ANALYST, JANUARY 1992, VOL. 117 13 Mutual Separation and Preconcentration of Vanadium(v) and Vanadium(iv) in Natural Waters With Chelating Functional Group Immobilized Silica Gels Followed by Determination of Vanadium by Inductively Coupled Plasma Atomic Emission Spectrometry Kazuo Hirayama, Susumu Kageyama and Nobuyuki Unohara Depa rtm en t o f In d us tria I Ch em is tr y, Co lleg e o f Engineering, Nih o n Un ive rsit y, Ko ri yam a , Fu kus h ima 963, Japan The mutual separation and preconcentration of Vv and VIv using chelating functional group immobilized silica gels (CISs) has been studied, based on inductively coupled plasma atomic emission spectrometric detection. Ethylenediamine-bonded (ED-CIS) and ethylenediaminetriacetate-bonded (ED3A-CIS) silica gels were used.A two-column system consisting of a first column of ED-CIS and a second column of ED3A-CIS was developed for on-line separation and preconcentration of Vv and VIv. In the pH range 2.5-3.0, ED-CIS retains Vv and separates it completely from VIv, whereas ED3A-CIS collects both Vv and VIv. The second column can be used to preconcentrate VIv in the breakthrough solution from the first column. Separation is achieved with recoveries of 91-105%. The recovery of both ions is 98% or better. An enrichment factor of 40 is obtained, with which the detection limit of V with inductively coupled plasma atomic emission spectrometry is improved down to 60 pg ml-l. The method was applied successfully to V speciation in natural waters. Keywords: Vanadium speciation; water analysis; separation and preconcentration of vanadium(v) and vanadium(iv); chelating functional group immobilized silica gel; inductively coupled plasma atomic emission spectrometry In recent years, the chemical speciation of trace metals has been focused on the interpretation of their roles in water surveys and in environmental and biochemical studies.I,* Vanadium has many oxidation states and ionic forms in aqueous solution. It usually exists in two different oxidation states, Vv and VIV, in well aerated natural and industrial waters; the significance of V speciation is that the two oxidation states have different nutritional and toxic proper- ties.3" Therefore, it is essential to be able to determine Vv and VIV in environmental and biological samples. In particular, the accurate determination and knowledge of the distribution of V species in complex matrices such as sea-water is required.The concentration of V in natural waters is very low and usually in the range 0.5-2.5 pg 1-1.7 In order to determine ultratrace amounts of V o r other elements in complex matrices by instrumental analysis, a separation and preconcentration technique is frequently required. The separation procedure eliminates sample matrix components that might interfere with the subsequent instrumental analysis, whereas the preconcentration technique concentrates the analytes of interest from a large volume of sample solution. Many preconcentration techniques for the determination of V have been proposed, including chelation and extraction,8-11 extraction, 12 precipitation, I 3 .l 4 coprecipi tation, 15-18 ion exchangel9-21 and the use of chelating ion-exchange resin+- 2s. These preconcentration techniques have been used with instrumental methods of analysis; viz., spectrometry,8121 densitometry,l9 flow injection,20 neutron activation,24 atomic absorption spectrometry,ll,12,15,1~,2s X-ray fluorescence (XRF)13.14,17.26 and inductively coupled plasma atomic emis- sion spectrometry (ICP-AES).9.10.18,2' Most of the preconcen- tration techniques, however, collect Vv and Vrv together, or one species of V in natural waters; no direct simultaneous separation and subsequent determination of the two oxidation states of V has been reported in the literature. Therefore, a