Micelle mediated extraction of magnesium from water samples with trizma-chloranilate and determination by flame atomic absorption spectrometry

Talanta 56 (2002) 415– 424 www.elsevier.com/locate/talanta

Micelle mediated extraction of magnesium from water samples with trizma-chloranilate and determination by flame atomic absorption spectrometry Dimosthenis L. Giokas, Evangelos K. Paleologos, Panayotis G. Veltsistas, Miltiades I. Karayannis * Department of Chemistry, Uni6ersity of Ioannina, 45110 Ioannina, Greece Received 29 June 2001; received in revised form 4 September 2001; accepted 7 September 2001

Abstract This article describes an analytical method for the determination of magnesium taking advantage of the cloud point phenomenon employing a suitable chelating agent (chloranilate) for Mg analysis. The method encompasses pre-concentration of the metal chelate followed by flame atomic absorption spectrometry (FAAS) analysis. The chelating agent chosen for this task is a newly synthesised salt of chloranilic acid, trizma-chloranilate, which reacts with Mg but at the same time has a very low affinity for other metallic cations like silicon, aluminium and sodium, which interfere with the determination of Mg in FAAS. The condensed surfactant phase with the metal chelate(s) is introduced into the flame of an atomic absorption spectrometer after its treatment with an acidified methanolic solution. In this way, complex and time-consuming steps for sample treatment are avoided while increased sensitivity is achieved by the presence of both methanol and surfactant in the aspirated sample. The analytical curve was rectilinear in the range of 5–220 mg l − 1 and the limit of detection was as low as 0.75 mg l − 1 with a standard deviation of 5.2%. The method was applied for the determination of Mg in natural and mineral waters with satisfactory results and recoveries in the range of 97–102%. © 2002 Elsevier Science B.V. All rights reserved. Keywords: Atomic absorption spectrometry; Cloud point; Micelles; Trizma-chloranilate

1. Introduction The determination of alkaline earth metals, particularly calcium and magnesium is of importance in environmental, biological and industrial appli-

* Corresponding author. Tel.: +30-651-98406; fax: + 30651-98796. E-mail address: [email protected] (M.I. Karayannis).

cations. Both elements belong to the physiologically essential elements and their concentration in natural waters determines the hardness of the water, which plays a significant role in the quality of drinking water and the toxicity of many metals [1]. In groundwater, the adsorption of metallic cations is decreased significantly and their transportation through soils is enhanced in the presence of Ca, due to the competition with Ca for the exchange sites of the soil [2]. Furthermore, the

0039-9140/02/$ - see front matter © 2002 Elsevier Science B.V. All rights reserved. PII: S 0 0 3 9 - 9 1 4 0 ( 0 1 ) 0 0 5 6 4 - 1

416

D.L. Giokas et al. / Talanta 56 (2002) 415–424

determination of calcium and magnesium, among several other cations and anions, is a common practice in industrialised areas with problems from acid rain incidents [3]. From a physiological point of view, calcium and magnesium, along with sodium and potassium, are the most important ions affecting cardiology, owing to their role in nervous impulse and cell contraction [4]. Recent studies have demonstrated an inverse relationship between the concentration of calcium and magnesium in drinking water and the risk of death from breast cancer [5]. Additionally, a significant protective effect of magnesium intake from drinking water and other dietary sources against the risk of prostate cancer development has been observed [6]. In agriculture, bivalent cations, especially calcium and magnesium, which are the most abundant, were shown to decrease the phytotoxicity of surfactants on plants [7], which are added in order to enhance the stability of pesticide suspensions. The significant role of calcium and magnesium has resulted in the publication of many methods for their determination in a wide variety of matrixes [8–10], especially in natural and drinking waters [11,12]. Several techniques such as atomic absorption spectrometry [8,13] atomic emission spectrometry [10], ICP-OES [14], capillary electrophoresis [12,15], spectrofluorimetry [16], liquid chromatography [17], spectrophotometry [18] combined with chemometrical methodologies [4,9] have been presented over the years for the determination of these elements. Flame atomic absorption spectrometry (FAAS) is one of the most sensitive and efficient techniques used for the determination of magnesium [19]. Although, apart from sodium, no significant spectral interferences have been reported in the determination of magnesium, chemical interferences have to be taken into account for the analysis of real samples [19]. Silicon, aluminium and phosphate suppress the signal of Mg in the air– acetylene flame through the formation of stable compounds. Therefore, the addition of releasing agents (lanthanum or vanadium) is necessary for the reduction of these effects. Nevertheless, the interferences from silicon and aluminium persist to this treatment [8]. The interference of sodium becomes detectable when it is present at ten times

