JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY APRIL 1991 VOL. 6 233 Flow Injection Flame Atomic Spectrometric Determination of Aluminium Iron Calcium Magnesium Sodium and Potassium in Ceramic Material By On-line Dilution in a Stirred Chamber Vincente Carbonell Angel Sanz Amparo Salvador and Miguel de la Guardia Department of Analytical Chemistry University of Valencia 50 Doctor Moliner Street 46 I00 Burjasot Valencia Spain ~~ The on-line dilution of ceramic samples previously fused with alkaline carbonates increases the dynamic range of the flame atomic spectrometric determination of aluminium iron calcium magnesium sodium and potassium. A double-channel manifold with a simultaneous double injector and a versatile system of dilution with stirred cham- bers allows direct determination of the six elements studied in different ceramic matrices i.e.clay kaolin quartz felspar varnish stoneware and porcelain rapidly and with low consumption of the reagents. The effect of the flow injection parameters on the analytical characteristics of the flame atomic spectrometric determination has been studied and the results obtained in the analysis of real samples have been compared with those found in batch analyses. Keywords Ceramic; flow injection; determination of aluminium iron calcium magnesium sodium and potas- sium; flame atomic spectrometry The analytical determination of the elemental composition of ceramic materials is very important in order to control the characteristics of raw materials chosen for use and to test the quality of finished products.'.' Both the major and minor components of ceramic matrices must be simply but selectively determined; flame and plasma spectrometric methods are ideal for this purpose.3-s Flame atomic absorption and flame atomic emission spectrometry are currently the most widely used techniques for the determina- tion of various elements e.g.aluminium iron calcium mag- nesium sodium and potassium in ceramic samples. However problems arise because flame spectrometry has a small dynamic range and the sensitivity of the technology is too high for the determination of the minor components hence for the practical analysis of ceramic materials a series of dilu- tions is required in order to obtain absorbance or emission readings that fall within the linear range of the calibration graph.The principles of flow injection analysis have been applied in recent years in all branches of analytical hemi is try.^.^ One area in which the application of flow injection (FI) methods improves the performance of the analytical technique is atomic spectrometry."-' The use of FI in conjunction with flame spectrometry can not only provide on-line pre-concentration dilution or treat- ment of the samples but can also extend the linear calibration range. A simple way to dilute the samples is by the injection of a small sample volume using normal-" or pulse-timed solenoid injectors,I3 with the use of a high volume of the transport medium. l 4 Other strategies that have been proposed in order to extend the dynamic range of the calibration graph are based either on the treatment of the absorbance peaks," or in the use of an appropriate manifold. In the latter different systems have been proposed such as a variable dispersion manifold with di- lution coils of different sizes;I6 the use of a multi-line FI system with zone sampling;I7 a cascade dilution;Ix a gradient chamber based on zone ampl ling;'^.^^) or split-zone flow injec- tion.I On the other hand it is true that the dilution occurring in a mono-line FI system is highly reproducible. Therefore by dis- crete dilution of the samples direct analysis of the minor com- ponents using simple manifolds is possible.22-2s In this paper a versatile manifold that allows on-line dilu- tion of samples for analysis by flame atomic spectrometry is proposed and the determination of aluminium iron calcium magnesium sodium and potassium in eight different ceramic materials is described.Experimental Apparatus and Reagents A Perkin-Elmer 5000 and a Perkin-Elmer 2380 flame atomic absorption spectrometer have been used to carry out the emission and absorbance measurements. With the latter calcium iron and magnesium hollow cathode lamps