ANALYST, MARCH 1986, VOL. 111 285 Studies on the Determination of Cadmium in Blood by Furnace Atomic Non-thermal Excitation Spectrometry Heinz Falk, Erwin Hoffmann and Christian Ludke Central Institute for Optics and Spectroscopy, Academy of Sciences of GDR, Rudower Chaussee 5, 7799 Berlin, GDR and John M. Ottaway and David Littlejohn Department of Pure and Applied Chemistry, University of Strathclyde, Cathedral Street, Glasgow GI IXL, UK Cadmium atomic emission can be detected in a FANES low-pressure Ar discharge at atomiser temperatures as low as 140 "C when the analyte is present as CdCI2. Cadmium chloride molecules vaporised at this temperature are dissociated by electron impact in the discharge, giving a substantial Cd atom concentration before thermal dissociation of the molecules becomes feasible.This results in a 100-fold greater tolerance towards chloride matrix chemical interferences than encountered for Cd in ETA-AAS with tube-wall atomisation. However, the determination of cadmium in deproteinised whole blood by FANES is not totally interference free and a standard additions procedure is required to give an accurate determination. The FANES instrument detection limit for Cd was calculated to be 0.04 pg I-'. For the deproteinisation procedure applied, the detection limit for cadmium in whole blood was 0.2 pg 1-1. Keywords: Atomic emission spectrometry; low-pressure discharge; electrothermal atomisation; electron- impact molecular dissociation; blood matrix interferences The determination of cadmium in biological samples by atomic absorption spectrometry with electrothermal atomis- ation (ETA-AAS) is well established in clinical analysis.However, as with most volatile elements, the measurement of Cd AAS signals is subject to severe chemical and spectral matrix interference effects when conventional tube-wall atomisation procedures are applied. As cadmium has a comparatively low atom appearance temperature, it is not normally possible to ash biological samples at temperatures that allow the complete removal of organic and inorganic matrix constitutents prior to the atomisation stage. Hence, cadmium atom formation tends to be suppressed by the presence of a large excess of chloride salts, and substantial non-specific background absorption invariably occurs. To minimise the influence of these effects in the determination of Cd in whole blood, a variety of regimes has been implemented including the use of matrix matching, matrix modification, platform atomisation and Zeeman-effect background correc- tion.Stoeppler and Brandtl reduced the mass of carbonaceous matter injected into the atomiser tube by deproteinisation of 50-200-yl volumes of whole blood with 1 M HN03. Pleban and Pearson2 simply diluted whole blood with 5% V/V HN03 for the determination of Cd by Zeeman effect - ETA-AAS using a standard additions procedure. Matrix modification with (NH4)2HP04 was applied by Subramanian and Meranger3 to determine Cd directly in blood. A similar procedure was reported by Delves and Woodward,4 but oxygen was added to the atomiser gas flow during the ashing stage to assist removal of the blood matrix.Matrix modification was also used by Hinderberger et al. 5 in conjunction with platform atomisation, and Claeys-Thoread employed platform atomisation and Zeeman-effect background correction for the determination of Cd in blood diluted 1 + 9 with Triton X-100. It is generally accepted that the mechanism of cadmium atomisation depends on the chemical nature of the sample matrix. When present as the chloride, CdC12, cadmium atoms are formed by thermal dissociation of gaseous CdCl molecules produced on vaporisation of CdCI2. In the presence of an excess of chloride salts, however, atom formation is impaired, as reported by Barnard and Fishman7 for solutions containing 1% mlV NaCl, KCI, MgC12 or CaC12. Similar observations for tube-wall atomisation have been reported by other workers .a9 When cadmium is present as an oxy-anion two mechanisms of atom formation are feasible.In an early study, Campbell and Ottawaylo suggested that CdO was reduced on the graphite surface to produce Cd atoms. Salmon and HoIcombell also supported the oxide reduction mechanism and have postu- lated that metallic Cd is formed on the graphite tube. In contrast, L'vov and Ryabchuk12 believe that CdO is dis- sociated in the vapour state. This mechanism was also suggested by Sturgeon and Chakrabarti13 and has been supported by more recent work by Sturgeon and co- workers. 