Anal. Chem. 2004, 76, 1302-1309
Metal Ion Determination by Flame Atomic Absorption Spectrometry through Reagentless Coacervate Phase Separation-Extraction into Lamellar Vesicles Dimosthenis L. Giokas,* George Z. Tsogas, Athanasios G. Vlessidis, and Miltiades I. Karayannis
Department of Chemistry, University of Ioannina, 451 10, Ioannina, Greece
The phase separation of lamellar vesicles of anionic surfactants in aqueous solutions and its application as a novel liquid coacervate extraction procedure was examined. Solutions of lauric acid sodium salt separate into two phases in the presence of alkaline earth metals and a water miscible cosurfactant. It is proven that the surfactant phase is built of a perplexed network of multilamellar vesicles consisting of densely packed bilayers. Several factors affecting the formation of this new phase as well as its analytical utility in the preconcentration of metallic ions were assessed on the basis of better exploitation of this new nonspecific extraction technique. In essence, although the procedure to arrive at the optimum conditions seems laborious, the delivered method is straightforward, alleviating the requirement for prereaction with a complexing agent and highly reproducible under the optimum experimental conditions. As an analytical demonstration, the method was successfully applied to the determination of Cd2+ and Zn2+ in natural waters. Recoveries were higher than 95%, and detection limits as low as 3 µg L-1 were accomplished by preconcentrating only 10 mL of sample volume in the presence of 0.45% (w/v) anionic surfactant. Of particular interest in the context of analytical chemistry separations are procedures employing the solubilizing properties of surfactant molecules in aqueous solutions.1,2 To date, the majority of the applications cited in the literature deal with the liquid-liquid-phase separation of nonionic or zwitterionic surfactant micelles (named cloud point phase separations) due to the simplicity of the operation which is involved in the clouding behavior of these surfactants.3-6 By altering the properties of the * Corresponding author. Phone: +30-26510-98401. Fax:+30-26510-44831. E-mail:
[email protected]. (1) Paradkar, R. P.; Williams, R. P. Anal. Chem. 1994, 66, 2752-2756. (2) Quina, F. H.; Alonso, E. O.; Farah, J. D. S. J. Phys. Chem. 1995, 99, 1170811714. (3) Garcia Pinto, C.; Perez Pavon, J. L.; Moreno Cordero, B. Anal. Chem. 1994, 66, 874-881. (4) Carabias-Martinez, R.; Rodriguez-Gonzalo, E.; Dominguez-Alvarez, J.; Herna´ndez-Me´ndez, J. Anal. Chem. 1999, 71, 2468-2474. (5) Giokas, D. L.; Paleologos, E. K.; Tzouvara-Karayanni S. M.; Karayannis, M. I. J. Anal. At. Spectrom. 2001, 16, 521-526. (6) Katsaounos, C. Z.; Giokas, D. L.; Vlessidis, A. G.; Paleologos, E. K.; Karayannis, M. I. Sci. Total Environ. 2003, 305, 157-167.
