JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 115 Application of Isotope Dilution Analysis-Inductively Coupled Plasma Mass Spectrometry to the Precise Determination of Silver and Antimony in Pure Copper* Koichi Chiba lsamu lnamoto and Masao Saeki Materials Characterization Laboratory Nippon Steel Corporation I6 18 Ida Nakahara-ku Kawasaki 2 1 I Japan isotope dilution analysis combined with inductively coupled plasma mass spectrometry was applied to the determination of ultra-trace levels of silver and antimony in pure copper. The precision and sensitivity for the analysis of pure metallic materials was investigated. This technique gives analytical values very close to those of the conventional standard additions method but has higher precision. By using a preconcentration technique it is possible to determine 20 ng g-l of silver and 5 ng g-l of antimony in pure copper while keeping the analytical errors to less than 10%.Keywords inductively coupled plasma mass spectrometry; isotope dilution analysis; pure copper analysis; silver determination; antimony determination Recently there has been a concerted effort to analyse high- purity metallic materials with high sensitivity and high precision. Copper is one of the most important components in electronic devices and the precise determination of ultra- trace levels of impurities especially silver and antimony is required in order to improve the quality of the devices. Inductively coupled plasma mass spectrometry (ICP-MS) makes it possible to determine sub-ng ml-l levels of elements in the aqueous phase.The remarkably high sensi- tivity of the technique has attracted the attention of many researchers since it was introduced by Houk et al. in 1980.' Isotope dilution analysis is noted for its high precision. It does not require calibration graphs for elemental analysis and is not affected by losses and errors during sample treatment procedures. The method has been used primarily with thermal ionization mass spectrometry (TIMS)2 and spark-source mass spectrometry (SSMS).3 Thermal ioniza- tion requires skilful instrument operation and has the limitation that some elements are undetectable. Spark- source MS needs complicated sample preparation for precise measurements. Isotope dilution analysis has found application primarily in the geological field.The combination of ICP-MS and isotope dilution analy- sis is expected to make possible the determination of ultra- trace levels of elements with high precision. Isotope dilution analysis-ICP-MS has begun to be applied to the analysis of environmental ~a'mples,~~~ biological sample^,^*^ food samples,* geological ~amples,~JO water and acids' l - I 2 and reference material ~amp1es.l~ However there have been few applications of the technique to the highly precise analysis of industrial m a t e r i a l ~ . ~ ~ J ~ The primary objective of this work was to determine ultra- trace levels of silver and antimony in pure copper using isotope dilution analysis-ICP-MS. Coprecipitation tech- niques were also applied to isotope dilution analysis in order to increase the sensitivity of this technique.This is valid because isotope dilution analysis is not affected by either recoveries or losses during the coprecipitation procedures. Experimental Instrumentation The ICP-MS instrument used was a PlasmaQuad (VG Elemental Winsford Cheshire UK). The operating condi- * Presented at the XXVII Colloquium Spectroscopicum Interna- tionale (CSI) Bergen Norway June 9- 14 199 l . Table 1 Operating conditions ICP- Power 1 . 