were used. The experimental parameters employed are listed in Table 1 . In order to provide sample introduction and on-line dilu- tion of the samples the manifolds shown in Fig. 1 were em- Table 1 Instrumental parameters employed for the determination of aluminium iron. calcium magnesium sodium and potassium Element Parameter Al Fe Ca Mg Na K Wavelength / nm 309.3 248.;1 422.1 285.2 589 766.5 Technique FP* AA ' AA AA FP FP Flame N,GC,H Air-C2HI Air-C,H2 Air-C2H2 Air-C,H Air-C,H Burner height (arbitrary units) 6 6 5 5 5 5 Burner angle (") 180 I80 180 180 I80 180 * FP Flame photometry. AA Atomic absorption.234 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY APRIL 1991 VOL.6 I - 1 I I I ( C) ( d ) I I I I r\ - 1 R U I Fig. 1 Manifolds employed to carry out the on-line dilution of ceramic samples prior to their analysis by FAAS. P. Peristaltic pump; I. injection valves; C three-way connector; MI and M2. mixing chambers of 1 and 3 ml volume. respectively; S magnetic stirrer; and R recorder ployed. A Gilson Minipuls HP4 peristaltic pump ensured a constant carrier flow of water through a double channel. A double injector allowed the simultaneous introduction of both sample and lanthanum solution.A merging point pro- moted the mixing of both channels before the streams passed through a stirred mixing chamber. This system pro- vided on-line dilution of the samples. If a higher dilution was required two chambers were used. After dilution samples were fed to the atomization system directly and the absorbance peaks were recorded. A Rheodyne injection valve Type 50 with various fixed volume loops was used for sample introduction. A Y-shaped three-way connector (Omnifit) was used as the merging point. The mixing chambers were laboratory-built from polytetrafluoroethylene (PTFE) according to a model previously patented for the on-line treatment of samples.'h The chambers were composed of two parts a lower cylindrical portion with a 12.2 mm i.d.and a height of either 16.8 or 24.3 mm and an upper part with a slight dome which ensured Y r 1 Fig. 2 Geometry of the mixing chambers employed Table 2 Preparation of the lo00 pg I-' stock solutions Standard Masslg Salt Acid Mg 3.9 I60 MgC 1 ,.6H,O - 25 ml 1 mol dm- HCI Ca 2.4973 CaCO Fe 1.oooO Fe 25 ml concentrated HNO 25 ml concentrated HCI+ Al 1.oooO Al K 1.9070 KCl - Na 2.5420 NaCl - drops of HNO the escape of any air bubbles. Fig. 2 shows the geometry of the mixing chambers used. All the connecting tubes were of PTFE with 0.8 mm i.d. Stock solutions of magnesium calcium iron potassium and sodium (lo00 mg 1 - I ) were prepared by dissolving the appro- priate analytical-reagent grade salt in de-ionized water as summarized in Table 2 and a 1 % lanthanum solution was pre- pared from LaCI,.7HI0.General Procedure Digestion of the sample Weigh 0.3 g of sample into a platinum crucible and add 1.5 gJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY APRIL 1991 VOL. 6 235 Table 3 Normal content of aluminium iron calcium magnesium sodium and potassium in different ceramic matrices Concentration (for 0.3 g of sample in 250 ml)/mg I-' Matrix Clay Kaolin Quartz Felspar Varnish Mayolica Stoneware Porcelain Al 110 225 2 108 64 95 116 146 Fe Ca Mg Na K 7 0.9 2 4 17 4 1 0.7 2 7 0.4 0.9 0.1 1.2 0.8 1 3 0.4 17 105 1 8.7 5 10 39 4 102 4 3 9 10 2 2 7 26 2 1 1 11 34 Table 4 Effect of the manifold employed on the dynamic range of the determination of magnesium by FI-flame atomic absorption spectrometry. Manifold configurations employed ( h ) 1 ml mixing chamber; (c) 3 ml mixing chamber; (6) 1 ml plus 3 ml on-line mixing chambers; and ( e ) 3 ml plus 1 ml on-line mixing chambers.Calibration graph equations are given in absorbance units and the concentration expressed in mg I-'. The dynamic range has been established from the characteristic concentration obtained in the experimental conditions in the upper part of the calibration graph Manifold Dynamic range/ Calibration Regression mg I-' graph equation coefficient ( h ) 0.17-5 A = O.OOO9 + 0.0251. 