JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 393 Determination of Cadmium and Zinc Isotope Ratios in Sheep's Blood and Organ Tissue by Electrothermal Vaporization Inductively Coupled Plasma Mass Spectrometry* Invited Lecture D. Conrad Gregoire Geological Survey of Canada 607 Booth Street Ottawa Canada KIA OE8 Julian Lee New Zealand Pastoral Agricultural Research Institute Ltd. Fitzherbert West Private Bag 7 7008 Palmerston North New Zealand A method is described for the determination of Cd and Zn isotope ratios in sheep's blood and organ tissue. Samples were digested with nitric acid using a microwave oven. Cadmium and Zn were separated from matrix components using adsorption chromatography prior to isotope ratio measurement by electrothermal vaporization inductively coupled plasma mass spectrometry.A concentration factor of 35 was achieved. Limits of detection for the determination of Cd and Zn in blood were 0.34 and 0.40 pg g-' respectively. Cadmium isotope ratios ("'Cd '06Cd; "'Cd '"Cd) were determined with a precision of 2-3% for both peak height and area count measurements. Zinc isotope ratios (68Zn 67Zn; 68Zn 66Zn) were determined with a precision of 2% for peak height measurements and 1% for peak area count measurements. Keywords Inductively coupled plasma mass spectrometry; electrothermal vaporization; isotope ratios; blood liver and kidneys A characteristic feature of Cd metabolism in animals is its poor homeostatic control and accumulation in the kidneys and liver. This may pose a potential problem since these organs cannot be exported when their Cd concentration exceeds 1 mg kg-' of fresh tissue.The Cd concentration in New Zealand soils and pastures (which are naturally low) has gradually increased as a result of the regular and extended use of imported phosphatic fertilizers containing varying trace amounts of naturally occurring Cd. Cadmium is a non-essential metal which because of its toxic properties has been widely studied in small animals and humans. However the few studies reported for ruminants are based on the intake of Cd at concentrations much higher than those encountered by grazing animals. The main impediment to the completion of more realistic studies involving whole body infusion of enriched stable isotopes has been the lack of adequate analytical techniques capable of measuring Cd iso- tope ratios in small samples at concentration levels below 0.2ng g-'.Pool sizes of Cd in most tissues especially blood are low and therefore it is desirable that additions of enriched isotopes in this study lo6Cd be kept at low levels to minimize changes to physiological concentrations. This imposes further constraints on analysis. In recent years inductively coupled plasma mass spec- trometry (ICP-MS) has been applied to human nutrition and metabolic studies. Serfass et a1.4 reported that Zn isotope ratios in human faecal material could be determined with a precision of better than 1 %. Janghorbani and co-worker~~-'~ have reported extensively on the determination of isotope ratios of several elements in biological materials by ICP-MS including blood and faeces.Isotope ratios of Fe,5-8 Cu,6v8 Zn,6*8 Br9 and Lila were determined with a precision ranging from 1 to 2%. Delves and Campbell" have reported a precision of better than 0.5% for the determination of major isotope ratios of Pb in human whole blood. More recently Viczian et a1.12 deter- mined Pb isotope ratios in blood and environmental materials * Presented in part at the XXVIII Colloquium Spectroscopicum Internationale (CSI) Post-Symposium 5th Surrey Conference on Plasma Source Mass Spectrometry Durham UK July 4-6 1993. GSC publication No. 17793. to identify potential environmental sources of childhood lead poisoning. Smith et determined B isotope ratios in a variety of biological samples with a precision ranging from 0.4 to 1.5% depending on B concentrations.In ruminants Cu and Se metabolism has been recently studied using stable isotope tracer methodology and ICP-MS.14*15 Solution nebulization (SN) as a means of sample introduc- tion in ICP-MS requires from 5 to 15 ml of sample solution for the accurate and precise determination of isotope ratios. Although the use of a direct injection neb~lizer'~ significantly reduces the volume of sample solution required (0.5-1 ml) these volumes can still be too great for applications involving scarce samples such as blood and/or analytes such as Cd which occur at ppb to sub-ppb concentration levels in blood and tissue. Electrothermal vaporization (ETV) requires only microlitre volumes of sample solution and allows for the determination of analytes at parts per trillion (ppt) concentration I e ~ e l s .