JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 199 Matrix Separation by Chelation to Prepare Biological Materials for Isotopic Zinc Analysis by Inductively Coupled Plasma Mass Spectrometry* Steven F. Durrant Antoaneta Krushevska Dula Amarasiriwardenat Mark D. Argentine Sabine Romon-Guesnier and Ramon M. Barnes University of Massachusetts Department of Chemistry Lederle Graduate Research Center Towers Amherst MA 0 1003-0035 USA Following an evaluation of three chelating resins [Chelex-1 00 poly(dithi0carbamate) (PDTC) and carboxyme- thylated poly(ethy1eneimine)-poly(methylenepolypheny1ene) isocyanate (CPPI)] a procedure was established with the last of these for the separation of Zn from biological matrix elements prior to 70Zn:68Zn isotopic analysis by inductively coupled plasma mass spectrometry (ICP-MS).The method was verified by establishing Zn recoveries and by determining its effectiveness in removing CI and Na from buffered test solutions. Calcium Na and Zn concentration data were determined by inductively coupled plasma atomic emission spectrometry. Chlorine was measured by electrothermal vaporization ICP-MS. The efficacy of the technique was demonstrated by the determination of zinc isotope ratios in bovine milk and human urine. Results compared favourably with those obtained using a previously established extraction procedure. Keywords Zinc isotope ratios; inductively coupled plasma mass spectrometry; matrix separation; bovine milk; human urine Inductively coupled plasma mass spectrometry (ICP-MS) is a widely accepted reliable method for elemental and isotopic determinations in diverse The multi-elemental capability of the technique is particularly useful for environ- mental samples and ICP-MS accounted for 1% of the total number of the analyses reported between 1986 and 1989.4 Both ICP-MS and isotopic'-'' analyses of biological materials provide unique information.Biological applications have recently been reviewed.''*12 The advantages of ICP-MS have been do~umented;'-~ how- ever a number of difficulties are encountered in the analysis of practical samples. Spectral interferences from oxides doubly- charged or polyatomic species of elements present in the sample and reagents13-" and non-spectroscopic interferences from solutions containing relatively high (about 0.2% m/m) dissolved solids typically observed as a suppression of analyte responses,'&'* hamper convenient analyses.A number of strategies have been developed to minimize spectral interferences. These include mathematical correction," addition of a molecular gas such as nitrogen to the argon plasma to modify the plasma chemistry,20 and separation of the analytes from interfering elements prior to analysis.21 The last method is also effective in reducing or eliminating non- spectroscopic interferences. Separation techniques for spectro- chemical analysis including detection by ICP-MS have been reviewed by Horvath et aLZ2 Resins have long been used to separate analytes from matrix elements prior to elemental analysis. For example Kingston et used a Chelex-100 resin to separate Cd Co Cu Fe Mn Ni Pb and Zn from alkali and alkaline earth elements in sea-water prior to their determination by electrothermal atomic absorption spectrometry. Resin separation procedures have been automated.For example Wang and Barnes24 used two chelating resins to preconcentrate Cu and Zn for the analysis of water samples by flow injection inductively coupled plasma atomic emission spectrometry (ICP-AES). A previous study described an acid digestion of biological samples with microwave heating." Urine red blood cells and plasma were prepared by extraction with ammonium * Presented at the XXVIII Colloquium Spectroscopicum t Present address Hampshire College School of Natural Science Internationale (CSI) York UK June 29-July 4 1993. Amherst MA 01002 USA.pyrrolidin-1-yldithioformate (APDC) and carbon tetrachloride (CCl,) for the measurement of 70Zn 68Zn isotope ratios by ICP-MS. In the present report an equally effective but quicker and less labour intensive method was sought. The motivation for studying 70Zn 68Zn ratios lies in their use in zinc bioavail- ability studies of