DECEMBER, 1973 THE ANALYST Vol. 98, No. 1173 Application of the Carbon Cup Atomisation Technique in Water Analysis by Atomic-absorption Spectroscopy BY F. DOLINSEK AND J. STUPAR (The " Jofej Stefan" Institute, University of Ljubljana, Ljubljana, Yugoslavia) A modified, laboratory-made carbon cup (small-scale Massmann) atomiser is described, with particular reference to the atomic-absorption determination of copper, lead and cadmium in water samples. Several parameters such as sample volume, time and temperature of the atomisation steps, and sample composition, have been investigated. It was found that injection of a 10-pl sample in one portion is the most con- venient technique with respect to sensitivity and speed of operation. Addition of EDTA causes an enhancement of sensitivity, which is considerable when determining lead.The adsorption of these elements on the polyethylene containers has also been examined in order to evaluate possible errors that may arise after sample storage. The detection limits are 0.45 ng ml-l of lead, 1-7 ngml-l oi copper and 0.04 ng nil-I of cadmium, and the average precision is f3 per cent. in a single measurement. The method permits the direct and rapid determination of these elements in various water samples, which determinations are frequently required in pollution control. CONCENTRATIONS of several heavy metals such as lead, cadmium and copper in public water supplies are limited to certain safe levels because of the toxic character of these metals. For example, by several national criteria for water quality the following levels are permitted: lead 50, cadmium 5 to 10 and copper 10 to 1OOOngml-l.Different industrial wastes, particularly such as those from plating, zinc and lead smelting, etc., discharge significant amounts of these elements in various forms. The concentration of these metals in waters may therefore rise to a level that can be hazardous to livestock. Monitoring of natural waters and industrial effiuents is thus becoming increasingly important, particularly in highly industrialised areas. The determination of metals at nanogram levels, however, presents a difficult task. In addition to the need for high sensitivity, a rapid analytical method is required that will enable the large number of samples involved in water pollution control to be dealt with. Among the various analytical methods available, anodic stripping polarography,l neutron a~tivation,~*~ flame atomic-absorption spectro~copy~-~~ and spectroph~tometryl~-~~ seem to be the only suitable techniques for the determination of lead, cadmium and copper at the parts per billion ( lo9) level.However, pre-concentration of solutions (for flame atomic- absorption spectroscopy) or chemical treatment prior to the determination (for polarography and spectrophotometry) restrict the lower limit of detection because of the relatively high blank values. Neutron activation requires expensive instrument facilities and is not sensitive enough to permit direct determination of cadmium and lead at the desired concentration. The possibility of direct analysis and the high sensitivity achieved by flameless atomic-absorption spectroscopy make this technique particularly advantageous for the determination of these elements in liquid samples.Fernandez and Manning15 determined twelve elements directly in natural waters by using a Massmann-type graphite atomiser (HGA-70); 50 to 100 pl of sample were used for a single measurement. The standard addition method was used in order to calculate the results. Pickford and Rossi16 designed an automatic sampling system in conjunction with a graphite atomiser (HGA-70). A high degree of sensitivity was reported as a result of the large sample volumes (100 to 500 pl) taken. The aim of the present work was to develop the carbon cup technique into a speedy and reliable tool for the determination of these elements in actual water samples, concen- trating especially on the parameters that affect its accuracy and sensitivity.841 Also, the accuracy and speed of analysis are seriously affected. @ SAC and the authors.842 EQUIPMENT- Hollow-cathode lamps (lead, cadmium, copper and hydrogen) were operated at the currents recommended by the manufacturer (Varian). The burner was replaced by a locally made carbon cup attachment that was similar to the Varian carbon cup atomiser, Model 63. A Perkin-Elmer 165 chart recorder (10 mV, full-scale response time 1 s) was connected