~ ~ ~ ' ' This high sensitivity is essential to the successful determination of Cd isotope ratios an element whose concentration is in the ppt range in blood. Microlitre sample volumes enable precon- centration techniques to be used even on millilitre volumes of blood or gram amounts of tissue. This paper reports on the application of ETV-ICP-MS to the determination of Cd and Zn isotope ratios in blood and tissue samples. Samples were obtained from several sheep which were part of a lo6Cd and 67Zn whole-body infusion study." Sheep were continuously infused for several days with enriched stable isotopes at the rate of 1.5 pg h-' for lo6Cd and 30pg h-l for 67Zn. At regular intervals blood samples were withdrawn by syringe for analysis. Although Zn occurs in blood at ppm concentration levels and its isotope ratios have been successfully determined using SN-ICP-MS this element was included in this study in order to compare the performance of ETV- and SN-ICP-MS for the determination of Zn iso- tope ratios.Experimental Sample Collection and Preparation Sheep aged 6-8 months were housed indoors in metabolism crates and fed cut ryegrass from overhead feeders at hourly394 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 intervals. After a period of adaption to the diets stable isotopes [ 80.8 Yo Io6Cd and 9 1.1 YO 67Zn (Isotec Miamisburg OH USA)] in sterile physiological saline were infused into one side of a bilateral catheter. At regular intervals blood samples were withdrawn by syringe from the other side of the catheter into chilled polypropylene tubes containing 0.1 ml of 1667 nkat ml-' of heparin (checked for Cd and Zn impurities).Whole blood samples chilled to 4 "C were centrifuged for 15 min at 2000g to separate plasma from red cells. Tissue samples were obtained following the slaughter of the animals and freeze-dried prior to analysis. All samples were stored at - 10 "C prior to final processing. Plasma or red cells (7 g) were wet digested with a minimum volume (10 ml) of concentrated HNO (Aristar). Digestions were performed in a Milestone MLS 1200 microwave oven system. The digest was evaporated to near dryness on a hot- plate and the sample was then made up to 10 ml with 1% HNO,. Liver and kidney samples were prepared in an ana- logous manner using 0.4 g of freeze-dried sample.Prior to the determination of Cd and Zn isotope ratios by ETV-ICP-MS both analytes were separated from matrix com- ponents using off-line adsorption chromatography. The acidity of the prepared blood or organ tissue digest was adjusted to a pH of 6.0 with ammonia solution (Aristar) before being passed through a 1 ml column containing silica-immobilized 8-hydro~yquinoline.'~ The resin was eluted with 10 ml of 1 moll-' HNO,-O.l moll-' HCl and evaporated to dryness in the microwave oven. The residue was dissolved in 0.2 ml of 1% HNO and divided into two 0.1 ml portions. The first portion was used for the determination of Cd isotope ratios and the second was diluted to 5 ml (or more) for the determi- nation of Zn isotope ratios.For the determination of Cd isotope ratios a preconcentration factor of 35 was achieved increasing the concentration of Cd in solution from approxi- mately 0.2 ngg-' in plasma to 7 ngg-' in the analytical sample. The much higher concentration of Zn in plasma (0.7-1.0 pg g-') and in whole blood (4-5 pg g-') required dilution of the concentrates before measurement of the isotope ratios was possible. Both Cd and Zn levels in organ tissues are at much higher concentrations compared with blood levels and these samples also required dilution prior to analysis by either ETV- (Cd Zn) or SN- (Zn) ICP-MS. Reagent blanks were treated as samples and taken through the entire digestion and preconcentration steps. Recoveries of Cd and Zn for the overall procedure including digestion and adsorption chroma- tography (each column separately assessed) steps were