0.9996 ( 0.13-10 A = 0.002 + 0.034~ 0.9998 (4 0.12-30 A = 0.002 + 0.036~ 0.9998 (el 0.16-50 A = 0.01 + 0.028~ 0.9997 (4 0.1 1-40 A =-0.03 + 0.039~ 0.998 of sodium carbonate (Na,CO,) or potassium carbonate (K2C0,) and 0.3 g of boric acid.Introduce the crucible into a muffle furnace at 1100 OC after 40 min allow to cool to 700- 800 "C. Transfer the contents of the crucible into a 250 ml por- celain evaporating basin add 50 ml of HCI (1+1 v/v) and stir until the solids are completely dissolved. Place the basin under an infrared (IR) lamp and evaporate to dryness at 130 "C. Add 5 ml of concentrated HCI to re-suspend the silica (SiO,) and then dilute to 100 ml with HCI (1+1 v/v). Filter using a red- band without ash Schuller filter-paper and dilute to 250 ml. Batch analysis Samples previously fused were diluted to an appropriate volume in order to obtain absorbance readings within the linear part of the calibration graph of each element. Hence di- lutions of from 1+4 to 1+99 of the previously obtained solu- tions were made.Alternatively sodium or potassium could be added as an ionization buffer and for the determination of calcium and magnesium lanthanum could be added as an in- terference buffer. Flow injection Appropriate volumes of the previously digested samples were injected into an FI manifold simultaneously with an equal volume of I % lanthanum solution. For each element in each matrix the manifold and the volume of sample injected can be modified in order to obtain a linear relationship between absorbance readings and the concentration of the element. Results and Discussion Effect of the Manifold Employed on the Dynamic Range The content of aluminium iron calcium magnesium sodium and potassium in ceramic materials obviously varies in different matrices. Table 3 summarizes the normal con- centration of each element found in various ceramic materi- als (in mg I-') for a dilution level of 0.3 g of sample in 250 ml.From these data and taking into account that the upper level of the calibration graphs in flame spectrometry corresponds to 100 mg I-' of aluminium 5 mg I-' of iron and calcium 2 mg I-' of potassium I mg I-' of sodium and 0.5 mg 1-' of magnesium it can be concluded that it is necessary to carry out different di- lutions for the determination of each element in each type of sample. One of the properties of the FI manifolds is that their com- ponents can be easily changed and so a simple manifold as indicated in Fig. 1 can be arranged in order to use ( a ) a dilu- tion coil 2 m in length without the use of a mixing chamber; ( h ) two short coils and a small ( 1 ml) stirred mixing chamber; (c') two short coils and a stirred chamber with a high volume (3 Table 5 Calibration graph data obtained for the determination of aluminium iron calcium sodium and potassium using different manifolds for the on- line dilution of samples. Dynamic range has been established from the characteristic concentration obtained in the experimental conditions in the upper part of the calibration graph Element Manifold Injected volume/pl Carrier flow/ml min-' Dynamic range Calibration graph equation* Regression coefficient 8.2 8-250 E= 0.002 + 0.00055~ 8.2 4-250 E= 0.002 + 0.00 1 I c 8.2 3.7-250 E= 0.005 + 0.012~.8.4 0.61 2 A= 0.0002 + 0.0074~ 8.4 0.1-12 A= 0.005 + 0.032~ 8.4 0.3-1 2 A=-0.0007 + 0.0 17~' 8.4 0.5-10 A= 0.006 + 0.0096~ 8.4 0.2-10 A= 0.0 19 + 0.020~ 8.4 2.2-1 26 A= 0.001 + 0.002~ 0.9996 0.9994 0.9995 0.9998 0.999 1 0.9996 0.9997 0.999 1 0.9999 Na (('1 100 8.2 0.2-14 E= 0.0 19 + 0.020~ 0.9993 8.2 2.2-140 E= 0.002 + 0.002~ 8.2 0.7-50 E= 0.0009 + 0.0064~ 8.2 0.3-20 E= 0.002 + 0.014~ 8.2 0.2-10 E= 0.003 + 0.023.0.9995 0.9996 0.9997 0.9998 * A Absorbance in absorbance units; E emission in emission units.236 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY APRIL 1991 VOL. 6 D 5 min - c- Time Fig. 3 Effect of the dilution levels on the magnesium peak shapes A 4; B 8; C 16; and D 32 mg I-'. The manifold employed in each instance in- cludes A a 1 ml volume mixing chamber [Fig. I(h)]; B. a 3 ml mixing chamber [Fig. 1 (c)]; C I ml plus 3 ml on-line mixing chambers [Fig. I(d)]; and D 3 ml plus I ml on-line mixing chambers (Fig.I(e)l ml); (6) two on-line stirred chambers the first being that with the lower volume ( 1 ml) the second with the higher volume (3 ml); and ( e ) two on-line stirred chambers the first being the higher volume chamber (3 ml) and the second the lower ( 1 ml) (see Fig. I.) The five