the concentration of magnesium in an air–acetylene flame and a monochromator of bandpass 0.5 A, [19]. However, such selectivity does not solve the problems of spectral interference of sodium when seawater is analysed, as the concentration of sodium ion in seawater is about 8– 26 times higher than that of the magnesium ion [14]. Common analytical methods applied for the determination of magnesium with FAAS employ solvent extraction as a pre-concentration step, partly because of the low detection limits available, but primarily because of the relatively high abundance of this element in many sample types [20]. The complex of 8-hydroxyquinoline with magnesium, among several other complexes like magneson or Eriochrome black T, has been widely utilised for its determination with FAAS [20]. The complex was subsequently extracted in an organic solvent (mostly 4-methylpentan-2one) in order to remove the analyte element from the initial matrix [20]. The disadvantage of solvent extraction (toxicity, required volume, etc.) following the formation of the metal–chelate complexes in the usual spectrometric applications has initiated a wide research activity on the development of alternative methods like the use of adsorptive columns [21], co-precipitation [22], cloud point extraction [23], and even micellar-enhanced ultrafiltration [24]. Among the simplest and most versatile methods is the cloud point extraction. This technique is based on the property of most non-ionic surfactants in aqueous solutions to form micelles and become turbid when heated to a temperature, known as the cloud point temperature (CPT). Above this temperature, the solution separates into two distinct phases: a surfactant rich phase composed almost totally of the surfactant; and an aqueous phase in which the surfactant concentration is close to the critical micellar concentration. The mechanism by which this separation occurs is attributed to the rapid increase in the aggregation number of the surfactant’s micelles as the temperature is increased or to critical phenomena [25]. The presence of surfactant-based organised assemblies is able to change the physical properties (e.g. density, viscosity, and surface tension) of liquid samples. The use of such assemblies in

D.L. Giokas et al. / Talanta 56 (2002) 415–424

analytical atomic spectrometry was extensively studied in the literature [26] and constitutes a known approach for improving sample transport to the atomiser. In a different function, species that can interact with micellar systems either directly or after being derivatized become concentrated in a small volume of the surfactant-rich phase which can subsequently be analysed using classical analytical systems (FIA, AAS, HPLC, GC, and CE) [27,28]. Such applications exploiting the capability of micellar systems for the determination of metallic species after derivatization with the appropriate ligand have been widely presented over the past years [29,30]. In more recent applications, the technique has been utilised for the speciation analysis of metals [31,32] as well as for the removal of interferences from usual spectrochemical applications utilising FAAS measurements [33]. Taking advantage of the analytical merits of the CPT approach we put forward the concept of using the technique for accomplishing the determination of magnesium, using the appropriate chelating agent in order to form water insoluble or sparingly soluble complexes. In this respect, the proposed analytical method rests on the formation of hydrophobic complexes of magnesium with chloranilate, which is subsequently separated from the bulk aqueous phase with a non-ionic surfactant, prior to its determination by FAAS. Interestingly, based on the results obtained, this effort presents a simple, sensitive and solvent-free method for the accurate determination of magnesium in natural and drinking water without interferences.

2. Experimental

2.1. Materials All reagents used were of analytical grade. Stock solutions of magnesium and calcium and those used for the interference study were prepared by dissolving appropriate amounts of their respective salts in doubly distilled water. The nonionic surfactant Trinton X-114 (Fluka Chemie AG, Switzerland) was used without further purifi-

417

cation. Trizma-chloranilate (Trizma-Chl) was synthesised according to the procedure described below and standard solutions were prepared in doubly distilled water. Trizma base, chloranilic acid, NaOH and HCl were obtained from Sigma– Aldrich Ltd, Greece.

2.2. Apparatus A GBC 2000 (GBC Ltd, Victoria, Australia) atomic absorption spectrometer with a hollow cathode lamp operating at 4 mA was used throughout the measurements made at 285.2 nm. An adjustable capillary nebuliser and supplies of acetylene and air (in the ratio of 2.5:8) were used for the generation of aerosols and atomisation. The output signals were processed with a time constant of 1.0 s in the peak area mode. A thermostated bath maintained at the desired temperature was used for cloud point temperature experiments and phase separation was assisted using a centrifuge (Hettich, Universal).