deter- mined using spikes (1 ng ml-' of Cd and 2 pg ml-' of Zn) prepared from stock standard solutions (Spectrosol Merck Poole Dorset UK).Recoveries were 111 & 18% for Cd and 95 f 10% for Zn (n = 30) for inter-batch analyses. The ICP-MS signal intensities for Cd and Zn in reagent blanks were low accounting for about 1% of the sample signal intensity. Blank intensities for analyte isotopes were subtracted from sample intensities prior to calculation of isotope ratios. Instrumentation A Perkin-Elmer SCIEX Elan 5000 ICP mass spectrometer equipped with an HGA-600MS electrothermal vaporizer and Model AS-60 autosampler was used for multi-isotope mass spectral analysis.The experimental conditions for both the Elan 5000 and the HGA-600MS are given in Table 1. Optimization of plasma and mass spectrometer conditions was accomplished using solution nebulization sample introduc- tion. The HGA-6OOMS was interfaced to the argon plasma via a 0.8m length of 6mm (i.d.) Teflon tubing. Operation of the HGA-600MS was completely computer controlled. During the drying and charring stages of the temperature programme opposing flows of argon gas (300 ml min-') originating from both ends of the graphite tube removed water and other vapours through the dosing hole. During the high temperature Table 1 Instrumental operating conditions and data acquisition parameters ICP mass spectrometer- R.f. power/W 1000 Outer argon flow rate/l min-' Intermediate argon flow rate/ml min - ' Carrier argon flow rate/ml min-' Sampler/skimmer Nickel 15.0 850 900 HGA-600MS electrothermal vaporizer- Sample volume/pl 20 Drying stage (10 s ramp) 90°C for 40 s 300 Charring stage (3 s ramp) 300°C for 10 s 300 Vaporization temperaturerc 2300 Internal argon flow rate/ml min-' Internal argon flow rate/ml min-' Heating rate/"C s-' 2000 Time at maximum temperature/s 5 Data acquisition- Dwell time (ETV)/ms Dwell time (SN)/ms Measurements per isotope (SN) Scan mode Points per spectral peak Isotopes monitored per measurement cycle Signal measurement mode Cd isotope ratios Zn isotope ratios 20 400 100 1 2 Peak-hopping Intensity maximum Integrated signal pulse or vaporization step a graphite probe was pneumatically activated to seal the dosing hole.Once sealed a valve located at one end of the HGA workhead directed the carrier argon flow originating from the opposite end of the tube directly to the argon plasma at a flow rate of 900ml min-'. Pyrolytic graphite coated graphite tubes were used throughout. Procedure for Measurement of Isotope Ratios Although the isotope ratios of both Cd and Zn could be measured during the same vaporization cycle the isotope ratio for each element was determined separately. This was done to reduce the cycle time between the intensity measurements of the two isotopes for each analyte. The fast transient nature of the ETV-ICP-MS signal requires a relatively short duty cycle to ensure good precision of the isotope ratio measurement.Fig. 1 illustrates the transient nature of the ETV-ICP-MS signal obtained for the vaporization of 200 pg of Cd and 50 pg of Zn. Using a dwell time of 20 ms (Table 1) approximately 40 intensity readings were obtained for each isotope during the high temperature vaporization step. Isotope ratios could 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 Time/s Fig. 1 ETV-ICP-MS signal pulses for 200 pg of Cd and 50 pg of ZnJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 395 be calculated using either peak height (maximum intensity) or peak area (integrated) counts. Except where otherwise noted isotope ratios are reported as the mean of five separate measurements. For ETV-ICP-MS determinations the addition 10 pl of NASS-3 (diluted 1 in 500) as chemical modifier was added to both standard and sample solutions.The reference material NASS-3 Open Ocean Seawater is available from the National Research Council of Canada (Ottawa). The addition of this solution provides 0.7 ng of salt containing Na Cl Ca Mg and Sr which act as a physical carrier,l7 ensuring efficient transport of vaporized analyte from the graphite tube to the argon plasma. Prior to dilution and use the NASS-3 sea-water was purified of any traces of Cd and Zn by adsorption chromatography .Ig For SN-ICP-MS isotope ion intensities for three Zn isotopes were monitored during the same