Anal. Chem. 2000, 72, 1328-1333
Permeation Liquid Membrane Metal Transport: Studies of Complex Stoichiometries and Reactions in Cu(II) Extraction with the Mixture 22DD-Laurate in Toluene/Phenylhexane Franc¸ ois Guyon, Nalini Parthasarathy, and Jacques Buffle*
CABE, Department of Inorganic, Analytical and Applied Chemistry, University of Geneva, 30 quai E. Ansermet, CH-1211 Geneva 4, Switzerland
The role of lauric acid (LAH) in the transport of copper(II) through a permeation liquid membrane (PLM) comprising 1,10-didecyldiaza-18-crown-6 (22DD) and lauric acid (ratio 1:1) in 1:1 v/v toluene/phenylhexane has been investigated by determining the stoichiometry of metal extraction and of the metal complex formed in the organic phase by performing 1H NMR and liquid/liquid and liquid/membrane extraction measurements. In the absence of copper(II), the 1H NMR data suggest that there is a strong interaction between the proton of LAH and the nitrogen of the 22DD macrocycle but no interaction between the aliphatic long chains of LAH and 22DD. Thus, in the organic solution, the two compounds are associated as (22DD-H)+-LA-, the laurate being away from (22DD-H)+. The signal intensity of the acidic proton was found to decrease when the metal Pb(II) was incorporated by the carrier after its extraction from the aqueous phase. Additionally, liquid/liquid as well as liquid/ membrane extraction results reveal that Cu(II) extraction proceeds via the loss of two protons from the organic phase. The Cu(II) is found to be located in the 22DD cavity and the stoichiometry of the complex in the organic phase is (22DD-Cu)2+-2LA-. Metal extraction is governed by 22DD and laurate acts only as counteranion. An unexpected feature was observed in the liquid/liquid extraction which was that, at low 22DD and LAH concentrations, the slope for log(Kp) ) f(pH) was 2 whereas it was much lower at high carrier concentration. This unexpected result seems to stem from impurities present in 22DD: only 0.1 mol % of impurity can indeed influence the exchange ratio of Cu(II) and H+. This type of anomaly, however, is not found in the normal procedure of liquid/ membrane extraction possibly due to the lower carrier/ metal molar ratio which is used in the classical PLM conditions. Permeation liquid membranes (PLM) for metal ion transport consist of a hydrophobic membrane, impregnated with an organic solvent containing a metal complexing carrier, separating a source solution containing the target species and a feed solution with a complexing agent stronger than the carrier. The attractive feature of this technique is the simultaneous separation and preconcen1328 Analytical Chemistry, Vol. 72, No. 6, March 15, 2000
tration of the test metal ions and the selectivity due to a tailormade carrier for a given element. This technique has been widely used for organic compound separation and wastewater treatment.1-6 However, few applications of this technique have been made for trace metal analysis in natural water,7 and very few studies have been performed on trace metal speciation. A PLM system has been developed for trace metal speciation under natural water conditions (in particular neutral pH in the source and feed solution) comprising a Celgard membrane impregnated with an equimolar mixture of 1,10-didecyl-1,10-diaza-18-crown-6 ether (22DD) and lauric acid (LAH) in solution in toluene/phenylhexane (1:1 v/v).8-12 Although the transport properties of this membrane has been studied, no study has been reported up to now on the mechanism of metal exchange at the source/solvent interface. This is essential for improving long-term stability and designing new tailor-made selective carriers for other trace metal or organic compounds of environmental importance. Previous study of this system showed that Cu(II) transport is aided by a sodium countergradient and that the transport-limiting step is the diffusion through either the membrane or both the membrane and the aqueous Nernst layer, depending on the membrane thickness and stirring rate.13 Previous studies on Cu(II) transport across this PLM showed that practically no transport occurs when 22DD alone is used as the carrier whereas in the presence of LAH almost all the Cu(II) (1) Tsukube, H. In Liquid membrane application; Arachi, T., Tsukube, H., Eds.; CRC Press: Boca Raton, FL, 1990. (2) Danesi, P. R. Sep. Sci. Technol. 