3. Synthesis of the trizma-chloranilate salt The commercially available chloranilic acid (C6Cl2O4H2) has a relative solubility in water (9.09× 10 − 3 M in 13 °C and 6.69× 10 − 2 M in 99 °C), which was not adequate for the application described in this study. Therefore, a more soluble compound, which maintains the properties of chloranilate as a ligand and at the same time is highly soluble in water, was synthesised. The free base (Trizma base, Sigma, pKb1 3.27) dissolved in a water–alcohol solution (1:1) reacts with a stoichiometric amount of chloranilic acid dissolved in warm acetone, towards a dark nutbrown crystal compound almost quantitatively (95%), according to the reaction: 2C4H11O3N+ C6Cl2O4H2 “ [C4H12O3N]2(C6Cl2O4) The product is washed thoroughly with warm acetone and dried under air stream and subsequently under vacuum in the presence of phosphorous pentoxide. This salt is readily dissolved

D.L. Giokas et al. / Talanta 56 (2002) 415–424

418

in DMSO, DMF and FAM and its density was estimated as 1.61 g cm − 3. It is also readily dissolved in water (up to 1 M) and it is stable and non-hydroscopic under standard conditions. Its aquatic solutions (up to 1 M) are almost neutral. In Table 1 the IR, Vis and UV spectra, as well as some properties of the Trizma-Chl salt are presented [34]. This compound can be used as a solid primary standard for chloranilate solutions, prepared by simple weighing.

phase. After removing the bulk aqueous phase the remaining micellar phase was treated with 200 ml of the methanolic solution of 1 mol l − 1 HNO3 to reduce its viscosity. The final solution was aspirated directly into the flame of AAS. The conventional procedure used for comparison was carried out using the same AAS manifold, employing an acetylene–air stoichiometric flame at the wavelength of 285.2 nm. Magnesium determinations were carried out under deuterium background correction. The samples were diluted beforehand with a lanthanum solution.

4. Procedure The final volume of the surfactant-rich phase was determined by measuring the phase after cloud point extraction along with the added methanolic solution. The values reported are the average of triplicate measurements. Cloud point temperature of Triton X-114 was taken from Ref. [26]. Typically, a cloud point extraction experiment was performed as follows: an aliquot of 10 ml of a solution containing both Mg and Ca (to simulate real conditions), 1 g l − 1 Triton X-114 and 6 mg l − 1 Trizma-Chl was adjusted to the appropriate pH value (pH 10) with diluted NaOH. Subsequently, it was shaken for 1 min and left to stand in a thermostated bath for 10 min at 70 °C. Separation of the phases was achieved by centrifugation for 10 min at 3500 rpm. The phases were cooled in an ice bath to increase the viscosity of the surfactant-rich

5. Real samples Samples of river water and seawater were collected from the Ipirus region (Greece). A commercially available mineral water was purchased from the local store. All samples were stored in glass bottles and acidified for preservation. Prior to analysis the samples were filtered through a Whatman No. 40 filter to remove any suspended solids. They were then subjected to the cloud point extraction described in Section 4. The certified water reference material IMEP-9 (acidified river water) obtained from the Institute for Reference Materials and Measurements (IRMM, Belgium) was used to examine the accuracy of the proposed method.

Table 1 Cumulative experimental data of conventional analytical measurements for bis-(tromethaminium) chloranilate (Trizma-Chl) compound MP 195 °C, normal melting procedure to deep-red oily drop, without any apparent indication of disintegration and charring Theoretical

Experimental

%C

%N

%H

%Cl

%C

%N

%H

%Cl

37.23

6.20

5.32

15.71

37.12

6.12

5.24

15.65

Infra red spectrum

385s, 570–595–625s, 840–1035–1060vs, 1130–1200w, 1290s, 1380–1400m, 1540vs (broad), 1625vs, 2080w, 2580–3210vs (broad) log m1= 2.57 (504.8 nm); log m2 = 4.42 (329.6 nm); log m3 =4.42 (318.8 nm); log m4 = 3.67 (255.6 nm) in DMSO C14H24O10Cl2N2 and [C4H12NO3]2·[C6Cl2O4]

UV–visible spectrum Observed tentative molecular and syntactic formula

D.L. Giokas et al. / Talanta 56 (2002) 415–424

Fig. 1. Effect of pH on the extraction of Mg. Mg 200 mg l − 1; Trizma-Chl 5 ppm; TX-114 0.1%; temperature 65 °C; incubation time 10 min.