measurement cycle. A measurement cycle consisted of 100 separate sequential measurements (dwell time = 400 ms) for each analyte isotope. Isotope ratios were calculated from the mean of the intensities obtained from five such measurement cycles.Instrumental mass discrimination for both ETV and SN work was generally less than the precision of the isotope ratio measurement itself. Correction for mass discrimination was accomplished using as a reference the mean of ten separate ETV-ICP-MS isotope ratio measurements for a pure aqueous standard of Cd or Zn (High-Purity Standards Charleston NC USA). Comparison of this value with the accepted value for the natural abundance ratio2' was used to calculate the appropriate correction factor. Corrections for mass discrimi- nation for SN-ICP-MS measurements were carried out in an analogous manner using intensities measured during ten separ- ate measurement cycles. Results and Discussion Selection of Isotopes and Analytical Figures of Merit Cadmium has eight stable isotopes ranging from m/z 106 to 116.Of these isotopes only "'Cd (natural abundance 12.86%) is free of isobaric interferences from neighbouring isotopes of other elements and was therefore selected as the reference isotope for Cd isotope ratio determinations. The Io6Cd isotope (1.22%) is isobaric with lo6Pd (27.10%) lo8Cd (0.89%) is isobaric with Io8Pd (26.7%) and "'Cd (12.43%) is isobaric with 'IoPd (13.5%). The '12Cd isotope (23.79%) is isobaric with '12Sn (0.95%) '13Cd (12.34%) is isobaric with '131n (4.16%) '14Cd (28.81 %) is isobaric with '14Sn (0.65%) and '16Cd (7.66%) is isobacric with 'I6Sn (14.24%). For the spike isotope either Io6Cd or Io8Cd was suitable as both have low natural abundances. The Io6Cd isotope was selected as the spike isotope purely on the basis of cost of the enriched stable isotope.Zinc has five stable isotopes ranging from m/z 64 to 70. Of these isotopes 66Zn (27.81%) 67Zn (4.11%) and 68Zn (18.56%) are free of isobaric interferences from isotopes of other elements. The major isotopes of zinc 64Zn (48.9%) is isobaric with 64 Ni (1.16%) and 70Zn (0.63%) is isobaric with 70Ge (20.52%). Interference from the molecular ion 40Ar14N14N+ at m/z 6821 was not observed. Because of its low relative abundance and freedom from is0 baric interferences 67Zn was selected for use as the spike isotope. Although 70Zn has a much lower natural abundance than 67Zn and would therefore be better than 67Zn as a spike isotope the use of this isotope is prohibitively expensive considering the pool size for Zn of an animal with a mass of 50 kg.Table 2 summarizes the analytical figures of merit for the determination of Cd and Zn by ETV-ICP-MS. The blank matrix used to measure the background for each element was 10 pl of diluted purified NASS-3 solution. The relatively high blank obtained for Zn was attributed to Zn contamination Table 2 Analytical figures of merit for the determination of Cd and Zn by ETV-ICP-MS Parameter "'Cd "Zn NASS-3 background (n = 10) 185 3162 Standard deviation 32 191 RSD (Yo) 17.3 6.0 Integrated counts per 100 pg/s 42 0oO 214 000 RSD (Yo n=10) 2.9 1.9 LOD ( 3 4 absolute/pg 0.23 0.27 LOD ( 3 4 relative (20 pl)/pg ml-' 12 14 originating partly from the NASS-3 solution and partly from the graphite tube.This background diminished only slightly during the course of the experiments. The relative standard deviation (RSD) for the measurement of peak area counts for both elements was of the order of 2-3%. The absolute limit of detection for Cd and Zn was 0.23 and 0.27 pg respectively. For a 20 pl sample volume relative limits of detection of 12 and 14pgml-1 were obtained for Cd and Zn respectively. Taking into account a preconcentration factor of 35 the limit of detection in blood and tissue was 0.34pg g-' for Cd and 0.40pg g-' for Zn. For Cd this concentration is about 600 times below natural levels of Cd in sheep's plasma. Measurement of Isotope Ratios Tables 3 and 4 summarize the analytical results for the determination of Cd and Zn isotope ratios in a standard reference solution.Isotope ratios were determined on 200 pg of Cd and 50pg of Zn as nitrate in the presence of NASS-3. Table 3 Precision of Cd isotope ratio measurement (corrected for mass discrimination) by ETV-ICP-MS Run 1 2 3 4 5 6 7 8 9 10 Mean SD RSD (Yo) Integrated Peak height 10.33 10.56 10.75 10.45 10.90 10.27 10.07 10.58 10.94 10.55 10.37 10.06 10.97 10.64 10.69 10.94 9.84 10.38 11.03 10.53 10.54 10.54 0.27 0.39 2.61 3.71 Integrated Peak height 1.013 1.038 1.034 1.033 1.019 1.01 1 1.053 1.057 1.060 1.030 1.007 1.027 1.049 1.012 1.007 1.044 1.05 1 1.065 1.05 1 1.033 1.035 1.035 0.017 0.02 1 1.67 2.03 Table 4 Precision of Zn isotope ratio measurement (corrected for mass discrimination) by ETV-ICP-M S Run 68Zn "Zn 68Zn "Zn 1 2 3 4 5 6 7 8 9 10 Mean SD RSD (Yo) Integrated Peak height 4.527 4.527 4.450 4.541 4.638 4.453 4.483 4.530 4.448 4.565 4.372 4.388 4.462 4.490 4.735 4.425 4.614 4.580 4.509 4.581 4.516 4.516 0.060 0.113 1.33 2.50 Integrated Peak height 0.664 0.665 0.666 0.665 0.663 0.668 0.667 0.669 0.677 0.672 0.658 0.642 0.667 0.667 0.658 0.675 0.678 0.671 0.677 0.673 0.667 0.667 0.006 0.01 1 0.87 1.72396 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL.9 Isotope ratios were corrected for mass discrimination and calculated using both peak height intensities and peak area counts. Isotope pairs were selected to include both the reference and spike isotope used in the actual study. A third isotope for each element was also selected such that a comparison could be made of the precision of measurement for isotope ratios ranging from unity to some higher value.For Cd and Zn isotope ratios peak height measurements gave poorer precision than ratios calculated using peak area counts. This is probably due to the uncertainty introduced into the peak height measurement from the sequential measure- ment of isotope intensities. A reduction in the dwell time could provide more readings per signal pulse but the shorter measurement time would degrade the counting statistics. Both elements exhibited the same trend with respect to the measure- ment precision of ratios ranging from 1 to 10. The isotope ratio for ' W d "'Cd (1.035) was measured with a precision that was about 1% better than the precision of the "'Cd lo6Cd (10.54) isotope ratio. For zinc the measurement precision for the 68Zn:66Zn (0.667) ratio was about 0.5% better than the precision of the 68Zn 67Zn (4.516) isotope ratio measurement.Comparison of Tables 3 and 4 shows that in general the Zn isotope ratio can be determined with a better precision than the Cd isotope ratio. Other studies completed in this laboratory17 have shown that the vaporization and transport of Cd is complicated by some factor related to the operation of the electrothermal vaporizer or the vaporization/transport process. Further studies are underway to investigate this problem in detail and to determine if this is the source of signal instability. Tables 5 and 6 summarize results for the determination of 111Cd:106Cd and 68Zn:67Zn isotope ratios in a number of representative plasma red cell and organ tissue samples not necessarily obtained from the same sheep.The determination of the ll'Cd lo6Cd isotope ratio was complicated by a Pd contamination resulting in an isobaric interference from lo6Pd (27.10%). The source of the Pd has not been determined but Table 5 Cd isotope ratios in sheep's blood and organ tissue Sample Plasma (2 h) Plasma (4 h) Plasma (8 h) Plasma (17 h) Plasma (24 h) Plasma (50 hj Plasma (56 hj Plasma ( 120 h) Red cells Kidney Liver 111Cd. 106cd 1.531 f0.028 1.263 & 0.045 0.987 f 0.014 0.758 k 0.045 0.69 1 k 0.003 0.624 f 0.01 8 0.621 0.030 0.61 5 k 0.01 7 1.639 f 0.033 0.313 k0.003 0.070 & 0.001 Table 6 Comparison of Zn isotope ratios in sheep's blood and organ tissue determined by ETV- and SN-ICP-MS Sample 68Zn %n Plasma (2 h) Plasma (4 h) Plasma (8 h) Plasma (17 h) Plasma (24 h) Plasma (42 h) Plasma (66 h) Plasma (72 h) Red cells Kidney Liver ETV 4.000 f 0.076 3.984 _+ 0.056 3.937 & 0.059 3.817 f 0.069 3.650 f 0.073 3.289 f 0.022 3.096 -t 0.062 2.985 -+_ 0.045 4.5 19 f 0.036 3.51 1 f0.046 3.379 f 0.022 SN 4.132 f 0.066 4.049 f 0.065 3.937 k0.020 3.876 0.041 3.636 f 0.028 3.145 f0.033 3.075 fO.010 2.857 k 0.03 1 4.510f0.018 3.533 f 0.026 3.349 k0.014 probably originated from acids used in the sample prep- aration step.Fig. 2 illustrates the vaporization characteristics of Pd rela- tive to that of two Cd isotopes. Clearly Pd is less volatile relative to Cd and is vaporized at a time corresponding to the decaying portion of the Cd analyte