1 4 ~ 5 In this work, the atomisation of cadmium from chloride and nitrate matrices has been studied as part of an investigation aimed at the development of a method for the determination of Cd in whole blood by furnace atomic non-thermal excitation spectrometry (FANES).The characteristics of furnace atomisation with non-thermal excitation have been described in recent publications. 16-19 The technique involves conventional electrothermal atomisation of samples in a tube atomiser in which a low-pressure gas discharge is simul- taneously generated using the graphite tube as the cathode. The system therefore combines the efficiency of electrother- mal vaporisation and atomisation with the high excitation capability of a hollow-cathode type discharge. Although thermal excitation in a conventional electrothermal atomiser allows the measurement of atomic emission signals for many elements with high sensitivity,20>21 energy levels greater than 4 eV are not significantly populated.In contrast, the FANES source is ideally suited to the excitation of metals and non-metals with high excitation potentials (e.g., cadmium, zinc, selenium and halogens). Hence FANES combines many of the attractive features of ETA-AAS and plasma emission spectrometry, not least of which are sub-pg 1-1 detection limits and the ease of operation in a simultaneous multi-element mode. In previous FANES studies, the instrumental features of the system have been described and the analytical characteristics of the source established with respect to analyte detection limits.1619 However, to date, the atomiser has been applied to relatively few analytical problems and a primary aim of this work was to assess the suitability of the FANES method for286 ANALYST, MARCH 1986, VOL.111 the determination of a relatively volatile element such as cadmium in a complex matrix such as whole blood. As the FANES atomiser is operated at low pressure and the sample is vaporised into a low-pressure discharge, it was expected that the analytical behaviour of the source would be different to that of a conventional electrothermal atomiser as applied in atomic absorption spectrometry for this analysis. The influ- ence of alkali and alkaline earth metal salts on the vaporis- ation and atomisation of cadmium in the FANES atomiser has been studied and compared with observations reported in the ETA-AAS literature. The results of the study indicated that the presence of a high-energy discharge assists the dissociation of vapour-phase molecules and for this reason the influence of NaCl, KCl, MgC12 and CaC12 on Cd atom production is less pronounced than in a conventional electrothermal atomiser.Experimental Instrumentation The FANES source, which has been decribed in detail in previous publications,1”19 consisted of a sealed atomiser chamber and power supply unit with separate functions that controlled the heating of the graphite tube and the establish- ment of the discharge. The atomiser was water cooled, and connected to a mechanical pump for evacuation and an argon gas supply system for purging of the tube and formation of the low-pressure Ar discharge. The FANES atomiser tube was similar in dimensions to that used in the Perkin-Elmer HGA-500 graphite furnace.With the exception of the discharge pressure (1-20 Torr) and current (15-60 mA), which were set manually, all parameters were under microprocessor control. A series of up to ten temperature stages could be selected via the instru- ment’s microcomputer. At each step in the temperature programme the operator selected a temperature (up to 3000 “C), a linear ramp rate (0-2000 “C s-1) and a hold time. On the basis of the selected temperature and ramp rate, the microprocessor computed the ramp time. The selection of atomiser gas flow-rate at