more sensitive and selective separation and preconcentration technique for V speciation would be expected to have widespread application.Chelating functional group immobilized silica gel (CIS) has great potential for the preconcentration of trace metals in water.27-30 Leyden and co-workers27,28 used an ethylene- diamine-immobilized silica gel (ED-CIS) and its derivatives (e.g., dithiocarbamate bonded) for the preconcentration of heavy metals in water. Sturgeon et a1.29 developed a separa- tion procedure that involved adsorption of trace metals onto silica-immobilized 8-hydroxyquinoline. In this laboratory, an ethylenediaminetriacetate-immobilized silica gel (ED3A- CIS)3" has been developed for the enrichment of heavy metals and their subsequent determination by XRF. There are two possible modes for the mutual separation of Vv and V'" with CIS, because the two species have different molecular structures and ionic charges in an acidic solution: (1) an ion-pairing mode, which can retain the negatively charged V species by anion exchange, for example, the ammonium ion (i.e., -NH3+) on the ion-exchange site of ED-CIS; and (2) a chelation mechanism in which ED3A-CIS can act as a chelating ligand in the same way as ethylenediamine- tetraacetic acid (EDTA).In the present paper, the development of a method for the mutual separation and preconcentration of Vv and VIV is described using ED-CIS and ED3A-CIS. The two-column system, consisting of an upper column of ED-CIS and a lower column of ED3A-CIS, permits the complete separation and simultaneous preconcentration of Vv and V'".The method was applied to the quantification of these species by ICP- AES. The application of the method to water analysis is also discussed. The method is simple and rapid, and has adequate sensitivity for practical applications. Experimental Apparatus The inductively coupled plasma atomic emission spectrometer used was an SPS-1100 instrument (Seiko Electronic Co., Tokyo, Japan). The operating conditions are summarized in Table 1 together with the emission lines used throughout this work. Measurements of solution pH were made with a Hitachi-Horiba glass electrode pH meter (Model F-7).14 ANALYST, JANUARY 1992, VOL. 117 Table 1 Operating conditions for ICP-AES measurements* Radiofrequency power 1.2 kW Reflected power 4 . 0 W Plasma gas flow rate Nebulizer gas flow rate Auxiliary gas flow rate Observation height Repetition 5 times Analytical line: 16 I min-l 0.4 1 min-1 0.6 1 min-l 12.0 mm above load coil Integration time 1.0s V 292.40 nm c o 228.62 nm c u 324.77 nm * A concentric glass nebulizer was used.Reagents Distilled, de-ionized water (DDW) was used. Sulfuric acid (97%0), HCI (20%) and CH3C02H (100%) were of Super Special Grade (Nacalai Tesque, Kyoto, Japan). Other chem- icals were of guaranteed-reagent grade. A commercial atomic absorption standard solution (Wako Pure Chemicals, Osaka, Japan) containing 1000 mg 1-1 of Vv (NH4V03 in 0.5 rnol dm-3 H2SO4) was used. A stock solution of VIv (1000 mg 1-1) was prepared by dissolving VOS04 in 0.05 rnol dm-3 H2SO4. Each working solution was prepared by diluting the standard solutions with DDW and adjusting the H2SO4 concentration to 0.01 mol dm-3.Acetate buffer solutions were prepared by mixing 1 rnol dm-3 CH3C02Na and 1 rnol dm-3 CH3C02H appropriately. Silica gel (Wako Pure Chemicals, Column Chromatography Grade C-100,40-100 mesh) was first purified by soaking it in 6 rnol dm-3 HCI for 2 d and then washing repeatedly with DDW. Finally, the silica gel was dried in an oven at 110 "C. Syntheses of ED-CIS and ED3A-CIS A 100 ml volume of 20% v/v N-(2-aminoethyl)-y-amino- propyltrimethoxysilane (Tokyo Chemical, Tokyo, Japan) solution and 10 ml of CH3C02H solution were added to 100 g of the dried silica gel and the mixture