantage of lower operating costs than reverse osmosis. Low cost membrane materials have also contributed to its spread, making it less expensive than the other two methods mentioned earlier (Pontie´ et al., 1995). Since reverse osmosis cannot be used for partial and/or selective demineralization (Colon, 1985), nanofiltration or electrodialysis is more suitable for producting drinking water directly without the need for remineralization. (Rautenbach and Groschl, 1984; Tanghe et al., 1992). One of the main differences between nanofiltration and electrodialysis is that the solvent passes through the membrane with more or less selective solute retention in nanofiltration (Rumeau and Pontie´, 1998), whereas the solutes pass through the membrane with more or less selective transfer in electrodialysis. There are also no ion exchange resins or specific membranes for fluorides, whereas there are for nitrates and borates. Thus, the selective removal of fluorides is best done by nanofiltration, while borates and nitrates are best removed by electrodialysis.
3260
Selective demineralization of water
Retention is also often due more to the dissolution properties (Rumeau, 1995) in water than to the membrane itself in these processes, since the fluoride ion is the smallest monovalent anion, it is the most soluble in water and hence passes through the membrane with difficulty. By contrast, nitrates and borates, are only slightly soluble in water and are less well retained. This study should also provide a clearer understanding of how solutes are transferred in nanofiltration, which would lead to a better prediction of the possibilities of application and to identification at the optimal operating conditions. Thus, assumptions such as the passage from solubilization–diffusion to convection and the influence of the polarization layer and solvation of the ions could explain the variations in performance of the pilot unit.
3261
Rejection rates were calculated using the flowing equation: R ¼ 1 Cp =C0 ; where Cp is the permeate concentration and C0 is the initial concentration. The ions were assayed with specific electrodes calibrated for fluoride, chloride and iodide ions. RESULTS AND DISCUSSION
Permeate concentration Analysis of the sodium halide retention as a function of pressure showed (Fig. 2) that small ions like fluoride are better retained than large ions. This was especially true at lower pressure. The fluoride permeate concentration tended to reach a plateau at the highest pressures. Hence the passage of salts and selectivity were higher at low pressure, where the chemical parameters are predominant.
MATERIALS AND METHODS
The test solutions were obtained by adding sodium halide salts (Prolabo Rectapur) to the tap water nanofiltered. The nanofiltration pilot plant (Fig. 1) supplied by Europeenne de Traitement des Eaux (ETE) consisted of a feed tank, a pump and a spiral module containing a Filmtec NF70-2540 membrane (DOW Chemical, Denmark). The membrane of 180 Da tested was a thin-film composite of polyamide on a polysulfone support which is designed for tangential filtration. We investigated the retention of monovalent fluoride, chloride and iodide ions by the pilot plant using 5.67 mM solutions of NaCl, NaF and NaI(solution I) and the binary solution was composed of 17 mM NaCl and 0.24 mM NaF (solution II). We analyzed the performance of the membrane with fluorides, chlorides and iodides, as well as their interactions on the selectivity of the membrane during the filtration of mixtures of these ions. The study should help to provide a better understanding of the way retention changes with the operating parameters: pressure and conversion rate. The conversion rate Y was monitored by flowmeters placed at the outlet of the retentate and permeate. Operations thus proceeded at a constant conversion rate of 50%. Y ¼ Qp=Q0 ; where Q0 and Qp are the initial and permeate flow rates, respectively.
Permeate flow rate The flow rates varied with the ion concerned (Fig. 3). The flow rate of iodide solution increased faster than that of fluoride at high pressures. Thus, the change in osmotic pressure was not the same for these two ions. The following equations can explain these differences: P ¼ai RT
P ¼gi ci RT
where P is the osmotic pressure, ai the activity of salts, ci the molar concentration of salts, and gi the activity coefficient of ions. While the concentrations were the same, the activity coefficients gi are different because the ionic radii are not the same. Thus, gi has the following form for our concentration range: pffiffiffi X 0:5 I pffiffiffi Log gi ¼ with I ¼ 1=2 ci ; 1 þ bi I where I is the ionic strength and bi is the ionic radius. The activity coefficient of the ions modifies the activity coefficient of the water. Hence the latter
Fig. 1. Pilot plant flow chart.