atmospheric pressure, the initiation of the evacuation stage and the establishment of the discharge were also under computer control and the required conditions were selected when compiling the various steps in the atomiser programme.In this study, the FANES source was operated in conjunc- tion with a laboratory-contructed 1.5-m Rowland circle polychromator equipped with a 2000 grooves mm-1 grating. All cadmium atomic emission measurements were made at the Cd 228.8-nm resonance line wavelength with a spectral band pass of 0.017 nm. Emission signals were recorded on a K 201 Jenoptik chart recorder. For atomic absorption measure- ments, a Cd hollow-cathode lamp was positioned such that the lamp radiation was focused through the FANES atomiser and on to the polychromator entrance slit. FANES Operational Procedure The operation of the FANES source is analogous to that used in conventional atomisation. The main difference in the furnace programme occurs prior to and during the atomisation stage.A sample aliquot (typically 10 or 20 ~ 1 ) was injected into the graphite tube by means of a micropipette and dried and ashed at atmospheric pressure with the injection port lid open and a purge gas flow-rate of 80 1 h-1 to remove matrix vapours from the atomiser. The injection port lid was then closed, the atomiser evacuated to a pressure of about lo-* Torr, the argon discharge gas pressure established (1-20 Torr) and the discharge formed. This sequence was initiated by selection of the appropriate parameters in the programme step prior to the intended atomisation stage. The conditions of pump-down could be arranged to allow a continuation of the ashing step during the evacuation procedure. As in a conventional atomisation regime, the temperature and ramp rate of the atomisation stage were normally selected to give rapid vaporisation and atomisation of the analyte.In the discharge the analyte atoms were excited and the resulting emission signal detected with the polychromator. When atomic absorp- tion measurements were required with the FANES atomiser, the programme was modified to either prevent formation of a discharge or eliminate the evacuation - discharge sequence entirely. The optimum atomiser programme devised for the determi- nation of cadmium in blood is given in Table 1. Modified versions of the programme employed in the chloride salt interference studies are indicated in the appropriate section of the text. Preparation of Whole Blood Samples To minimise the mass of carbonaceous material atomised in the FANES source a deproteinisation procedure similar to that developed by Stoeppler and Brandtl was applied.A 200-y1 volume of whole blood was mixed with 200 yl of water or aqueous calibration solution, 50 yl of concentrated nitric acid were added and the mixture was centrifuged for 7 min to produce a clear supernatant solution over the protein precipi- tate. Volumes of the deproteinised solution (10 or 20 yl) were then injected into the FANES atomiser for the determination of cadmium. Table 1. FANES Step Dry? . . . Ash . . . . . Purge$ . . . Evacuate . . . Pre-atomisation11 Atomise . . . Clean . . . programme for the determination of cadmium in deproteinised blood Temperature/ “C . . . 150 . . . 350 . . . 35 . . . 35 .. 100 . . . 600 . . . 2000 Ramp rate/ “Cs-* 3 100 NPO 0 30 600 1000 Ramp time*/ 45.0 2.0 6.7 0 2.1 0.8 1.9 S Hold time/ 5 28 5 45 90 9 3 S Total step time/ S 50 30 11.7 45 92.1 9.8 4.9 Argon pressure/ Functions Torr 760 760 760 E457 19 19 19 760 Gas flow- rate/ 1 h-I 80 80 80 0 0 0 80 * Calculated from temperature difference and ramp rate selected; for the first step an ambient temperature of 15 “C was assumed. t Conditions for 10-pl injection volumes. $ Additional step used in this application to remove matrix vapour without tube heating. 