children and pregnant ~ o m e n . ~ ' * ~ ~ Three chelating resins were tested for their ability to separate biological matrix components for subsequent Zn isotope ratio measurements using ICP-MS. Owing to their contribution to spectroscopic and non-spectroscopic interferences C1 and Na were the principal target elements; the former as 35Cl is coincident with 70Zn+ while the latter if present at a suffic- iently high concentration causes signal suppression.The three chelating resins investigated were Chelex-100 poly(dithi0carbamate) (PDTC) and carboxymethylated poly (ethyleneimine) - poly (methylenepolypheny1ene)isocyanate (CPPI). Preliminary studies involved the recovery of Zn from buffered solutions in the pH range 4.8-10.2. Based on these studies and further experiments to measure the effectiveness of the resins in removing Na C1 and Ca from test solutions the CPPI resin was selected for testing with real samples of bovine milk and human urine. In urine Cl and Na are present typically at > 3000 pg ml - and > 1500 pg ml- ' respectively;26 thus Zn in urine analysis is a challenging test sample. Owing to the lower Cl and Na concentrations expected in bovine milk (1000 and 500 pg ml-' re~pectively),~~ milk analysis is an example of a relatively straight-forward measurement.Total Zn recoveries determined by ICP-AES and a pre- established APDC-CC14 extraction are compared with those obtained using the new separation technique employing the CPPI resin. Zinc (70Zn 68Zn) isotope ratio measurements obtained after sample preparation by the two methods were also obtained. Experimental Instrumentation ICP-AES The Ca Na and Zn concentrations of the test solutions were determined by ICP-AES using a sequential spectrometer with two monochromators and internal reference channels (Perkin- Elmer Plasma 11 Norwalk CT USA). Operating conditions are given in Table 1. The nebulizer flow rate and viewing200 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL.9 Table 1 Instrumental operating parameters for ICP-AES Parameter Generator frequency/MHz Forward r.f. plasma power/kW Reflected power/W Argon gas flow rates/] min-' Outer Intermediate Nebulizer Observation height/mm Read delay/s Integration time/ms Replicates Analytical wavelengths/nm 27.12 1.0 <5 Value 15.0 1 .o 0.7 11 20 100 3 Ca I1 393.366 Co I1 228.616 Cu I 324.754 Mn IT 257.610 Na I 589.592 Sc I1 424.683 Y I1 371.030 Zn 1213.856 height were optimized for Zn I emission signal-to-background ratio. Other parameters are default conditions recommended by the manufacturer. Final solutions after extraction or resin separation were 2mol 1-' in nitric acid and their analysis was possible with acid-matched standards. However the acid concentration of the initially digested milk or urine varies with the nature and amount of the reagents used and their evaporation rates during digestion. Thus matching of the sample and standard acid concentrations is impractical. However differences in viscosity between unmatched solutions can lead to deterioration in analytical precision and accuracy.Therefore Myers-Tracy signal compensation was used for the determinations in the digested material with 50 pg ml-' of Sc added as a reference element. 28 ICP-MS Zinc isotope ratios were determined by ICP-MS (SCIEX Elan Model 250 Thornhill Ontario Canada). The mass spec- trometer settings and plasma conditions were optimized with a 200 ng ml-' solution of Zn prior to isotope ratio measure- ment. The instrument operating conditions and data collection parameters for the isotope ratio measurements are listed in Table 2.As only small sample volumes were available after the isotope ratio and elemental measurements chlorine was deter- Table 2 Instrumental operating parameters for zinc isotope ratio ("Zn 68Zn) determinations by ICP-MS Parameter Generator frequency/MHz Forward r.f. plasma power/kW Reflected power/W Distance/mm Load coil-sampler orifice Torch injector tip-sampler orifice Torch (Sciex 'long')/mm Spray chamber Nebulizer Argon plasma gas flow rate/ min-' Outer Intermediate Nebulizer Resolution Measuring mode Measurements per peak Measurement time/s Dwell time/ms Cycle time/s Replicates per integration Solution