Metal ion transport through microemulsion liquid membranes

Metal ion transport through microemulsion liquid membranes

Colloids and Surfaces, 48 (1990) 243-258 Elsevier Science Publishers B.V., Amsterdam 243 Metal Ion Transport Through Microemulsion Liquid Membranes ...

986KB Sizes 0 Downloads 67 Views

Colloids and Surfaces, 48 (1990) 243-258 Elsevier Science Publishers B.V., Amsterdam


Metal Ion Transport Through Microemulsion Liquid Membranes A. DEROUICHE

and C. TONDRE*

Laboratoire d’Etudes des Solutions Organiques et Collloidales (L.E.S.O.C.), Unite' Associie au C.N.R.S. No 406, Universitk de Nancy Z, B.P. 239,54506 Vandoeuvre-lb-Nancy Cedex (France) (Received 26 July 1989; accepted

14 March 1990)

ABSTRACT Reversed micelles of AOT and Ci2E04 are used as mobile carriers to transport metal ions between two liquid-liquid interfaces. For this purpose w/o microemulsions are replacing classical liquid membranes separating two aqueous compartments, one “source” compartment containing the solute to be transported and one “receiving” compartment in which ion transfer can be monitored. Two types of detection have been used: UV-visible absorption (picrate transport) and atomic absorption ( Ni2+ transport). Different parameters influencing the transport rate are examined: nature and composition of the microemulsion, nature of the salt (anion effect), pH of the receiving phase. The transport mechanism is shown to be different for the Ci2E04 and the AOT systems and models are proposed for the interfacial transfer step. Coupled transport experiments are also reported in which two different carriers are operating in the liquid membrane: reversed micelles and extractant molecules (crown-ethers or Kelex 100). Synergy effects for metal ion transport have been found in some cases but not in others. These results are discussed in terms of diffusion of the carriers in the stagnant layers. They provide useful guides for further investigations concerning the development of liquid membrane separation processes based on microemulsion systems.


A liquid membrane is usually a simple organic solvent separating two immiscible liquid phases, generally of aqueous nature [ 11. Sometimes the liquid membrane is impregnated on a solid porous support (those are the so-called “supported liquid membranes”), in other circumstances, as for instance in the experiments reported here, “bulk liquid membranes” are used [ 2,3]. Transport of chemical compounds through liquid membranes is not only important in biological systems [ 4-121 but it is also playing an increasing part in the development of practical separation processes [13-E]. It usually re*Author to whom correspondence


should be addressed.

0 1990 -

Elsevier Science Publishers



quires the presence of a mobile carrier (macrocyclic compounds or lipophilic complexing agents for instance) solubilized in the organic phase, which facilitates the diffusion of a particular solute in the concentration gradient. In some specific instances, the transport of the particular solute may be driven by the back transport of another species [ 161. Over the last years a certain amount of work has been done in our laboratory, aimed at demonstrating that microemulsion droplets also behave as carrier agents, capable of transferring substances from one liquid-liquid interface to another one [ 17-211. To make possible such experiments we resorted to biphasic systems consisting of either an o/w microemulsion in thermodynamic equilibrium with an external oil phase (Winsor I system) or a w/o microemulsion in equilibrium with an external water phase (Winsor II system). The designation of these systems according to Winsor’s definitions [22] is very commonly adopted. Our interest in this paper will be focused on the transport of metal ions by w/o microemulsion droplets. Contrary to the case of macrocyclic carriers, metal ion transport by w/o microemulsions is not selective, but it is much faster. From a practical point of view, the coupling between selective extractants and microemulsion carriers may thus lead to situations in which one takes advantage of both a selective (due to the extractant) and a fast (due to the microemulsion) transport of metal ions. This would overcome the major drawback of bulk liquid membranes for commercial development, i.e., the very slow transport. Besides this applied aspect, liquid membrane transport studies using a w/o microemulsion in place of the classical organic phase are also interesting from a fundamental point of view: the rate of transfer and the transfer mechanism which is operative, can provide information on the structure of the microemulsion phase and on the behavior of the surfactant film at the liquid-liquid interfaces. The data reported here concern the two aspects just mentioned. EXPERIMENTAL

Chemicals and biphusic systems The surfactants tetraethyleneglycol dodecylether (C,,EO,) and sodium bis(2-ethyl-hexyl) sulfosuccinate (AOT) were obtained from Nikko Chemicals (Japan) and from Sigma, respectively. They were used as received. n-Decane (purum quality) and hexanol-1 (puriss) were bought from Fluka. Potassium picrate was prepared as in previous publications [ 18,201. The salts used were of analytical grade. Kelex 100 (Schering, F.R.G. ) was purified by chromatographic separation [ 231 to get 7- (4-ethyl-1-methyloctyl) -&hydroxyquinoline (C,,-HQ ) . Deionized, doubly-distilled water was utilized throughout.