0 NP implies “no power.” fl When the “E” function is selected evacuation begins at the start of the step, with discharge on after 45 s in this instance; 30 mA current 11 Step used to stabilise discharge before atomisation; could have been redued to 20-30 s.selected; discharge on until end of atornise step.ANALYST, MARCH 1986, VOL. 111 287 Reagents Stock solutions of CdC12, NaCl, KCl, MgC12 and CaC12 were prepared by dissolving salts of the highest available purity in distilled water. High-purity nitric acid was used for the deproteinisation of blood samples and for the addition to CdC12 standard solutions. Results and Discussion Vaporisation Characteristics of Cadmium Salts in the FANES Atomiser It was considered probable that the vaporisation charcteristics of cadmium salts in FANES would be different to those observed in conventional ETA-AAS as atomisation occurs at low pressure and in the environment of a low-pressure discharge. To investigate the influence of both conditions on cadmium atom production in the FANES atomiser, atomic absorption and atomic emission measurements were obtained following vaporisation of cadmium from chloride and nitrate media.In initial experiments, cadmium atomic emission and atomic absorption signals were obtained for the vaporisation of 10-y1 volumes of a 100 pg 1-1 cadmium solution prepared from CdCI2. Emission or absorption signals were obtained for FANES atomisation temperatures in the range 100-1000 "C, at a linear ramp rate of 600 "C s-1. The emission measure- ments were obtained in the presence of an Ar discharge at 17 Torr, and absorption measurements were made without a discharge at 17 Torr or at atmospheric pressure. At an argon pressure of 17 Torr the maximum cadmium atomic emission and atomic absorption signals were obtained at a similar temperature of about 520 "C as measured with a thermo- couple.The first appearance of a cadmium atomic emission signal occurred at a tube temperature of 140 "C. In contrast, however, without the presence of a discharge, atomic absorp- tion signals at 17 Torr were not measured until temperatures above 300 "C. Plots of the relative integrated cadmium atomic emission and atomic absorption signals obtained at 17 Torr for atomisation temperatures in the range 100-1000 "C are illustrated in Fig. 1, together with the corresponding plot for cadmium atomic absorption at atmospheric pressure. The fact that the cadmium emission signal appeared at a considerably lower temperature than the atomic absorption signal at the same pressure suggests that at 17 Torr vaporisation of CdC12 begins at temperatures as low as 140 "C and that dissociation of gaseous CdCl or CdC12 molecules by electron impact in the discharge results in the formation of cadmium atoms at atomiser temperatures much lower than would normally be expected.Without the assistance of the discharge, dissociation of vaporised CdCl - CdC12 molecules i9 a purely thermal process and at 17 Torr did not occur until the atomiser tube had reached a temperature of approximately 300 "C. At atmospheric pressure significant cadmium atom formation did not occur until about 360-380 "C and the maximum AAS signal was obtained at about 680 "C. When an excess of chloride salts is present in solution, the vaporisation characteristics of CdC12 are altered noticeably.Fig. 2 shows the vaporisation curves for cadmium atomic emission obtained at 10 and 17 Torr in the presence of 0.1% mlV NaCl and 0.05% mlV MgC12 when the atomiser tube was heated to temperatures in the range 100-1000 "C at a ramp rate of 600 "C s-1. At 17 Tori the first appearance of cadmium atomic emission occurs at about 320 "C, probably because the evaporation of CdC12 is impaired owing to the occlusion of the cadmium salt in the NaCl - MgC12 matrix. When the discharge pressure is reduced to 10 Torr, evaporation of the chloride salt matrix occurs at a lower temperature and the first appearance of cadmium atomic emission occurs at 200 "C. Fig. 2 also indicates the vaporisation curve for cadmium in the presence of 0.01% V/V HN03. Although the cadmium solution was prepared from CdC12, the presence of an excess of oxy-anion probably ensured the formation of CdO, which has a higher vaporisation temperature than CdC12.A comparison of the vaporisation curves for CdCI2 and "CdO" in Figs. 1 and 2, respectively, (17 Torr) suggests that the dissociation of the gaseous CdO molecules by electron impact in the discharge is less efficient than dissociation