configurations of the manifold allow different dilu- tions of the same sample and so can provide different dynamic ranges for the determination of the elemental compo- sition of a sample by flame atomic spectrometry. Magnesium has been used as a test system in order to deter- mine the effect of the FI manifold on the upper level of the cal- ibration graphs. By using an injected sample volume of 100 pI and a carrier flow-rate of 8.2 ml min-I a series of standards was injected into the manifold and the corresponding calibra- tion graphs calculated from the data obtained.Table 4 sum- marizes the equations found and shows that the dynamic range can be increased from 5 mg I-' using a dilution chamber with a 1 ml volume to 50 mg I-' using the 3 ml plus the 1 ml on-line dilution chambers. It is interesting to note that using manifold configuration (6) (in Fig. l) a good linear relationship is obtained for 0.12-30 mg I-' of magnesium but a poor relationship for 0.1 1-40 mg I-' therefore the use of the 3 ml chamber before the 1 ml chamber [Fig. l(e)] is preferable as this arrangement provides a greater sample dilution and hence a better dynamic range. The changes in the manifold can be made more easily and hence are more convenient than the use of different flame pa- rameters.They therefore allow the absorbance measurements to be performed using the optimal experimental conditions. However it is necessary to indicate that the use of greater dilu- tion levels requires a longer time delay and that corresponding- ly the peak shapes are broader (see Fig. 3). From the data in Table 3 and the results obtained for magne- sium the appropriate manifolds have been selected for the de- termination of the other five elements. The details from the corresponding calibration graphs obtained are summarized in Table 5. Effect of the Injected Sample Volume Table 5 shows that the injection of larger sample volumes im- proves the sensitivity of the analytical method.A study of the influence of sample volume on the absorbance values obtained for all the elements determined has been carried out using the most appropriate manifold configuration. The results obrained (Fig. 4) show that a dramatic increase in absorbance or emission is produced when the injected volume is increased from 100 to 500 p1 but the dynamic range can also be reduced. However this effect varies for the differ- ent elements considered. The loss of linearity in the calibration graphs are shown by the poorer regression coefficients. This effect is most marked for the manifold in which there is no di- lution chamber. Effect of the Carrier Flow-rate In order to obtain stable and reproducible results in FI-flame atomic spectrometry it is necessary to work with a pump flow- rate that is greater than the natural aspiration rate of the nebu- lizer." The analytical sensitivity in FI-flame atomic spectro- metry has been previously shown to increase when the pump flow-rate increases.2x By using the most appropriate manifold configuration for each element considered the influence of the pump flow-rate on the absorbance has been studied.Fig. 5 shows that sensitivity can be increased by increasing the carrier flow-rate so that the time taken for the analysis can be improved without reducing the dynamic range of the cali- bration graph. 1 A 100 500 Volume injected/pl 1 ooc Fig. 4 Effect of the volume of sample injected on the absorbance of A 20 mg I-' of magnesium; B I0 mg I-' of iron; F 10 mg I-' of calcium and on the emission of C 5 mg I-' of sodium; D 200 mg I-' of aluminium; and E 10 mg I-' of potassium I I 6 9 0- 3 Carrier flow-rate/mI min ' Fig.5 Effect of the carrier flow on the absorbance of F. iron (10 'mg ml-I); C . calcium (26 mg ml-I); and D magnesium (20 mg m1-I). and o n the emission of A aluminium (250 mg mi-'); B. sodium (8 mg ml-I); (and E. potassium (I0 mg ml-')JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY APRIL 1991 VOL. 6 237 Analysis of Real Samples Samples of ceramic materials of different types have been treated as described under General Procedure and analysed by both flame atomic spectrometry using conventional off-line di- lution of the samples with continuous aspiration into the flame and the proposed FI procedure. The