uptake rate/ml min-' 27.12 1.2 <5 Value 27 6 123 Scott double-pass Perkin Elmer cross-flow 11 1.4 1 u at 10% Peak hop 1 1 .ooo 20 0.85 6 0.6 1.05- 1.15 mined by electrothermal vaporization (ETV)-ICP-MS.For these determinations the central channel flow rate was optim- ized by maximizing the response at m/z 35 from a 150 pg ml-' C1 solution. The analysis conditions are given in Table 3. Reagents Sub-boiled nitric acid and isothermal ammonia solution were prepared. Ammonium acetate buffers (0.1 mol I-') at pH values in the range 4.8-10.2 were prepared from glacial acetic acid and ammonium acetate (Fisher Scientific Fairlawn NJ USA). The buffer was cleaned by passing it through a CPPI column. Distilled de-ionized water (DDW; NanoPure Sybron/Barnstead Boston MA USA) was used throughout to make-up solutions. For the ICP-AES determinations standard solutions of Zn Ca and Na were prepared by dissolving Zn metal (Johnson Matthey Ward Hill MA USA) CaCO and NaCl respect- ively in sub-boiled HNO and diluting to the appropriate volume with DDW.For the determination of C1 by ETV- ICP-MS standard solutions were prepared in the range 3-1 500 pg ml - ' from NaCl (Fisher Scientific Fairlawn NJ USA) dissolved in DDW. Cobalt at 0.2 pg ml-' was used as an internal reference. Columns Columns of Chelex-100 PDTC and CPPI were prepared. The Chelex- 100 was a commercial material (Biorad Laboratories Richmond CA USA). Preparation of the PDTC and CPPI resins has been described by Barnes29 and Horvath and Barnes,,' respectively. Resins were soaked in 2mol 1-1 nitric acid (made from sub-boiled nitric acid) for a few hours and then rinsed with DDW prior to use.Each resin (200-400mg) was packed in plastic funnels (capacity about 5 ml) fitted with glass wool retaining plugs. Typically natural flow rates were about 0.25 ml min-'. Table 3 Instrumental operating parameters for the determination of chlorine by ETV-ICP-MS Parameter Argon plasma gas flow rate/] min-' Outer Intermediate Central channel Forward r.f. plasma power/kW Reflected powerp Measurement parameters Resolution Dwell time/ms Replicates Replicates/reading Measuring mode m/z Sample volume/jd Electrothermal vaporization furnace Start/"C FinalPC 20 700 700 700 700 1100 1100 1100 1100 1200 1200 1200 1200 2400 2400 2400 2400 20 Ramp Sequence- Value 11+0.1 o2 1.4 1.25 1.2 t 5 High 20 300 (1 point/peak) 1 Peak hop 35 10 Perkin-Elmer HGA-400 Time/s 10 30 5 10 5 10 2 6 5JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL.9 201 Digestion Procedures Urine A bulk urine sample (24 h collection) was digested in a microwave oven (CEM MDS-8 1 CEM Corporation Matthews NC USA) according to a procedure given pre- viously.l0 The resulting final volume after disgestion was half of the original representing approximately a doubling of the elemental concentrations compared with the original sample. Scandium was added to give 50 p ml-' in the final solution. Milk A 10ml volume of commercial pasteurized bovine milk (Vitamin D milk grade A homogenized Maple Hill Farms Bloomfield CT USA) was prepared in an ashing furnace (Fisher Isotemp Model 497).1° A few millilitres of sub-boiled nitric acid together with 10 ml of 250 pg ml-' of Sc solution were added to the residue and the final solution made up to 50 ml with DDW.Procedure Resins The ability of the three resins to extract metals from buffered solutions (0.1 mol 1-l ammonium acetate) was studied. The columns were prepared by washing with 2mol 1-l nitric acid followed by 5 ml of buffer. A 5 ml volume of the buffered test solution was then added. The standard buffered solutions contained Zn Cu Co Mn and Y at 0.5vgml-'. The pH values studied were 4.8 6.7 8.9 9.4 and 10.2. Elution was with 5 ml of 2 mol 1-l nitric acid. The separation of Na and C1 from Zn was also investigated. Ammonium acetate buffer (5 ml) at pH 5.5 containing 5000 pg ml-' of Na 200 pg ml-' of Ca and 0.5 pg ml-' of Zn was passed through the columns and eluted.Additional wash- ing with 5 ml of buffer preceded the elution step. CPPI resin with milk and urine Extractions of the urine and milk digests were performed in a class-100 clean room according to the already established procedure." For urine 