3 kW Coolant gas flow rate Auxiliary gas flow rate Carrier gas flow rate 12.4 dm3 min-I 0.48 dm3 min-' 0.82 dm3 min-' Ag Sb Mass spectrometer- Mass range 106.0- 1 10.0 1 19.90- I3 1.56 Channel No. 512 512 Scanning No. 1000 I500 Exposure time per channel 80 ps 80 ps tions are given in Table 1. The total isotope ratio measure- ment time is 40 and 60 s for silver and antimony respectively.The measurement was repeated ten times for each sample and the analytical errors in the isotope ratio measurements were defined as a la deviation. Reagents All acids were of ultra-pure grade from Tama Chemical (Japan) and were used as received. The water used for all solutions was prepared by double distillation of deionized water. Enriched nuclides of silver and antimony were purchased from Oak Ridge National Laboratory (Oak Ridge TN USA). The enriched nuclide of silver consisted of 99.1 1% of Io7Ag and 0.89% of Io9Ag and that of antimony consisted of 97.48% of 123Sb and 2.52% of l2ISb. The spike solutions were prepared by dissolving the enriched nuclides in 1 mol dme3 nitric acid. In the preparation of a stock solution of silver about 0.1 g of silver was dissolved in 20 ml of nitric acid ( I + 1).The test solutions for the investigation of the analytical preci- sion of the isotope ratio measurement were prepared by diluting the silver stock solution. Samples Standard reference materials (SRMs) of copper from the National Institute of Standards and Technology (NIST) (Gaithersburg MD USA) were used as test samples. The SRMs 393 395 and 396 were analysed for silver and SRMs 393 395 and 398 for antimony. Two other copper samples were used to examine the validity of the method 99.99%116 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 Copper sample (1 g) e- HNO (1+1) (10 ml) Spike addition +Te4'; 50 mg addition I Decomposition Evaporation to dryness I - HCI (20 ml) Dissolution -a- SnCI2-2H,O-HCI I Reduction to precipitation Filtration I (Filtrate) (Precipitate) Rejection -1 - Water - HN03(1+1) (10 ml) Dissolution Test solution (50 ml) ICP-MS Fig.1 Preconcentration procedure for isotope dilution analysis of silver in copper and 99.9999% copper samples purchased from Koch-Light (Colnbrook Buckinghamshire UK). Sample Preparation Procedures for Copper Samples When the concentrations of silver or antimony were expected to be more than 1 pg g-l about 0.1 g of copper sample was dissolved in 10 ml of nitric acid (1 + 1) and diluted to 100 ml. The sample solutions were measured directly by ICP-MS. A coprecipitation technique was applied for preconcen- tration of silver and antimony before analysis when the concentrations were expected to be less than 1 pg g-l.The coprecipitation procedures for silver and antimony are summarized in Figs. 1 and 2 respectively. For the determination of silver samples (about 1 g) were decomposed in 10 ml of nitric acid (1 + 1) and diluted to 150 ml. A suitable amount of spike solution was added to the sample solutions followed by 10 ml of 1% tellurium(1v) chloride solution which contained about 50 mg of Te4+ as a coprecipitating reagent. This solution was evaporated to dryness in order to remove the nitric acid because it interferes with precipitation in the next step. The residue was dissolved again in hydrochloric acid and diluted to 100 ml. An aliquot (20 ml) of 20% tin(@ chloride in 6 mol dm-3 hydrochloric acid was added to the sample solution in order to reduce Te3+ for the formation of a metallic tellurium precipitate.The precipitate was filtered and then dissolved again in 10 ml of nitric acid (1 + 1) and diluted to 50 ml. This was the final test solution for ICP-MS measurements. The recovery of the over-all procedure was more than 70%. The concentration of silver was about ten times higher than that of the test solution without the preconcentration treatment. For the determination of antimony samples (about 1 g) were decomposed in 10 ml of nitric acid (1 + 1) and diluted to 50 ml. If some residue remained it was filtered and dissolved in a sulfuric acid-nitric acid mixture. The two solutions were combined to form the sample solution. A Copper sample (1 g) I - HNO (1+1) (10 ml) Decomposition Filtration - H,SO + HNO - Spike addition - La3+; 10 mg addition Heating pH adjustment to precipitation Filtration 1 I c N H l a q ) (1+1) (Filtrate) I (Precipitate) I +-NO ( i + i ) (10 mi) Dissolution c- Water I 1 Rejection Test solut,ion (50 ml) ICPiMS Fig.2 Preconcentration procedure for isotope dilution analysis of a.ntimony in copper suitable amount of spike solution was added followed by I0 ml of 0.2% lanthanum(II1) chloride solution which contained 10 mg of La3+ as a coprecipitating agent. The pH of the solution was adjusted to 9 in order