1302 Analytical Chemistry, Vol. 76, No. 5, March 1, 2004
solution, such as pressure, temperature, or salt content, they can rapidly form very large aggregates which separate from water while they scatter the visible light passing through the solution, causing the solution to become turbid (hazy). On the other hand, there is still a dearth of information regarding the phase separation of ionic surfactant solutions (called coarcevation). Not until recently, Casero et al.7 exploited the ability of sulfonic or sulfonate anionic surfactants to undergo phase separation under extreme acidic conditions. In the presence of redundant hydrogen ions, anionic surfactants are protonated toward their respective molecular form, which separate from the aqueous solution. Kumar et. al.8 reported a similar behavior for the anionic surfactant sodium dodecylbenzesulfonate in the presence of tetra-n-butylammonium bromide, but no analytical demonstration has been made. In a like manner, only one example of analytical utilization of the coacervate approach using cationic surfactants has been reported.9 In this study, cationic surfactants of linear structure exhibited phase separation after the addition of high amounts of salt (400 g L-1 of NaCl) and a cosurfactant (1-octanol). Hitherto, the analytical utility of ionic surfactant phase separation procedures is still limited. Although the works of PerezBendito and co-workers7,10,11 have demonstrated the usefulness of anionic surfactant acid-induced phase separation methods in various analytical applications, the coarcevate approach practically lacks of any other uses, especially with regard to the extraction and preconcentration of metal ions. On the other hand, the coarcevation of linear cationic surfactants is limited by the sharp dependence of the volume of the surfactant-rich phase on the volume of the cosurfactant added. It has been reported that the volume of the surfactant-rich phase was reduced ∼10 times when the cosurfactant volume was increased from 0.05 to 0.08%, resulting in low reproducibility.9 Research on the properties of surfactants has shown that the introduction of divalent counterions into ionic surfactant systems (7) Casero, I.; Sicilia, D.; Rubio, S.; Pe´rez-Bendito, D. Anal. Chem. 1999, 71, 4519-4526. (8) Kumar, S.; Sharma, D.; Khan, Z. A.; Kabir-ud, D. Langmuir 2002, 18, 42054209. (9) Jin, Z.; Zhu, M.; Conte, E. D. Anal. Chem. 1999, 71, 514-517. (10) Merino, F.; Rubio, S.; Pe´rez-Bendito, D. J. Chromatogr., A 2002, 962, 1-8. (11) Merino, F.; Rubio, S.; Pe´rez-Bendito, D. J. Chromatogr., A 2003, 998, 143154. 10.1021/ac0303517 CCC: $27.50
© 2004 American Chemical Society Published on Web 01/27/2004
has a multitude of effects on the phase behavior and physical properties of these systems.12 The behavior observed is different from that of surfactants with monovalent counterparts, that is, lower cmc values, increased tendency toward micellar growth, and formation of vesicular and lamellar phases.12,13 In these systems, the presence of cosurfactants (mostly medium-chain alcohols) can markedly affect the shape and size of micellar aggregates,14,15 leading to the formation of lamellar phases, usually at high surfactant and cosurfactant concentrations. These properties have been exploited by Gra¨bner et al.13 who reported the formation of a novel solidlike phase between laurylamidomethyl sulfate with alkaline and alkaline earth metals as counterions and a cosurfactant (n-hexanol). Further experiments have shown that this phase consists of a network of polydispresed multilamellar vesicles, which are changed to stacked membranes with increasing cosurfactant concentration. In principle, the formation of this novel phase is based on the partial neutralization of the anionic headgroups of the surfactant by alkaline earth metals and subsequent precipitation in the presence of the cosurfactant. The importance of this finding is that this new phase is formed in low surfactant and cosurfactant concentrations while the volume fraction of this novel phase increases with increasing cosurfactant concentration (in contrast to the previously reported case with cationic surfactants). These unique features alleviate the problems of a declining surfactant-rich phase with increasing cosurfactant concentration in ionic surfactant solutions. In this study, we report on the formation of a novel lamellar phase between an anionic surfactant, alkaline earth metals, and methanol as extraction and preconcentration media of metal species. We found that solutions of lauric acid sodium salt (C12H23NaO2) separate into two phases in the presence of alkaline earth metals and methanol toward the formation of a new type of dispersed lamellar vesicles, which act as a coacervate phase able to entrap metal species. To our knowledge, this surfactantmediated liquid-liquid phase separation has seemingly never been reported before. To confirm our assumption, the new phase separation procedure was investigated with the view of preconcentration of metallic cations from natural waters. From an analytical standpoint, this is the first study reporting on the application of liquid coacervate extraction for the preconcentration of metal ions from aqueous matrixes. EXPERIMENTAL SECTION Reagents. All reagents were of analytical grade and were employed as supplied. The anionic surfactant lauric acid sodium salt (sodium dodedanoate, LA) was purchased from Aldrich. Working solutions of 5.5% (w/v) were prepared in spectroscopicgrade methanol (Carlo Elba, Italy) and heated mildly to achieve complete dissolution of the surfactant. Standard metal solutions used for the experiments as well as for the interference study were prepared by sequential dilution of 1000 µg mL-1 of Titrisol stock solutions (Merck). Ammonium pyrrolidineditiocarbamate (APDC, Sigma-Aldrich Ltd, Greece) was prepared in doubly (12) Zapf, A.; Beck, R.; Platz, G.; Hoffmann, H. Adv. Colloid Interface Sci. 2003, 100-102, 349-380. (13) Gra¨bner, D.; Matsuo, T.; Hoinkis, E.; Thunig, C.; Hoffmann, H. J. Colloid Interface Sci. 2001, 236, 1-13. (14) Hoffmann, H. Adv. Mater. 1994, 6, 116-129. (15) Mizushima, H.; Matsuo, T.; Satoh, N.; Hoffmann, H.; Graebner, D. Langmuir 1999, 15, 6664-6670.