15 ml aliquots of the digested material were extracted. The final volume was 5 ml (2 mol I-' HN03). For milk 1 ml aliquots were extracted and the final volume was 5 ml (2 mol 1-l HNO,). Aliquots of the digested urine (15 ml) were evaporated to dryness in pre-leached quartz beakers. The residue was taken up in sub-boiled nitric acid neutralized with isothermal ammonia solution and buffered with ammonium acetate buffer (final buffer concentration about 0.1 mol l-' pH 5.5). The resulting solution was passed through a CPPI column and eluted with 5 ml of 2 mol 1-' nitric acid into a vial (12 ml number 6133 Spex Industries Metuchen NJ USA) prior to analysis by ICP-MS.A 3-fold preconcentration resulted com- pared with the digested material. An identical procedure was employed for milk except that the initial aliquots were 1 ml. Only 1 ml of digested milk was required since it contains relatively high Zn concentrations (typically 1 pg ml - I). Results and Discussion ICP-AES Optimization Optimization of the Zn determination by ICP-AES has been described re~ently.~' The Zn 1213.856 nm line exhibited greater signal-to-background ratios than the Zn I1 202.548 nm emis- sion. Consequently the optimal conditions were established with the Zn I line based on the 3a detection limit (DL) and the background equivalent concentration (BEC).The DL and BEC for Zn at various nebulizer flow rates and measuring heights are given in Table4. The lowest DL and BEC values occur at a nebulizer flow rate of 0.7 1 min-'. This flow also corresponds to the best correlation between the intensities of the Zn 1213.856 nm and the Sc I1 424.683 nm line as a function of viewing height. The Sc I1 line was used for signal compensation. No significant difference exists in the BEC for viewing heights between 10 and 15 mm above the induction coil but the DL decreases with decreasing viewing height to a minimum at 7 mm. Signal-to-noise ratios are high at viewing heights above 15 mm or below 10 mm. The optimal viewing height was 11 mm at which the best peak precision was also found.At a nebulizer flow rate of 0.7 1 min-' and different viewing heights signal precision was tested with solutions containing 20 pg ml-' of Zn 50 pg ml-I of Sc and 2 10 and 20% v/v of nitric acid. The relative signals for Zn and Sc were estimated against solutions in 2% nitric acid with and without Myers-Tracy compensation As a check the Sc I1 424.683 nm emission was measured as an analytical line normalized to itself as measured by the second monochromator. The results are illustrated in Table 5. Investigation of the correlation between the Sc I1 424.683 nm line with atomic and ionic lines at different nebulizer flow rates and viewing heights has been presented recently.32 The best correlation of Zn I 213.586 nm with the Sc reference line over a broad range of viewing heights occurred at a nebulizer flow rate of 0.71 min-'.When the nitric acid concentration is in the range 1-20% v/v 10% v/v nitric acid can be used to prepare the standard solutions. The Myers-Tracy compensation also improves the precision by a Table 4 Detection limit (30 measured in pg m1-l) and background equivalent concentration (BEC) for Zn I 213.856 nm (n= 10) Nebulizer flow rate) min-' 0.7 1 .o 1.3 Viewing height/mm DL BEC DL BEC DL BEC 20 18 15 14 13 12 11 10 9 8 7 6 0.019 0.01 8 0.0099 0.0094 0.0078 0.0077 0.0070 0.0065 0.0065 0.0061 0.0056 0.0066 0.10 0.027 0.14 0.093 0.44 0.10 0.073 0.01 5 0.088 0.060 0.33 0.077 0.074 0.076 0.073 0.080 0.0 10 0.073 0.044 0.27 0.083 0.087 0.084 0.1 1 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -202 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL.9 130 110 - E 90- 2 > 2 7 0 - a 50 30 Table 5 nebulizer flow rate of 0.7 1 min-' Percentage relative signal for Zn and Sc with (MT) and without Myers-Tracy (NMT) signal compensation (n= 10 RSD = 2%) at a Acid concentration/% HN03 - - I I I I I ; 6 7 8 9 10 10 20 Zn s c Zn s c Viewing height/mm MT NMT MT NMT MT NMT MT NMT 20 96 88 99 89 93 81 99 83 18 97 87 99 89 94 82 99 83 15 98 82 99 83 96 81 100 83 14 98 93 99 94 99 93 99 93 93 100 93 96 94 100 96 13 97 12 97 90 100 96 95 91 99 94 11 98 86 99 87 96 82 99 84 10 98 91 99 91 96 91 99 94 9 98 84 99 85 95 73 99 76 8 98 78 99 74 95 73 99 85 7 98 80 99 79 96 77 99 82 factor of 2-5.32 Frequent recalibration is unnecessary because the responses exhibit good long-term stability. Fig.1 Recovery of elements following exchange on Chelex-100 A Mn; B Zn; C Cu; D Y; and E Co. Original solutions buffered in 0.1 mol 1-' ammonium acetate buffer. All elements originally at 0.5 pg ml-' 120 1 t 1. 