was stirred for 24 h at 90 "C. The silylated silica gel was washed with DDW, filtered and then dried at 110 "C.ED-CIS ED3A-CIS A 20 g amount of ClCH2COzH and 100 ml of DDW were added to 20 g of ED-CIS and the mixture was stirred for 1 h at 90 "C; during this time, NaOH solution (6 rnol dm-3) was added to keep the solution pH at 9-10. The mixture was then placed in a drying oven at 90 "C for 24 h. The ED3A-CIS obtained was washed repeatedly with DDW until no Cl- ion was detected in the washing water (AgN03 test). Hot DDW was found to be very effective at removing the CI- ion adsorbed on the silica gel. As the amino group of the ED-CIS reacts easily with salicylaldehyde to give a yellow Schiff's base, the completion of the iminoacetate reaction of ED-CIS was checked by using the above reagent. Finally, the ED3A-CIS obtained was dried at 110 "C. Separation and Preconcentration Procedure A glass column for chromatography (250 X 10 mm i.d.) equipped with a porous glass support at the bottom and a PTFE stopcock was used.The length of the packed column was about 30 mm. A small amount of glass wool was placed on top of the packed silica gel to hold it in place during the passage of the solution. 100 80 60 40 - 20 s - C .- 2 0 3 100 F 80 60 40 20 0 1 2 3 4 5 6 7 PH Fig. 1 Effect of pH on the collection of V species by (a) ED-CIS; and (b) ED3A-CIS: A , V"; and B, V'". Vanadium concentration in both instances, 0.2 pg ml-1; volume of sample solutions used, 50 ml. The sample flow rate was fixed at 10 ml min-1 A two-column system was employed: the first column was packed with 1.5 g of ED-CIS and the second column with 1.5 g of ED3A-CIS. The columns were connected with SPC joints (19/19).The sample solution, after adjustment of the pH to 2.5-3.0 with acetate buffer solution and 1 rnol dm-3 HCl, was percolated through the columns. The columns were then washed with two 20 ml aliquots of DDW adjusted to the same pH. After the columns had been separated, the Vv adsorbed on the first column and the VIv adsorbed on the second column were each eluted with 20 ml of 6 mol dm-3 HCl into separate 25 ml calibrated flasks. Then, 25 pg of Coil were added to each flask to give a concentration of 1 pg ml-1 and the final volume was adjusted to 25 ml with DDW. The eluates were analysed by ICP-AES. Results and Discussion Separation and Collection of Vv and Vrv by CISs The collection of the species on the CISs was examined by using the single column mode.Fig. l ( a ) shows that Vv is collected quantitatively from solution over the pH range 2.5-7; however, VIv is not collected at all at a pH of 3.0 or less using ED-CIS. For ED3A-CIS [Fig. l(b)], both Vv and VIv were retained completely in the pH range 2.5-7. These results indicate that when the sample solution is adjusted to a pH in the range 2.5-3.0, the ED-CIS column separates Vv from ViV and the VIV in the effluent from this column can be collected on the ED3A-CIS column. The collection of the V species with the CISs in the optimum pH range was 98% or better. The two-column system consisting of a first column of ED-CIS and a second column of ED3A-CIS was used to facilitate a simple and rapid on-line separation. Table 2 demonstrates that the two species of V are completely separated and recovered quantitatively using this system, even when Vv is present at ten times the concentration of V'".The proposed system offers the unique advantage of being capableANALYST, JANUARY 1992, VOL. 117 15 Table 2 Mutual separation of Vv and VLv using the two-column system. Sample volume: 100 ml VlPg Recovery Species Added Found ("/.I 10 9.97 99.7 0 0.00 - VV 0 0.12 - V'V 10 9.74 97.4 VV 10 9.48 94.8 V1V 1 1.05 105 VV V'" 1 0.99 99 10 9.32 93.2 VV 1 0.95 95 V1V 0 0.00 - VV 0 0.01 - V1V 1 0.97 97 VV 1 1.02 102 V1V 1 0.94 94 VV* 1 0.91 91 V1V 