of CdCl or CdCI2 by the same process. Influence of Chloride Salts on Cadmium Atomic Emission Intensity in FANES From ETA-AAS studies based on tube-wall atomisation,9 it is known that a matrix of 0.01% mlV NaCl and 0.005% mlV MgC12 causes a 10% depression of the Cd AAS signal. As electron impact in the FANES discharge apparently assists the dissociation of gaseous CdCl - CdCI2 molecules, it was expected that chloride interference effects would be less severe in FANES.Solutions containing 2, 5 , 10 and 20 pg 1-1 of Cd as CdC12 were prepared in aqueous, 0.01% VlV HN03 and 0.1% mlV NaCl - 0.05% mlV MgC12 solutions. For each solution, 10-yl volumes were injected into the FANES atomiser, dried at 150 "C with a ramp rate of 5 "C s-1 and then atomised into the discharge under atomisation conditions of 750 "C and 600 "C s-1. Cadmium atomic emission signals obtained for an Ar pressure of 19 Torr and a discharge current of 30 mA are illustrated in the form of calibration graphs in Fig. 3. No significant difference was observed in the Cd atomic emission intensities obtained for each solution matrix. The slight curvature in the calibration graph observed at 20 pg 1-1 of Cd was probably due to self-absorption of the Cd line profile.As there was no apparent interference of 0.1% mlV NaCl and 0.05% mlV MgC12 on the Cd atomic emission signal, the interferent concentrations were increased to determine the ? 00 200 300 400 500 600 7 Atomiser temperature/"C Fig. 1. Relative integrated atomic emission and atomic absorption signals for 10 p1 of 100 pg 1-1 of Cd at different atomisation temperatures. (A) FANES, 17 Torr Ar, 30 mA; (B) AAS, 17 Torr Ar; (C) AAS, 1 atm Ar. Plotted relative to individual maximum signals obtained at 520 "C for A and B and 680 "C for C 200 300 400 500 600 700 Y - 100 Atomiser temperaturei'c Fig. 2. Relative integrated FANES intensities for 10 p1 of 100 pg I-' Cd in different matrix solutions and at different atomisation tempera- tures.(A) 10 Torr Ar, 30 mA, 0.1% mlV NaCl - 0.05% m/V M CI2; (B) 17 Torr Ar, 30 mA, 0.1% mlV NaCl - 0.05% mlV MgCI,;((!) 17 Torr Ar, 30 mA, 0.01% VlV HN03. Plotted relative to individual maximum signals at 580 "C for A and 650 "C for B and C288 ANALYST, MARCH 1986, VOL. 111 interference-free limit for FANES Cd determinations. Solu- tions containing 10 pg 1-1 of Cd and various concentrations of NaCl, KCI, MgC12 and CaCI2 were prepared, up to levels in excess of the expected concentrations of chloride salts in undiluted whole blood. As shown in Fig. 4, a significant reduction in the Cd atomic emission intensity occurred only at a combined chloride concentration of 1.0% mlV NaCl, 1.0% 80 I 1 Concentration/pg I - Fig.3. FANES calibration raphs for Cd in (0) aqueous solution; ( X ) 0.01% V/VHN03; and (A) 0.1% mlVNaCl- 0.05% mlVMgC1,; conditions as indicated in the text E 50 I .$ 1 C w . + - 0.125 0.25 0.5 1.0 NaCl and KCI 4- plus Z 0.063 0.125 0.25 0.5 MgCI, a n d CaCI2 Matrix concentration, Ol0 m/V Fig. 4. Effect of various combined concentrations of NaCl, KCI, MgCI2 and CaCI, on the FANES atomic emission intensity for 10 yl of 10 pg 1-1 of Cd 50 I I A 40 I / \ C 3 2 30 E e 4- .- 10 0 10 20 30 40 50 60 Ash timeis Fig. 5.. Influence of ashing time on the FANES atomic emission intensity for 10 p1 of 20 pg 1-l Cd at ashing temperatures of (A) 300, (B) 350 and ( C ) 400 "C mlV KCl, 0.5% mlV MgC12 and 0.5% mlV CaC12, about two orders of magnitude higher than the onset of chloride interference in conventional ETA-AAS.9 Determination of Cadmium in Whole Blood Although the deproteinisation procedure described previ- ously was used to minimise the mass of carbonaceous material injected into the atomiser, sufficient of the blood matrix remained to merit the inclusion of an ashing step in the analytical programme.Ashing temperatures of 300, 350 and 400 "C were investigated. The Cd atomic emission signals obtained for 10-pl injection volumes of a 20 pg 1-1 solution after ashing at the above temperatures for 2-60 s are given in Fig. 5. Optimum ashing conditions of 350 "C for 30 s were selected as indicated in Table 1. The cadmium atomic emission vaporisation curve