data obtained are given in Table 7; both procedures provided similar results.However the FI method is less time consuming avoids the accumulation of salts in the nebulizer because of the dilution of the samples and the use of the merging point significantly reduces the lan- thanum consumption and hence reduces the cost of the analy- sis. Table 6 Analytical parameters of the FI determination of Al Fe Ca Mg Na and K in ceramic materials. Sensitivity has been calculated from the slope of the calibration graph Element Manifold Sensitivity LOD*/ RSDt mg 1-I (%) A1 ( 0 ) 0.55 a.u.e. mg-' 1' 1.5 0.84 Fe ( 0 ) 0.92 cm mg-' 1 0.09 0.63 Ca ( h ) 0.15 cm mg-I 1 0.31 1.34 Mg (6) 0.47 cm mg-'l 0.17 0.69 Na ((') 0.37 cmmg-' 1 0.24 0.75 K ( c ) 0.1 I cm mg-' 1 0.73 2.12 * LOD limit of detection fork = 2.t RSD relative standard deviation. t a.u.e. arbitrary units of emission. Use of the Dilution Profile for Calibration Purposes The use of dilution chambers has been shown to allow the use of an absorbance-time profile for a single standard as a contin- uous calibration Fig. 6 shows as an example the absorbance-time profile of a concentrated standard of magnesium and the absorbance peaks corresponding to various other standards. The first part of the absorbance profile cannot be used for calibration pur- poses but the absorbance values corresponding to different times in the tailing part of the graph can be used. The relationship between concentration and time corre- sponds to an exponential equation c=c,[ l-exp(-ut/V )] ( 1 ) where c is the effluent concentration at time f; c is the concen- tration of the calibration standard; u is the flow-rate; and V is the volume of the manifold (including the mixing chamber).' However a series of practical problems related to the exact determination of the experimental parameters disturbs the use of the theoretical dilution profiles.Empirical approaches of the Analytical Parameters The analytical parameters such as sensitivity limit of detec- tion and precision have been established for each of the elements considered using the most appropriate manifold configuration and FI parameter values. These data are sum- marized in Table 6. The sensitivity has been determined from the slope of the re- gression line of the calibration graph. The limit of detection corresponds to the concentration of the element in the digest- ed solution which provides an absorbance value equal to twice the standard deviation of the blank readings.The preci- sion has been derived from the relative standard deviation of ten measurements of a single sample containing a concentra- tion corresponding to the central part of the dynamic range. The data obtained show that the proposed procedure gives adequate sensitivity and precision. Table 7 Li2C0 Analysis of ceramic samples. ( 1 ) Samples fused with K,CO,; (2) samples fused with Na2C0,; (3) certified value; and (I) samples fused with Concentration found (%) CaO Na20 Sample* FI Batch FI Batch FI Batch FI Batch FI Batch FI Batch Clay sI (1) 29.17 27.25 (2) 29.18 27.56 0.17 0.15 0.19 0.15 0.30 0.33 0.30 0.33 0.41 0.48 - - 1.08 1.07 1.10 1.07 - - 1.88 1.85 - - Clay s2 (4) Kaolin (4) 34.01 33.82 0.17 0.1 1 0.23 0.34 0.82 0.90 0.14 0.15 0.13 0.11 5.72 4.01 0.32 0.31 0.35 0.37 0.12 0.13 0.04 0.02 0.18 0.13 0.06 0.04 Felspar sI (2) 18.95 17.30 0.51 0.49 0.05 0.06 11.40 11.24 0.13 0.14 Felspar s2 ( 1 ) 17.59 17.32 (2) 17.10 17.30 0.72 0.55 0.69 0.52 0.05 0.03 0.05 0.02 2.67 2.92 - - 0.17 0.10 0.13 0.1 1 - - 9.91 11.03 Felspar (3) 17.7 0.54 0.04 11.20 2.83 0.10 9.58 10.48 - - Varnish (4) 0.46 0.50 0.13 0.13 0.41 1.12 Stoneware s ( I ) - - 0.35 0.23 0.16 0.21 0.54 0.62 1.17 1.12 Stoneware s2 ( 1 ) 18.80 18.02 (2) 18.46 17.38 0.25 0.28 0.27 0.28 0.21 0.23 0.21 0.23 0.66 0.65 - - 1.13 1.12 1.15 1.19 - - 2.36 2.49 Porcelain s (4) 22.83 23.60 1.05 1.10 0.13 0.20 3.01 2.56 0.87 0.96 0.30 0.31 Porcelian s2 ( 1 ) 25.82 24.90 (2) 27.48 25.65 1.31 1.16 1.30 1.16 0.17 0.18 0.17 0.19 1.37 1.34 - - 0.33 0.26 0.33 0.28 - - 4.10 4.50 * s Sample 1; s2 sample 2.238 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY APRIL 1991 VOL.6 - Time A A J 5 min A' I I C ~ Fig. 6 Absorbance-time profiles of A a concentrated magnesium stan- dard (40 mg I-' of