1 0.95 95 * A 500 ml volume of artificial sea-water was used. Table 3 Capacities of ED-CIS and ED3A-CIS. The capacity was determined by the batch method.A sample solution (50 ml) containing the metal ion at an initial concentration of 0.1 mg ml-l was used. A 100 mg amount of CIS was added to the solution and collection was effected by stirring the mixture for 15 min. The solution was then filtered and the metal ion in the filtrate was determined Capacity/mmol g-1 PH VV- 3 6 9 v1v- 3 6 9 Cull- 3 6 9 ED-CIS ED3A-CIS 0.64 0.05 0.39 0.04 0.10 0.02 0 0.07 0.32 0.09 0.14 0.10 0 0.11 0.04 0.28 0.15 0.32 of in situ separation, as a portable unit can easily be carried into the field. The applicability of the method to field use is very attractive because the two V species might be unstable in the sample solution and the determination of trace amounts of V frequently involves preconcentration of large volumes of samples, which might be troublesome to transport back to the laboratory.Collection Mechanism and V Species Vanadium(v) and VIV have many complex forms in water that change in accordance with the solution pH and their concen- trations. It is known that in the pH range 2-6 the main species of Vv is the orange decavanadate anion VI0O2&, which can exist in several protonated forms, and which changes to the dioxovanadium(v) ion V02+ below pH 2.31 In contrast, VIV exists as the blue oxovanadium(1v) ion VO2+ in acidic solution; this cation readily changes to the anion V1x04212- at about pH 4.31,32 Table 3 summarizes the ion-exchange capacities of Vv, V" and Cu" with ED-CIS and ED3A-CIS at solution pH values of 3, 6 and 9. Copper(I1) was used to determine the chelating ability of the CISs.The ED-CIS has a lower chelating ability at lower pH values as shown by the results for the capacity of Table 4 Preconcentration of Vv in highly concentrated salt solutions with ED-CIS. A 100 pg amount of Vv was added to a 50 ml portion of the salt solutions used Salt concentration Relative Salt ("/.I V foundpg error (% ) Na2S04 1 98.6 -1.4 + MgS04 5 99.1 -0.9 NaCl 1 101.3 +1.3 + KCl 5 97.4 -2.6 NaN03 1 98.8 -1.2 + CaCI2 5 100.4 +0.4 Cu"; however, the capacities of Vv and VIV with ED-CIS increase with a decrease of pH. These results suggest that ED-CIS facilitates collection by an ion-pair interaction between the ammonium ion of ED-CIS and the anionic species of V in acidic solution; the collection curves of Vv and VIV shown in Fig. l(a) correspond to the formation of their respective anionic species.Further, ED-CIS can collect other oxometal anions such as those of MoV1 and Crvl, and the VIV-EDTA complex (i.e., the [VO-EDTA12- anion) from acidic solution. Vanadium(1v) has no affinity for ED-CIS at pH 3 because it is present as VO*+. The ED3A-CIS has lower capacities for Vv and VIV in comparison with those of ED-CIS; however, no significant differences between their capacities were observed at pH 3. The higher capacity values of ED-CIS for V species imply that ED-CIS adsorbs polymeric anionic species of V. Many metal cations passed through the ED-CIS column in the pH range 2.5-3.0; however, the ED3A-CIS column not only collected VIV, but also Cu", Zn", Co", Ni", FeIr1 and Pb" at pH 3, because it acted as a chelating agent in the same way as EDTA and had almost the same stability constants with Vv, VIv and Cu" (i.e., log K = 18-19).The advantages of the use of ED-CIS and ED3A-CIS columns in tandem are that the two-column system permits not only V speciation, but also multi-element analysis. Recovery Efficiency A dilute HCI solution is expected to be an effective eluent for Vv and VIV retained on the columns, because these species are not adsorbed by the CISs below pH 1. The effect of HCI concentration on the elution of Vv and VIV was investigated by using the