signal pulse. For this reason it was decided to calculate Cd isotope ratios for the blood and tissue samples using peak height measurements when only a small correction for lo6Pd isobaric interference was necessary. The correction to the Io6Cd ion count rate corresponded to the intensity of the lo5Pd ( x 1.199 to convert to lo6Pd) ion at the time corresponding to the maximum (or peak time) of the Cd signal pulse.Table 5 shows that the ' W d lo6Cd isotope ratio in blood and organ tissues varied from 0.070 to 1.531. For plasma samples the Cd isotope ratio decreased by 60% during the course of the infusion study. The use of these data to calculate enrichment factors and other indices to assess adsorption tissue entry rates and excretion of Cd and Zn are in progress. A measurement precision of 2% (peak height) for the "'Cd lo6Cd isotope ratio was more than adequate to accu- rately monitor the change in this ratio with time.Table 6 shows that the 68Zn:67Zn isotope ratio in blood and organ tissue varied from 2.985 to 4.519. For plasma samples the Zn isotope ratio decreased by 34% during the course of the infusion study. Although the change in the Zn isotope ratio was much smaller than for Cd a measurement precision of 0.87% (peak area counts) was sufficient to success- fully carry out metabolic studies. Following the measurement of 68Zn 67Zn by ETV-ICP-MS an aliquot of the remaining sample was diluted with distilled deionized water for isotope measurements using SN-ICP-MS. The results given in Table 6 show good agreement between ETV and SN isotope ratio measurements.The precision of the SN-ICP-MS isotope ratio measurement was about 0.6% and was limited by the amount of sample solution available ( 5 ml). Conclusion This work has shown that ETV-ICP-MS can successfully measure isotope ratios of Cd and Zn in blood and tissue samples at naturally occurring concentrations for these elements. The precision obtainable for isotope ratio measure- ments was of the order of 30 to 150 times smaller than the variation in isotope ratios observed during the course of these studies. Allowing 2 min per isotope ratio determination and five replicates per sample approximately 50 samples could be analysed by ETV-ICP-MS during an 8 hour day. The graphite tube could be used for approximately 250 firings before requir- ing replacement.0 0.5 1 .o 1.5 2.0 Timels Fig.2 Comparison of vaporization characteristics for Pd and Cd isotopes in spiked blood sampleJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 397 The loan of the HGA-600MS from the Perkin-Elmer Corporation is gratefully acknowledged. References Smith R. M. Griel L. C. Muller L. D. Leach R. M. and Baker D. E. J. Anim. Sci. 1991 69 4078. Van Der Veen H. G. and Vreman K. Netherlands J. Agric. Sci. 1986 34 145. Wentink G. H. Wensing T. Baars A. J. Van Beek H. and Zeeywen A Bull. Environ. Contam. Toxicol. 1987 39 131. Serfass R. E. Thompson J. J. and Houk R. S. Anal. Chim. Acta 1986 188 73. Janghorbani M. Ting B. T. G. and Fomon S . J. Am. J. Hematol. 1986 21 277. Ting B. T. G. and Janghorbani M. Anal. Chem. 1986,58 1334. Ting B. T. G. and Janghorbani M. Spectrochim. Acta. Part B 1987 42 21. Ting B. T. G. and Janghorbani M. J. Anal. At. Spectrom. 1988 3 325. Janghorbani M. Davis T. and Ting B. T. G. Analyst 1988 113 405. 10 11 12 13 14 15 16 17 18 19 20 21 Sun X. F. Ting B. T. G. Zeisel S . H. and Janghorbani M. Analyst 1987 112 1223. Delves H. T. and Campbell M. J. J. Anal. At. Spectrom. 1988 3 343. Viczian M. Lasztity A. and Barnes R. J. Anal. At. Spectrom. 1990 5 293. Smith F. G. Wiederin D. R. Houk R. S. Egan C. and Serfass R. E. Anal. Chim. Acta 1991 248 229. Koening K. M. Buckley W. T. and Shelford J. A. Can. J. Anim. Sci. 1991 71 175. Buckley W. T. Can. J. Anim. Sci. 1991 71 155. Gregoire D. C. Lamoureux M. Chakrabarti C . L. and Byrne J. J. Anal. At. Spectrom. 1992 7 579. Sturgeon R. E. Willie S. N. Zheng J. Kudo A. and GrCgoire D. C. J. Anal. At. Spectrom. 1993 8 1053. Lee J. unpublished data. Sturgeon R. E. Berman S . S. Willie S. N. and Desaulniers J. A. H. Anal. Chem. 1981 53 2337. De Bievre P. Gallet M. Holden N. E. and Barnes I. L. J. Phys. Chem. Ref. Data 1984 12 809. Tan S. H. and Horlick G. Appl. Spectrosc. 1986 40 445. Paper 310404 7 J Received July 12 1993 Accepted September 15 1993