In order to assess any possible matrix effects, the samples were spiked with appropriate amounts of the metal.

6. Results and discussion Any parameter (pH, surfactant and chelating agent concentration, temperature and incubation time) affecting the proposed reaction and micelle formation was included in the optimisation experimental design to estimate the importance of each. The pH was the first parameter examined for its effect on the extraction of magnesium. The results illustrated in Fig. 1 reveal that at pH 10 maximum extraction efficiency was attained. This value was therefore selected as the working pH value. The concentration of the chloranilate-chelating agent was subsequently studied for its effect on the extraction of magnesium and the results are presented in Fig. 2. As shown, a concentration of 5– 6 mg l − 1 was sufficient for the total complexation of 200 mg l − 1 of magnesium in the presence of equivalent concentration of Ca. However, the placement of an excess of chelating agent was found to diminish the extraction efficiency of the method. It was observed that the excess of reagent prevented the good separation of the two phases especially at very high concentrations. As noted, the transportation of both neutral and charged

419

species in the coacervate phase, including the target analytes and depending on the pH of the aqueous solution, can alter the cloud point temperature [35]. Owing to the fact that phase separation was evident at low and medium concentrations of chloranilate it was assumed that higher concentrations simulate the action of salting in electrolytes, which increase the cloud point, therefore prevent the good separation of the two phases. A successful cloud point extraction procedure should maximise the extraction efficiency by minimising the phase volume ratio, thus improving its pre-concentration ability. Triton X-114 was chosen for the formation of the surfactant-rich phase due to its low cloud point temperature and high density of the surfactant-rich phase, which facilitates phase separation by centrifugation. Fig. 3 highlights the effect of the surfactant concentration on the analytical signal. It was proved that Triton X-114 effectively extracts the derivatized metal species from liquid samples at concentrations of 1 g l − 1 almost quantitatively using a single-step extraction procedure. Larger quantities of the surfactant can be used at the expense of detection limits, as higher concentrations deteriorate the atomic signal due to the increase in the surfactant volume [36]. The equilibration temperature above the CPT and the incubation time were the parameters optimised next. It was desirable to employ the shortest incubation time and the lowest possible equilibration temperature, which compromise

Fig. 2. Effect of Trizma-Chl concentration on the extraction of magnesium. Mg 200 mg l − 1; pH 10; TX-114 0.1%; temperature 65 °C; incubation time 10 min.

420

D.L. Giokas et al. / Talanta 56 (2002) 415–424

Fig. 3. Effect of surfactant concentration on the extraction of Mg. Mg 200 mg l − 1; pH 10; Trizma-Chl 6 ppm; temperature 65 °C; incubation time 10 min.

with the completion of the reaction and the efficient separation of phases. The results illustrated in Figs. 4 and 5 indicate that the reaction is quantitatively and reproducibly completed after 10 min of heating at 70 °C. This temperature favours the extraction of the Mg– chloranilate complex as the equilibration temperature increases, reduces the surfactant rich phase [37] and thereby increases the pre-concentration factor. Other parameters such as centrifugation time for phase separation and ultrasonication during the reaction and cloud point extraction do not have any effect on the analytical characteristics [23]. A centrifugation time of 15 min was selected for the entire procedure, since the analyte extraction at this time period is almost quantitative [31,33]. In order to increase the accuracy and the reproducibility of the analytical measurements, the final surfactant-rich phase volume (ca. 100 ml), was directly aspirated into the nebuliser of the AAS after its uptake with 200 ml of a methanolic solution containing 1 mol l − 1 HNO3. As it has al-

ready been proved, the results are reproducible within the experimental error once the procedure for water removal from the concentrated micellar phase is always followed in a consistent way [31]. The presence of methanol and surfactant in the aspirated sample solutions could change the physical properties of the liquid sample by altering the solution environment. As reported, methanol

Fig. 4. Effect of temperature on the extraction of Mg. Mg 200 mg l − 1; Trizma-Chl 6 ppm; TX-114 0.1%; incubation time 10 min.

D.L. Giokas et al. / Talanta 56 (2002) 415–424

421

8. Interferences

Fig. 5. Effect of incubation time (heating time) on the extraction of Mg. Mg 200 mg l − 1; Trizma-Chl 6 ppm; TX-114 0.1%; temperature 70 °C.

brings about some enhancement to the signal, which is more pronounced when combined with the surfactant, therefore increasing the sensitivity of the method. This can be attributed to the fact that the organic solvent in the flame increases the atomisation temperature, while both the organic solvent and the surfactant can promote the generation of small droplets during nebulisation causing positive surface effects [20].