100 - A 80 - 2 ? = c 60- 40 - T A 5 6 7 8 9 10 PH Fig.2 Recovery of elements following exchange on CPPI A Mn; B Zn; C Cu; D Y; and E Co. Original solutions buffered in 0.1 mol 1-' ammonium acetate buffer. All elements originally at 0.5 pg ml-' Matrix Removal Using a test solution containing Ca Na and Zn in 0.1 mol 1-l ammonium acetate buffer (pH about 5.5) the effectiveness of Chelex- 100 and CPPI for matrix separation was examined. Calcium is present at relatively high concentrations in bovine milk (about 1200pgml-1)33 and may compete with Zn for binding sites. Calcium and Na also cause analytical signal suppression with ICP-MS.The concentrations of Na and Ca in the eluent from the resin columns are shown in Table 6. The CPPI resin is very effective at removing Na. Although some Ca is captured and eluted from the resin no indication exists that this interferes with the exchange of Zn under these conditions. Recoveries of the latter are always about 90%. Although present at several thousand pg ml-' in the starting solution C1 concentrations in the eluents were negligible. In spite of the relatively poor ionization of C1 in the ICP (=0.9%0) a detection limit of <lOpgrnl-' was obtained based on the response using a 10 p1 sample. Other chlorine species (i.e.37Cl 51C10 and 75ArC1) were measured but 35Cl ion was the most sensitive. This result is expected because the 35Cl isotope is the most abundant (75.8%) and suffers no significant interference. Chlorine was evolved above 1200 "C; therefore data acqui- sition began as soon as this temperature was reached and lasted 30 s. The response peak typically observed at about 12JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 203 Table 6 Concentrations (pg m1-I) in the eluent following separation; original solution 5000 pg ml-' of Na and 200 pg ml-' of Ca in 0.1 mol 1-' ammonium acetate buffer at pH 5.5 Eluent Na Ca c1 ~ ~~~ Chelex-100- Column 1 62 182 15 Column 2 332 161 19 Column 3 40 176 <lo* Column 1 1.9 93 < 10 Column 2 0.4 112 < 10 Column 3 3.3 157 < 10 CPPI- *DL= 30 of blank measured in pg m1-l.and 23 s was integrated for calibration purposes. The cali- bration exhibited good linearity (r2 = 0.988) in the range 10-1500 pg ml-'. Based upon the successful removal of Na and Cl from synthetic solutions the effectiveness of the CPPI separation with milk and urine samples was examined. Chelex- 100 appears less capable and suffers from considerable volume changes depending on the ambient pH. The 70Zn 68Zn ratio was determined in digested milk from which the matrix had been extracted by the established extrac- tion procedure and digested milk from which the matrix had been extracted by the chelating method using the CPPI resin. The data obtained are given in Table 7. The isotope ratios determined in the untreated milk digest differ significantly from those determined after matrix separ- ation.This is expected because the former contains both C1 and Na which give rise to the interference 70Cl; and matrix suppression respectively. Thus the measured 70Zn 68Zn ratio may be elevated or depressed depending on the concentrations of Cl and Na. However no statistically significant difference in the isotope ratios is obtained for the two treatment pro- cedures. Precision of the isotope ratios of the treated milks are a little poorer than those of the untreated milk. This may be due in part to the dilution of the treated samples and the resulting lower signal count rate. Excellent Zn recovery (99%) is obtained with the resin separation method. Concentrations of elements present in the test samples (and related to the determination of Zn but not reported in Table 7 include Ca Na and C1.The concentration of Ca in the digested milk was about 260 pgml-'. After column separation the concentration