The Winsor II biphasic systems were allowed to separate in a thermostatted bath regulated at 20’ C. The composition of the separated phases can be found in preceding publications for the two systems used: C,,E04/n-decane/hexanol-l/water [18] and AOT/n-decane/water (0.25 M KBr) [21]. The concentration of added substances was always kept low enough so as not to visibly alter the initial phase equilibrium. Techniques Two different kinds of setup have been used for the liquid membrane transport measurements depending on whether UV-visible spectrophotometric detection or atomic absorption detection was considered. In the former situation (detection of picrate transport) the absorption change of the receiving phase with time was monitored in a continuous way, whereas in the latter case (detection of nickel transport) samples were withdrawn at controlled time intervals. The first setup has been fully described previously [ 18,191. A schematic representation of the transport cell and of the sampling device used in the second case is shown in Fig. 1. The microemulsion phase M is in contact with the water excess phase in both arms of an inverted U-tube. In the left arm is the source phase S in which is solubilized the substance to be transported. In the right arm is the receiving phase. The liquid samples are thrown out by moving a plunger so as not to perturb the liquid-liquid interfaces. The diluted samples were analyzed for nickel content with a Varian AA-1275 atomic absorption spectrophotometer. In the case of UV-visible detection the sampling device shown in Fig. 1 was simply replaced by a circulating spectrophotometric cuvette. In fact one should be aware that two different transport cells were utilized. Since they were not

Fig. 1. Schematic drawing of the setup used for the transport experiments with atomic absorption detection. M = microemulsion liquid membrane, S = source aqueous phase, R = receiving aqueous phase.


strictly identical, comparisons of data are meaningful only when obtained with the same cell. The plots giving the concentration of solute transported in the receiving compartment versus time are usually characterized by a time lag, which corresponds to the time necessary for the liquid membrane (microemulsion phase) to reach an equilibrium state, and by a linear portion (steady-state). The slope of the straight line gives the transfer rate. The flux is expressed in moles transported per unit of time and per unit of area of the interface (3.14 cm2). The water content of the microemulsion phase was determined by Karl Fischer titration. Conductivity measurements were performed with a Wayne Kerr B 331 autobalance precision bridge (IX= lo4 rad 8-l ) equipped with a microelectrode Tacussel CM 05.55 G. The cell constant was redetermined prior to each series of measurements. All experiments were run at 20” C. RESULTS AND DISCUSSION

A. Transport of potassium picrate by AOT reversed micelles In the presence of a large concentration of salt, mixtures of AOT, decane and water separate into two clear phases. The upper phase contains the AOT reversed micelles which behave as carrier agents for the transport of potassium picrate, the lower phase being constituted of excess water. The quantity of water incorporated in the upper phase can be varied by changing the amount of AOT in the initial mixture. As shown in Fig. 2, the flux of potassium picrate goes through a maximum when the water content in the microemulsion phase is increased. There is thus an optimum composition for fast transfer rate. A flux maximum was also observed for the transport of picrates by w/o microemulsions made of C,2E04/alcohol/decane/water [ 18,201, but the physical significance of this maximum appears to be different in the two cases considered. For the C,,E04/hexanol system the origin of the maximum was attributed to a structural change of the dispersed phase from simple hydrated aggregates to real droplets with water pools. For the AOT system, the viscosity behavior and the variation of the water diffusion coefficient of the microemulsion phase suggested possible cluster formation at volume fraction of dispersed phase &, larger than 0.2 [ 211. The conductivity change which we have also represented in Fig. 2 is in agreement with this explanation. The conductivity values are always largely inferior to 1 $S cm-‘. Whereas, in the case of droplet percolation the conductivity should increase by orders of magnitude, on the contrary it levels off for Qr, larger than 0.2 (i.e., above the flux maximum). This is likely to be the result of two antagonistic effects: first the increase of