Despite the high selectivity provided by flame AAS, Mg absorbance was reported to be suppressed by elements like silicon, aluminium and sodium [7,11], while similar suppressing effects have also been reported for phosphates [19]. This fact, along with the behaviour of chelating agents, which are not selective and react favourably with a multitude of elements, should be taken into consideration. Although the consumption of chloranilate due to the presence of Ca can be overcome by the addition of a calcium masking agent like ethylene glycol-bis-(b-aminoethyl ether)N,N,N%,N%-tetraacetic acid (EGTA) [16] it was decided to include Ca in the experimental design procedure to minimise the reagent consumption and experimental effort. The reaction of chloranilate with Ca has been widely applied for its determination [38,39] and the possibility of the simultaneous determination of both elements (magnesium and calcium) will be discussed later. Table 2 Analytical characteristics of the method Parameter

Mg

Phase volume ratio Pre-concentration factor Extraction Concentration factor LOD (mg l−1) RSD (%) (n =3, 200 mg l−1) Regression equation

0.02

7. Figures of merit The calibration curve was constructed by preconcentrating 10 ml of standard solutions with Triton X-114. Table 2 features the analytical characteristics of the method. Under the specified experimental conditions, the calibration curve for magnesium was rectilinear from 5 to 220 mg l − 1. The pre-concentration factor of about 50 obtained by pre-concentrating a 10 ml of sample volume can be considered highly satisfactory compared with the other CPE methodologies [29– 31] where usually larger amounts of sample volumes are used. The limit of detection was sufficiently low as compared with the other studies [13,15,17]. Further improvement is also feasible, either by pre-concentrating larger amounts of the sample solution, using the respective higher concentrations of both the surfactant and the chelating agent, or by diluting the surfactant-rich phase to a smaller volume of the methanolic solution.

Correlation coefficient (r)

50

1 0.75 5.2% A =0.382 ( 90.014)+5.88×10−3( 91.2×10−4)×C (C in mg l−1) 0.9984

Phase 6olume ratio: the ratio of the final volume of surfactantrich phase to that of the aqueous phase; pre-concentration factor: the ratio of the concentration of analyte after pre-concentration to that without pre-concentration giving the same absorbance peak area; LOD: limit of detection, defined as three times the signal-to-noise ratio; extraction concentration factor: the ratio of the amount in the surfactant-rich phase to that in the original solution.

422

D.L. Giokas et al. / Talanta 56 (2002) 415–424

Several advantages emerge from the use of chloranilate for the determination of magnesium, primarily because it forms polar complexes with a multitude of metallic cations (including Al), which are not extracted in the surfactant micelles, but basically from the fact that it does not react with silicon and sodium [34]. Therefore, the final solution aspirated to the flame of the AAS does not contain any of the interferences, which suppress the signal of magnesium. Even if too high concentrations of sodium are present in the sample, the required dilution step and the poor affinity of sodium for chloranilate result in zero or minor transportation of Na in the condensed micellar phase. Even in the latter case, it is highly unlike that sodium concentrations can be as high as ten times that of magnesium so that interference can occur. Another advantage of the proposed method is the high pH level, which favours the formation of the Mg –chloranilate complex. In pH 10, most metallic cations are precipitated as their relative oxides that are not entrapped in the micelles of the surfactant. Negative effects from the entrapment of phosphates in the micelles have been reported in a similar micellar-enhanced extraction procedure utilised for the spectrophotometric determination of Uranium [40]. However, no adverse effects in FAAS measurements have been reported [31,33]. Taking into consideration that a dilution factor of at least 1000 may be required for the determination of Mg following the methodology described here, phosphates did not posse any adverse negative effects at the very low concentrations present in the final solution. In Table 3 a variety of interfering cations and anions are presented for their effect on the analytical signal of magnesium. As we can observe, no adverse effect was imposed at the concentration levels investigated.