to the instrument. The power unit supplied a maximum of 600 A at 1OV to the carbon cup atomiser. In Fig. 1 the circuit diagram of the power unit is shown. In order to achieve the various temperatures required for atomisation of different samples, the current passing through the atomiser can be varied by changing the voltage on the primary coil of a step-down transformer. Five different temperatures can be selected for drying, five for ashing and five for the final atomi- sation.The timers automatically control the sequence of operations. DOLINSEK AND STUPAR: CARBON CUP ATOMISATION TECHNIQUE IN [Analyst, Vol. 98 EXPERIMENTAL The atomic-absorption instrument used was the Varian AA-5 model. 220-V auto-transformer 7 carbon cup I heating I n I I step-down L - - - - - - _ _ - _ A transformer b b To carbon cup Fig. 1. Circuit diagram of the power unit The details of the carbon cup attachment are shown in Fig. 2. The graphite cup (Ringsdorf, RW OOZ), of 6.5 mm 0.d.and 9 rnrn in height with 1 6 m m wall thickness, is held by two support rods clamped between massive water-cooled stainless-steel blocks (1) , which also serve as electrical terminals. The distance between the blocks can be varied by means of a screw (6) , which enables the graphite cups to be exchanged. Spring-loaded support rods ensure good and constant electrical contact. The carbon cup and part of the support rods are continuously purged with argon so as to minimise the oxidation of the carbon. The pro- tective action of argon was observed to be significantly improved by mounting a curved quartz roof (4) above the cup. Normally, about 200 measurements can be made with a single cup but the number of measurements depends on the operating temperature in the atomisation step.By using a quartz roof more than 600 measurements were possible. The carbon cup device was found to be very convenient for solution analysis, although the carbon tube or rod has normally been used for this purpose. Larger volumes (up to 30 p1) can be introduced into the cup if higher sensitivity is required. This sampling pro- cedure is more reproducible as the injected solution always collects at the tapered bottom of the cup. A Hamilton 10-pl syringe and an Oxford 10-p1 micropipette were used for sampling solutions.n 2 3 'p Fig. 2. Carbon cup atomiser: 1, stainless-steel blocks; 2, water-cooling system; 3, electrical connections; 4, quartz roof; 5, argon inlet tube; and 6, adjusting screw r, 03 n IuDecember, 19731 WATER ANALYSIS BY ATOMIC-ABSORPTION SPECTROSCOPY SELECTION OF OPERATING CONDITIONS- Instrument parameters used in the measurements are summarised in Table I.TABLE I OPTIMAL INSTRUMENT SETTINGS 84 3 Element Wavelength/nm Spectral band pass/nm Lamp currentlmil Copper . . .. .. .. 324.8 0-33 3 Cadmium .. . . .. .. 228.8 0.66 3 Lead .. .. .. .. 217.0 0.66 6 Hydrogen background corrector. . - - 3 to 15 Several factors such as the volume and the volatility of the sample were considered in selecting the appropriate conditions for atomisation of water samples. The following procedure for atomisation of a 10-pl sample, introduced in one portion, was found to be the most suitable: drying at approximately 90 "C for 30 s, ashing at 300 "C for 15 s and atomisation at 2500 "C for 2-2 s.With copper, the duration of the atomisation step was extended to 3 s. It is preferable to introduce the sample solution into a slightly warm cup so as to minimise the risk of soaking of the solution into the porous carbon walls. On the other hand, too high a temperature during the drying period causes rapid disintegration of the sample solution inside the cup, which may result in sample losses and a non-reproducible absorption signal. ABSORPTION MEASUREMENTS ON AQUEOUS LEAD, COPPER AND CADMIUM SOLUTIONS- Appropriate solutions were prepared from lead, copper and cadmium stock solutions (1 mg ml-l, prepared by using Specpure grade metal) by dilution with doubly distilled water. Water was taken as the blank solution. A linear relationship between the concentration and absorbance was observed for all three elements in the range up to 50 per cent.absorption. Because the sensitivity can be increased by taking larger sample volumes, an examination of the relationship between sample volume and absorbance was made. Four different lead solutions were measured for this purpose; two (5 and 10 ng ml-l of lead) in the range 10 to 30 p1 and the other two (25 and 50 ng ml-l of