for a deproteinised blood solution containing 10% VlV HN03 was obtained and compared with the corresponding curve for an aqueous Cd solution containing the same HN03 concentra- tion.The cadmium concentration in both solutions was 20 pg 1-1 and separate 10-pl injection volumes were dried, ashed and then atomised at a ramp rate of 600 "C s-1 to various atomisation temperatures in the range 100-700 "C at 100 "C intervals. Although, as indicated in Fig. 6, there appears to be no significant difference in the vaporisation curves obtained for the two solutions, closer inspection in comparison with the vaporisation curves at 17 Torr in Fig. 2 suggests that the chloride salts present in the deproteinised blood solution may still influence the vaporisation of cadmium even though HN03 is present in a large excess. From the results presented in Figs.2-6, it was concluded that the chloride and organic constituents of the deproteinised blood sample were unlikely to exert a severe interference on the determination of cadmium by FANES. However, as indicated in Fig. 7, there was a substantial difference in the slopes of the deproteinised blood standard additions graph and the aqueous cadmium calibration graph when both sets of solutions contained an equal concentration of HN03. For Cd concentrations in the range 2.5-20 pg 1-1, the atomic emission intensity for the deproteinised blood sample was only 20% of the equivalent signal for the HN03 solution. The deproteini- sation process may not have released all the Cd attached to blood protein, which would result in a lower than expected concentration of the metal in the supernatant liquid injected into the FANES atomiser.However, previous experience with the protein precipitation procedure suggests that at a nitric acid concentration of 10% V/V incomplete release of Cd is unlikely. An alternative explanation is that residual organic components and inorganic species, other than chloride salts, I ' I / 1 I 1 4 200 300 400 500 600 700 Atom iser temperatu re/"C Fig. 6. Relative integrated FANES intensities for 10 pl of 20 pg 1- of Cd in (A) deproteinised whole blood and (B) 10% V/V HN03 at different atomisation temperatures. Conditions as in text and Table 1 ; results plotted relative to individual maximum signals at 750 "CANALYST, MARCH 1986, VOL. 111 289 vaporised during the atomisation stage altered the discharge conditions, which may have had an adverse effect on the cadmium atom excitation.As it was not possible to analyse deproteinised blood samples by direct comparison with cadmium standard solu- tions prepared in nitric acid, a standard additions procedure was used to determine the Cd concentration of two whole blood samples supplied by the Biochemistry Department at Glasgow Royal Infirmary. The values obtained are given in Table 2, together with the concentrations determined by ETA-AAS at the hospital. Good agreement was achieved between the two techniques. The Cd detection limit of the FANES standard additions procedure was calculated to be 0.2 yg 1-1 on a 20 basis. For aqueous solutions containing 10% V/V HN03 the Cd detection limit was a factor of 5 lower, 0.04 yg 1-1.The relative standard deviation of the standard additions procedure was about 10% for the two blood samples analysed. This is a precision of at least a factor of two poorer than would normally be expected for conventional ETA-AAS with manual pipetting. The precision was impaired by the spreading of the nitric acid solution droplets in the tube during the drying sequence, and was also influenced by the blank correction required to take account of the cadmium content of the nitric acid available during this study. Conclusions The measurements reported in this work confirmed the widely held opinion that when present as CdCI2, cadmium atom formation proceeds through the dissociation of gaseous CdCl 20 0 5 10 15 Concentrationivg I - Fig. 7. (B) deproteinised whole blood.Conditions as in text and Table 1 FANES calibration graphs for Cd in (A) 10% V/VHN03 and Table 2. Determination of cadmium in deproteinised whole blood by FANES and ETA-AAS Concentrationipg 1- * Sample FANES ETA-AAS" 