single-column technique. Both Vv and VIV adsorbed on the columns could be recovered quantitatively by using 20 ml of 6 mol dm-3 HCI at a flow rate of 5 ml min-1. The recoveries were 95% or better. The effects of sample volume and flow rate were also examined. The proposed method could be applied to a 1000 ml sample solution at a maximum flow rate of 15 ml min-1.An enrichment factor of 40 was obtained, and the detection limit of V using ICP-AES could be improved to 60 pg ml-1 with the preconcentration technique. The CISs are stable and can be regenerated by passing 100 ml of water through the columns after elution with acid. The separation procedure was carried out ten times on the same column with no loss of efficiency of the extraction and removal of v . Effect of Foreign Ions In order to evaluate the feasibility of the method for water analysis, the effect of a salt-water matrix was studied. Solutions containing two different salts were prepared and the concentration of each salt was adjusted to 1 or 5% m/v.Table 4 shows that Vv is collected quantitatively from the highly concentrated salt solutions by ED-CIS. Vanadium(1v) was16 ANALYST, JANUARY 1992, VOL. 117 Table 5 Linear regression of calibration data for the determination of V by ICP-AES Slope Intercept Correlation Detection limit/ Calibration (R)/ml pg-1 (R) coefficient ngml-l range/pgml-l Internal standard method*- 6 mol dm-3 HCl 4.467 k 0.031 0.0178 f 0.0176 0.9999 2.4 0-5 DDW 4.469 k 0.050 0.0399 k 0.0266 0.9997 1.6 0-5 6 mol dm-3 HCl 1.080 k 0.029 0.0135 k 0.0164 0.9986 2.6 0-5 DDW 1.133 k 0.017 0.0031 k 0.0092 0.9994 1.7 0-5 Direcr calibration method?- * R is defined as the ratio of the intensity of V to that of Co (Iv : Ice). [CO~~]: 1 pg ml-1. t R x 105 is the emission intensity (counts) of V.100 80 s - 60 .- w 2 fi 40 W 20 0 1 2 3 4 5 6 7 PH Fig. 2 Effect of Fell1 concentration on the collection of V*" by ED-CIS. Fell1 concentration: A, 2; and B, 0.2 pg ml-1. Vanadium(1v) concentration, 0.2 pg ml-1. The broken line shows the original collection curve also recovered from the spiked salt-waters by ED3A-CIS. Chelex-100 resin is commonly used for the enrichment of metals in sea-water.22.33 However, it also preconcentrates Ca and Mg cations in addition to trace metals in sea-water; both Ca and Mg cations cause spectral and physical interferences in ICP-AES mea~urements.22~3~ The advantage of the proposed method is that the CISs have no affinity for alkaline earth metals in acidic solution; in addition, the method is applicable to the direct separation .and preconcentration of V species from sea-water and their subsequent determination by ICP- AES.Fig. 2 shows the effect of the Fell' concentration on the collection of V'" by ED-CIS. The adsorption of V'" in the pH range 2-5 appeared to be enhanced as the Fell' concentration increased. However, when 1 ml of 0.1% m/v ascorbic acid solution was added to the sample solution in order to reduce Fel" or Vv, the collection curve of VIv was not affected by Fe"'. This indicates that V'" is oxidized to Vv by Fellr in weakly acidic solution, and that eventually a redox requilibration between VIV and Vv is reached. Calibration Data In ICP-AES measurements, acids in the solution to be analysed are sometimes prone to spectral and physical interferences.35 In the present study, the internal standard method was utilized to overcome the difficulty caused by the variation of the HCI concentration that was used to elute V ions adsorbed on the CISs.Cobalt was chosen as the internal standard, Fig. 3 shows that the emission intensity of V decreased gradually with increasing HCI concentration; however, the ratio of the intensity of V to that of Co was constant and independent of the HCI concentration over the range 0-8 rnol dm-3. As the concentration of Co" in natural waters is usually very Iow,18,29 the intensity of Co from natural waters can be neglected. 