lead) in the range 2 to 10 pl of sample volume. The solution was injected into the cup either in one portion or in successive small portions (10 or 2 p1, respectively). These experiments are illustrated in Figs. 3 and 4. o'20 I Sam p I e vo 1 u me/,uI Fig. 3. Variation of absorbance with sample volume a t constant lead concentra- tion: A, A' 25 and B,B' 50 ng ml-1 of lead; A,B, solution injected in one portion; and A,'B,' solution injected in several 2-p.1 portions in succession 0 10 20 30 Sample volume/pl Fig.4. Variation of absorbance with sample volume a t constant lead concentra- tion: C,C' 5 and D,D' 10 ng ml-l of lead; C,D, solution injected in one portion; and C',D', solution injected in several 10-pl portions in succession844 DOLINSEK AND STUPAR: CARBON CUP ATOMISATION TECHNIQUE I N [Autabst, vol. 98 It can be seen that a linear relationship between sample volume and absorbance is generally established, although the injection of sample in successive small portions (2 or 10 pl) always results in a higher sensitivity (curves A', B', C', D' in Figs. 3 and 4). A satisfactory explanation of the latter phenomenon can be found by considering sample losses to result from penetration of the solution into the cup walls during the drying step.It can be assumed that these losses are approximately proportional to the retention time of the solution in the cup, and to the surface area of the cup in contact with the solution. The results for the evaporation times of particular sample volumes are summarised in Table 11. TABLE I1 DURATION OF THE DRYING STEP AS A FUNCTION OF THE SAMPLE VOLUME Sample volume/pl Duration of the drying step*/s 2 to 4 15 6 19 8 23 10 30 20 80 to 90 30 110 to 120 2 x 10 60 3 x 10 90 * Temperature below 100 "C. Increasing the temperature of the drying step to slightly above 100 "C would certainly reduce the evaporation time significantly, but the figures given in Table I1 will still retain their relative values.A sample volume of 3Opl was considered for practical purposes to be the upper limit for use with the carbon cup atomiser, although the technique of successive injection of 10-pl portions can be successfully used. EFFECT OF EDTA AND VARIOUS ORGANIC REAGENTS ON THE SENSITIVITY OF THE DETER- MINATION- Measurements of copper, lead and cadmium in various streams and relatively unpolluted river waters by using the standard addition method have shown that poor sensitivity is obtained in comparison with doubly distilled water, particularly for cadmium and copper. Acidification of the water prior to analysis has been frequently adopted,12 which slightly improves the sensitivity but also increases the risk of contamination. The effect of different organic reagents (EDTA, ammonium tetramethylenedithiocarbamate and sodium diethyl- dithiocarbamate) that form metal chelates was investigated for this purpose.A substantial enhancement of absorbance was obtained for lead and copper in natural water samples, while only lead showed a considerable improvement in distilled water. During measurements of the effect of these organic reagents on the peak height of the absorption signals of the elements concerned, appreciable day-to-day variations were found, which may be ascribed to variation in instrument parameters. Table 111 summarises some typical results of these experiments ; TABLE I11 RELATIVE ABSORBANCE OF LEAD, CADMIUM AND COPPER IN DISTILLED AND NATURAL 10-pl sample volume WATER, AS INFLUENCED BY THE ADDITION OF DIFFERENT ORGANIC REAGENTS Distilled water r--------h--? 0.2 ml of 1 per cent.Element EDTA (concen- No solution tration/ng ml-l) addition per 50 ml Cd (2.5) 100 115 Pb (50) 100 358 Cu (150) 100 110 Mountain stream water r A 0.2 ml of 1 per cent. 0.2 ml of 2 per cent. No solution solution addition per 50 ml per 50 ml 64 112 115 78 113 120 95 332 343 EDTA APDC - 0-2 ml of 1 per cent. NaDDTC solution per 50 ml 110 109 338 APDC = Ammonium tetramethylenedithiocarbamate. DDTC = Sodium diethyldithiocarbamate.Decern ber, I 97313 WATER ANALYSIS BY ATOMIC-ABSORPTIOS SPECTKOSCOP\- 846 the relative values given are true atomic absorbances. Correction for non-specific absorption was made a t the resonance line wavelength by using a hydrogen hollow cathode.A 1 per cent. solution of EDTA (free acid) was selected as an additive to water samples and standard solutions in order to normalise the sensitivity. The concentration of EDTA in the water sample that was necessary to cause