1 19.5 k 2.0 19.6 2 8.5 t 0.9 8.8 * Samples and values provided by Glasgow Royal Infirmary, Department of Clinical Biochemistry. molecules. The dissociation is assisted in the FANES dis- charge by electron impact and Cd atoms are formed at temperatures as low as 140 "C at 17 Torr. Without the action of the low-pressure discharge, thermal dissociation of CdCl molecules does not occur until around 300 "C at this pressure. Although an excess concentration of alkali and alkaline earth metal chloride salts retards the vaporisation of CdCI2, no significant chemical interferences were encountered until the combined concentrations of NaC1, KCI, etc., had reached 2-3% mlV. In an oxy-anion medium ( e .g . , HN03), it is likely that CdO is formed at some stage in the atomisation process. The vaporisation studies conducted with FANES for Cd in HN03 do not prove conclusively that Cd atoms are formed by the dissociation of gaseous CdO molecules. However, at low pressure the volatility of CdO will undoubtedly increase and it is possible that electron impact in the discharge will assist the thermal dissociation of CdO. Additional measurements of both cadmium atomic absorption and atomic emission signals are required in this instance to give a clear indication of the atomisation mechanism in the FANES atomiser.The chemical interferences encountered in the analysis of deproteinised blood may be due to the action of organic and inorganic constituents (other than chloride salts) on the nature of the discharge. Further fundamental studies of the excitation conditions during atomisation are required in order to assess the influence of the blood matrix in the determination of Cd. Experiments of this nature are in progress. However, the FANES atomiser clearly has potential advantages over conventional ETA-AAS with regard to its tolerance to chloride salt interferences. This work was made possible by the Cultural Exchange Agreement between the Royal Society in the UK and the Academy of Sciences of the GDR. The authors are extremely grateful for the opportunity for collaborative study and for the financial support provided through the Exchange Scheme.Financial support from the Pye Foundation (for D. L.) is also gratefully acknowledged. The authors thank Dr. G. S. Fell and Dr. D. J. Halls, Department of Clinical Biochemistry, Glasgow Royal Infirmary, for the provision of blood samples. 1. 2. 3. 4. 5 . 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. References Stoeppler, M., and Brandt, K., Fresenius 2. Anal. Chem., 1980, 300, 372. Pleban, P. A., and Pearson, K. H., Clin. Chim. Acta, 1979,99, 267. Subramanian, K. S . , and Meranger, J. C., Clin. Chem., 1981, 27, 1866. Delves, H. T., and Woodward, J., At. Spectrosc., 1981, 2, 65. Hinderberger, E. J., Kaiser, M. L., and Koirtyohann, S. R., At. Spectrosc., 1981, 2, 8. Claeys-Thoreau, F., At. Spectrosc., 1982, 3, 188. Barnard, W. M., and Fishman, M. J . , At. Absorpt. Newsl., 1973, 12, 118. Cruz, R. B., and Van Loon, J. C., Anal. Chim. Acta, 1974,72, 231. Campbell, W. C., and Ottaway, J. M., Analyst, 1977,102,495. Campbell, W. C., and Ottaway, J. M., Talanta, 1974,21,837. Salmon, S. G., and Holcombe, J. A., Anal. Chem., 1982, 54, 630. L'vov, B. V., and Ryabchuk, G. N., Spectrochim. Acta, Part B , 1982, 37, 673. Sturgeon, R. E., and Chakrabarti, C. L., Prog. Anal. At. Spectrosc., 1978, 1, 132. Sturgeon, R. E., Siu, K. W. M., and Berman, S . S., Spectrochirn. Acta, Part B, 1984, 39, 213. Sturgeon, R. E., and Berman, S. S . , Anal. Chem., 1985, 57, 1268.290 ANALYST, MARCH 1986, VOL. 111 16. 17. 18. 19. Falk, H . , Hoffmann, E., and Ludke, Ch., Spectrochim. Acta, Part B , 1981, 36,767. Falk, H., Hoffmann, E., and Ludke, Ch., Fresenius 2. Anal. Chem., 1981, 307, 362. Falk, H . , Hoffmann, E., Ludke, Ch., Ottaway, J. M., and Giri, S. K., Analyst, 1983, 108, 1459. Eichardt, K., and Falk, H., Jenaer Rundsch., 1983, 28, 118. 20. 21. Bezur, L., Marshall, J., Ottaway, J. M ., and Fakhrul-Aldeen, R., Analyst, 1983, 108, 553. Giri, S. K., Littlejohn, D., and Ottaway, J. M., Analyst, 1982, 107, 1095. Paper A51271 Received July 23rd, 1985 Accepted October 9th, 1985