Determination of Atmospheric Trace Metal Concentrations by Isotope Dilution Inductively Coupled Plasma Mass Spectrometry after Separation from Interfering Elements by Solvent Extraction TAKUNORI KATOH MASAYUKI AKIYAMA AND HIDEYUKI OHTSUKA Hokkaido Institute of Environmental Sciences Kita-ku Sapporo 060 Japan SEIJI NAKAMURA Muroran Institute of Technology Muroran 050 Japan KENSAKU HARAGUCHI Hokkaido National Industrial Research Institute Sapporo 062 Japan KUNI H I KO AK AT SUK A* Kitami Institute of Technology Kitami 090 Japan An isotope dilution ICP-MS method for the determination of six elements (Ni Cu Zn Cd TI and Pb) in atmospheric particulate samples is described. The method involves a complexation/solvent extraction step to separate these elements from potentially interfering elements prior to ICP-MS analysis.Accuracy of the method was demonstrated by analysis of a Vehicle Exhaust Particulates CRM. The method was successfully applied to the analysis of atmospheric particulates collected in a remote mountainous area of Hokkaido Island Japan. Keywords Isotope dilution inductively coupled plasma mass spectrometry trace metals atmospheric particulates solvent extraction The measurement of atmospheric concentrations of metals is gaining importance as the pollution of the atmosphere is growing world-wide. As samples of filter-collected atmospheric particulate matter are usually very small (ranging from several to a few tens of milligrams) a very sensitive analytical method is required to analyse them for background levels of trace metals.'-' Isotope dilution (ID) ICP-MS has become the preferred method of elemental McLaren et aL6 discussed the advantages of ID ICP-MS where elemental concentrations are determined by measurement of an isotope ratio rather than an absolute ion intensity.A great advantage of ID ICP-MS over other types of ICP-MS analyses is that its accuracy is not degraded by multiplicative interferences such as matrix suppression and isotope ratios are less affected by instrumental drift than are ion sensitivities. However additive-type inter- ferences such as spectroscopic overlaps can degrade the accu- racy of ID ICP-MS. The multi-element analysis of atmospheric particulate matter samples with ICP-MS particularly the determination of Ni Cu and Cd by an isotope dilution method can be complicated because of isobaric interferences on several of their most abundant isotopes by polyatomic species i.e.the interference * To whom correspondence should be addressed. Journal of Analytical Atomic Spectrometry of 40Ar180 and 40Ca'80 on 58Ni 44Ca160 on 60Ni and 40Ar23Na and 47Ti160 on 63Cu respectively. When Cd concen- trations are very low Sn MOO and MoOH may contribute significantly to the ion signals measured at m/z 110 111 112 113 114 and 116. Hence the prior separation of the trace metals from these matrix elements is necessary. In this paper an ID ICP-MS method is presented for the determination of concentrations of trace metals (Ni Cu Zn Cd T1 and Pb) in atmospheric particulates collected monthly from November 1989 to October 1990 in a mountainous area of Hokkaido Island (Hidaka Japan).Solvent extraction with dithizone was found to reduce the levels of interfering matrix elements adequately taking advantage of the high extractability of the heavy metal ions and the low affinity of Mo Sn V Ti and alkali and alkaline earth elements. The aim of this work was to evaluate the applicability of an ID ICP-MS method for the determination of total Ni Cu Zn Cd TI and Pb in airborne particulate samples after clean-up of dissolved samples by solvent extraction with dithizone. EXPERIMENTAL Apparatus A standard Model PMS-2000 inductively coupled plasma mass spectrometer (Yokogawa Analytical Systems Tokyo Japan) was used. The instrument was operated at an rf power setting of 1.2kW. The argon plasma gas flow was set 141min-'.Auxiliary gas flow and nebulizer gas flow were 0.5 and 0.8 1 min-l respectively. A sampler and skimmer both made of copper with orifice diameters of 1 and 0.5 mm respectively were employed. Sample solution was introduced by using a peristaltic pump which is standard equipment for the PMS-2000 at a delivery rate of about 0.5mlmin-' to the nebulizer. Airborne particulate matter samples were collected with a low-volume air sampler (Shintaku Amagasaki Japan). Reagents All high-purity acids ammonia solution and chloroform were prepared by sub-boiling distillation of analytical-reagent grade Journal of Analytical Atomic Spectrometry January 1996 Vol. 1 1 (69-71) 69reagents in quartz and Teflon stills. Doubly distilled water (DDW) was prepared by sub-boiling distillation of distilled water feedstock in quartz stills.Dithizone was purified by recrystallization from chloroform. Stable isotope-enriched spikes for the elements were pur- chased from the Oak Ridge National Laboratory (Oak Ridge TN USA). Stock solutions of the spikes were prepared by dissolving them in HNO or aqua regia followed by dissolution in 1 moll-' HN03 and their concentrations were checked by reverse spike ID ICP-MS. Samples Airborne particulate matter samples were collected on cellulose nitrate membrane filters (0.8 pm pore size; 47 mm diameter filter) at Hidaka (Hokkaido Japan) for 30 d with a low-volume air sampler. This sampler has a cyclone-type classifier which