the enhancement of sensitivity was deter- mined from curve A in Fig. 5; 0.2 ml of the EDTA solution in 60 ml of sample was chosen, as this value lies well on the plateau of the curve and should not cause appreciable contamina- tion. In order to determine any measurable contamination that may have been caused by the addition of EDTA solution, different volumes of the latter were added to the stream-water sample and the absorbance of lead was measured (Fig. 5, curve 13). The constant absorbance observed over the range 0.3 to 4 ml of EDTA solution indicated that no risk of lead con- tamination should be expected when 0-2 ml are added to 50 ml of the sample.Similar tests were made with copper and cadmium and no detectable amounts of these elements were found in the EDTA solution. 0 0.5 1.0 " 4.0 EDTA solution/ml Fig. 5. Variation of lead absorbance with volume of added 1 per cent. aqueous EDTA solution: A, stream water, 100 ng ml-1 of lead; and 13, stream water The enhancing action of organic reagents, which proved to be particularly strong in the determination of lead and copper in both natural and distilled water, can be explained by the formation of clielates. These compounds, although stable in solution, are easily decomposed at elevated temperature during the atomisation step, which results in a transient, high atomic population inside the cup.SENSITIVITY, PRECISION AND DETECTION LIMITS- Sensitivity, defined as the concentration that gives 1 per cent. absorption, was derived from the calibration graphs obtained for lead, copper and cadmium standard solutions (distilled water) in the presence of EDTA. Detection limits were calculated on the basis of the following relationships- Ac AA Detection limit = kub - ( k = 3) (n = 11, 15) where CTb is the standard deviation of the blank solution (distilled water PLUS EDTA) and &/AA is the reciprocal value of the slope taken from the calibration graphs. A scale expansion846 DOLIXSEK AND STUPAR: CARBON CUP ATOMISATION TECHNIQUE I N [AfialySt, VOl. 98 (factor of 2) was used for determination of the detection limits.The results obtained are presented in Table IV. TABLE IV SENSITIVITIES AND DETECTION LIMITS OBTAINED WITH CARBON CUP GRAPHITE ATOMISER AT lo-$ SAMPLE VOLUME Sensitivitylng ml-l per Detection Element 1 per cent. absorption limit*/ng ml-1 Lead . . .. .. 1.2 Copper . . .. .. 4-4 Cadmium . . .. .. 0.1 * Scale expansion 2 x . 0.45 1.7 0.04 The precision of the determination with the carbon cup atomiser was evaluated by taking eleven to fifteen successive measurements of 50 ng ml-l of lead, 2.5 ng ml-l of cadmium and 300 ng ml-l of copper solutions (distilled water plus EDTA). Coefficients of variation of 2.3, 1-9 and 3.9 per cent., respectively, were calculated, which are reasonable for the atomisa- tion technique used.The recorder trace shown in Fig. 6 demonstrates the precision typical of the carbon cup atomiser; a 10-pl sample volume injected in one portion was used for these measurements. W L 30- c 0 - .- w ,n 20- a - 2 10- - - - - - - - - - L - --. -. 1 . 0 - Fig. 6. Reproducibility of measurement typical of the carbon cup atomiser (sample volume 10 pl, 50 ng ml-l of lead, 2-5 ng ml-l of cadmium and 300 ng ml-l of copper) INTERFERENCES- Although the time available for atomisation in the carbon cup is significantly longer in comparison with the flame technique, several chemical interferences could still be observed. In addition, molecular absorption and light scattering often present a serious problem. In our experiments the influence of different salts commonly present in natural waters was studied.The effect of the addition of EDTA was also investigated. The results are sum- marised in Tables V and VI. Values given in these tables are relative absorbances produced by particular kinds of atoms, while those in parentheses are relative absorbances measured with the continuum light source. All values for a particular element are expressed relative to the absorbance of a pure solution containing no EDTA. Non-specific absorption was frequently observed when different salts were present in the solution; in some instances it was easily resolved from the atomic peak. Addition of EDTA had no, or very little, influence upon the absolute value of the non-specific absorption, although considerable changes were observed in the interference effects of various salts.The interferences of the various salts investigated in the determination