ANALYST, MARCH 1993, VOL. 118 303 Determination of Trace Amounts of Aluminium in Natural Waters by Solid-phase Spectrofluorimetry Jose Luis Vilchez, Alberto Navalon, Ramiro Avidad, Trinidad Garcia-Lopez and Luis Ferm in Ca pitan-Val hey* Department of Analytical Chemistry, University of Granada, E- 18071 Granada, Spain A spectrofluorimetric method for the determination of trace amounts of aluminium was developed, based on solid-phase spectrofluorimetry. Aluminium reacted with salicylidene-o-aminophenol t o form a fluorescent complex that was adsorbed on a dextran-type cation-exchange gel. The fluorescence of the gel, packed in a 1 mm silica cell, was measured directly with use of a solid-surface attachment. The applicable concentration range was from 0.20 to 14.00 pg 1-1, with a relative standard deviation of 1.0% and a detection limit of 0.02 pg 1-1.The method was applied to the determination of aluminium in natural waters. The method is more sensitive and selective than that based on salicylidene-o-aminophenol alone. Keywords: Salicylidene-o-aminophenol; aluminium determination; solid-phase spectro fluorimetry; natural water During the last decade, there has been an increased interest in the biological importance of aluminium because of evidence for the systemic toxicity of the element and the development of accurate analytical techniques. Aluminium is used in building and vehicle construction, and in the manufacture of paint, electrical equipment, packaging containers and cooking utensils. It is used therapeutically as an antacid, in drinking-water purification and as an antiperspir- ant.Aluminium toxicity has been linked to encephalopathy,'-3 osteomalacy and osteodistrophy,4.~ anaemia,"7 gastrointesti- nal symptoms8 and possible cardiotoxicityg features. Analyses for aluminium are mainly used to determine the blood and urine levels in individuals who work with aluminium and in patients subjected to haemodialysisl() or receiving total parenteral nutrition. 11 Monitoring of patients on haemodialysis is critical. Sequen- tial determination of aluminium in serum is required and even the water used in dialysis must be checked to ensure that levels of aluminium remain low.12 Aluminium determination by fluorescence measurements has been widely studied, and numerous methods have been proposed.13 One of the best known involves salicylidene-o- aminophenol as a reagent.13-16 This method yields a sensitive and relatively selective procedure, although there are several disadvantages regarding practical applications. 1.5 Solid-phase spectrofluorimetry (SPF) combines the measurement of solid-surface fluorescence with the use o€ a solid support (e.g., an ion-exchange gel) to preconcentrate the analyte, which has been rendered fluorescent by the use of an appropriate reagent. This approach has been found to be useful for the analysis of very dilute solutions, such as water. 17-22 In this paper, a method for the determination of trace amounts of aluminium by SPF is described, which was applied satisfactorily to the determination of aluminium in natural waters.By using this methodology, a higher sensitivity, a lower detection limit and a lower interference level than in solution are obtained. Experimental Reagents All the reagcnts used were of analytical-reagent grade unless stated otherwise. * To whom correspondcncc should be addressed. Sephadex CM C-25 cation-exchange gel (Pharmacia Fine Chemicals, Uppsala, Sweden) was used in the sodium form and without pre-treatment in order to avoid contamination. Salicylidene-o-aminoptienol (SOA Ph) . This was synthe- sized as described earlier19 and used as a 1.0 X 10-3 mol 