Table 3 Effect of foreign cations and anions on the recovery of 200 mg l−1 of Mg in the presence of equivalent concentration of Ca Cations

Concentration (mg l−1)

Recovery (%)

Cr Zn Co Cu Pb Cd Fe Si Al Na

200 200 200 200 200 200 200 200 200 3000

99 99 100 101 99 101 97 98 100 99

Anions

mg l−1

PO4 NO3 HCO3

2.5 2.5 2.5

99 100 100

sented [38,39]. However, under the micellar environment described in this study this reaction occurred in an uncontrollable manner. Experiments for the simultaneous determination of magnesium and calcium, involving the previously described variables were carried out. The pH values that favoured the formation of the complex under the micellar environment applied were in the range of 9–10. However, the results obtained were found to possess a high RSD= 16.3% and the linear area of the calibration curve was in the close range of 180–220 mg l − 1. Based on these findings, the use of the methodology described here for the determination of calcium ions would be laborious and time-consuming especially if magnesium ions are to be determined simultaneously. That is, because special care should be given to the dilution step in order to bring calcium concentrations within the dynamic measuring range of the detector under the specified experimental conditions.

9. Simultaneous determination of calcium and magnesium

10. Analysis of real samples

Many analytical methodologies exploiting the capability of chloranilate to react with calcium yielding the CaC6Cl2O4 complex have been pre-

Seawater, river water and mineral water were analysed to assess their content in Mg with the proposed and a conventional procedure to assess

D.L. Giokas et al. / Talanta 56 (2002) 415–424

423

Table 4 Determination of Mg in certified material, real and spiked samples Sample

Measured (mg l−1)

Seawater River water Mineral water

1050 7.4 1.9

Spiked (mg l−1) 200 0.5 0.5

Measured (mg l−1) a

Measured (mg l−1) b

Seawater River water Mineral water

1030 7.2 1.9

1050 7.4 1.9

CRM

Certified (mg l−1)

Measured (mg l−1) a

IMEP-9 a b

8.64

8.58

Found (mg l−1)

Recovery (%)

1213 7.91 2.39

97 102 98

Deviation (%) 1.9 2.7 0.0 Measured (mg l−1) b 8.51

Deviation (%) a 0.7

Deviation (%) b 1.5

Conventional procedure. Proposed method.

the accuracy of the method further. In order to verify the absence of the matrix effect, selected samples were spiked with appropriate amounts of the analyte. The certified water reference material IMEP-9 (acidified river water) was also analysed for Mg. The results given in Table 4 indicate that the proposed method can be used reliably for the determination of magnesium in various water matrixes. As observed, there are no relative error higher than 2.7% between the conventional and the proposed procedure, which further supports its utility for the analysis of magnesium in water samples.

11. Conclusions The combined advantages of the cloud point methodology and the use of chloranilate as a ligand for Mg were utilised for its determination in water samples. The method rested on the formation of a hydrophobic complex of Mg with a newly synthesised salt of chloranilate, TrizmaChl, which was subsequently separated in the micelles of the non-ionic surfactant TX-114, on increasing the solution temperature to 70 °C. The method gives a very low limit of detection, good RSD and the solvent-free extraction of the ele-

ment from its initial matrix following a single-step extraction procedure without interferences. The method was verified with real samples (natural and mineral waters) and it was proven satisfactory for the determination of Mg in a variety of water matrixes.

References [1] P.G. Welsh, J. Lipton, G.A. Chapman, T.L. Podrabsky, Environ. Toxicol. Chem. 19 (2000) 1624. [2] W.-Z. Wang, M.L. Brusseau, J.F. Artiola, J. Contam. Hydrol. 25 (1997) 325. [3] K. Tanaka, K. Ohta, P. Haddad, J. Fritz, A. Miyanaga, W. Hu, K. Hasebe, J. Chromatogr. A 884 (2000) 167. [4] F.B. Go´ mez, F.B. Reig, P.C. Falco´ , Talanta 49 (1999) 155. [5] C.H. Yang, H.F. Chiu, B.H. Cheng, T.Y. Hsu, M.F. Cheng, T.N. Wu, J. Toxicol. Environ. Health A 60 (2000) 231. [6] C.H. Yang, H.F. Chiu, S.S. Tsai, M.F. Cheng, M.C. Lin, F.C. Sung, J. Toxicol. Environ. Health A 60 (2000) 17. [7] B.A. Uhlig, A.H. Wissemeir, Crop Protection 19 (2000) 13. [8] Z. Arslan, J. Tyson, Talanta 50 (1999) 929. [9] O. Herna´ dez, F. Jime´ nez, A.I. Jime´ nez, J.J. Arias, J. Havel, Anal. Chim. Acta 320 (1996) 177. [10] D.T. Thuy, D.D. Weever, W.T. Kok, P. Luan, T.V. Nghi, Anal. Chim. Acta 295 (1994) 151. [11] S. Motellier, S. Petit, P. Decambox, Anal. Chim. Acta 410 (2000) 11.