1.3 0 3 0 c.' ul z 1.2 > m al c.' .- c.' - 1.1 Fig. 3 1 5 \- I I I 1 ' 1 1 0 2 4 6 8 [HCl]/rnol dm-3 Effect of HCl concentration on the emission intensitv of V (x), and the correction of emission intensity by the internal stindard method (B).Vanadium and Co concentrations, 1 pg ml-l each. The abbreviations I , and Zco represent the intensities of V and Co, respectively The linear regression calibration data for the determination of V by ICP-AES are presented in Table 5. The slope obtained with the use of 6 rnol dm-3 HCl agreed very closely with that obtained with DDW by using the internal standard method. The calibration data measured on different days were reproducible, and the relative standard deviation of the slope was 1.2% (n = 20). The detection limit (defined as the average of the blank value plus three times its standard deviation) was found to be 2.4 ng ml-1 using 6 rnol dm-3 HCI. Applications to Water Analysis The method was applied to the determiantion of Vv and V'" in natural waters. Sea-water and river water samples were filtered through 0.45 ym pore size membrane filters (Milli- pore, Type HAWP 047XX) immediately after sampling.The filtrate was then acidified to pH 2.5-3.0 with HCI (1 mol dm-3) and acetate buffer solution. The results are given in Table 6. By using the proposed method, sub-ng ml-1 levels of V species could be separated and determined with good reproducibility. In order to calculate the recovery, known amounts of Vv and V" were added to the sample solutions. The recoveries of added V'" were lower than those of added Vv for the sea-water (1) and the river water sample; however, the recoveries of total added V were found to be quantitative.As pointed out by Cole et al.," the oxidation states of V change in aqueous solution; it appears that the redox equilibration, VV-VIv, is not particularly stable in natural waters and is closely related to matrix elements such as Fellr and organic substances. The values of the total concentration of V in the acidified and filtered sea-water samples obtained in this work were similar to the reported values.9~~2~~9 It has been ascertained that natural waters contain Vv and Vlv at similar concentration levels, although many workers tend to deter- mine only Vv.ANALYST, JANUARY 1992, VOL. 117 17 ~~~~ ~ Table 6 Determination of V in natural waters. Sea-waters (1) and (2) were coastal sea-waters and were collected at Toyoma Shore and Usuiso Beach, Iwaki, respectively.River water was collected from the Abukuma River, Koriyama. All locations are in Fukushima Prefec- ture, Japan. All determinations were performed by using 500 ml sample solutions. The reproducibility was calculated from five replicate determinations Sample* Sea-water (1)- VV VIV VV V1V VV V1V VV VIV V t VV V1V VV V1V V t VV V'V VV VIV vt Sea-water (2)- River water- V Added/ pg 0 0 1 2 0 2 2 0 0 0 0 1 1 0 0 0 1 2 0 Foundlpg 0.54 _+ 0.054 0.37 ? 0.053 1.95 1.97 1.20 1.78 1.80 1.03 1.05 0.50 t 0.03 0.28 k 0.03 1.49 1.31 0.77 0.31 k 0.03 0.26 k 0.03 1.52 2.06 0.86 V in sample/ ng ml- 1.08 * 0.11 0.74 k 0.11 1.90 2.40 - - - 2.06 2.10 1.00 4 0.06 0.56 k 0.06 0.98 0.62 1.54 0.62 k 0.06 0.52 k 0.06 1.04 0.12 1.72 Total V concentra- tion/ ng ml-1 1.82 1.84 1.96 1.66 2.10 1.56 1.60 1.54 1.14 1.16 1.72 * Samples were filtered and then adjusted to pH 2.5-3.0 im- 7 Samples were acidified to pH 1 and then filtered.The total mediately after collection. amount of V was determined using ED3A-CIS at pH 3. In general, V exists in soluble, insoluble and organic complex forms in natural waters. The determination of the oxidation state of V is important in water surveys. When the natural water samples were acidified to pH 1 immediately after