1 - 1 solution in acetone, which was stable for at least 1 week. Solutions of lower concentrations were prepared fresh each day. Standard aluminium(rir) stock solution, 1 .O mg ml- 1 . Prepared from A1(N03)3-9H20 (Merck, Darmstadt, Ger- many) in 0.1 mol 1-1 nitric acid and standardized by titrimetry with ethylenediaminetetraacetic acid (EDTA) (Xylenol Orange as indicator).Working solutions were prepared by appropriate dilutions with doubly distilled water. Buffer solutions. Solutions of the required pH were prepared from 1.0 mol 1 - 1 sodium acetate (Merck) solution and 1.0 rnol 1-1 acetic acid (Merck). Apparatus All spectrofluorimetric measurements were performed with a Perkin-Elmer LS 5 luminescence spectrometer (Norwalk, CT, USA), equipped with a xenon discharge lamp (9.9 W) pulsed at line frequency, Monk-Gillieson F/3 monochromators, a Rhodamine 101 counter to correct the excitation spectra, a Hamamatsu R928 photomultiplier, a Houston Omnigraphic x-y recorder (Houston Instruments, Houston, TX, USA), a variable-angle solid-surface accessory, designed and construc- ted in-house (see Fig.l),23 and a Braun Melsungen Thermo- mix 1441 thermostat (B. Braun, Germany). In order to compare all the spectrofluorimetric measurements and ensure reproducible experimental conditions, the LS 5 spectrometer was checked daily with a sample of the fluorescent polymer standard p-terphenyl (1.0 x 10-7 mol I - I ) having a relative fluorescence intensity of 90% at he, = 340 nm, he, = 295 nm; the slit-widths were 2.5 and 2.5 nin, and the sensitivity factor was 0.594. The LS 5 spectrometer was interfaced with an IBM PS/2 30-286 microcomputer, with RS 232C connections for spectral acquisition and subsequent calculation of the excitation- emission matrices.24 The contour plots in the excitation- emission plane were produced by linking points of equal fluorescence intensity.A Canon RJ-300 printer (Canon, Tokyo, Japan) was used for graphical representation. A Crison 501 digital pH meter (Crison Instruments, Barcelona, Spain) with a combined glass-saturated calomel electrode and an Agitaser 2000 rotating agitator (Tecnotrans, Barcelona, Spain) were also used.304 ANALYST, MARCH 1993, VOL. 118 f hem I 31 5 135 225 r I L Fig. 1 Variable-angle solid-surface accessory Fluorescence Measurements The measured relative fluorescence intensity (RFI) of the gel beads, containing the fluorescence products and packed into a 1 mm silica cell, was the diffuse transmitted fluorescence emitted from the gel at the unirradiated face of the cell.The optimum angle between the cell plane and the excitation beam was 45" in all instances.23 Procedures Basic procedure A 500 ml water sample containing 0.2&14.00 pg 1 - 1 of Al"' was transferred into a 1 1 polyethylene bottle, and 2.5 ml of 1.0 mol 1-1 acetic acid-acetate buffer (pH 5.80), 2.5 ml of 1.0 X 10-3 mol 1-1 SOAPh and 100 mg of Sephadex CM C-25 gel were added. The mixture was shaken mechanically for 10 min. Afterwards, the gel beads were collected by filtration under suction and, with the aid of a pipette, were packed into a 1 mm cell, together with a small volume of the filtrate. A blank solution containing all the reagents except aluminium was prepared and treated in the same way as described for the sample. The fluorescence intensities (20.0 5 0.5"C) of the sample and blank were always measured 5 min after loading the samples at A,,, = 508 nm, with A,, = 410 nm.A calibration graph was established in the same way, with use of aluminium solutions of known concentration. Procedure for natural waters The above-mentioned reagents were added to a volume of natural water sample containing an adequate