1984-85, 19, 857. (3) Jonsson, J. A.; Mathiasson, L. Trends Anal. Chem. 1992, 11, 106. (4) Danesi, P. R. Sep. Sci. Technol. 1987, 261 (Science and Application of SLM, Proceedings of the 58th Engineering Foundation Conference; Li, N., Strathamann, H., Eds.). (5) Chiarizia, R.; Horwitz, E. P.; Rickert, P. G.; Hodgson, K. M. Sep. Sci. Technol. 1990, 25, 1571. (6) Bartsch, R. A.; Way, J. D. Chemical separation with liquid membranes: an overview; Bartsch, R. A., Way, J. D. Eds.; ACS Symposium Series 642; American Chemical Society: Washington, DC, 1996; Chapter 1. (7) Djane, N. K.; Ndung’u, K.; Malcus, F.; Johansonn, G.; Mathiasson, L. Fresenius J. Anal. Chem. 1997, 358, 822. (8) Parthasarathy, N.; Buffle, J. Anal. Chim. Acta 1991, 254, 1. (9) Parthasarathy, N.; Buffle, J. Anal. Chim. Acta 1991, 254, 9. (10) Parthasarathy, N.; Buffle, J. Anal. Chim. Acta 1994, 284, 649. (11) Parthasarathy, N.; Pelletier, M.; Buffle, J. Anal. Chim. Acta 1997, 350, 183. (12) Parthasarathy, N; Buffle, J.; Gassama, N.; Cue´noud, F. Chem. Anal. (Warsaw) 1999, 44, 455. (13) Guyon, F.; Parthasarathy, N.; Buffle, J. Anal. Chem. 1999, 71, 819. 10.1021/ac991047+ CCC: $19.00
© 2000 American Chemical Society Published on Web 02/16/2000
is transported across the membrane at neutral pH.10 This difference in behavior was attributed to the higher partition coefficient resulting from the presence of lipophilic counteranion in the organic phase to form neutral metal complex. To better understand the role of laurate in the Cu(II) transport across PLM, the nature of the complexes formed in the liquid membrane and the extraction stoichiometry have to be studied. In this paper, the pH and lauric acid concentration effects on the Cu(II) distribution ratio between the source solution and the PLM have been studied by means of liquid/membrane (Liq/membrane) and liquid/liquid (Liq/Liq) extractions. 1H NMR studies have also been made to evaluate LAH/22DD and metal/carrier interactions. The combination of these results provide information on metal/carrier complex stoichiometry, the metal location in the complex, and the electroneutrality balance. MATERIALS AND METHODS 1. Reagents. All the chemicals used were reagent grade, i.e., 2-(N-morpholino)ethanesulfonic acid (MES, Sigma), trans-1,2diaminocyclohexane-N,N,N′,N′-tetraacetic acid monohydrate (CDTA, Fluka), copper(II) and lead(II) nitrate (Merck), lead(II) acetate (Merck), sodium hydroxide (Merck), lithium hydroxide (Merck), lauric acid, LAH (Fluka), and 1,10-didecyl-1,10-diaza-18-crown-6 ether (Kryptofix 22DD > 99%, Merck). The solvents (toluene, >99.5% and phenylhexane, ∼97%) were Fluka products. NMR solvents, CDCl3, D2O, and CD3OD, were purchased from Dr Glaser AG. MilliQ water was used for preparing all the aqueous solutions. Membrane. Celgard 2500 (Celanese Plastic, Charlotte, NC) polypropylene hydrophobic membrane (porosity Θ, 0.45; thickness l, 25 µm; pore diameter, 0.04 µm i.d.) was used as the liquid membrane support. Apparatus. Copper(II) and sodium(I) concentrations were measured by flame atomic absorption spectrometry (AAS) using a Pye Unicam SP9 spectrophotometer. NMR spectra were run on a Varian Gemini 300 (300 MHz). 