distilled water. 3-(4-Methylbenzyldene)camphor was supplied by Merck (Darmstadt, Germany). NaNO3 and humic acid (HA) used for the interference study was obtained from Fluka Chemie AG (Switzerland). Instrumentation. A Shimandzu AA-6800 flame atomic absorption spectrophotometer with hollow cathode lamps operating at 8 and 7 mA for Cd2+ and Zn2+, respectively, was used throughout the measurements, which were made at 228.80 and 213.86 nm, respectively. An adjustable-capillary nebulizer and supplies of acetylene and air at a ratio of 1.8 for Cd2+ and 2.0 for Zn2+ were used for the generation of aerosols and atomization. The output signals were collected and processed in the continuous peak height mode. Scanning electron microscopy (SEM) examinations were performed with a JEOL JSM-6300 instrument equipped with an Oxford ISIS 300 energy dispersive X-ray spectrometer (EDS). Samples. River and lake water samples were collected in glass bottles from the Louros River and Lake Pamvotis (Epirous region, Greece). The samples were filtered through a Whatman no. 40 filter to remove the suspended solids and stored at 4 °C. Procedure. A typical procedure of liquid coacervate extraction into lamellar vesicles was employed as follows: An aliquot of 10 mL of a cold (10 °C) water solution containing the analytes was spiked with 100 mg L-1 of Ca2+, 50 mg L-1 of Mg2+, and an appropriate amount of NaNO3 to yield a final concentration of 0.3 M. Lauric acid dissloved in methanol was then added to a final concentration of 4.4 g L-1, and the mixture was shaken for 1 min. Subsequently, 40 µL of 4 M NaOH was added to increase the pH above 9, and the mixture was churned and allowed to stand for 15 min at 10 °C in an ice bath. Separation of the phases was accomplished by mere centrifugation for 30 min at 4000 rpm. The bulk aqueous phases was decanted by inversion of the vial, and the vesicular phase was treated with a methanolic solution of 1 M HNO3 in order to dissociate the lamellar structure and redissolve the white solidlike material remaining at the bottom of the vial. The final solution was directly aspirated into the flame of AAS. MECHANISTIC STUDIES Synthesis and Characterization of Lamellar Vesicles. A 2-mL portion of a methanolic solution of 0.247 M (5.5% w/v) lauric acid sodium salt were diluted at a 1:1 ratio with doubly distilled water. Ca2+ and Mg2+ nitrate salts at a concentration of 500 mg L-1 were added, and the mixture was vigorously shaken and left to stand for 30 min, then shaken again and centrifuged at 4000 rpm for 35 min. The water was carefully removed with a Pasteur pipet, and the cotton-white solidlike material was stored in an Ependorrf vial. The white powder was air-dried. Samples of the examined materials were gold-coated prior to optical and SEMEDS examination. The SEM micrograph of the new phase with Ca2+ and Mg2+ as counterions is shown in Figure 1. In the absence of counterions, the single-chain ionic surfactant molecule of LA cannot form bilayers or vesicles due to the relatively large area occupied by the charged hydrophilic headgroup.16 The characteristic of this sample is an indication that LA forms a three-dimensional network (16) Israelachvili, J. N. Intermolecular and Surface Forces; Academic Press: London, 1985; Chapter 16.