amount of All1', levelled off at 500 ml with doubly distilled water and placed in a 1 1 polyethylene bottle. The subsequent steps were as in the basic procedure. The calibration graph method was used for calibration purposes. Sample treatment Natural waters (preserved by addition of 0.25 ml of concen- trated nitric acid per litre of sample) were passed through a 600 W 350 I I --- 450 500 550 Em ission/n m 600 Fig.2 ( a ) Projected three-dimensional spectrum of the aluminium- SOAPh complex fixed on Sephadcx CM C-25 gel in an acetic acid-acetate buffcr (pH 5.80). Increments in excitation wavelengths were 2 nm for each emission scan and scan speed was 240 nm min-1. (b) Contour plot of the excitation-emission matrix of the aluminium- SOAPh complcx fixed on Sephadex CM C-25 gel in an acetic acid-acetate buffer (pH 5.80). The contours join points showing the same relative fluorescence intensity filter-paper with a pore size of 0.45 pm (Millipore, Milford, MA, USA) and the filtrates were collected in polyethylene containers that had been cleaned carefully with nitric acid. The samples were stored at 4 "C until analysis. Analyses were performed with the least possible delay.The usual precautions were taken to avoid contamination .25 Results and Discussion Spectral Characteristics The reagent SOAPh reacts with All1', originating in solution as a 1 : 1 fluorescent chelate at slightly acidic pH (approximately 6). 1 6 1 6 In the prescence of Sephadex cation-exchange gel, the complex, probably cationic, is adsorbed on the gel, as the complex is not adsorbed on anion-exchange gels. A CM C-25 dextran-type gel was selected as it was found to have a lower background fluorescence. In Fig. 2(a), the three-dimensional spectrum of the alumi- nium-SOAPh complex adsorbed on the gel (after the contri- bution of the blank has been subtracted) is represented as anANALYST, MARCH 1993, VOL. 118 305 isometric projection, where the emission spectra at stepped increments of the excitation wavelength have been recorded and plotted.Computer software allows the spectrum to be examined from a high or low excitation wavelength. In Fig. 2(b), the three-dimensional spectrum has been transformed into a contour plot in the excitation-emission plane, in order to ascertain both excitation and emission maxima. The peak wavelengths in the excitation spectra of the SOAPh-AI"' system are identical for the immobilized and solvated systems (410 nm). The maxima of the emission spectra for the two systems differ, being located at 520 nm in solution and at 508 nm in the gel phase. The modification of the features of the fluorescence spectra was considered to be a result of the modification of the surrounding environment of the complex in the gel phase with respect to solution. In addition, it was observed that a decrease in the excitation slit-width (Slitcx) or an increase in the emission slit-width (Slitem) increased the fluorescence signal.A similar effect has been reported by other workers.26 For optimum excitation and emission, slit-widths of 2.5 nm were selected in both instances. From a study of the half-life time of the excited state of the complex in the solid phase at different temperatures, it was concluded that the luminescence process was fluorescence (T< 5 x 10-6 s). Optimization of Variables p H dependence First, a buffer solution from those proposed in the literature was chosen on this system in solution. The ammonium acetate-hydrochloric acid buffer16 cannot be used, because it alters the gel.The sodium acetate-acetic acid buffer solution was found to afford the best results. The optimum pH value for the formation and fixation of the