2. Liquid/Membrane Experiments for Cu(II) Extraction. Distribution coefficients were determined using two methods: Liq/Liq extractions and Liq/membrane extractions. In both cases, the source solution was 10-1 M MES buffer solution. The organic phase consisted of a mixture of toluene/phenylhexane (1:1 v/v) containing various ratios of lauric acid/22DD. In most experiments, the Cu(II) (5 × 10-5 mol L-1) and 22DD (5 × 10-2 mol L-1) concentrations were kept constant. Extraction experiments were carried out (a) by keeping the lauric acid concentration constant and varying the pH from 4.5 to 6.9, the pH being adjusted with LiOH, and (b) by varying the LAH/22DD ratio in the range 0-4, keeping pH constant at 6.0. A single Celgard membrane (area ∼6.4 cm2) impregnated with the organic solution, rinsed with milliQ water to remove the solvent excess, was placed in a beaker containing 40 mL of Cu(II) aqueous solution, and it was gently shaken overnight to ensure equilibrium. The membrane turns blue due to the extraction of copper(II) in the organic phase. After rinsing it with water, the membrane was dipped in 10-3 M CDTA solution (pH 6.2 adjusted with NaOH) to back-extract the copper(II). The Cu(II) concentrations in the membrane, [Cu]m, and in the sample before ([Cu]i), and after ([Cu]aq) extraction were determined by AAS. In all the experiments, the copper(II) mass balance was checked using,
Ni ) Nm + Naq, where Ni, Nm, and Naq are respectively the initial, membrane, and aqueous copper(II) number of moles. The distribution coefficient, KD, is defined as
KD ) [Cu]m/[Cu]aq
(1)
3. Determination of the Partition Coefficient by Liquid/ Liquid Extractions. The experimental conditions are similar to those described above for Liq/membrane extractions. A 2 mL aliquot of the organic phase was added to 2 mL aliquot of the Cu(II) aqueous solution placed in a polypropylene tube. After the mixture was shaken for 2 h, it was centrifuged (4000 rpm) for 30 min. An aliquot of the aqueous solution was withdrawn by piercing a hole through the bottom of the tube with the needle of a syringe to prevent any entrainment of the organic-phase solution. The copper(II) concentration was determined in the aqueous solution before and after extraction, and the partition coefficient, Kp, was computed from
Kp ) [Cu]org/[Cu]aq
(2)
where [Cu]org, the Cu(II) concentration in the solvent, is given by [Cu]org ) (Ni - Naq)/Vorg, Vorg being the volume of the organic solution. 4. 1H NMR Spectroscopy. The interactions between 22DD, lauric acid, and metal in nonaqueous solvent were investigated by means of 1H NMR. The spectra were run in deuterated chloroform (CDCl3); the observed chemical shifts are reported in Table 1. A 1H COSY NMR was run to assign the chemical shifts to the 22DD protons in the mixture 22DD/LAH. NOEDIFF experiments were performed to detect 1H-1H proton coupling in the space. 1H NMR were run to elucidate the stoichiometry of the complex formed between 22DD and LAH (in CDCl3) as well as 22DD and Pb(OAc)2 (in CD3OD). Instead of Cu(II), lead(II) ion was used as the extracted metal as it is diamagnetic. Since Pb(II) is transported through the PLM10 at a rate similar to that of Cu(II), it is expected that it forms with the carrier a complex similar to that of Cu(II). RESULTS 1. Determination of Interactions between Lauric Acid and 22DD by 1H NMR. NMR has been used for studying metal crown ether complexes,14-17 and in this paper this technique has been used to study the interactions between LAH/22DD and Pb(II)/carrier. In deuterated chloroform, the chemical shift of H1 and H2 proton types of the 22DD and the acidic proton HA of LAH are located at 3.60, 3.60, and 10.9 ppm, respectively (Table 1). A homonuclear shift correlated 2D NMR study (COSY) was performed to assign each signal of the spectrum for an equimolecular mixture of 22DD and LAH as HA is shifted downfield at 4.6 ppm and H2 upfield at 3.65 ppm. Only chemical shifts of protons H2, H3, and H4 are affected by the addition of LAH to 22DD, (14) Dietrich, B.; Lehn, J. M.; Sauvage, J. P. Tetrahedron Lett. 1969, 34, 2889. (15) Popov, A. I.; Lhen, J. M. in Coordination Chemistry of Macrocyclic Compounds; Nelson, G. A., Ed.; Plenum Press: New York, 1979; Chapter 9, p 537. (16) Andre´s, A.; Bazzicalupi, C.; Bencini, A.; Bianchi, A.; Fusi, V.; Garcia-Espan ˜a, E.; Paoletti, P.; Valtancoli, B. Inorg. Chem. 1994, 33, 617. (17) Hiratani, K.; Takahashi, T.; Sugihara, H.; Kasuga, K.; Fujiwara, K.; Hayashita, T.; Bartsch, R. A. Anal. Chem. 1997, 69, 3002.