to form a lanthanum(II1) hydroxide precipitate. The precipitate was dissolved in 10 ml of nitric acid (1 + 1) and diluted to 50 ml. This is the final test solution for ICP-MS measurements. The recovery of the over-all procedure was more than 80%.The concentration of antimony was at least ten times higher than that of the test solution without the preconcentration treatment. Isotope Dilution Analysis In isotope dilution analysis the amount of an element is calculated from the isotope ratio using the following equation (1) where x is the molar concentration of the target element in the sample P is its molar concentration in a spike solution R is the measured isotope ratio in the mixture of the sample and the spike a and b are the natural isotopic abundances and A and B are the isotopic abundances in the spike. The analytical error in isotope dilution analysis [the error magnification factor F(Oh)] is defined as follows (2) where R is the measured isotope in the sample solution r is the deviation when the isotope ratio is observed as R and X(R-tr) and X(R) are the molar concentrations of the element determined from the isotope ratio R k r and R according to eqn.( l ) respectively. From eqn. (2) it can be seen that the error magnification factor i.e. the precision in the isotope dilution analysis is affected both by the values of the isotope ratio and by the x= P(A - BR) /(bR - a) F= 1 OO[X(R k r) - X(R)] /X(R)JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 10 E n a 1.0 v) 117 - - W -8 - i 0 10 20 30 40 50 60 Isotope ratio ( R ) Fig. 3 Relationship between error magnification and isotope ratio (lo7Ag:'OgAg) in the determination of silver. RSD in measuring isotope ratio 1 5; 2 I; 3 0.5; and 4 0.1% 10 - ' I .- 0 10 20 30 Isotope ratio ( R ) Fig.4 Relationship between error magnification and isotope ratio ( 123Sb:121Sb) in the determination of antimony. RSD in measuring isotope ratio I 5 ; 2 1; 3 0.5; and 4 0.1% deviations in the measurement of the isotope ratio. The dependences of the error magnification factors on R and r for the determination of silver and antimony are shown in Figs. 3 and 4 respectively. There is an isotope ratio which makes the error magnification factor a minimum when the deviation of the isotope ratio measurements is assumed to be constant because F is a non-linear function of both R and r. The results suggest that the isotope ratio in a mixture of a sample and spike solution should be adjusted by the addition of the correct amount of spike solution in order to increase the precision of the method.Results and Discussion Mass Discrimination Mass discrimination which is the discrepancy between the natural isotope ratio and measured isotope ratio was also investigated. There may be two main causes of mass discrimination effects matrix effects and the bias of the detection system itself. Matrix effects of copper on the isotope ratio measure- ments of silver and antimony are summarized in Table 2. When the amount of copper matrix increases from 0 to 1000 pg ml-l the measured isotope ratios of both silver and antimony remain constant. A copper matrix of 1000 pg ml-1 was thought to be the upper limit of matrix concentration because shot noise was observed. This noise is thought to be due to the neutral particles of copper and/or copper oxide which may be condensed at the end of the plasma.The stability of ICP-MS measurements also de- creased when a solution with a higher dissolved solids 0.1 1 .o 10 100 Ag concentrationtng ml-' Fig. 5 Dependence of precision of isotope ratio measurement on the concentration of silver content was introduced into the plasma. It is concluded that there is no mass discrimination effect due to the copper matrix in measurements of silver and antimony isotope ratios,' because the mass of copper is much less than that of silver or antimony.16 The averages of the measured isotope ratios of silver and antimony were 1.044 and 0.7783 respectively their natural abundances being 1.076 and 0.7463 respectively. It was found that there were small but constant differences between the natural and the measured isotope ratios.These