Analytical Chemistry, Vol. 76, No. 5, March 1, 2004
1303
Figure 1. SEM micrographs of the bilayer and vesicular aggregates of lauric acid sodium salt with alkaline and alkaline earth metals: (a) bilayers and globular aggregates with Mg2+ counterion; (b) multilamellar vesicles and spongelike aggregates formed with a Ca2+ counterion; (c) multilamellar vesicles and bilayers with both Ca2+ and Mg2+ as counterions.
of crystalline very thin fibers, a feature that has also been observed with other surfactants.17 Due to the fact that LA was available as a sodium salt, planar structures could be detected, even in the absence of other counterions, but only at 50 µm and above (since practically all crystalline surfactant phases have a bilayer structure).18 However, they were not adequate to promote phase separation. Increasing Na+ content up to 0.8 M induced phase (17) Hoffmann, H.; Thunig, C.; Miller, D. Colloids Surf. A 2002, 210, 147-158.
1304 Analytical Chemistry, Vol. 76, No. 5, March 1, 2004
separation, but the precipitated material remained undissolved in the methanolic-nitric acid solution, thus making it unfavorable for analytical applications. The addition of Mg2+, a heavier counterion, changed this situation, and the globular or fiberlike structure of LA-Na+ grew and changed to a lamellar structure aggregating toward the formation of bilayer and multibilayer (18) Laughlin, R. G. The Aqueous Phase Behavior of Surfactants; Academic Press: London, 1994.
structures (Figure 1a). Elemental analysis revealed that multibilayer structures (Figure 1a-i) contain approximately a double quantity of Mg2+, as compared to globular (or planar) structures where Mg2+ was approximately equal to or less than Na+ (Figure 1a-ii). These findings can be explained by considering the difference of average curvature in the oriented molecular layer of each structure; in other words, they have different ratios of planar to curved parts. In the presence of Na+ counterparts, the ratio of LA anions to nonionic amphiphiles is larger. Hence, the average area for the hydrophilic group and the average curvature of the oriented molecular layer are larger, too.19 Consequently, the molecular assemblies have a larger area of high curvature. As Mg2+ ions are incorporated in the assemblies, the amount of nonionic amphiphiles is increased (and so does the hydrophobicity of the assembly), resulting in a larger planar part and a smaller curved part (Figure 1a). In the case of Ca2+, though, a different behavior is observed. The micrograph of Figure 1b clearly shows the formation of lamellar-like structures, which is in agreement with previous findings for other surfactant and cosurfactant mixtures.13 The formation of this structure is attributed to the decline of anionic LA molecules, a situation that favors the formation of mutlilamellar vesicles. As the concentration of Ca2+ increases, behavior similar to that in the case of Mg2+ is observed. However, since Ca2+ is a heavier counterion, the mutlibilayer structure becomes larger and larger, so that its ends are flexible enough to combine with each other and take advantage of the reducing area.19 In that manner, the formation of closed bilayers or multilamellar vesicles is accomplished. In a mixture of both Ca2+ and Mg2+ (Figure 1c), the lamellar structures prevail, but in several parts, there is a curved part which is larger than those observed with Ca2+ only, forming a complex network of organized surfactant assemblies containing both multilamellar and bilayer vesicular aggregates. As we can observe, the pores of Ca-LA vesicles are no longer present, whereas in many places, the lamellar vesicles resemble the appearance of bilayers observed with Mg counterions disrupting (or normalizing) the complex spongelike appearance of CaLA vesicles at 1 µm. The above findings are similar to those reported in previous studies with other surfactant/cosurfactant mixtures that lead to the formation of lamellar structures, thus providing solid proof of the formation of a new type of lamellar vesicle. Unambiguously, these lamellar vesicles are different in composition and structure from those known to date,13,19 although no SEM micrographs seem to be available in the literature depicting such vesicular aggregates. In any case, this novel phase of multilamellar vesicles appears to have a great analytical potential, which until now has not been exploited. Mechanism of Extraction. To investigate the potential mechanism involved, a series of preliminary experiments were conducted. Solutions of 50 mg L-1 of Cd2+ and Zn2+ were extracted with and without the presence of 100 mg L-1 Ca2+. Interestingly, all solutions exhibited phase separation, which verifies the notion that not only alkaline earth metals but also other metal species can lead to the formation of organized molecular assemblies with (19) He, X.; Zhu, B.-Y.; Huang, J.; Zhao, G.-X. J. Colloid Interface Sci. 1999, 220, 338-346.