Analytical Chemistry, Vol. 72, No. 6, March 15, 2000
1329
Table 1. Proton Chemical Shift of the 22DD, Lauric Acid, and Lead(II) Acetate and the Mixtures (in ppm) at 298 K
confirming that only the nitrogen atoms of 22DD interact with HA and not the oxygen atoms. At a ratio 22DD/LAH ) 1:2, HA is shifted downfield, indicating that this proton becomes more acidic. These results confirm that for the ratio 1:1 the proton HA of lauric acid is delocalized between the two nitrogen. To determine whether interactions of 22DD and LAH may occur through aliphatic long chains, NOEDIFF experiments were performed. This technique allows the detection of proton/proton spatial interactions such as interactions between the protons of LAH and 22DD aliphatic chains. Under these experimental conditions, no such correlation was observed. This result may be interpreted as follows: either interaction is not detected by this technique because the gap between the two long chains is larger than 3 Å or, after the complexation of the LAH proton by 22DD, the laurate moiety does not remain close to (22DD-H)+ and acts only as a free counterion. These results suggest that, in absence of metal, 22DD and LAH are present in the solvent as (22DDH+)(LA-). The association of H+ with the macrocycle may make easier the exchange between H+ and Cu(II) (or other metals) at the interface. 2. Lead/Carrier Interaction. To determine the metal/carrier stoichiometry, 1H NMR spectroscopy was used. The experiments were performed in deuterated methanol because of the low solubility of lead(II) acetate in chloroform. The spectral changes with increasing [Pb(II)]/[22DD] ratio (0 < [Pb(II)]/[22DD] E 1) are shown in Figure 1. At a 1:1 ratio corresponding to the stoichiometry of the Pb/22DD complex, a maximum intensity for 22DD/Pb is observed and the protons signal of free 22DD disappears. For [Pb(II)]/[22DD] > 1, no further spectral modification is observed, indicating the formation of only a 1:1 Pb(II)/22DD complex in solution. The fact that two distinct sig1330 Analytical Chemistry, Vol. 72, No. 6, March 15, 2000
Figure 1. 1H NMR spectra of 22DD at various lead(II) acetate concentrations in deuterated methanol. Pb(II)/22DD ratio: (a) 0; (b) 0.5; (c) 1. *, solvent peaks.
nals corresponding to the free 22DD and the complex Pb(II)/ 22DD can be observed indicates that the 1:1 Pb(II)/22DD complex is highly stable; the high stability constant prevents any fast exchange to be observed within the NMR time scale. This is in agreement with previous results which show that log β(Cu2+/22DD) ) 8, where β is the stability constant defined as β ) [ML]/[M][L] (L ) 22DD).8 The chemical shifts of H5, H6, and
H7 proton types (Table 1), far away from the macrocycle cavity, are not modified and those of the H1, H2, and H3 proton types are shifted downfield. It has been shown that the oxygen atoms of 22DD do not interact with the proton of LAH. They do however participate in metal complexation as the chemical shift of the protons close to the oxygen is modified. Thus, these experiments confirm that the stoichiometry of the extracted metal complex is one metal to one 22DD and that the metal is complexed by the nitrogen and oxygen atoms in the macrocycle cavity, which is in accordance with the results reported for similar complexes.18,19 Metal carrier complexation in the presence of both 22DD and LAH was also studied by 1H NMR. In a NMR tube, 0.2 mL of D2O was poured on 0.7 mL of CDCl3 solution containing 10-3 M 22DD and LAH (1:1 ratio), and then spectra were run for 1 h to detect any exchange between the acidic proton in the organic phase and the deuterium of D2O, which might level off its signal. As no change occurred, solid Pb(NO3)2 was added to the deuterated aqueous solution. After 5 min, the signal corresponding to the acidic proton from the 22DD/LAH interaction was half of its initial value, and after 50 min, it disappeared completely. This observation clearly indicates that the uptake of metal is accompanied by an exchange of the protons from the organic phase. 3. Determination of Cu(II) Distribution Coefficient during Liquid/Membrane Extraction. By considering the results in sections 1 and 2, the electroneutrality requirement and the fact that a side reaction of metal/carrier complex adsorption on polypropylene has been observed,13 Cu(II) extraction reaction in the membrane can be written as
Figure 2. Effect of pH on membrane/liquid extraction of 5 × 10-5 mol L-1 Cu(II) in MES buffer by 5 × 10-2 mol L-1 22DD and LAH (ratio 1:1) in 1/1 phenylhexane/toluene mixture. Inset: Plot of log(KD) vs pH.