Fig. 2. Transport of picrate by AOT reversed micelles. Flux of picrate transported ( 0, left scale) and conductivity of the microemulsion phase ( + , right scale) versus percent water in the liquid membrane. Initial potassium picrate concentration: 1.87. 10m3 M.

droplet concentration and, second, cluster formation producing a decrease of the droplets mobility. Figure 2 also shows that at low water content the flux values in the microemulsion phase extrapolate at a finite value, whereas extrapolation to zero was found for the C,,EO, system. Another noticeable difference between the C,,EO, and the AOT systems was relative to the time lag preceding the establishment of the steady-state transport across the liquid membrane. At fixed composition of the microemulsion, with the first mentioned system the time lag was increasing when the initial concentration of picrate in the source compartment decreased (see Fig. 3 in Ref. [ 181)) whereas with the second one the time lag was independent of the initial concentration (see Fig. 2 in Ref. [ 211) .

B. Transport ofNz? by AOT reversed micelles We had used so far picrates as an indirect means to detect metal ion transport, for the evident reason of easy detection. With the new device described in this paper we can directly measure the transport of metal ions, and Ni2+ was chosen because of its interest in hydrometallurgy (see also below, coupled transport in the presence of extractants). Due to the ionic character of AOT, and although Winsor II systems were obtained in this case in the presence of a large excess of 1: 1 electrolyte, ion exchange may affect the transport process as demonstrated by the pH effect reported below. The possible influence of the anion associated with Ni2+ has also been checked. For all the experiments the composition of the microemulsion system was chosen so as to be close to the optimum flux (Fig. 2)) but slightly before it in order to avoid cluster formation. The initial mixture contained (in weight percent): AOT (7) /decane (4)/water,

0.25A4 1:1 electrolyte (53) and the microemulsion phase itself (after phase separation) contained 8.47% of water. Effect of changing the pH of the receiving phase The source phase contained nickel nitrate 4*10m2 M. Figure 3 shows the change of Ni2+ concentration in the receiving compartment (in mg 1-l ) versus time for different values of pH at the beginning of the experiment. The proton concentration is changing with time in both compartments due to countertransport. The results can be interpreted as indicating that the de-extraction step at the second liquid-liquid interface is facilitated by Ni2+/H+ exchange on the polar heads of AOT. The pH change in the water phase is expected to affect the water content of the microemulsion phase as would be the case with any added cation [ 241. In fact, due to the high concentration of background salt, the change of water content remains small enough not to perturb the liquid-liquid equilibrium in the transport cell. We found from Karl Fischer titration that, at final equilibrium (which is far from being attained in these experiments) the amount of dispersed water increased by about 6% of its initial value when fixing the water pH at 1.0.

Fig. 3. Transport of Ni’+ by AOT reversed micelles. Concentration of Ni’+ in the receiving phase versus time for different values of pH: (0 ) 6.44, (0 ) 5.0; (A ) 4.0; (0 ) 1.05 and ( + ) 0.5. Initial nickel nitrate concentration: 4. lo-’ M.


Effect of the nature of the-anion associated with NP Different nickel salts have been dissolved in the source phase (always at a concentration of 4010-~ M) and the results obtained are represented in Fig. 4. They show that the flux is insensitive to the nature of the anion whatever chloride, nitrate, sulfate, perchlorate or acetate. One should not forget that the water phase contained in this case 0.25 M KBr. The large excess of bromide may thus be responsible for the absence of any effect. For this reason we have also measured the transport of Ni (ClO,), replacing KBr by NaClO, and the transport of NiC12 replacing KBr by NaCl. The results are shown in Fig. 5. In this case only one anionic species (either ClO, or Cl- ) is present in the system in addition to AOT itself. The fluxes obtained are, again, perfectly identical. C. Transport of N?+ by C,,EO, reversed mice&s With the C,,EO, system we get rid off the problem related with the presence of a background salt. Consequently, the interpretation of the effect of the nature of the anion associated with Ni2+ should be unambiguous. The results are represented in Fig. 6 for the same anion series as for the AOT system (Fig. 4). A tremendous effect of the anion can be seen in this case, with a flux approximately 45 times as large for ClO, (fastest transferring anion) than for SOi- (slowest transferring anion). As already pointed out [ 251, if the anions









time (mid 100




Fig. 4. Transport of NP by AOT reversed micelles. Concentration of NP in the receiving phase versus time, varying the nature of the anion: (0 ) NO, ; (0 ) C10~ ; (0 ) Cl-; (A ) SOi- ; ( + ) acetate. Initial concentration of nickel salt: 4.lo-’ M.