anionic surfactants.16,20 At lower metal concentrations (0.05 mg L-1), the presence of Ca2+ counterions was mandatory, since no phase separation occurred, and therefore, no absorbance signal was recorded. From these results, it can be inferred that both Cd2+ and Zn2+ are involved in the formation process of the lamellar vesicles and, thus, are co-extracted with Ca2+. To explore whether the proposed procedure also comprises hydrophobic interactions, 3-(4-methylbenzyldene)camphor (a UV filter with log Kow ) 5.47) was extracted with LA-calciummethanol, and the final extract was measured spectrophotometrically against a solution of the same final concentration. A clear peak at 300 nm was obtained, revealing the existence of hydrophobic encapsulation in the new extraction scheme. On the basis of these results, we can propose that the complexes between LA and Cd2+ or Zn2+ can be incorporated in the vesicular formulations through both hydrophobic encapsulation and via co-extraction in the surfactant lamellar structures. Due to the fact that Cd2+ and Zn2+ also induced phase separation, the latter mechanism seems to be the pertinent extraction mechanism. From an analytical point of view, this observation provides experimental evidence of a coacervate-based separation procedure that facilitates direct interaction with polar analytes. In the main, the coexistence of both mechanisms renders several analytical ramifications that will be addressed further below. RESULTS AND DISCUSSION The ability of the proposed phase separation procedure to extract and preconcentrate metal species from aqueous solutions was then investigated. All parameters that can influence the efficiency of the proposed scheme were meticulously studied and are presented below. Effect of Metal Counterions. As previously mentioned, the concentration of counterions present in solution plays an important role in the phase separation process. In natural waters, the most abundant metal species are calcium and magnesium; therefore, their effect on the proposed extraction scheme was initially investigated. The results depicted in Figure 2 reveal that calcium concentrations up to 150 mg L-1 enhance the extraction, but for magnesium, a maximum peak is attained at 50 mg L-1. These results indicate that surfactant-counterion reactions cannot be explained by point-charge electrostatics without taking into account specific bonding that causes different behavior for two kinds of ions having the same charge.13 The unusual behavior between these two bivalent metal ions can be attributed to the different radius of the hydrated cation, which seems to influence counterion condensation.21 However, in natural waters, both metals are present in excess, as compared to other ion species; therefore, the examination of their ratio to the yield of the extraction is relevant to this context. The results of Figure 2c reveal that maximum extraction is obtained for a Ca2+/Mg2+ ratio between 2 and 5, which is close to the values usually observed in natural waters.22 A value of 2 was finally selected for the subsequent experiments. In all of the above cases, a lower concentration of (20) Kralchevsky, P. A.; Danov, K. D.; Broze, A.; Mehreteab, A. Langmuir 1999, 15, 2351-2365. (21) Conway, B. E. Ionic Hydration in Chemistry and Biophysics; Elsevier: Amsterdam, 1981. (22) Reimann, C.; Bjorvatn, K.; Frengstad, B.; Melaku, Z.; Tekle-Haimanot, R.; Siewers, U. Sci. Total Environ. 2003, 311, 65-80.
Analytical Chemistry, Vol. 76, No. 5, March 1, 2004
1305
Figure 3. Influence of lauric acid concentration in the performance of the method. [Cd2+] ) [Zn2+] ) 50 µg L-1, Ca2+/Mg2+ ) 2, [NaNO3] ) 0.1 M, temp ) 20 °C, reaction time ) 10 min, centrifugation time ) 35 min.