The extraction coefficient, Kex, is also defined as
Kex )
[22DDCu(LA)2]org[H+]2aq [22DD]org[LAH]2org[Cu]aq
(6)
Combining eqs 1 and 4-6 and rearranging gives
Kex )
KD[H+]2aq
1 ′,MC [22DD]org[LAH]2org Kads +1
(7)
Equation 7 can be also transformed by considering that not only the Cu/carrier complex but also LAH and 22DD adsorb on the ′,LAH polypropylene supports with apparent constants Kads and ′,22DD Kads . The expression of KD is then given by eq 8:
log(KD) ) 2 log([LAH]m) + log([22DD]m) + 2pH + ′,MC ′,22DD log(Kex) + log(Kads + 1) - log(Kads + 1) -
where 22DDCu(LA)2,p is the carrier metal complex adsorbed on the polypropylene membrane. KMC ads is the corresponding adsorption constant.13 Assuming that the concentration of adsorption sites is much larger than that of the adsorbed species, X, the adsorption constant, KXads can be expressed as
KadsX ) ΓX/[X]org
(4)
where ΓX is the surface concentration of X on the polypropylene support. Since the surface area, A, of the polypropylene support inside the membrane, and the volume V of solvent are constant, ′,X the apparent adsorption constant Kads can also be defined as
[X]p [X]m - [X]org A ′,X Kads ) KXads ) ) V [X]org [X]org
(5)
where [X]p and [X]m are the concentrations of X on the polypropylene support and in the whole membrane, expressed as number of moles per volume of solvent. (18) Herceg, M.; Weiss, R. Inorg. Nucl. Chem. Lett. 1970, 6, 435. (19) Bradshaw, J. S.; Izatt, R. M. Acc. Chem. Res. 1997, 30, 338.
′,LAH + 1) (8) 2 log(Kads
In eq 8, the only variables are [LAH]m, [22DD]m, and pH. Thus, if the extraction proceeds according to eq 3, then plots of (i) log(KD) ) f(pH) at [LAH]m ) const and (ii) log(KD) ) f([LAH]m) at pH ) const should yield straight lines of slopes equal to 2. The results obtained for the effect of pH and [LAH]m are shown in Figures 2 and 3. As can be seen, KD values increase with increasing pH and lauric acid concentrations. Two observations are made: (1) KD values are much higher than the values obtained using Liq/Liq extraction (see below) probably due to the side reaction, i.e., adsorption of the metal complex at the membrane polypropylene pore surfaces as previously reported.13 (2) Plots of log(KD) vs (pH) and vs log([LAH]m) (inset Figures 2 and 3) yield straight lines of slopes 1.94 ( 0.02 and 1.9 ( 0.3, respectively, as expected from eq 8. These results suggest that, during the Cu(II) extraction by the organic phase, two protons are released and 2 mol of lauric acid are required for 1 mol of copper(II) extracted to maintain charge neutrality as predicted by eq 3. Since the aliphatic chains of LA and 22DD do not seem to interact (section 1), the nature of the complex in the membrane solvent is thus (22DD-Cu)2+‚2(LA)-. Analytical Chemistry, Vol. 72, No. 6, March 15, 2000
1331
Figure 3. Effect of [LAH]m/[22DD]m ratio on KD values for membrane/ liquid extraction. Inset: Plot of log(KD) vs log([LAH]m). Conditions used: Cu(II) concentration: 5 × 10-5 mol L-1, pH 6.0, MES buffer, fixed 22DD concentration 5 × 10-2 mol L-1, and various values of [LAH]m. (Error bars calculated for four replicate measurements)
4. Determination of the Cu(II) Partition Coefficient by Liquid/Liquid Extraction. In liquid/liquid extraction, the Cu(II) equilibrium reaction can be described by eq 3, where the adsorption side reaction then disappears. Hence, Kex (eq 6) can be expressed as a function of Kp (eq 2) as follows:
Kex )
Kp[H+]2aq [22DD]org‚[LAH]2org
(9)
Figure 4. (a) pH dependence of Kp for liquid/liquid extractions of Cu(II). (b) Plot of log(Kp) vs pH. Conditions used: (9) [LAH]org ) 5 × 10-2 mol L-1 alone. (b) 1:1 mixture of [LAH]org/[22DD]org both at 5 × 10-2 mol L-1. [Cu(II)] ) 5 × 10-5 mol L-1 in 10-1 MES buffer.