of copper, cad- mium and lead, which manifested themselves in an enhancement or decrease of the peak absorption value, can generally be attributed to the change in the rate of atomisation. ADecember, 19731 WATER ANALYSIS BY ATOMIC-ABSORPTION SPECTROSCOPY TABLE V EFFECT OF VARIOUS SALTS (CATIONS) ON THE RELATIVE ABSORPTION SIGNAL 847 O F LEAD, CADMIUM AND COPPER Lead (200 ng ml-l) Cadmium (5 ng ml-l) Copper (300 ng ml-I) & +--7 +- Concen- 1 per cent. 1 per cent. 1 per cent. tration of EDTA EDTA EDTA 0.2 ml of 0.2 ml of 0.2 ml of Salt - NaCl KCl CaCl, MgCb MgSO, FeCI, -41*(S04), cation/ No solution NO solution N O pgml-1 EDTA in 50 ml EDTA in50ml EDTA - 100 360 100 116 100 200 87 3 60 150 152 95 100 73 260 102 126 103 200 266 168 170 138 106 200 81 72 66 61 103 200 31 216 116 161 79 60 105 62 95 62 120 50 95 370 179 148 103 * Non-specific adsorption easily resolved from the atomic peak.(12) (12) (250)* (250) * (0) (0) (0) (17) (17) (0) (106) (106) (131)* (131)* (0) (37) (37) (34) (34) (0) (40) * (40) * (34) (34) (0) ( 7) (71 (17) (17) (0) (13) (13) (17) (17) (0) detailed study of these salt systems should perhaps be undertaken by using a fast-response instrument so as to obtain more knowledge of the mechanism of interference. However, any resultant interference effect that may be expected in measurements of actual water samples cannot be predicted, as the concentration and chemical character of particular matrix elements differ widely.TABLE VI EFFECT OF VARIOUS SALTS (ANIONS) ON THE RELATIVE ABSORPTION SIGNAL OF LEAD, CADMIUM AND COPPER Lead (200 ng ml-l) Cadmium (5 ng ml-I) r------h--? - r---A--7 Copper (300 ng ml-l) 0-2 ml of 0.2 ml of 0.2 ml of Salt - NaCl Na,CO, NaNO, Na,SO, Na,HP04 Concen- 1 per cent. sodium/ No solution No pgml-I EDTA in50mI EDTA - 100 170 100 50 100 170 99 50 55 157 39 50 100 157 101 60 26 77 89 50 98 170 83 tration of EDTA (0) (0) U3)* (0) (0) (2) (0) (0) (2) (0) (0) (2) (0) (0) (2) * Non-specific absorption easily resolved 1 per cent. EDTA solution NO in 50 ml EDTA 110 100 114 104 77 64 114 105 115 101 108 83 from the atomic peak. (13)* (0) (2) (0) (2) (0) (2) (0) (2) (0) 1 per cent. EDTA solution in 50 ml 100 110 (0) 85 (0) 108 (0) 110 (0) 95 (0) DETERMINATION OF LEAD, COPPER AND CADMIUM IN NATURAL WATERS AND INDUSTRIAL Preliminary studies with the carbon cup atomiser showed that direct determination of these elements in a variety of water samples should be possible at the nanogram per millilitre WASTES-848 [Analyst, Vol.98 level. Several river and stream waters, effluents from different industries and tap water samples were analysed in order to demonstrate the feasibility of the carbon cup flameless atomic-absorption method. SAMPLING PROCEDURE- In the determination of nanogram amounts of an element, special care should be taken in sample collection if significant errors from losses or contamination are to be prevented. The effect of adsorption of lead on polyethylene was therefore investigated. For this purpose, two water samples (1000 ml) containing different amounts of lead were collected in 1-litre polyethylene bottles and stored at 4 "C for 3 weeks.The lead concentration was determined periodically. The absorbance of the sample (A,) was divided by the average absorbance of two standard lead solutions (Ast) containing 60 ng ml-l of lead, which were prepared freshly each time from the lead stock solution. The ratio ASIAs, was plotted against time and is shown in Fig. 7. Curves A and B have a slight negative slope, which indicates that the lead concentration in the solution decreased slowly with time. Although the absolute amount of lead lost in 3 weeks is approximately the same in both solutions, the relative error produced by this loss is significant only at low lead concentrations (see Table VII).DOLINSEK AND STUPAR: CARBON CUP ATOMISATION TECHNIQUE IN 0.1'- A I I I I I I I I I ~ 2.1 r 1 0.2 1 TABLE VII RELATIVE ERROR CAUSED BY LOSS OF LEAD FROM SOLUTION AFTER 3 WEEKS' STORAGE IN POLYETHYLENE BOTTLES Lead concentration in the solution/ng ml-l -7 Relative Sample collection collection per cent. 1 hour after 3 weeks after loss of lead, Tap water . . .. .. .. . . .. .. 