Figure 2. Effect of counterion concentration on the extraction yield of the proposed method: (a) effect of Ca2+ concentration, (b) effect of Mg2+ concentration, and (c) effect of the Ca2+/Mg2+ ratio. [Cd2+] ) [Zn2+] ) 50 µg L-1, [LA] ) 2.8 g L-1, [NaNO3] ) 0.1 M, temp ) 20 °C, reaction time ) 10 min, centrifugation time ) 35 min.
alkaline earth metals reduced the absorbance signal, possibly due to inadequate formation of lamellar vesicles, whereas at higher concentrations, the antagonistic action against the target metal species deteriorated the analytical signal. Moreover, with increasing Mg2+ concentration, the formation of bilayers is favored, promoting the gradual disintegration of multilamellar vesicles,23 1306 Analytical Chemistry, Vol. 76, No. 5, March 1, 2004
which contributes to the reduced extraction, possibly through inadequate phase separation. Effect of Surfactant Concentration. The dependence of the method on the concentration of LA was the next parameter investigated. The results of Figure 3 show that for both metals, a similar pattern is observed, yielding maximum absorbance in the concentration range between 3 and 5 g L-1. For lower LA concentrations, the complete counterbalance of LA micelle charge was surmised to induce inadequacy of reactive sites or resist phase separation (swelling of the lamellar phase), as would normally be expected for uncharged or low-charged vesicular membranes.13 However, it must be noted that with varying LA concentration, there was a respective variance in the methanol content of the solution. According to previous reports, low amounts of cosurfactant do not aid the formation of the lamellar phase,13 which advocates reduced extraction. On the other hand, with increasing cosurfactant concentration, a transition from vesicles to stacked membranes was suspected to occur.13 In that case, the precipitate swells to fill the entire volume of the sample (hazy solution),13 producing a minor precipitate at the bottom of the vial, which reasonably contains only a small quantity of the extracted species (inadequacy of extraction media). Effect of Mixture pH. The influence of hydrogen or hydroxyl concentration present in the sample was subsequently studied. The final mixture had a pH value around 6.5, so this value was taken as a point of reference. Appropriate amounts of 4 N HCl or NaOH were added in the mixture and churned vigorously. The effect of pH on the extraction performance is graphically demonstrated in Figure 4. The results illustrate an increase in the yield of the extraction with increasing pH, indicating the presence of a weak base. Considering the composition of the solutions, it is likely that this base is a carboxylate anion. In the alkaline pH range, the anionic group of LA is deprotonated, releasing H+ and COOin the aqueous phase, which aids complexation with metal species. A similar behavior in the reaction efficiency between metal species (23) Danino, D.; Talmon, Y.; Zana, R. Colloid Interface Sci. 1997, 185, 84-93.
Figure 4. Performance of the method at various pH values. [Cd2+] ) [Zn2+] ) 50 µg L-1, Ca2+/Mg2+ ) 2, [LA] ) 4.4 g L-1, [NaNO3] ) 0.1M, temp ) 20 °C, reaction time ) 10 min, centrifugation time ) 35 min.
and complexing surfactants in solvent-water media has been reported elsewhere.24 Effect of Salt Addition. In aquatic surfactant solutions, when salt concentration is increased, the micelle size and aggregation number are increased, keeping the micellar concentration constant.25 In the proposed extraction scheme, though, a more complex mechanism is involved. The influence of salt addition on the extraction performance was, therefore, examined by adding NaNO3 at concentrations ranging from 0.05 to 0.8 M. The results depicted in Figure 5 show an escalated increase in the signal up to 0.2-0.4 M. In the presence of Na+ counterions, the repulsion between the LA molecules in the oriented molecular layer is decreased; thus, the transformation from micelles to bilayers or vesicles is favored. As the NaNO3 concentration increases, the excess of sodium cations could cause the anionic form of LA to become inert through antagonistic complexation with Ca2+ and Mg2+ cations. In that manner, the replacement of Mg2+ and Ca2+ ions by Na+ could disintegrate the vesicular assemblies, leading to the formation of multibilayer or bilayer structures (planar or bent, depending on the concentration of counterions). Effect of Temperature. The reaction temperature for the studied system should compromise both completion of the reactions and efficient separation of the two phases. From the results of Figure 6, it can be inferred that an increase in the solution temperature is detrimental for the analytical signal, but cold conditions favor the extraction. The reason for this unique feature probably lies at the alteration of the viscosity of the surfactant-rich phase, which reasonably increases as the solution temperature decreases. At the low surfactant concentrations applied, the reduction of the temperature could lead to the formation of a viscoelastic phase,17 which was better separated from the aqueous phase. At higher temperatures, the increased solubility of the lamellar phase in water made phase separation (24) Schwuger, M. J.; Subklew, G.; Woller, N. Colloids Surf. A 2001, 186, 229242. (25) Hinze, W. L.; Pramauro, E. Crit. Rev. Anal. Chem. 1991, 63, 2520-2525.