magnesium) and B C and D corresponding to standards containing 30 20 and 10 mg ml-' of magnesium respectively Table 8 Equations of dilution obtained for each manifold Manifold Element Maximum Equation of Regression standard concen- dilution coefficient tration/mg ml-' ( h ) Ca I26 c = 1.08~0.91' 0.996 (($1 Na 14 c. = 1.14 x 0.92' 0.96 (4 Mg 30 c = 2.15 x 0.96' 0.992 (el K 140 c = 1.12~0.96' 0.990 type proposed by Olsen et alez9 can be used more easily.The empirical relationship between the absorbance at each time value of the dilution profile using the most concentrated standard and the experimental absorbance values found by discrete injection (in the same manifold) of standards with a known concentration provides a dilution equation c=AB' (2) in which A and B are empirical coefficients where A is directly proportional to the concentration of the standard injected and B depends on the manifold volume and the flow-rate employed and t is the time (in seconds). Using this empirical equation the dilution absorbance- time profile can be used as the calibration graph to relate the absorbance values obtained to the concentration of the sample. The different dilution equations obtained for each manifold configuration considered are shown in Table 8.It has been previously reportedz5 that the same empirical equa- tion for the dilution profile could be used for different ele- ments when using the same manifold with the same injected volume and carrier flow while taking into account the differ- ence between the concentration of the standard used for cali- bration. Conclusions The studies carried out demonstrate that the appropriate selec- tion of a simple manifold permits a rapid determination of a range of elements at different concentration levels in the same sample without off-line dilutions and that this approach can be useful in applied analyses. 1 2 3 4 5 6 7 8 9 10 1 1 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 .2 7 .2 8 29 References Bennet H.and Hawley W. G.. Methods of Silicate Analysis. Aca- demic Press London 2nd edn. 1965. Singer F. and Singer S. S. Industrial Ceramic Chapman & Hall London 197 1. Buckley F. Ramsey M. H. Rooke J. M. Hughes H. and Norman P. J. Anal. At. Specworn. 1988,3 203R. Voinovitch I. Analyse des Sols Roches et Ciments Masson Paris 1988. Astraitsis J. Zachariadis G. A. Dinitrakonidi E. D. and Limeonov V. Fresenius Z. Anal. Chem. 1988,331,725. R;%ka J. and Hansen E. H. Flow! Irtjection Analysis Wiley New York 1988. Valcarcel M. and Luque de Castro M. D. Flow Injection Analysis. Principles and Applications Ellis Horwood Chichester 1987. Tyson J. F. Analyst 1985 110,419. RiiiEka J. Fresenius Z. Anal. Chem. 1986,324. 745. Tyson J. F. Anal. Chim. Acta 1988,214,57.Flow Injection Atomic Spectroscopy ed. Burguera J. L. Marcel Decker New York 1989. Burguera J. L. Burguera M. and Alarcon 0. M. J . Anal. At. Spec- ti'orn. 1986 1.79. Burguera J. L. Burguera M. Rivas C. de la Guardia M. and Sal- vador A. Anal. Chim. Acta 1990,234,253. Basson W. D.. and van Staden J. F.. Fresenius Z. Anal. Chem. 1980 302. 370. Tyson J. F. Analvst 1984 109 3 19. Tyson J. F. Mariara J. R. and Appleton J. M. H. J. Anal. At. Spec- trom. 1986 1,273. Reis B. F. Jacintho J. Moratatti J. Krug F. J. Zagatto E. A. G. Bergamin F. H. and Pessenda L. C. R. Anal. Chim. Acta 1981 123 22 I . Whitman D. A.. and Christian G. D. Talanta 1989,36,205. Toei J. Anal. Lett. 1988 21 1633. Garn M. B. Gisin M. Gross H.. King P. Schmidt N. and Thommen C. Anal. Chim. Acta 1988,207,225. Clark G. D. RhziEka J. and Christian G. D. Anal. Chem. 1989 61 1773. RkiEka J. and Hansen E. H. Anal. Chim. Acta 1978 99,37. Stewart K. K. and Rosenfeld A. G. Anal. Chem. 1982,534,2368. Gisen M. Thommen. C. and Mansfield K. F. Anal. Chim. Actu 1986,179 149. de la Guardia M. Carbonell V.. Morales. A. and Salvador A Fresenius Z. Anal. Chem. 1989,335975. de la Guardia M. Morales A. Carbonell V. and Salvador A. Spanish Pat. 9002644 1990. Brown M. W. and RkiEka J. Analjist. 1984 109 1091. Carbonell V. de la Guardia M. Salvador A. Burguera J. L. and Burguera. M. Microc~hem. J. 1989,40 233. Olsen. S. RiziEka. J. and Hansen E. H. Anal. Chim. A m 1982 136. 101. Paper- 0104 I86F Recei\?ed September- 13th I990 Accepted December- loth 1990