250 ml CNi~*(ppm)


t ime (min) 300



Fig. 5. Transport of Ni’+ by AOT reversed micelles. Concentration of Ni2+ in the receiving phase versus time. (0 ) transport of Ni (C101)2r background salt 0.25 A4 NaClO,; (0 ): transport of NiC12,background salt 0.25 M NaCl.


are classified according to the flux values (see Table 1) the series obtained is reminiscent of the well known Hofmeister (or lyotropic) ion series [ 26,271. D. Coupled transport by reversed micelles and extra&ant molecules Transport of alk&metalpicrates by C,,EO, reversed micelles in the presence of crown-ethers It was previously shown [ 18,191 that the flux of alkali-metal picrates shows a synergy effect when two carriers are operating independently in the liquid membrane, namely the reversed micelles and dicyclohexano-1%crown-6. We said independently because the synergy effect seems to exist only when the two types of carriers do not belong to a single diffusing entity, i.e., when they are allowed to diffuse independently. Table 2 recalls some of the flux values measured, showing that the flux obtained with the coupled system was larger than the sum of the fluxes obtained for the reversed micelles and the crown-ether taken separately. Transport of N? + by AOT reversed micelles in the presence of oxine We have used for these experiments an extra&ant of industrial importance [ 28-311, which was the oxine constituting the major component of Kelex 100: 7- (4-ethyl-1-methyloctyl) -8-hydroxyquinoline, which will be abbreviated C,,HQ. This molecule is a complexing agent for Ni2+ at neutral pH. Decomplex-



- 150




Fig. 6. Transport of N?+ by CrzEO,/hexanol reversed micelles. Concentration of Ni2+ in the receiving phase versus time, varying the nature of the anion: (0 ) NOT ; (0 ) ClO; right scale; ( 0 ) Cl-; ( A ) SO:- ; ( + ) acetate. Initial concentration of nickel salt: 4. lo-’ M. TABLE 1 Transfer rate and flux of N?+ transported by ClzEO,/hexanol reversed micelles depending on the associated anion Anion

Transfer rate (ppm min-‘)

Flux” (mol cmm2min-’ )







ClAcetate so:-

0.0373 0.0341 0.0966

3.24*10-’ 2.96.10-’ 6.94.10-i’

“The flux calculation takes into account the volume of the receiving phase (16 cm3) and the interfacial area (3.14 cm2).

ation takes place in the acid range of pH. The transport process was thus expected to be pH-dependent for a double reason, one coming from the extractant, and the other one from the reversed micelles as we have shown above. Similar experiments, as those reported in Fig. 3, have been performed in the presence of C,,-HQ dissolved in the microemulsion phase at a concentration of 5. 10e3M. The corresponding bundle of curves has not been represented

TABLE 2 Flux of potassium

picrate transported

for varying compositions

of the liquid membrane

Liquid membrane

Flux (mol cm-’ min-‘)

Crown-etheP in decane Microemulsion alone Microemulsionb containing

2.23.10-’ 1.11.10-* 2.81*10-*


[ 191

“Dicyclohexano-18-crown-6. bSystem C,,EO,/hexanol/decane/water.

cNi?‘(ppm) +













, 6



Fig. 7. Coupled transport by AOT reversed micelles and CI1-HQ. Concentration of Ni2+ in the receiving phase after 125 and 400 min as a function of pH: ( 0 ) microemulsion alone; ( 0 ) microemulsion containing 5~10~~ M CI1-HQ; (m) decane containing 5.10P3 A4 C,,-HQ.