3262
A. Lhassani et al.
Fig. 2. Passage of sodium halide salts in single solute solution. The concentration was 5.67 mM.
Fig. 3. Permeate flow rate in the presence of sodium iodine and/or fluoride salts. The concentration was 5.67 mM.
decreases faster in the presence of fluoride ion than with iodide, because the water solvates the fluoride more strongly. This caused a decrease in osmotic pressure P and consequently increased the permeate flow rate. We therefore have Qp ¼AðDPDPÞ; where Qp is the permeate flow rate, A the membrane permeability, DP the difference in pressure applied to membrane, and DP the difference in osmotic pressure. It is also normal for the flow rate of the iodidefluoride mixture to be lower than that of iodide alone or fluoride alone, because of the overall increase in osmotic pressure. Ion retention Ions are transferred by two mechanisms in nanofiltration: (i) Convection. They are carried by the solvent stream as a function of the transfer coefficient. The larger ions are more retained (physical parameters). (ii) Solubilization–diffusion. A function of the solvation energies and the partition coefficient. The larger the ion, the less well it is retained (chemical parameters). The convection transfer mechanisms are modified by altering the physical parameters (pressure, conversion rate), without altering diffusion, which is influenced only by the chemical parameters (concentration, pH) and vice versa. Convection is low at low pressure and in contrast, the physical parameters predominate at high pressure, and the larger ions are better retained. Nevertheless, chemical selectivity is always much more important than physical selectiv-
Fig. 4. Retention of sodium halide salts in single solute solution, as a function of pressure. The concentration was 5.67 mM.
ity for separating ions. This means that selectivity is always higher at low pressure. The retention curve R ¼ f (pressure) of halide salts alone (Fig. 4), shows that the smaller ions are better retained by the membrane. This is particularly true at low pressure where the selectivity of the membrane is the greatest. The retention of ions also tends to plateau at higher pressures. Similarly, retention is generally higher at high pressure. This is because convection predominates over diffusion, whereas diffusion is most influential at low pressure. Fluoride retention is practically unaffected by pressure because the passage of this ion is mainly due to diffusion. Convection has virtually no effect, since a significant increase in pressure caused no significant change in retention. By contrast, iodide
Selective demineralization of water
3263
Table 1. Desalination of a fluorinated brackish water in Senegal, P¼ 6 bar; Y ¼ 10%
Untreated water Nanofiltered water
Total salinity (mg/l)
F (mg/l)
Total heterotrophic plate count cfu/100 ml
2025 230
13.5 0.7
6000 0
The retention rates were also analyzed as a function of conversion rates. An increase in the conversion rate caused a decrease in the retention rate for the same pressure and for a given ion (Fig. 6). This technique was used in a village in southwestern Senegal in the Kaolack region to produce water by desalination of fluorinated water (Table 1).
Fig. 5. Retention of sodium halide salts in ternary mixtures (F, C1, I). The concentration was 5.67 mM.
Fig. 6. Retention of chloride and fluoride in binary mixtures, as a function of conversion rate. P¼ 5 bar; C0 (NaCl)=17 mM; C0 (NaF)=0.24 mM.
and chloride retention was much more influenced by pressure because the passage of these salts is more due to physical than chemical forces. In fact, diffusion is governed by chemical parameters like the partition coefficient P, whereas convection is due to physical properties, such as the size of the ions. The same remarks are valid for mixtures of the three halides salts (Fig. 5), apart from some significant differences. There was a relative decrease in fluoride retention simultaneously with an increase in iodide and chloride retention. Similarly, selectivity between the latter two ions decreased as the pressure increased until the retention curves overlapped. Thus, the selectivity curves was actually reversed above a pressure of 8 bar.