species falls in the narrow range 5.50-6.00. At pH <3.5 or >7.5 the complex is not formed and/or not fixed on the gel (Fig. 3). The fluorescence is independent of ionic strength, adjusted with the buffer solution, NaCl or NaC104, up to 0.01 moll-'. For higher values, the fluorescence emission decreases according to the equation: RFT = 8.2 X p-1'2 (RFI = relative fluorescence intensity; p = ionic strength; r = 0.995). This effect can be attributed to the competition from other ions in the ion-exchange equilibrium .27 SOA Ph concentration The optimum SOAPh concentration for maximum fluores- cence intensity was 5.0 x 10-6 mol 1-1 for an SOAPh-to- aluminium ratio of about 11 : 1.At higher SOAPh concentra- tions, however, the quenching effect observed in solution appeared more pronounced in the gel phase, probably owing to the re-absorption effect of the solid matrix (SOAPh fixed on Sephadex) .2X 3 4 5 6 7 8 PH Fig. 3 Influence of H on relative fluorescence intensity. [SOAPh], 2.0 x 10-5 mol I - l ; r.411i1], 1.85 x 10-6 mol 1-1; acetic acid-acetate buffer solution; Sephadex CM C-25, 100 mg; sample, 500 ml; stirring time, 10 min; A,,, 410 nm; kern. 508 nm; and T , 20.0 k 0.5 "C Influence of temperature The effect of temperature on the ion-exchange process and hence, on the fluorescence emission, was studied. The ion-exchange process was independent of temperature in the range 0-4OoC, with measurement of RFI at 20.0 k 0.5"C.In the latter instance, the fixation of species was carried out at room temperature. On the other hand, RFI decreased when the temperature of the system increased, the effect being totally irreversible. The decrease of RFT was 0.4% at 1O"C, 0.8% at 20°C, 12.2% at 30"C, 43.1% at 40°C and 78.5% at 50 "C. All RFI measurements reported here were performed at 20.0 k 0.5"C. Other experimental conditions The stirring times necessary for maximum RFI development were 10, 15, 20 and 25 min for 500, 1000, 1500 and 2000 ml samples, respectively. As the use of a large amount of the gel lowered the RFI, only the amount required to fill the cell and facilitate handling, i.e., 100 mg, was used in all the measure- ments.With regard to the stability of the fixed complex, the RFI, after an initial increase during 2 min of 10% , remained constant for at least 3 h. The order of addition of reagents did not affect the results obtained. The order used was aluminium, buffer, SOAPh and gel. Effect of sample volume on sensitivity In previous papers,17-22 it was mentioned that one of the main advantages of SPF methods is the potential increase in sensitivity with increase in the sample volume taken for analysis. This effect can be assessed by measuring the RFl of Sephadex equilibrated with different volumes of solutions containing the same concentration of AI"' and proportional amounts of the other reagents. Plots of RFI versus sample volume show an increase in fluorescence signal with sample volume, tending asymptotic- ally to a constant RFI value above a certain volume.The shape of the graphs suggests a Langmuir-type isotherm, as is observed in some ion-exchange spectrophotometric studies.29 Stoichiometry of the Complex Fixed on the Gel The stoichiometry of the SOAPh-AI"' complex fixed on Sephadex CM C-25 gel was studied by continuous-variation and molar-ratio methods. In both instances, the ligand-to- metal ratio found was 1 : 1. This species is identical with the complex reported by Dagnall et al.15 and Morishige16 in solution. The cationic nature of the complex could justify its fixation on the gel. In this study the tendency of the gel to sorb complexes with a large number of ligands was not observed.This