because it looks very much like that of Fig. 3. We thought it more interesting to compare in Fig. 7 the concentrations of Ni2+ transported by the AOT droplets in the absence and presence of extractant, respectively. Also represented are some data relative to the transport of Ni2+ by the extractant alone (in this case the microemulsion phase was replaced by decane in which the extractant was solubilized). This was done for different values of the pH initially adjusted


in the receiving phase and for two values of the elapsed time (125 and 400 min). The results show that instead of improving the transport of Ni2+ the presence of C,,-HQ has a negative effect, unless the pH becomes very acidic (pH < 1) . This means that in a large pH range the extra&ant contributes to retain more Ni2+ ions in the liquid membrane. Due to the counter-transport of H+ ions quite a low pH is required in the receiving compartment for the deextraction to be facilited enough to result in a positive effect. The effect of the extractant was not improved by increasing its concentration up to 2. 10m2M, meanwhile decreasing the Ni2+ concentration to 10e2 M. It is thus much less spectacular than in the case of C12E04 reversed micelles with crown-ethers. E. Discussion of possible transport mechanisms There is no unique and simple mechanism able to explain all the preceding data. Different mechanisms are likely to operate depending on the microemulsion system considered and on the nature of the transported solute. We will forget in a first step the case of coupled transport in the presence of extractant molecules which we will discuss at the end. In previous papers of this laboratory [ 17,181, it was demonstrated that solute transport by microemulsion droplets was involving slow diffusion of the colloidal particles in the stagnant layers adjacent to the liquid-liquid interfaces. This can be considered to hold true for the other systems. The differences are likely to be found in the mechanism of solute transfer at the interface between phases. Interfacial transfer mechanism In the case of C,,EO, reversed micelles (uncharged system), the following mechanism was shown to provide a satisfying description of the flux behavior in the case of picrate transport [ 181. First

t MS +







M+ A-



[email protected]


i Second





0 M+A-





+ CDM =









(M+ is the metal ion, A- the associated anion, M+A- an ion pair, the circle symbolizes a colloidal particle; the subscripts S, R, cpM refer to source phase, receiving phase and continuous phase of membrane, respectively. ) This mechanism was postulating the interfacial transfer (from water-continuous to oilcontinuous) of the salt before uptake by a droplet. The effect of the anion (see Fig. 6) brings about additional evidences supporting the above mechanism. The equilibrium constant characterizing the reaction of ion-pair formation which takes place at the liquid-liquid interfaces is indeed expected to depend on the nature of the anion for a defined metal ion. The data are thus in agreement with the predictions of the above model. In the case of AOT reversed micelles several observations have shown a particular behavior suggesting a different mechanism takes place. These observations concern: (i) the absence of anion effect in Ni2+ transport (see Figs 4 and 5); (ii) the insensitivity of the time lag to changes of the initial solute concentration in picrate transport [ 211; and finally (iii) the non-zero value of the extrapolated flux at zero percent water in the microemulsion phase (see Fig. 2). The experimental results shown in Fig. 5 rule out a possible effect of the background salt since the same flux was obtained for Ni ( C1O4)2and NiC12 with no other anion present apart from AOT. The latter is thus responsible for the absence of any anion effect, but there exists several possible reasons, corresponding to different interfacial transfer mechanisms. The simplest explanation would be that the AOT anions (which are in large excess compared to Ni2+), act as counter-ions for Ni2+. In that case a model similar to that described above would still be valid except that Ni2+ would be transferred in the continuous phase of the microemulsion as a (AOT- )2Ni2+ complex. However this ion exchange process is unable to explain the time lag constancy characterizing picrate transport for different initial picrate concentrations. This constancy means that the time necessary for the liquid membrane to reach an equilibrium state is independent of the picrate concentration, which is in contradiction with the type of interfacial transfer postulated above. Indeed, the rate of transfer should increase with the concentration of the species. If it does not, it very likely means that the rate is governed by the droplet concentration which is kept constant in this case (due to the negative charge of picrate ions, transfer in association with AOT cannot be invoked in that case). The preceding remarks have led us to propose another mechanism for explaining the results observed in the case of AOT. It assumes a direct transfer inside the droplets which would be made possible by their opening by coalescence with the interfacial surfactant film: Fig. 8. Such coalescence processes are known to occur between droplets at least when they have an attractive character [ 32 1. Temperature-induced percolation at AOT/water microemulsion droplets have also been reported [ 331. Although, in Winsor II systems, the droplets are known to behave as hard spheres [ 341, transient coalescences


Fig. 8. Schematic view of the assumed coalescence of an AOT droplet with the interfacial surfactant film (W=water, O=oil).