CONCLUSION
This pilot study shows that ions of the same valency can be removed selectively by adjusting the operating conditions. The ion transfer mechanisms operating in nanofiltration are related to chemical and physical factors, whereas only chemical parameters are involved in reverse osmosis. The transfer coefficients regulate the physical factors and partition coefficients control the chemical factors. The chemical parameters predominate at low pressures, whereas physical factors predominate at higher pressures. The same remark applies to diffusion and convection. Hence diffusion is related to chemical parameters whereas convection depends on physical factors. The selectivity of the membrane for ions, which is of chemical origin, is also better at low pressure. The results indicate that the smaller the ion, the better it is retained, mainly at low pressure, when chemical selectivity predominates. This is derived from the solvation energy of the ions by water. Since fluoride ions are more solvated, they are better retained than chloride and iodide ions. REFERENCES
Colon W. (1985) Pilot field test data for prototype ultra low pressure reverse osmosis element. Desalination 56, 203–226. Diey A., Mar C. and Rumeau M. (1994) Les proce´de´s de de´fluoruration des eaux de boisson. Tribune l’eau 568, 27–34. Pontie´ M. (1996) Phe´nome`nes electrocine´tique et transferts ioniques dans les membranes poreuses a` faible seuil de coupure. Application au traitement des eaux saumaˆtres. The`se, Univ. Tours France. Pontie´ M., N’diaye M., Chevallier S., Rumeau M., Lemordant D. and Anselme C. (1995) Nanofiltration: a new method against fluoride ion poisoning in Senegal. Euromembrane 1995, Bath (R.U.), 18–20 September, DP8, pp. I–529. Pontie´ M., Sissoko H., Rumeau M. and Mar-Diop C. (1994) Dessalement se´lectif des eaux saumaˆtres fluorure´es du bassin du Se´ne´gal par nanofiltration}1e`res. Journe´es Internationales Interfiltra-Intermembrane, Paris 94, 101. Rautenbach R. and Groschl A. (1984) Separation potential of nanofiltration membranes. Desalination 77, 73–84.
3264
A. Lhassani et al.
Rumeau M. (1995) La nanofiltration: Principes, performances et application. Information chimie, no. 373, November 1995. Rumeau M. and Pontie´ M. (1998) Potabilisation d’une eau saumaˆtre hyperfluorure´e du Se´ne´gal par
de´mine´ralisation se´lective. Hydrotop. Marseille, 21–23 Avril. Tanghe N., Kopp V., Dard S. and Faivre M. (1992) Application of nanofiltration to obtain drinking water. Euro membrane 92, Paris Vol. 6, pp. 67–72.
Wat. Res. Vol. 35, No. 13, pp. 3260–3264, 2001 # 2001 Elsevier Science Ltd. All rights reserved Printed in Great Britain 0043-1354/01/$ - see front matter
SELECTIVE DEMINERALIZATION OF WATER BY NANOFILTRATION APPLICATION TO THE DEFLUORINATION OF BRACKISH WATER A. LHASSANI1,2, M. RUMEAU2*, D. BENJELLOUN1 and M. PONTIE3 2
1 De´partement de Chimie, Faculte´ des Sciences Dhar Mahrez. B.P. 1796 Fe`s, Maroc, France; Hydrosciences, Place Eugene Bataillon-Case MSE, UM 2, F-34095 Montpellier Cedex 5, France; and 3 Electrochimie et chimie analytique, ENSCP, F-75005 Paris, France
(First received 1 December 1999; accepted in revised form 28 December 2000) Abstract}Nanofiltration is generally used to separate monovalent ions from divalent ions, but it is also possible to separate ions of the same valency by careful application of the transfer mechanisms involved. Analysis of the retention of halide salts reveals that small ions like fluoride are the best retained, and that this is even more marked under reduced pressure when selectivity is greatest. The selectivity desalination of fluorinated brackish water is hence feasible and drinking water can be produced directly at much lower cost than using reverse osmosis by optimizing the pressure for the type of water treated. # 2001 Elsevier Science Ltd. All rights reserved Key words}nanofiltration, defluorination, selective demineralization, transfer mechanisms
INTRODUCTION