was reduced to <20 pg ml-'. Sodium concen- tration was reduced from 60pgml-' in the digest to < 1 pg ml-' following column separation. Chlorine concen- trations were also low (about 20 p ml-') in the final solutions. For urine the Zn isotope ratios determined after extraction or matrix separation are not significantly different (Table 7). However the expected difference between treated and untreated samples is significant for the same reasons given above for the milk analysis. The precision of the Zn isotope ratios are better for the treated urine samples than for the untreated digest which might be due in part to the preconcen- tration of Zn that accompanies the urine treatment.A Zn recovery of 76% was obtained with the resin procedure. Typically Ca concentrations in the digested urine were about 200 pg ml-l. These were reduced to < 15 pg ml-' fol- lowing column extraction despite the 3-fold concentration produced by the separation. Similarly Na concentrations were reduced from 3800pgml-' in the untreated digest to < 1 pg ml-' in the solution subjected to column matrix separ- ation. Chlorine concentrations in the final treated solutions were typically < 15 pg ml-'. The lower concentrations of interferent elements and the higher Zn concentrations found in milk make the determi- nation of Zn in milk straightforward compared with that of urine.Moreover the digestion of milk is simple and rapid while the digestion of urine requires more manipulations and thus introduces a greater probability of elemental loss or contamination. Precipitation also can occur in the urine digest and requires the addition of distilled de-ionized water and re-heating of the sample. These factors perhaps explain the relatively low (76%) Zn recovery for urine with the resin procedure. Conclusions Although care must be taken with the handling of the digested material both to minimize contamination and to ensure adequate redissolution of the residue following drying of the digested material the proposed resin separation procedure is simpler and less labour intensive than the extraction methods previously Saturation of the CPPI resin would lead to low Zn recovery and sufficient resin should be placed in the column to avoid this.A similar limitation exists in the conventional CCl extraction. For example recovery of only approximately 50% was obtained in preliminary experiments with bovine milk and the conventional extraction resulting from too large an aliquot volume used for the extraction. Any error depends on the final Zn Concentration obtained. A poor recovery implies poor sensitivity and precision. However in principle the isotope ratio measurement should not be affected if the Zn concentration remains significantly above the limit of determination. In practice measuring the Zn concentration in the digested material is prudent.This allows identification of samples with high Zn content and permits determination of Zn recovery from the separation procedures. Thus an alternative sample treatment for matrix separation of biological materials has been established for subsequent Zn isotope ratio measurements by ICP-MS. The procedure is simpler than the conventional extraction procedures.10734 Moreover the procedures established earlier"*34 involve the extraction of Zn into a solution containing CCl and either APDC or diethylammonium diethydithiocarbamate (DDDC) respectively. Subsequently the Zn is extracted into nitric acid. As CC14 must not be transferred into the final solution careful and time-consuming sample handling is required. This intrinsic disadvantage of working with chlorine-containing reagents is not shared by the resin extraction method.The next stage in development is to employ an automated Table 7 70Zn 68Zn isotope ratios determined in biological samples by ICP-MS; analysis of six replicates Sample Ratio Standard deviation Zn recovery (%) Bovine Milk Solution- Digest 0.03478 0.00002 Conventional extraction 0.03742 0.00045 Column separation 0.03706 0.00076 Digest 0.1 1035 0.00395 Conventional extraction 0.03921 0.001 10 Column separation 0.04117 0.000 3 8 Human Urine Solution- 89 99 - 94 76204 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 