discrepancies are due to the efficiency of the mass detection system for each element. The ratio of the natural and the measured isotope ratios was defined as the mass discrimina- tion factor. Fortunately these types of mass discrimina- tions are easily corrected for by multiplication by the mass discrimination factors. Precision of Isotope Ratio Measurement The analytical precision of the isotope ratio measurement using ICP-MS was investigated. The dependence of the relative standard deviations of isotope ratio measurements on the concentration of silver is shown in Fig. 5 . In common with other analytical methods the precision is in inverse proportion to the concentration.For the determina- tion of silver the relative standard deviation (RSD) of measurement at the 10 ng ml-l level is about 1% and that at the 1 ng ml-I level is about 10%. These deviations together with the measured isotope ratio define the precision of isotope dilution analysis. 0.1 ' I I 0.1 1 .o 10 Ag concentration lpg g-' Fig. 6 Relationship between analytical precision and concentra- tion of silver in the analysis of copper. A Isotope dilution analysis-ICP-MS; and B standard additions method118 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1992 VOL. 7 Table 2 Mass discrimination effect Cu matrix/pg ml-1 Concen tration/ Element ng ml-' 2 5 10 20 Average Natural abundance Io7Ag:'O9Ag Sb 10 20 Average Natural abundance 123Sb:121Sb 0 100 500 1000 1.064 f 0.053 1.057 f 0.046 1.023 f 0.046 1.033 -t 0.070 1.046 f 0.03 1 1.042 f 0.034 1.037 f 0.02 1 1.02 1 +- 0.033 1.048 f 0.0 1 7 1.05 1 2 0.02 1 1.046 f 0.007 1.039 f 0.009 1.048 f 0.01 1 1.049 2 0.0 1 3 1.044f0.011 1.076 0.778 1 f 0.0098 0.7780 f 0.0060 0.7795 ? 0.0069 0.7788 -+ 0.0096 0.7787 f 0.0039 0.776 1 f 0.0059 0.7798 f 0.0075 0.7773 k 0.0059 0.7783f0.0012 0.7467 1.053 f 0.021 1.047 f 0.02 1 Table 3 Determination of silver in copper samples.Mean results in pg g-l in= 10) Without preconcentration- Method* 99.99% SRM 395 SRM 396 copper sample I D-ICP-MS 12.19 f 0.20 3.15 f 0.06 10.36k0.08 ICP-MS (standard additions) lo7Ag 11.9f0.4 3.0 f 0.2 9.8 f 0.4 Certified value 12.2 3.30 - ImAg 12.1 f0.4 3.1 f 0.2 10.1 f0.5 With preconcen t ra t ion- Method* 99.9999% SRM 393 copper sample I D-ICP-MS 0.106 f 0.004 0.2 14 f 0.006 ICP-MS (standard additions) Io7Ag <O.i! 0.26 f 0.07 lwAg <O.i! 0.4 1 f 0.08 ETAAS 0.20 Certified value 0.10 f 0.02 - ID= isotope dilution analysis; ETAAS= electrothermal atomic abscwption spectrometry (graphite furnace). ~~ Table 4 Determination of antimony in copper samples.Mean results :in pg g-I (n= 10) Without preconcentration- Method* ID-ICP-MS Certified value With preconcentration- Method* ID-ICP-MS Certified value ID=isotope dilution analysis. SRM 395 SRM 398 8.0 7.5 7.82 1 f 0.07 1 7.429 f 0.072 SRM 393 39.99% copper sample 99.9999% copper sample 0.128 f 0.04 0.25 f 0.05 - - 0.256 k 0.0 1 1 0.0056 f 0.0006 Analytical Precision of Isotope Dilution Analysis The analytical precision of isotope dilution analysis as a function of silver concentration was investigated and the results are shown in Fig.6. In these experiments the copper samples were analysed after dissolution in order to investi- gate the precision of the over-all analytical procedure. The isotope ratio of the test samples was adjusted to be about 4 by the addition of a suitable amount of spike solution because this isotope ratio gives the best precision according to Fig. 3. The precision is compared with that of the standard additions method in Fig. 6. As can be seen the isotope dilution method has a much higher precision. When a 1 pg g-' concentration of silver in a copper sample is measured by ICP-MS the isotope dilution method gives an RSD of about 3% whereas using the standard additions method an RSD of about 15% is obtained.Isotope dilution analysis has another major advantage when it is necessary to concentrate a target element before determination isotope dilution analysis is free from prob- lems due to recoveries and losses in preconcentration processes. This is because the isotope ratio in a test solution never changes once it has reached a state of