or as
log(Kp) ) log(Kex) + 2 log([LAH]org) + log([22DD]org) + 2pH (10)
The effect of pH on Kp is shown in Figure 4a using LAH alone ([LAH]org ) 5 × 10-2 mol L-1) and a 1:1 mixture of [LAH]org/ [22DD]org (both at 5 × 10-2 mol L-1) in toluene/phenylhexane (1:1). A plot of log(Kp) vs pH, using LAH alone, yields a straight line of slope 1.96 ( 0.01 (Figure 4b), indicating that Cu(II) extraction occurs with a release of two protons as expected. As the pKa of LAH is 5.3,11 at pH < 4.3, less than 10% of Cu(II) is extracted by LAH alone due to the competition of H+ and Cu(II) for LA-, whereas at pH > 6, up to 95% of Cu(II) is extracted. A similar behavior with a much less marked effect of pH on Kp is observed with [LAH]org/[22DD]org mixture: e.g., only 70 and 90% of copper(II) is extracted at pH 5.5 and 6.85, respectively. A plot of Kp vs [LAH]org/[22DD]org ratio (Figure 5) shows that up to a [LAH]org/[22DD]org ratio of 1, Kp values increase rapidly and then begin to level off; but at ratios of >2, Kp values increase again. This behavior is expected since LAH forms a 2:1 stable complex with 22DD (see section 1). Indeed at [LAH]org/[22DD]org ratios of e2, all the acidic protons from LAH are bound to 22DD and Cu(II) is extracted by exchange with (22DD-H)+. For [LAH]org/[22DD]org ratios of >2, additional Cu(II) may be extracted directly by the excess of LAH. log(Kp) vs pH and log(Kp) vs log([LAH]org) in the presence of 22DD ([22DD]org ) 5 × 10-2 mol L-1) yield linear plots (Figure 1332 Analytical Chemistry, Vol. 72, No. 6, March 15, 2000
Figure 5. Effect of [LAH]org/[22DD]org molar ratio on Kp for liquid/ liquid extractions. Inset: Plot of log(Kp) vs log([LAH]org). Conditions: [22DD]org ) 5 × 10-2 mol L-1 ) const, [Cu(II)]tot) 5 × 10-5 mol L-1 ) const, in MES buffer pH 6.0. (Error bars calculated for four replicate measurements).
4b and inset in Figure 5) with slopes 0.57 ( 0.03 and 0.53 ( 0.02, respectively, which are much lower than the values of 2 expected by eq 10. However, the slope of the log(Kp) ) f(pH) plot increases with decreasing [22DD]org and [LAH]org concentrations at constant ratio of [LAH]org/[22DD]org ) 1 and at constant Cu(II) concentration in the aqueous solution () 5 × 10-5 mol L-1). For [LAH]org and [22DD]org e 5 × 10-4 mol L-1, i.e., a concentration value not more than 10 times that of the Cu(II) concentration, a value of the slope log(Kp) ) f(pH) close to the theoretical value of 2 is observed (Figure 6).
Figure 6. Effect of the 22DD and LAH concentration at 1:1 ratio on the slope of log(Kp) ) f(pH). Conditions used are the same as those given in Figure 3.