system for the column separation. Similar procedures for both off- and on-line preconcentration with ICP-AES have been described.354’ A semi-automated procedure was applied to preconcentrate and separate trace metals from high-salt matrices prior to analysis by ICP-MS.21 On-line approaches reduce the time of analysis and the risk of sample contamination.We thank C. Amarasiriwardena for valuable discussions of the digestion procedures and ICP-MS analyses and P. Kandola for technical assistance. This research was supported by the ICP Information Newsletter and some equipment was provided by The Perkin-Elmer Corporation. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 References Applications of Inductively Coupled Plasma Mass Spectrometry eds. Date A. R. and Gray A. L. 1988 Blackie Glasgow. Hieftje G. M. and Vickers G. H. Anal. Chim. Acta 1989 216 1. Jarvis K. E. Gray A. L. and Houk R. S. Handbook of Inductively Coupled Plasma Mass Spectrometry 1992 Blackie Glasgow.Braun T. and Zsindely S. TrAC Trends Anal. Chem. 1992 11 267. Ward N. I. Abou-Shakra F. R. and Durrant S . F. Biol. Trace Elem. Rex 1990 26 177. Durrant S . F. Analyst 1992 117 1585. Amarasiriwardena C. J. Gercken B. Argentine M. D. and Barnes R. M. J. Anal. At. Spectrom. 1990 5 457. Ting B. T. G. Mooers C. S. and Janghorbani M. Analyst 1989 114 667. Cary E. E. Wood R. J. and Schartz R. J. Micronutr. Anal. 1990 8 13. Amarasiriwardena C. J. Krushevska A. Foner H. Argentine M. D. and Barnes R. M. J. Anal. At. Spectrom. 1992 7 915. Ward N. I. Abou-Shakra F. R. Durrant S. F. Thompson J. Havercroft J. M. and Yadegarian L. in Monography Proceedings Round Tables and Discussions of the 7th Int. Symp. on Trace Elements in Man and Animals (TEMA-7) ed.Momcilovic B. IMI Edinburgh Scotland 1991 pp. 33. Durrant S . F. TrAC Trends Anal. Chem. 1992 11 68. Zhu G. and Browner R. F. Appl. Spectrosc. 1987 41 349. Date A. R. Cheung Y. Y. and Stuart M . E. Spectrochim. Acta Part B 1987 42 3. Gray A. L. and Williams J . G. J. Anal. At. Spectrom. 1987 2 599. Gregoire D. C. Spectrochim. Acta Part B 1987 42 895. 17 18 19 20 21 22 23 24 25 Tan S. H. and Horlick G. J. Anal. At. Spectrom. 1987 2 745. Kim Y.-S. Kawaguchi H. Tanaka T. and Mizuike A. Spectrochim. Acta Part B 1990 45 333. Lyon T. D. B. Fell G. S. Hutton R. C. and Eaton A. N. J. Anal. At. Spectrom. 1988 3 265. Branch S. Ebdon L. Ford M. Foulkes M. and O’Neill P. J. Anal. At. Spectrom. 1991 6 151. Heitmar E. M. Hinners T. A. Rowan J. T. and Riviello J.M. Anal. Chem. 1990 62 857. Horvath Z. Lasztity A. and Barnes R. M. Spectrochim. Acta Reu. 1991 14 45. Kingston H. M. Barnes I. L. Brady T. J. Rains T. C. and Champ M. A. Anal. Chem. 1978 50 2064. Wang X. and Barnes R. M . J. Anal. At. Spectrom. 1989,4 509. Egan C. B. Smith F. G. Houk R. S. and Serfass R. E. Am. J.-Clin. Nutr. 1991 53 547. Ward. N. 1.. J. Micronutr. Anal. 1986. 2 211. 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 Rose,’ M. S The Foundations of Nutrition 3rd edn Macmillan New York 1938 p. 391. Myers S. and Tracy D. Spectrochim. Acta Part B 1983,38 1227. Barnes R. M. Biol. Trace Elem. Res. 1984 6 93. Horvath Z. and Barnes R. M. Anal. Chem. 1986 58 1352. Krushevska A. Barnes R. M. Amarasiriwardena C. J. Foner H. and Martines L. .I. Anal. At. Spectrom. 1992 7 851. Krushevska A. Barnes R. M. and Martines L. 1992 Winter Conference on Plasma Spectrochemistry Paper TP29 San Diego California January 6-11 1992. Emmett S . E. J. Anal. At. Spectrom. 1988 3 1145. Patterson K. Y. Veillon C. Moser-Veillon P. B. and Wallace G. F. Anal. Chim. Acta 1992 258 317. Barnes R. M. and Genna J . G. Anal. Chem. 1979 51 1065. Barnes R. M. and Lu S.-j. Can. J. Spectrosc. 1983 28 139. Knapp G. Muller K. Strunz M. and Wegscheider W. J . Anal. At. Spectrom. 1987 2 611. Prakash N. Csanady G. Michaelis M. R. A. and Knapp G. Mikrochim. Acta 1989 111 257. Schramel P. Xu L.-Q. Knapp G. and Michaelis M. Mikrochim. Acta 1992 106 191. Schramel P. Xu L.-Q. Knapp G. and Michaelis M. Fresenius’ J. Anal. Chem. 1993 345 600. Israel Y. Krushevska A. P. Foner H. Martines L. J. and Barnes R. M. J. Anal. At. Spectrom. 1933 8 467. Paper 3/0391 OB Received July 6 1993 Accepted September 1 1993