with the interfacial surfactant film cannot be excluded, considering the time scale of the present experiments. Such “pockets” opening up at the interface would be able to pick up any solute present in the source phase, which would explain all the results relative to the AOT system. Of course, depending on the electric charge of the solute to be transported, electrostatic interactions between the solute and the AOT polar heads inside the droplets will have to be considered, but the electrostatic potential is expected to be considerably reduced owing to the presence of a background salt. This 1: 1 electrolyte is in very large excess compared to the Ni’ + concentration. Even if, for this reason, the ion exchange equilibrium at the AOT polar heads might not be so much in favor of the divalent ion, the pH effect clearly shows its importance during the de-extraction step. Similar effects have been reported for microemulsions containing sodium dodecyl sulfate [ 351. A very schematic view of the transport mechanism, operating in the case of AOT (neglecting the electrostatic effects) could thus be as follows: Firstinterface



Diffusion of















1 e



ME -6 Ai

The vertical bar represents the surfactant monolayer which is known to exist at the interface [36,37]. The other symbols have the same meaning as in the previous paragraph. Diffusion in the stagnant luyers in case of coupled transport When a second carrier (extractant molecule) is present in the system we have seen that a synergy effect may show up or not. For what concerns nonionic systems, we have already emphasized the importance of the solubilization site of the extractant molecule [ 191. If it is sol-


ubilized in the droplet both carriers will necessarily diffuse as a single entity. In this case it is unlikely to find a significant synergy effect. On the other hand, when the extractant is preferentially solubilized in the continuous phase (this has of course to be evaluated in terms of the partition coefficient), the two carriers can diffuse independently and synergy effects become more likely for the following reasons. A factor of lo-100 exists between the diffusion coefficients of the small extractant molecules and that of the droplets. Fundamentally, the extractant molecules are thus expected to move in the stagnant layers much faster than the droplets, so that they should reach the second liquidliquid interface first. After it has released the transported solute in the receiving phase, the free extractant molecule will be driven back to the first interface by the concentration gradient. On the way back there may be an exchange of solute between the charged droplets and the free extractants. In that case the concentration gradient will force the empty droplet to move in the direction of the first interface before it has reached the second one, whereas the charged extractant will move in the opposite direction. This is certainly an oversimplified picture, but it suggests that, when the two carriers are independent, the effective thickness of the diffusion layers may be reduced for both carriers. This would explain the increase of the flux, which is inversely proportional to the thickness of the diffusion layers [ 17,181. In the case of the AOT&-HQ system the situation is even more complex due to the fact that both the surfactant and the extractant bear electric charges. The lack of significant synergy effect may indicate, as suggested by the preceding case, that both carriers diffuse as a single entity. The extractant in its neutral form is highly soluble in decane, but the ionized forms are tensioactive and may prefer a solubilization in the surfactant layer of the AOT droplets. The exact ionization state is very difficult to ascertain due to the existence of local pH’s [ 38,391. In the vicinity of the AOT polar heads, the pH is expected to be even more acidic than in the water core, leading undoubtedly to the protonation of C1-HQ. This is likely to be too simple an explanation. The major drawback of C,,HQ compared to the crown-ether case previously investigated is the requirement for a low pH during the stripping step (not speaking of the flux dependence with pH for Ni2+ transport by the AOT droplets alone). Due to proton counter transport the pH gradient decreases during the time necessary for a transport experiment. This, in addition to the NP/H+ exchange process on the AOT polar heads, may be responsible for the absence of any synergistic effect, except at very low pH. CONCLUSION

In conclusion, we have shown that the mechanism of metal ion transport between two liquid-liquid interfaces through microemulsion liquid mem-


branesmay considerablydepend on the type of systemconsidered.Much effort will still have to be made before liquid membraneprocesses can be optimized by the use of microemulsionscoupled with selectiveextractants.The present work was intendedto indicatesome of the principlesto take into consideration. It is hoped to provide some useful guidesfor furtherinvestigationsin this field. ACKNOWLEDGMENTS

We are indebtedto J.L. Vasseur,L. Leclerc and J, Clementfor theirtechnical contributionsto the realizationof the setup shown in Fig. 1. REFERENCES 1 2 3 4 5 6