Figure 5. Effect of mixture ionic strength on the performance of the procedure. [Cd2+] ) [Zn2+] ) 50 µg L-1, Ca2+/Mg2+ ) 2, [LA] ) 4.4 g L-1, NaOH ) 1.5 × 10-2 M, temp ) 20 °C, reaction time ) 10 min, centrifugation time ) 35 min.
Figure 6. Absorbance signal as a function of the mixture temperature. [Cd2+] ) [Zn2+] ) 50 µg L-1, Ca2+/Mg2+ ) 2, [LA] ) 4.4 g L-1, [NaNO3] ) 0.25 M, NaOH ) 1.5 × 10-2 M, reaction time ) 10 min, centrifugation time ) 35 min.
more cumbersome. A temperature of 10 °C was finally selected for the subsequent experiments. Effect of Reaction and Centrifugation Time. Other pertinent experimental parameters that were examined for their effect on the performance of the method were the centrifugation and the reaction times. The data suggest that the reaction is kinetically satisfied within 15 min, but for longer times, a gradual decline is observed for both metals. The centrifugation time necessary to ensure complete differentiation of the two phases was above 15 min, with only insignificant gains being made at longer times. Effect of Chelating Agents. From an analytical standpoint, the feasibility for the direct application of the method (as described thus far) offers a significant simplification, since no additional reaction or derivatization steps are required. Nevertheless, in real samples, the excess of cations may cause a deficiency of comAnalytical Chemistry, Vol. 76, No. 5, March 1, 2004
1307
Table 1. Effect of Increasing APDC Concentration on the Recovery of Increasing Amounts of the Target Metal Species APDC (%, w/v)
Cd2+/Zn2+ added (µg L-1)
0
50/50 250/250b 500/500b
100.2 ( 0.9; 98.4 ( 0.8 100 ( 1.1; 100 ( 1.0 99.4 ( 0.9; 99.7 ( 0.8
0.25
0.5
0.75
Recoveries (%)a 24.2 ( 2.0; 95.2 ( 0.9 14.8 ( 2.5; 73.2 ( 1.8 100.2 ( 1.2; 99.6 ( 1.1 79.5 ( 1.7; 85.6 ( 1.9 100.0 ( 1.1; 99.6 ( 1.0 96.9 ( 1.0; 99.3 ( 0.8
16.8 ( 2.7; 43.2 ( 1.9 77.6 ( 2.0; 60.0 ( 2.2 94.3 ( 1.3; 98.7 ( 1.3
a Mean ( 2SD (p ) 0.95) of triplicate measurements. b Appropriate dilutions were made in order to bring the concentrations within the linear range of the calibration curve.
Table 2. Analytical Features of the Method parameter
Cd2+
Zn2+
phase volume ratioa enhancement ratiob extraction concentration factorc LOD (µg L-1)d RSD (%) (n ) 6, 50 µg L-1) regression equation
0.1 8.5 ∼1.0 3.1 0.97 A ) -3.4 × 10-3 ( 2.1 × 10-3 + 2.0 × 10-3 C ( 3.5 × 10-5 0.9989 (confidence level 95%)
0.1 10 ∼1.0 2.4 2.14 A ) 4.6 × 10-2 ( 4.1 × 10-3 + 5.2 × 10-3 C ( 7.4 × 10-5 0.9997 (confidence level 95%)
correlation coefficient (r)
a Phase volume ratio: the ratio of the final volume of surfactant-rich phase to that of the aqueous phase. b Enhancement ratio: the ratio of the concentration of analyte after preconcentration to that without preconcentration giving the same absorbance peak area. c Extraction concentration factor: the ratio of concentration in the surfactant-rich phase to that in the original solution. d LOD: Limit of detection, defined as three times the signal-to-noise ratio.