6.6 5.7 13.5 Doubly distilled water left overnight in a polyvinyl tube 120.5 120.1 0-3December, 19731 WATER ANALYSIS BY ATOMIC-ABSORPTION SPECTROSCOPY a49 0 n 6Q Concentration of added lead/ng ml-l Calibration graphs of lead in different river and industrial waste waters : A, River MeZa (contamination from lead smelter); B, River Sava; C, Stream Voglajna (highly polluted by different industrial effluents) ; and D, effluent from zinc plant (five times diluted) Fig.8. a zinc plant. It can be seen that the latter two curves have a significantly lower slope (curve C 1/2 and curve D 1/3 of the slope of curves A and B) due to interference effects caused by the different sample compositions. Therefore, it is obviously necessary to prepare a calibration graph for each particular type of water if accurate results are desired. TABLE VIII LEAD, COPPER AND CADMIUM CONTENTS OF SOME REPRESENTATIVE WATER SAMPLES ANALYSED BY CARBON CUP ATOMIC-ABSORPTION AND COMPARATIVE METHODS Sample 21 23 11 27 31 34 38 24 Cadmium, ng ml-l, by & absorption absorption absorption spectroscopy spectroscopy spectroscopy * * - cup Flame graphy* cup Flame graphy* cup Flame graphy* activation Lead, ng ml-l, by Copper, ng ml-l, by L.I > 7Gz-- Atomic- Atomic- Carbon Polaro- Carbon Polaro- Carbon Polaro- Neutron 4.3 2.9 4.5 20.5 20 14 24 40 (6) (2) 0.30 0-5 - 5100 4600 - 10 6 0.17 0-5 - 2.2 <D.L. - 6 5 (16) (16) (9) 0.05 0.5 <0*3 8 6 5.5 21 25 (50) (5) (9) (2) 20.3 20 - 1040 1150 - 3.5 3 (2) (1.3) (2) (1.8) 0.40 0.5 0.5 9s 113 96 5 6 (1.9) (1.8) 0-35 0.5 0-3 48 48 45 4 3 4-5 160 240 5.1 5.0 3-4 6.0 (2.6) (9) (1.5) (0.9) (3.5) - (9) (17) * Anodic stripping polarography. D.L. denotes detection limit. Values given in parentheses are percentage standard deviation. 24 - 220 -850 DOLINSEK AND STUPAK In order to determine the accuracy of the method some representative samples were analysed by comparative analytical methods : flame atomic-absorption spectroscopy, polaro- graphy and neutron activation, with the first of which the use of an extraction procedure was necessary.The results obtained by these methods are summarised in Table VIII. Generally, satisfactory agreement was obtained between the results by the various analytical methods in those instances when comparison was possible. Unfortunately, the sensitivity of the comparative methods was not sufficient to enable them to be used for deter- mining cadmium in most of the water samples. The relatively high blank readings involved in flame atomic-absorption spectroscopy and polarography have a significant influence on the accuracy and detection limits of these techniques. CONCLUSION Carbon cup, flameless atomic-absorption spectroscopy has been demonstrated to be ;t rapid and sensitive technique that permits the direct analysis of natural waters and industrial wastes.Only a single set of standard solutions is necessary for each type of water sample. The precision of the method is as good as that given by other analytical methods with which it was compared, while its accuracy is satisfactory. The low cost of the equipment used and the extreme speed of analysis make the proposed method superior to these other methods for monitoring lead, copper and cadmium in various water samples. The authors thank the “Boris KidriE” Foundation for providing financial support, and are also indebted to the late Dr. I. Sink0 and to Dipl. Ing. V. Ravnik, who performecf the polarographic and neutron-activation analyses, respectively. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. REFEREKCES Siriko, I., and Dolefal, J., J . Electroanalyt. Chem., 1970, 25, 299. Perkin-Elmer Instyurn. News Sci. Ind., 1970, 21, 4. Das, H. A., and Van der Sloot, H. A., in “Proceedings of the Symposium on Nuclear Activation Robinson, J. L., Barnekow, G., jun., and Lott, P. F., Atom. Absorption Newsl,, 1969, 8, 60. Nix, J., and Goodwin, T., Ibid., 1970, 9, 119. Jones, J. L., and Eddy, R. D., Analytica Chim. Ada, 1968, 43, 165. Midgett, ill. Ii., and Fishman, M. J., Atom. Absorption Newsl., 1967, 6, 128. Mansell, 13. E., and Emmel, H. W., Ibid.. 1965, 4, 365. Sprague, S., and Slavin, W., Ibid., 1964, 3, 37. Kuwata, K., Hisatomi, K., and Hasegawa, T., Ibid., 1971, 10, 111. Mansell, K. E., Ibid., 1965, 4, 276. “Standard Methods for the Examination of Water,” A.P.H.A., A.W.W.A. and W.C.P.F., Geneva, Dozanska, W., Roczn. Pa&. Zakl. Hig., 1963, 14, 433. Saltzman, B. E., Analyt. Chem., 1953, 25, 493. Fernandez, F. J., and Manning, D. C., Atom. Absorfltion Newsl., 1971, 10, 65. Pickford, C. J., and Rossi, G., Analyst, 1972, 97, 647. Techniques in the Life Science,” Ljubljana, Yugoslavia, 10-14 April, 1972. Switzerland, Second Edition, 1963. Received Januavy 2&h, 1973 Accepted July rjth, 1973