occurrence has been described by several workers.3(&-32 Calibration and Precision The calibration graphs for samples treated according to the procedure described above are linear for the concentration ranges 0.20-14.00 pg 1-1 for 500 ml, 0.20-12.0 pg 1-1 for 1000 ml and 0.10-10.0 pg 1-1 for 1500 and 2000 ml sample volumes. The analytical parameters are summarized in Table 1. The reproducibility of the proposed method and of the packing of the gel in the 1 mm cell was determined. The precision was measured for an aluminium concentration of 1.00 pg 1-1 by performing ten independent determinations. The relative standard deviations (RSDs) (p = 0.05, n = 10) were 1.0, 0.9, 0.8 and 0.8% for 500, 1000, 1500 and 2000 ml samples, respectively. The precision (RSD) of the packing operation, calculated from ten measurements, was 0.9% for the aluminium-SOAPh complex fixed on the gel, 0.9% €or the gel blank (gel with SOAPh and buffer) and 0.8% for the gel only. It appears, therefore, that one of the main factors306 ANALYST, MARCH 1993, VOL.118 Table 1 Analytical parameters Sample volume/ml Parameter 500 1000 1500 2000 Slope 4.68 6.65 8.56 10.34 Linear dynamic Correlation Detection limit/ Determination RSD* (%) I .0 0.9 0.8 0.8 Intercept 0.1 0.3 0.2 0.2 range/pg I- 1 0.20-14.00 0.20-12.00 0.10-10.00 0.10-10.00 coefficient 0.999 0.998 0.998 0.998 Pg 1- 0.022 0.016 0.014 0.012 limit/pg 1-1 0.075 0.053 0.047 0.039 * RSD = relative standard deviation. Table 2 Methods for the spectrofluorimetric determination of aluminium Detection limit"/ Reagent I%-' SOAPh 0.27 Morin 0.27 Salicylidene-2-amino-3-hydroxyfluorene 0.2 6-(4-Methylsalicylideneamino)-rn-cresol 0.2 N-Salicylidene-2-hydroxy-4-carboxyaniline 0.2 Morint 0.1 2,6-Bis[(o-hydroxy)phenyliminomethyl]- l-hydroxybenzene 0.1 2,4-Dihydroxybenzaldehyde semicarbazone 0.08 N-Salicylidene-2-hydroxy-5-sulfoaniline 0.08 MorinS 0.02 SOAPh$ 0.02 * Or minimum concentration used for calibration.$ Ion-exchange spectrofluorimetry . Extraction procedure. Ref. 14 35 36 16 37 38 39 40 37 18 This work affecting the reproducibility is the packing of the gel. Centrifugation of the gel when packed in the cell did not lead to improved precision. Sensitivity and Detection Limit The sensitivity in SPF methods can be enhanced by increasing the volume of the sample.In practice, this increase can be calculated from the slope of the calibration graphs. The calculated values of the sensitivity ratio (S) for the samples analysed in this study are: = 2.21, S1soo/5oo = 1.83 and S1(H)oIsoO = 1.42, where the subscripts represent the sample volumes (ml). The non-linear dependence of sensitivity versus sample volume can be attributed to the decrease in the distribution coefficient with analyte concentration, as is usual in a non-linear isotherm. The increase in sensitivity obtained with the proposed method is substantial, particularly with respect to solution methods that involve use of SOAPh as a reagent. In order to compare this increase in sensitivity, the calibration graph for the determination of Al"' with SOAPh in solution, was established, i.e., for the method of Dagnall et aZ.14 Under our experimental conditions, the equation for the calibration graph was RFI = 0.16[Al"'] ( r = 0.999), the ratio of the slopes being 29 : 1.The IUPAC detection limits33 and the limits of determina- tion34 were calculated for 500, 1000, 1500 and 2000 ml sample volumes. The results are reported in Table 1. The proposed method was compared with methods de- scribed in the literature for the spectrofluorimetric determina- tion of aluminium. For comparison purposes those methods Table 3 Effect of foreign ions on the