R.D. Noble and J.D. Way (Eds), ACS Symp. Ser., 347 (1987). S. Schlosser and E. Kossaczky, J. Radioanal. Nucl. Chem., 101(1986) 115. R.M. Izatt, J.D. Lamb and R.L. Bruening, Sep. Sci. Technol., 23 (1988) 1645. H.L. Rosano, P. Duby and J.H. Schulman, J. Phys. Chem., 65 (1961) 1704. R. Ashton and L.K. Steinrauf, J. Mol. Biol., 49 (1970) 547. W.I. Higuchi, A.H. Ghanem and A.B. Bikhazi, Fed. Proc. Fed. Am. Sot. Exp. Biol., 29 (1970) 1327. 7 K.H. Wong, K. Yagi and J. Smid, J. Membr. Biol., 18 (1974) 379. 8 J.D. Lamb, J.J. Christensen, S.R. Izatt, K. Bedke, M.S. Astin and R.M. Izatt, J. Am. Chem. Sot., 102 (1980) 3399. 9 C.F. Reusch and E.L. Cussler, AIChE J., 19 (1973) 736. 10 Y. Kobuke, K. Hanji, K. Horiguchi, M. Asada, Y. Nakayama and J. Furukawa, J. Am. Chem. Sot., 98 (1976) 7414. 11 M. Kirch and J.M. Lehn, Angew. Chem. Int. Ed. Engl., 14 (1975) 555. 12 E. Pefferkorn and R. Varoqui, J. Coloid Interface Sci., 52 f 1975) 89. 13 N.N. Li and A. Shrier, in N.N. Li (Ed.), Recent Developments in Separation Science, Vol. 1, CRC Press, Cleveland, OH, 1972, p. 163. 14 N.N. Li, Ind. Eng. Chem., Process Des. Dev., 10 (1971) 215. 15 N.N. Li, AIChE J., 17 (1971) 459. 16 R.A. Bar&h, W.A. Charewicz, S.I. Kang and W. Walkowiak, ACS Symp. Ser., 347 (1987) 86. 17 A. Xenakis and C. Tondre, J. Phys. Chem., 87 (1983) 4737, 18 C. Tondre and A. Xenakis, Faraday Discuss. Chem. Sot., 77 (1984) 115. 19 A. Xenakis, C. Selve and C. To&e, Talanta, 34 (1987) 509. 20 A. Xenakis and C. To&e, J. Colloid Interface Sci., 117 (1987) 442. 21 A. Derouiche and C. Tondre, J. Chem. Sot. Faraday Trans. 1,85 (1989) 3301. 22 P.A. Winsor, Trans. Faraday Sot., 44 (1948) 376. 23 C. Tondre and M. Boumezioud, J. Phys. Chem., 93 (1989) 846; M. Boumezioud, P. Lagrange and C. To&e, Polyhedron, 7 (1988) 513. 24 E.B. Leodidis and T.A. Hatton, Langmuir, 5 (1989) 741. 25 C. Tondre and A. Derouiche, J. Phys. Chem., 94 (1990) 1624. 26 P. Firman, D. Haase, J. Jen, M. Kahlweit and R. Strey, Langmuir, 1 (1985) 718. 27 W.A.P. Luck, Top. Curr. Chem., 64 (1976) 113. 28 J.A. Ha&age, Trans. Sot., Min. Eng. AIME, (1969). 29 A.W. Ashbrook, Coord. Chem. Rev., 16 (1975) 285.

258 30 31 32 33 34 35 36 37 38 39

A. Leveque and J. HeIgorsky, Proc. Int. Solv. Extr. Conf. (1977) CIM Special Volume 21, 439. G.P. Demopouios and P.A. Distin, Hydrometallurgy, 11 (1983) 389. P.D.I. Fletcher, A.M. Howe and B.H. Robinson, J. Chem. Sot. Faraday Trans. 1,83 (1987) 985. S. Geiger and H.F. Eicke, J. CoIIoid Interface Sci., 110 (1986) 181. M.-J. Hou and D.O. Shah, Langmuir, 3 (1987) 1086. F.J. Ovejero-Escudero, H. Angelino and G. Casamatta, J. Disp. Sci. Technol., 8 (1987) 89. D. Guest and D. Languevin, J. Colloid Interface Sci., 112 (1986) 208. L. Tenebre, G. Haouche and B. Brun, in K.L. Mittal and P. Bothorel (Eds), Surfactants in Solution, Vol. 6, Plenum Press, New York, 1986, p. 1383. E. Bardez, E. Monnier and B. Valeur, J. Phys. Chem., 89 (1985) 5031. E. Bardez, B.-T. Goguillon, E. Keh and B. VaIeur, J. Phys. Chem., 88 (1984) 1909.