sampling and then filtered, the total V concentration in the river water was higher than that of the filtered sample. For sea-water, no significant difference was observed between the values obtained with the two processes. These results indicate that V exists mainly in soluble forms in sea-water; however, in river water some of the V is present as insoluble or colloidal forms and is filtered with particulate matter.The proposed method can provide information about soluble Vv and V'" in natural waters, particularly sea-water, whereas direct analysis using other techniques is difficult or not possible. References 1 2 Florence, T. M., and Batley, G. E., CRC Crit. Rev. Anal. Chem., 1980, 9, 219. Nakayama, E.. Suzuki, Y., Fujiwara, K., and Kitano, Y., Anal. Sci.. 1989, 5 , 129. 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 Shamberger, R. J., in Toxicity of Heavy Metals in the Environ- ment, ed. Oehme, F. W., Marcel Dekker, New York, 1979, pt. Lagerkvist, B . , Nordberg, G. F., and Vouk, V., in Handbook on the Toxicology of Metals, eds.Friberg, L., Nordberg, G. F., and Vouk, V. B., Elsevier, Amsterdam, 1986, vol. 11, pp. 638-663. Morrison, G. H., CRC Crit. Rev. Anal. Chem., 1979, 8,287. Sadler, P. J . , Higham. D. P., and Nicholson, J . K . , in Environmental Inorganic Chemistry, eds. Irgolic, K . J., and Martell, A. E., VCH, Deerfield Beach, FL, 1985, pp. 255 and 256. Bowen, H. J. M., Environmental Chemistry of the Elements, Academic Press, New York, 1979, pp. 17 and 23. Ramana Murthy, G. V., Sreenivasulu Reddy, T., and Brahmaji Rao, S., Analyst, 1989, 114, 493. Tao, H., Miyazaki, A., Bansho, K . , and Umezaki, Y., Anal. Chim. Acta, 1984. 156, 159. Sugiyama. M., Fujino, O., Kihara, S . , and Matsui, M., Anal. Chim. Acta, 1986, 181, 159. Monien, H., and Stangel, R., Fresenius Z.Anal. Chem., 1982, 311, 209. Shijo, Y., Kimura, Y., Shimizu, T., and Sakai, K., Bunseki Kagaku, 1983,32, E285. Hirayama, K., and Leyden, D. E., Anal. Chim. Acta, 1986, 188, 1. Saitoh, Y., Yoneda, A., Maeda, Y., and Azumi, T., Bunseki Kagaku, 1984. 33, 412. Fujiwara. K., Morikawa, T., and Fuwa, K . , Bunseki Kagaku, 1986, 35, 361. Shimizu, T., Uchida, Y., Shijo, Y., and Sakai, K . , Bunseki Kugaku, 1981,30, 113. Cole, P. C., Eckert, J. M., and Williams, K. L., Anal. Chim. Acta, 1983, 153, 61. Akagi, T . , Fuwa, K., and Haraguchi, H., Anal. Chim. Acta, 1985, 177, 139. Shriadah. M. M. A., and Ohzeki, K., Analyst, 1985, 110, 677. Fukasawa, T., Kawakubo, S . , Okabe, T., and Mizuike, H . , Bunseki Kagaku, 1984, 33, 609. Kiriyama, T., and Kuroda, R., Mikrochim. Acta, Part I , 1985, 405. Cheng, C. J., Akagi, T., and Haraguchi, H., Bull. Chem. SOC. Jpn.. 1985,58, 3229. Rueter, J . , and Schwedt, G., Fresenius Z. Anal. Chem., 1982, 311, 112. Greenberg, R. R.. and Kingston, H. M., Anal. Chem., 1983,55, 1160. Colella, M. B., Siggia, S . , and Barnes, R. M., Anal. Chem., 1980. 52,967. van Grieken, R.. Anal. Chim. Acta, 1982, 143, 3. Leyden, D. E., and Luttrell, G. H., Anal. Chem., 1975, 47, 1612. Leyden, D. E., Luttrell, G. H., Nonidez, W. K., and Werho, D. B., Anal. Chem., 1976,48, 67. Sturgeon, R. E., Berman, S. S., Willie, S. N., and Desaulniers, J . A. H., Anal. Chem., 1981, 53,2337. Hirayama, K., and Unohara, N., Bunseki Kagaku, 1980, 29, 452. Cotton, F. A , , and Wilkinson, G., Advanced inorganic Chemistry, Wiley, New York, 4th edn., 1980, pp. 708-719. Johnson, G. K . , and Schlemper, E. O., J. Am. Chem. SOC., 1978, 100, 3645. Kingston, H. M., Barnes, I. L., Brady, T. J., Rains, T. C., and Champ, M. A., Anal. Chem., 1978, 50,2064. Berman, S. S . , McLaren, J. W., and Willie, S. N., Anal. Chem., 1980, 52,488. Greenfield, S . , McGeachin, H. M., and Smith, P. B., Anal. Chim. Acta, 1976, 84,67. 2, pp. 745-751. Paper 1102685B Received June 5, 1991 Accepted August 19, 1991