determination of 1.00 pg I-' of aluminium Foreign ion o r species Tolerance level/pg I - I Table 4 Determination of aluminium in natural waters Water Amount found*/pg 1-1 156 t 2 37.6 If: 0.3 19.0 k 0.1 17.2 k 0.2 5.12 k 0.08 14.5 k 0.2 Tap water (Granada City) Raw water (Genil River) Raw water (Quentar Dam) Raw water (Aguas Blancas River) Mineral water (Lanjaron) Mineral water (Ortigosa del Monte) * Average value k standard deviation of three determinations.Table 5 Recovery study of aluminium in natural waters Amount added1 Water* M - l - Tap water (Granada City)/lO mlS 1 . 00 2.00 3.00 - Raw water (Genil River)/SO mlS 1 .oo 2.00 3.00 Raw water - (Quentar Dam)/100 mlS 1 .OO 2.00 3.00 - Raw water (Aguas Blancas River)/ 1 .00 100 ml$ 2.00 3.00 (Lanjaron)/250 mlS 1 .OO Mineral water - 2.00 3.00 Mineral water - (Ortigosa del Monte)/ 1 .oo 100 ml$ 2.00 3.00 * Final volume 500 ml in all instances. Amount foundt/ 3.12 4.10 5.20 6.10 3.76 4.74 5.70 6.68 3.80 4.90 5.75 6.92 3.44 4.48 5.40 6.50 2.56 3.52 4.54 5.60 2.90 3.96 4.85 5.80 % - ' t Data are the average values of three determinations.3 Initial sample volume in each instance. Recovery (Yo) - 99.5 101.6 99.7 99.6 99.0 98.8 102.1 99.1 101.8 100.9 99.3 100.9 98.9 99.6 100.7 101.5 99.0 98.3 - - - - - that were considered to be among the most sensitive reported to date were selected (Table 2). Effect of Foreign Ions A systematic study was carried out on the effect of foreign ions on the determination of All1' at the 1.00 pg 1 - 1 level. A 10 mg 1 - 1 level of potentially interfering ions was tested first, and if interference occurred the ratio was reduced progressively until interference ceased. Higher ratios were not tested. Tolerance was defined as the amount of foreign ion that produced an error not exceeding +5% in the determination of the analyte.The results are summarized in Table 3. Interference levels were lower than those found in solution methods.1416 On the other hand, the proposed method is more selective than those involving methods based on morin. 183.38ANALYST, MARCH 1993, VOL. 138 307 Determination of Aluminium in Natural Waters The method was applied to the determination of aluminium in water samples. Tap water from Granada, which is treated with aluminium compounds for flocculation purposes, raw water from Granada supplies to the city reservoirs (Quentar Dam, Genil River and Aguas Blancas River), and mineral water from Lanjaron (Granada) and Ortigosa del Monte (Segovia) natural springs were selected. The volume of water used for the analysis depended on the aluminium content: 10 ml of tap water; 50 ml of raw Genil River water; 100 ml of raw Quentar Dam water, raw Aguas Blancas River water and Ortigosa del Monte spring water; and 250 ml of Lanjaron spring water.The analysis was carried out by the calibration graph method. The average aluminium content (based on three determina- tions) in the samples studied is listed in Table 4. The aluminium content found in tap water was higher than that in raw waters. This is related to the use of commercial aluminium salts in the water-treatment plant, as previously stated. In order to check the accuracy of the proposed method a recovery study was carried out on the waters mentioned above. For this, various amounts of aluminium were added, and the percentage recovery was determined.Table 5 shows the results obtained for all the water samples. This study was funded by the Direccion General de Investigac- ion Cientifica y Ticnica (DGTCYT) del Ministerio de Educac- ion y Ciencia (Spain) (Project No. PS88-0101). 