equilibrium. Even when the concentration of silver is much lower than 0.1 pg g-l in a copper sample it can be determined with high precision if a suitable preconcentration technique is applied. Determination of Silver and Antimony in Copper Samples Silver and antimony in some copper samples were deter- mined by isotope dilution analysis-ICP-MS. Samples con- taining small amounts of silver or antimony were analysed after applying the preconcentration techniques shown inJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY.MARCH 1992 VOL. 7 119 Figs. I and 2 and the others were analysed without preconcentration. The results for silver and antimony are summarized in Tables 3 and 4 respectively. In the determination of silver the NIST SRMs 395 and 396 and the 99.99% copper sample were analysed without preconcentration. The results obtained by isotope dilution analysis and the standard additions method are in close agreement with each other and both results also agree well with the certified values. The SRM 393 and the 99.9999% copper sample were analysed after preconcentration. The result for SRM 393 obtained by isotope dilution analysis agrees well with the certified value.The results for the 99.9999% copper obtained by isotope dilution analysis the standard additions method and electrothermal atomic absorption spectrometry are in very close agreement with each other. Table 3 shows that the analytical precision of isotope dilution-ICP-MS is much higher than that of'the other methods. With isotope dilution-ICP-MS the analyti- cal error is within 5% even for the determination of 0.1 pg g-l of silver in copper. In the determination of antimony SRMs 395 and 398 were analysed without preconcentration. The SRM 393 and the 99.99% and 99.9999% copper samples were analysed after preconcentration. The results for the SRMs with and without preconcentration agreed well with the certified values. The other results show that isotope dilution-ICP- MS can be used to determine 0.005 pg g-* of antimony within a 10% RSD even after preconcentration.Conclusions Analysis by isotope dilution-ICP-MS was applied to the determination of ultra-trace levels of silver and antimony in high-purity copper samples and was shown to give the same accuracy and much better precision than the conventional standard additions method. It is very easy to combine isotope dilution with preconcentration methods because it is not affected by recoveries and losses in the preconcentra- tion procedures. While keeping analytical errors to less than lo% it is possible to determine 20 ng g-' of silver and 5 ng g-l of antimony in high-purity copper. Hence the combination of ICP-MS with isotope dilution analysis makes very precise measurements of low concentrations of impurities in high-purity materials possible. This method is capable of producing excellent results in a very short time. References 1 Houk R.S. Fassel V. A. Flesch G. D. Svec H. J. Gray A. L. and Taylor C. E. Anal. Chem. 1980 52 2283. 2 Saito T. Shimizu H. and Masuda A. Geochem. J. 1987,21 237. 3 Beary E. S. Paulsen P. J. and Lambert G. M. Anal. Chem. 1988,60 733. 4 Beauchemin D. McLaren J. W. Mykytiuk A. P. and Berman S. S. Anal. Chem. 1987 59 778. 5 Usztity A. Viczihn M. Wang X. and Barnes R. M. J. Anal. At. Spectrom. 1989 4 76 1. 6 Ting B. T. G. and Janghorbani M. Anal. Chem. 1986 58 1334. 7 Ting B. T. G. Mooers C. S. and Janghorbani M. Analyst 1989 114 667. 8 Dean J. R. Ebdon L. and Massey R. J. Anal. At. Spectrom. 1987 2 369. 9 Hall G. E. M. Park C. J. and Pelchat J. C. J. Anal. At. Spectrom. 1987 2 189. 10 Ward D. B. and Bell M. Anal. Chim. Acta 1990 229 157. 11 Paulsen P. J. Beary E. S. Bushee D. S. and Moody J. R. Anal. Chem. 1988 60 971. 12 Haraldsson C. Westerlund S. and Oehman P. Anal. Chim. Acta 1989 221 77. 13 Beary E. S. Brletic K. A. Paulsen P. J. and Moody J. R. Analyst 1987 112 44 1. 14 Makishima A. Inamoto I. and Chiba K. Appl. Spectrosc. 1990 38 697. 15 Umeda H. Inamoto I. and Chiba K. Bunseki Kagaku 199 1 40 109. 16 Tan S. H. and Horlick G. J. Anal. At. Spectrom. 1987 2 745. Paper 1/03950D Received July 30 1991 Accepted November 5 1991