Some brackish waters encountered in nature contain undesirable ions like fluorides. The regular ingestion of water with a fluoride ions concentration higher than 2 mg l1 causes serious ailments like dental and bone fluorosis (Pontie´, 1996). This investigation concerns the use of nanofiltration to remove fluorine from slightly brackish fluorinated water. The objective was to identify optimum condition for the operation of the pilot unit as a first step towards industrial application. Many constraints must be taken into account, including compliance with standards and minimization of operating cost, in order to select the operating parameters. Technologies using membrane processes are being used in many applications. Many processes today have progressed from the research stage to industrial application. Since the 1970 s, these separation processes have also spread considerably to other fields, particularly for brackish water desalination (Pontie´ et al., 1994). Reverse osmosis allows total desalination, which then entails remineralization, while electrodialysis can be used for partial and possibly selective desalination (Diey et al., 1994). Nanofiltration is a new process that is still little used in the water industry, but is beginning to
*Author to whom all correspondence should be addressed. Tel.: +33-467-143605; fax: +33-467-144774; e-mail: [email protected]
compete with the other two membrane techniques for the treatment of brackish water (Rumeau and Pontie´, 1998). Nanofiltration can provide selective desalination and is generally used to remove divalent ions, such as sulfates and calcium ions. But it is also important to determine whether this process can be used to separate ions of the same valency. This technology also offers the great advantage of lower operating costs than reverse osmosis. Low cost membrane materials have also contributed to its spread, making it less expensive than the other two methods mentioned earlier (Pontie´ et al., 1995). Since reverse osmosis cannot be used for partial and/or selective demineralization (Colon, 1985), nanofiltration or electrodialysis is more suitable for producting drinking water directly without the need for remineralization. (Rautenbach and Groschl, 1984; Tanghe et al., 1992). One of the main differences between nanofiltration and electrodialysis is that the solvent passes through the membrane with more or less selective solute retention in nanofiltration (Rumeau and Pontie´, 1998), whereas the solutes pass through the membrane with more or less selective transfer in electrodialysis. There are also no ion exchange resins or specific membranes for fluorides, whereas there are for nitrates and borates. Thus, the selective removal of fluorides is best done by nanofiltration, while borates and nitrates are best removed by electrodialysis.
3260
Selective demineralization of water
Retention is also often due more to the dissolution properties (Rumeau, 1995) in water than to the membrane itself in these processes, since the fluoride ion is the smallest monovalent anion, it is the most soluble in water and hence passes through the membrane with difficulty. By contrast, nitrates and borates, are only slightly soluble in water and are less well retained. This study should also provide a clearer understanding of how solutes are transferred in nanofiltration, which would lead to a better prediction of the possibilities of application and to identification at the optimal operating conditions. Thus, assumptions such as the passage from solubilization–diffusion to convection and the influence of the polarization layer and solvation of the ions could explain the variations in performance of the pilot unit.
3261
Rejection rates were calculated using the flowing equation: R ¼ 1 Cp =C0 ; where Cp is the permeate concentration and C0 is the initial concentration. The ions were assayed with specific electrodes calibrated for fluoride, chloride and iodide ions. RESULTS AND DISCUSSION
Permeate concentration Analysis of the sodium halide retention as a function of pressure showed (Fig. 2) that small ions like fluoride are better retained than large ions. This was especially true at lower pressure. The fluoride permeate concentration tended to reach a plateau at the highest pressures. Hence the passage of salts and selectivity were higher at low pressure, where the chemical parameters are predominant.