424

D.L. Giokas et al. / Talanta 56 (2002) 415–424

[12] Z.L. Chen, M.A. Adams, Talanta 47 (1998) 779. [13] A.N. Arau´ jo, R.C.C. Costa, J.L.F.C. Lima, B.F. Reis, Anal. Chim. Acta 358 (1998) 111. [14] C. Prohaska, K. Pamazal, I. Steffan, Fresenius J. Anal. Chem. 367 (2000) 479. [15] K. Fukushi, K. Hiiro, Fresenius J. Anal. Chem. 356 (1996) 150. [16] G. de Armas, A. Cladera, E. Becerra, J.M. Estela, V. Cerda´ , Talanta 52 (2000) 77. [17] B. Paull, M. Macka, P.R. Haddad, J. Chromatogr. A 789 (1997) 329. [18] Y.A. Zolotov, E.I. Morosanova, S.V. Zhalovannaya, S.S. Dyukarev, Anal. Chim. Acta 308 (1995) 386. [19] J.A. Dean, T.C. Rains, Flame Emission and Atomic Absorption Spectrometry, vol. 3, Marcel Dekker, New York, 1975. [20] M.S. Cresser, Solvent Extraction in Flame Spectroscopic Analysis, Butterworths, London, UK, 1978. [21] K.A. Tony, S. Kartkeyan, B. Vijayalakshmy, B. Prasada Rao, C.S. Padmanabha Iyer, Analyst 124 (1999) 191. [22] Q. Zhang, H. Minami, S. Inoue, I. Atuya, Anal. Chim. Acta 407 (2000) 147. [23] C. Garcı´a Pinto, J.L. Pe´ rez Pavo´ n, B. Moreno Cordero, Anal. Chem. 66 (1994) 874. [24] S. Akita, L.P. Castillo, S. Nii, K. Takahashi, H. Takeuchi, J. Membr. Sci. 162 (1999) 111. [25] M.J. Rosen, Surfactants and Interfacial Phenomena, Wiley/Interscience, New York, 1987. [26] M. Kodoma, S. Miyagawa, Anal. Chem. 52 (1980) 2358.

[27] R. Carabias-Martinez, E. Rodrı´guez-Gonzalo, B. Moreno-Cordero, J.L. Pe´ rez-Pavo´ n, C. Garcia-Pinto, E. Ferna´ dez-Laespada, J. Chromatogr. A 902 (2000) 251. [28] B. Fro¨ schl, G. Stangl, R. Niessner, Fresenius J. Anal. Chem. 357 (1997) 743. [29] C.G. Pinto, J.L. Pe´ rez Pavo´ n, B. Moreno Cordero, J. Anal. At. Spectrom. 11 (1996) 37. [30] M.C. Oliveros Cerrato, O. Jimenez de Blas, J.L. Pe´ rez Pavo´ n, B. Moreno Cordero, J. Anal. At. Spectrom. 13 (1998) 547. [31] E.K. Paleologos, C.D. Stalikas, S.M. TzouwaraKarayanni, G.A. Pilidis, M.I. Karayannis, J. Anal. At. Spectrom. 15 (2000) 287. [32] E.K. Paleologos, C.D. Stalikas, M.I. Karayannis, Analyst 126 (2001) 389. [33] D.L. Giokas, E.K. Paleologos, S.M. TzouwaraKarayanni, M.I. Karayannis, J. Anal. At. Spectrom. 16 (2001) 521. [34] P.G. Veltsistas, PhD Dissertation, University of Ioanninam, 1992. [35] J. Szymanowski, W. Apostoluk, J. Colloid Interface Sci. 228 (2000) 178. [36] F.H. Quina, W.L. Hinze, Ind. Eng. Chem. Res. 38 (1999) 4150. [37] T. Okada, Anal. Chem. 64 (1992) 2138. [38] E.H. Tyner, Anal. Chem. 20 (1948) 76. [39] P.G. Veltsistas, M.I. Karayannis, Analyst 122 (1987) 1579. [40] M. Fernadez Laespada, J.L. Perez Pavon, B. Moreno Cordero, Analyst 118 (1993) 209.