plexing sites, thus reducing the efficiency in the determination of the target species. Because hydrophobic interactions are also feasible with the proposed scheme, the addition of a chelating agent, able to react with the target species, was investigated. The addition of a complexing agent could provide further reactive sites for metal complexation to compensate for the consumption of the ligand from other metals amenable to such complexation. Ammonium pyridyldithiocarbamate (APDC) was selected due to its increased solubility in slightly alkaline media while it does not react with neither Ca2+ nor Mg2+; thus, it does not interfere with the phase-separation process. Evidently, increasing APDC concentration produced a considerable sink in the absorbance signal of both metals. This behavior was surprisingly opposite to the expected results, because one would normally expect APDC either to improve or at least not alter the performance of the method. The first reason stated to explain this behavior is that acidic rather than alkaline pH values aid the complexation of metals with APDC. However, reducing the solution pH did not improve the extraction. A possible explanation for the observed behavior could be the fact that with increasing chelating agent concentrations (APDC and LA), charged rather than uncharged complexes are formed, thus decreasing the extraction efficiency, since nonpolar species are preferably partitioned in the hydrophobic surfactant core.26 To verify the validity of this assumption, a set of experiments were conducted, and the results are presented in Table 1. As can be seen, the higher the amount of metals present in solution, the less reduction in the extraction yield of the procedure with increasing concentration of APDC. Figures of Merit. Calibration curves were obtained by preconcentrating 10 mL of standard solutions with 4.45 g L-1 of LA, (26) Da Silva, M. A. M.; Frescura, V. L. A.; Curtius, A. J. Spectrochim. Acta Part B, 2000, 55, 801-811.
1308 Analytical Chemistry, Vol. 76, No. 5, March 1, 2004
100 mg L-1 of Ca2+ and 50 mg L-1 of Mg2+ under the optimum experimental conditions established above. Table 2 features the analytical characteristics of the method. Under the proposed experimental protocol, the calibration curves were rectilinear in the range 10-100 µg L-1 for Cd2+ and 5-100 µg L-1 for Zn2+. The limits of detection were very satisfactory, considering the low phase volume ratio applied. It is conceivable that further improvement in the preconcentration factor is feasible, either by preconcentrating larger sample volumes or by diluting the surfactantrich phase to a smaller volume of methanolic solution. In the latter case, heating of the methanolic extract can be employed to enhance the solubility of the vesicular assemblies, thus enabling the application of a smaller volume of redissolution mixture. The time of analysis, although it may seem quite long, is adequate for the preconcentration of multiple samples, and depending on the capacity of the centrifuge, more than 40 samples can be analyzed in less than 1 h. Interferences. In a competitive reaction ambience, the chelating agent (in this case, LA) may react with various metallic species, in addition to Cd2+ and Zn2+, thereby reducing the extraction efficiency of the target metals. In this respect, the interfering effects of other cations on the determination of Cd2+ and Zn2+were investigated. The interfering cations were studied at metalto-interfering agent ratios of 1:1, 1:5, and 1:100, while the concentrations of Cd2+ and Zn2+ were set at 50 µg L-1 for both metal species, respectively. The interfering cations examined (Cr3+, Pb2+, Cu2+, Co2+, Fe3+, Al3+, Ni2+) were not found to impair the quality of the analytical data due to sufficient reagent for complete extraction in combination with the selectivity offered by FAAS. For too high a metal-to-interfering agent ratio, the required chelation capacity can be maintained by adding LA or by the addition of APDC. In the latter case, though, an especially
Table 3. Determination of Cd2+ and Zn2+ in Real Samples and Comparison with a Reference Methoda concn measb,c
a
ref methodc
rel error (%)
sample
Cd2+
Zn2+
Cd2+
Zn2+
Cd2+
Zn2+
river water lake water table water tap water
2.1 ( 0.4 0.8 ( 0.2