MATERIALS AND METHODS
The test solutions were obtained by adding sodium halide salts (Prolabo Rectapur) to the tap water nanofiltered. The nanofiltration pilot plant (Fig. 1) supplied by Europeenne de Traitement des Eaux (ETE) consisted of a feed tank, a pump and a spiral module containing a Filmtec NF70-2540 membrane (DOW Chemical, Denmark). The membrane of 180 Da tested was a thin-film composite of polyamide on a polysulfone support which is designed for tangential filtration. We investigated the retention of monovalent fluoride, chloride and iodide ions by the pilot plant using 5.67 mM solutions of NaCl, NaF and NaI(solution I) and the binary solution was composed of 17 mM NaCl and 0.24 mM NaF (solution II). We analyzed the performance of the membrane with fluorides, chlorides and iodides, as well as their interactions on the selectivity of the membrane during the filtration of mixtures of these ions. The study should help to provide a better understanding of the way retention changes with the operating parameters: pressure and conversion rate. The conversion rate Y was monitored by flowmeters placed at the outlet of the retentate and permeate. Operations thus proceeded at a constant conversion rate of 50%. Y ¼ Qp=Q0 ; where Q0 and Qp are the initial and permeate flow rates, respectively.
Permeate flow rate The flow rates varied with the ion concerned (Fig. 3). The flow rate of iodide solution increased faster than that of fluoride at high pressures. Thus, the change in osmotic pressure was not the same for these two ions. The following equations can explain these differences: P ¼ai RT
P ¼gi ci RT
where P is the osmotic pressure, ai the activity of salts, ci the molar concentration of salts, and gi the activity coefficient of ions. While the concentrations were the same, the activity coefficients gi are different because the ionic radii are not the same. Thus, gi has the following form for our concentration range: pffiffiffi X 0:5 I pffiffiffi Log gi ¼ with I ¼ 1=2 ci ; 1 þ bi I where I is the ionic strength and bi is the ionic radius. The activity coefficient of the ions modifies the activity coefficient of the water. Hence the latter
Fig. 1. Pilot plant flow chart.
3262
A. Lhassani et al.
Fig. 2. Passage of sodium halide salts in single solute solution. The concentration was 5.67 mM.
Fig. 3. Permeate flow rate in the presence of sodium iodine and/or fluoride salts. The concentration was 5.67 mM.
decreases faster in the presence of fluoride ion than with iodide, because the water solvates the fluoride more strongly. This caused a decrease in osmotic pressure P and consequently increased the permeate flow rate. We therefore have Qp ¼AðDPDPÞ; where Qp is the permeate flow rate, A the membrane permeability, DP the difference in pressure applied to membrane, and DP the difference in osmotic pressure. It is also normal for the flow rate of the iodidefluoride mixture to be lower than that of iodide alone or fluoride alone, because of the overall increase in osmotic pressure. Ion retention Ions are transferred by two mechanisms in nanofiltration: (i) Convection. They are carried by the solvent stream as a function of the transfer coefficient. The larger ions are more retained (physical parameters). (ii) Solubilization–diffusion. A function of the solvation energies and the partition coefficient. The larger the ion, the less well it is retained (chemical parameters). The convection transfer mechanisms are modified by altering the physical parameters (pressure, conversion rate), without altering diffusion, which is influenced only by the chemical parameters (concentration, pH) and vice versa. Convection is low at low pressure and in contrast, the physical parameters predominate at high pressure, and the larger ions are better retained. Nevertheless, chemical selectivity is always much more important than physical selectiv-
Fig. 4. Retention of sodium halide salts in single solute solution, as a function of pressure. The concentration was 5.67 mM.
ity for separating ions. This means that selectivity is always higher at low pressure. The retention curve R ¼ f (pressure) of halide salts alone (Fig. 4), shows that the smaller ions are better retained by the membrane. This is particularly true at low pressure where the selectivity of the membrane is the greatest. The retention of ions also tends to plateau at higher pressures. Similarly, retention is generally higher at high pressure. This is because convection predominates over diffusion, whereas diffusion is most influential at low pressure. Fluoride retention is practically unaffected by pressure because the passage of this ion is mainly due to diffusion. Convection has virtually no effect, since a significant increase in pressure caused no significant change in retention. By contrast, iodide
Selective demineralization of water
3263
Table 1. Desalination of a fluorinated brackish water in Senegal, P¼ 6 bar; Y ¼ 10%
Untreated water Nanofiltered water
Total salinity (mg/l)
F (mg/l)
Total heterotrophic plate count cfu/100 ml
2025 230
13.5 0.7
6000 0
The retention rates were also analyzed as a function of conversion rates. An increase in the conversion rate caused a decrease in the retention rate for the same pressure and for a given ion (Fig. 6). This technique was used in a village in southwestern Senegal in the Kaolack region to produce water by desalination of fluorinated water (Table 1).