1 2 3 4 5 6 7 8 9 10 11 12 References Alfrey, A. C., Legendre, G . R., and Kaehny, W. D . , New Engl. J . Med., 1976, 294, 184. Dunea, G . , and Mahurkan. S. D., Ann. Intern. Med., 1978,88, 502. Davism, A. M., Oli, H., and Walker, G. S . , Lancet, 1982, 2, 785. Wills, H. R., Clin. Chem. (Winston-Salem, N.C.), 1985, 31, 5 . Gardiner, P. E., and Ottaway, J . M., Anal. Chim. Acta, 1981. 128, 57. Fernandez Soto, I., Allende, M. T., and Diaz de GreAi, M. C., Nefrologia, 1986, 6, 71.Drueke, T., Touam, M., and Lacour, B., Nefrologia. 1986. 6, 67. Andreoti, S. P., Bcrgtcin, J. M., and Sherrad, D . J., New Engl. J. Med., 1984, 310, 1079. Siderman, S . , and Marior, D., Nephron, 1982, 31, 1. Joffc, P., Olscn, F., Hcaf, J . G.. Gammelgaard, B., and Podenphant, J . , J. Clin. Nephrol., 1989, 32. 133. Vargas, J. H.. Klein. G . L . , Ament. M. E., Ott. S. M., Sherrad, D. J., Horst, R. L., Berquist, W. E., Alfrey, A. C., Slatopolsky, E., and Coburn. J. W., Am. J . Clin. Nutr., 1988,48, 1070. Woolfson, A. D., and Gracey, G . M., J. Clin. Pharm. Ther., 1988, 13, 243. 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 Fcrnandez Gutierrez, A., and Muiioz de la PeAa, A., in Molecular Luminescence Spectroscopy: Methods and Applica- tions, Part I, ed.Schulman. S. G.. Wiley, New York. 1985, ch. Dagnall, R. M., Smith, R . , and West, T. S . , Chem. Ind. (Lowdon), 1965,34, 1499. Dagnall, R. M., Smith, R., and West. T. S . , Tulunta, 1966, 13, 609. Morishige, K., Anal. Chim. Acfa, 1980, 121, 301. Capitan, F., Manzano, E., Navalon, A . , Vilche7, J. L., and Capitan-Vallvcy, L. F., Analyst, 1989, 114, 969. Capitan, F., Manzano, E., Vilchez, J. L., and Capitan-Vallvey, L. F., Anal. Sci., 1989. 5, 549. Capitan, F., Navalon, A.. Vilchez, J. L., and Capitan-Vallvey. L. F., Talantu, 1990, 37, 193. Capitan, F., de Gracia, J . P., Navalon. A., Capitan-Vallvey, L. F., and Vilchez, J. L., Analyst, 1990, 115, 849. Capitan, F., Manzano. E., Navalon, A., Vilchcz, J . L., and Capitan-Vallvey, L.F., Talanta, 1992, 39, 21. Capitan, F., Sanchcz-Palcncia. G . , Navalhn, A.. Capitan-Vall- vey, L. F., and Vilchez, J . L., Anal. Chim. Acta, 1992,259,345. Manzano, E., Ph.D. Thesis, University of Granada, Spain, 1989. Oms, M. T., Cerda. V., Garcia-Sanchez, F., and Ramos, A . L., Talanta, 1988, 35, 671. American Public Health Association, American Water Works Association, Water Pollution Control Federation, Mktodos Normalizados para el Analisis dc Aguas Potables y Residuales, Diaz dc Santos, Madrid, 1992, pp. 1.45-1.47. Garcia-Vior, L. O., and Possidoni de Albinati, J. F., An. Quim., Ser. B , 1987, 83, 319. Waki. H . , Noda, S . , and Yamashita, M.. React. Polym., 1988, 7, 227. Laserna, J . J., Ph.D. Thesis, University of Malaga, Spain, 1980. Yoshimura, K., Ishii, N., and Tarutani. T., Anal. Chem., 1986, 58, 591. Yoshimura, K., Waki, H., and Ohasi, S . , Talunta. 1976,23,449. Toshimitsu, Y.. Yoshimura, K., and Ohasi, S . . Talanra, 1979. 26, 273. Capitan-Vallvey, L. F., Bosque-Sendra, J . M., and Valencia, M. C., Analusis, 1989, 17, 601. IUPAC, Nomenclature, Symbols. Units and Their Usage in Spcctrochcmical Analysis. Pure Appl. Chem., 1976, 45, 105. Guidelines for Data Acquisition and Data Quality Evaluation in Environmental Chemistry, Anal. Chem., 1980, 52, 2242. Will, F., Anal. Chem.. 1961, 33, 1360. White, C . E., McFarlanc, H. C. E., Fogt, J.. and Fuchs, R., Anal. Chem., 1967, 39, 367. Morishigc, K., Anal. Chim. Acta, 1974, 72, 295. Medina Escriche, J . , and Hernandez Hernandez, F . , Analyst, 1985, 110, 287. Capitan, F., Avidad, R . , Navalon, A., and Capitan-Vallvey, L. F., Mikrochim. Acta, 1992. 107, 65. Morishige, K . , J. Inorg. Nucl. Chem., 1978, 40. 843. 4, pp. 372-546. Paper 21041 45 F Received August 3, 1992 Accepted November 17, I992