DISCUSSION The rate-limiting step of Cu(II) transport by 22DD/LAH in toluene/phenylhexane PLM is governed by diffusion either through the membrane or both through the membrane and the aqueous Nernst layer. However, these transport experiments did not provide any information on the nature of the complex formed and particularly on the role of lauric acid in the extraction mechanism. In particular, a key question is whether laurate ions act as the metal carrier itself or only as the counteranion for the (Cu-22DD)2+ complex. The results of Figures 2-5 together with NMR show that LAacts as counteranion, 22DD forms a 1 to 1 complex with the metal, the metal being in the cavity of the crown ether, and that the metal does not bind covalently with laurate. Furthermore, literature data suggest that, in a mixture of two complexing agents, usually higher metal extraction is observed owing to synergetic effects.21,22 The present studies, in contrast, show that the opposite effect is observed; i.e., Kp ) 78 for LAH alone and Kp ) 10 for the 1:1 mixture LAH/22DD (pH 6.1, [carrier]org ) 5 × 10-2 mol L-1, [Cu(II)]aq ) 5 × 10-5 mol L-1), supporting the fact that LA- acts only as a counteranion. Only when the ratio LAH/22DD is >2, LAH may also play the additional role of carrier but with a much lower selectivity. It should be pointed out that both Liq/Liq and Liq/membrane extraction show the theoretical slope of 2 (eqs 8 and 10) for the change of Kp with [LAH] and pH; the Liq/Liq extraction, however, yields this value only for high enough metal/carrier ratio (Figure 6). At lower ratio, a much lower slope is obtained. This may be partly due to the fact that 22DD ()5 × 10-2 mol L-1) alone extracts 50% of the copper(II); for such an extraction, a co-anion to neutralize the charge of the 22DD-Cu2+ complex is required. However, the investigation of the effect of the nature of the coanion (Cl-, NO2-, ClO4-) does not account for this anomalous result. An alternative explanation is that metal could be extracted by some impurity present in 22DD as the slope of log(Kp) ) f(pH) increases with decreasing carrier concentration at constant metal concentration (Figure 6). Indeed, the analysis of 22DD show that (20) Nyre´n, V.; Back, E. Acta Chem. Scand. 1958, 12, 1305. (21) Izatt, R. M.; Clark, G. A.; Christensen, J. J. Sep. Sci. Technol. 1986, 21, 865. (22) Bond, A. H.; Chiarizia, R.; Huber, V. J.; Dietz, M. L.; Herlinger, A. W.; Hay, B. P. Anal. Chem. 1999, 71, 2757.
it contained 2 × 10-2 mol % Na+ implying the presence of an hydrophobic anion. Moreover, after centrifugation of 22DD solubilized in hexane, a brown viscous compound was found at the bottom of the tube (0.6% in mass); its nature, however, could not be easily characterized. This latter purification does not completely eliminate the extraction of Cu(II) by 22DD alone. Although, an unambiguous explanation could not be found, these results demonstrate that mechanistic interpretation of Liq/Liq extraction results cannot always be readily extrapolated to Liq/ membrane extraction where such anomaly was not observed. An explanation is that this difference is probably due either to the adsorption of this impurity on the pore wall of the polypropylene membrane as this also occurs with the 22DD/metal complex itself13 or to the lower carrier/metal molar ratio used in practice, because of the low volume of organic solvent in the membrane compared to that of the aqueous solution. It is important to bear in mind this difference in behaviors of Liq/Liq and Liq/membrane extractions when performing mechanistic studies, even though the impurity of 22DD has little effect on PLM transport under the conditions used8-13 in analytical applications, even at low concentration, because (i) it does not play a role in permeation membrane systems, as discussed above, and (ii) above all, Na+ and not H+ is used as countercation transport.13 Note that the values of Kp ()10, pH 6.1) for Liq/Liq extraction and KD ()1200, pH 6.1) for Liq/membrane found experimentally for [LAH]org ) [22DD]org ) 5 × 10-2 mol L-1 clearly indicate that the determination of Kp by Liq/Liq extraction cannot predict PLM transport activity of the studied carrier. CONCLUSION Despite of the metal extractant properties of lauric acid, laurate only acts as counterion in metal extraction in the mixture LAH/ 22DD at a molar ratio 1:1. The studies of variation of partition coefficient as a function of pH and [LAH]org, by Liq/Liq and Liq/ membrane extraction, show that two protons are exchanged during metal extraction and that two laurate, LA-, are necessary for neutralizing the charges. 1H NMR spectra show that the acidic proton of LAH is delocalized between the two nitrogen of 22DD macrocycle and that laurate does not seem to be close to (22DDH)+. The fact that metal is exchanged with two protons located in the 22DD cavity implies that the 22DD governs the metal extraction, the laurate acting only as counteranion. The stoichiometry of the complex formed in the organic phase is (22DDCu)2+/2(LA-). Moreover, during metal extraction, there is no coextraction of an anion of the aqueous phase to compensate the charges of metal extracted. This results clearly demonstrate that this system developed for PLM experiment can be used for environmental analysis as the metal is extracted by the 22DD which allows metal selectivity. ACKNOWLEDGMENT The authors thank Pr. H. P. van Leeuwen for helpful discussions. We thank HOECHST for kindly supplying us the Celgard membranes. Received for review December 28, 1999.
September
9,
1999.
Accepted
AC991047+ Analytical Chemistry, Vol. 72, No. 6, March 15, 2000
1333