Fig. 5. Retention of sodium halide salts in ternary mixtures (F, C1, I). The concentration was 5.67 mM.
Fig. 6. Retention of chloride and fluoride in binary mixtures, as a function of conversion rate. P¼ 5 bar; C0 (NaCl)=17 mM; C0 (NaF)=0.24 mM.
and chloride retention was much more influenced by pressure because the passage of these salts is more due to physical than chemical forces. In fact, diffusion is governed by chemical parameters like the partition coefficient P, whereas convection is due to physical properties, such as the size of the ions. The same remarks are valid for mixtures of the three halides salts (Fig. 5), apart from some significant differences. There was a relative decrease in fluoride retention simultaneously with an increase in iodide and chloride retention. Similarly, selectivity between the latter two ions decreased as the pressure increased until the retention curves overlapped. Thus, the selectivity curves was actually reversed above a pressure of 8 bar.
CONCLUSION
This pilot study shows that ions of the same valency can be removed selectively by adjusting the operating conditions. The ion transfer mechanisms operating in nanofiltration are related to chemical and physical factors, whereas only chemical parameters are involved in reverse osmosis. The transfer coefficients regulate the physical factors and partition coefficients control the chemical factors. The chemical parameters predominate at low pressures, whereas physical factors predominate at higher pressures. The same remark applies to diffusion and convection. Hence diffusion is related to chemical parameters whereas convection depends on physical factors. The selectivity of the membrane for ions, which is of chemical origin, is also better at low pressure. The results indicate that the smaller the ion, the better it is retained, mainly at low pressure, when chemical selectivity predominates. This is derived from the solvation energy of the ions by water. Since fluoride ions are more solvated, they are better retained than chloride and iodide ions. REFERENCES
Colon W. (1985) Pilot field test data for prototype ultra low pressure reverse osmosis element. Desalination 56, 203–226. Diey A., Mar C. and Rumeau M. (1994) Les proce´de´s de de´fluoruration des eaux de boisson. Tribune l’eau 568, 27–34. Pontie´ M. (1996) Phe´nome`nes electrocine´tique et transferts ioniques dans les membranes poreuses a` faible seuil de coupure. Application au traitement des eaux saumaˆtres. The`se, Univ. Tours France. Pontie´ M., N’diaye M., Chevallier S., Rumeau M., Lemordant D. and Anselme C. (1995) Nanofiltration: a new method against fluoride ion poisoning in Senegal. Euromembrane 1995, Bath (R.U.), 18–20 September, DP8, pp. I–529. Pontie´ M., Sissoko H., Rumeau M. and Mar-Diop C. (1994) Dessalement se´lectif des eaux saumaˆtres fluorure´es du bassin du Se´ne´gal par nanofiltration}1e`res. Journe´es Internationales Interfiltra-Intermembrane, Paris 94, 101. Rautenbach R. and Groschl A. (1984) Separation potential of nanofiltration membranes. Desalination 77, 73–84.
3264
A. Lhassani et al.
Rumeau M. (1995) La nanofiltration: Principes, performances et application. Information chimie, no. 373, November 1995. Rumeau M. and Pontie´ M. (1998) Potabilisation d’une eau saumaˆtre hyperfluorure´e du Se´ne´gal par
de´mine´ralisation se´lective. Hydrotop. Marseille, 21–23 Avril. Tanghe N., Kopp V., Dard S. and Faivre M. (1992) Application of nanofiltration to obtain drinking water. Euro membrane 92, Paris Vol. 6, pp. 67–72.