rejects particles larger than 10 pm in diameter. The sampling flow rate used was 20 1 min-'.The filter was kept at 20 "C for 2 d in an air-conditioned clean-room prior to weighing the filter without and with the aerosol sample. A Vehicle Exhaust Particulates CRM (NIES No. 8 National Institute for Environmental Studies Ibaraki Japan) was also used to check the accuracy of analyses. Digestion of the Sample Filter and Procedure The filter was folded and placed in a Teflon beaker. Appropriate amounts of enriched stable isotopes of the elements were added to the beaker for isotope dilution analysis and then 10ml of HNO and 5 ml of HC104 were added. The beaker was heated on a hot-plate at 200"C covered with a Teflon watch-glass. After the solution had clarified 2 ml of HF were added to the beaker and the mixture was re-digested on a hot-plate at 200°C until dry.The dried residue was subsequently re-dissolved in 2 ml of 1 + 1 HNO and made up to a volume of 20 ml. The sample solution was placed in a separating funnel after adjustment of the pH to 2. Then 30ml of 0.003% dithizone in chloroform were added. The mixture was shaken for 5 min to extract Cu into the chloroform layer. The aqueous phase was re-adjusted to pH 9 with ammonia solution and the mixture was again shaken for 5min. At this stage all of the elements (Cu Ni Zn Cd T1 and Pb) were extracted into the chloroform layer. The chloroform layer was then transferred into another separating funnel and 10 ml of 1 + 1 HNO were added to the funnel. The mixture was shaken for 5min to back-extract the elements into the aqueous phase. After the phases had been allowed to separate the aqueous phase was transferred into a Teflon beaker and 0.2ml of HC10 was added to the beaker. The solution was heated on a hot-plate at 200 "C until dry.The dried residue was re-dissolved in 0.5 ml of 1 + 1 HN03 and diluted with 10 ml of DDW. About 150 mg of the Vehicle Exhaust Particulates CRM (NIES No. 8) were digested and subsequently extracted by the same procedure. Blanks were prepared by carrying out the above procedure after the addition to the Teflon beaker in which the blank- filter was placed of enriched isotope spikes equivalent to one- tenth of the amounts added to the samples. Nickel Cu Zn Cd T1 and Pb were determined by ID ICP-MS using the following reference/spike isotope pairs 60Ni/61Ni; 63Cu/65Cu; 66Zn/68Zn; 111Cd/116Cd; 205Tl/203Tl; and 208Pb/206Pb.Weighed amounts of the spike solutions were added by means of adjustable micropipettes. In this work checks for mass discrimination were made with 50 pg1-l natural abundance solutions of each of the elements of interest as described in ref. 6. RESULTS AND DISCUSSION The described method for the collection of aerosol samples was first employed in 1988 in a mountainous area of Hidaka (Hokkaido Japan) about 100 km south-east of Sapporo in order to study the seasonal variability of atmospheric trace metal concentrations and to compare the data with those collected at Sapporo. During a 6 month experiment the amount of aerosol sample collected on each filter was in the range 4-15 mg when air was sampled at 20 1 min-l for 30 d.The aerosol samples contained levels of the trace analytes ranging from several tens of nanograms (Cd and T1) to several micrograms (Cu Zn and F'b) and much higher levels of Na Mg Al Ca Fe and Ti. Thus the prior separation of the trace metals from major matrix elements was necessary to perform ID ICP-MS analysis because of isobaric interferences by polyatomic ions of CaO CaOH ArNa and T i 0 on 60Ni 61Ni 63Cu and 65Cu. As concentrations of Cd in the samples are very low MOO and Sn probably affect the ratio of '"Cd 'I6Cd measured. When concentrations of Cr and V which mainly originate from anthropogenic emissions are high 51V160H and 52Cr160 may contribute significantly to the ion signals of Zn measured at m/z 68. Although solvent extraction with dithizone is a well estab- lished method and the metals that react with dithizone have been summarized by Sandell,' separation of the analytes from the major matrix elements of the sample was examined under the present extraction conditions.The recovery of all the analytes was greater than 99% except for Cd for which the Table 1 with dithizone extraction Recoveries of the matrix elements by the present procedure Element Recovery (%)* Element Recovery (%)* Ca 0.2f0.1 Mo 0.010 k 0.005 Mg 0.03 k 0.02 Sn'" 1.9k0.5 0.3 kO.1 Na 1.0k0.6 V Cr 0.1 1 k 0.04 Ti 1.8f0.7 *The procedure was carried out using 20ml of aqueous phase containing 1 mg of Ca Mg and Na and 200 pg of Cr Mo Sn V and Ti respectively. Table 2 Absolute blanks and total procedural blanks of the present method Element Absolute blank/ng Total procedural blank/ng m- 3* Ni 22.6 f 2.1 0.045 & 0.004 c u 9.3 f 0.7 0.049 f 0.004 Cd 0.21 f 0.04 0.00027 & 0.00005 T1 < 0.05 < 0.00005 Pb 0.82 f 0.05 0.0041 f 0.0002 Zn 9.0 f 0.05 0.18 kO.01 * The value was obtained by dividing the total blanks including the filter by the total flow-rate of sample collection.Table3 (NIES No. 8) ID ICP-MS analysis of Vehicle Exhaust Particulates CRM This work Element Found/pg Concentration/pg g - Ni 2.99 & 0.025* 19.4 f 0.19 c u 10.4k0.15 67.7 & 1.0 Zn 160f1 1040f 10 Cd 0.163 f0.002 1.06 f 0.02 T1 0.022 & 0.001 0.15 f 0.001 Pb 32.3 & 0.3 210f2 Certified value/pg g-' 18.5 f 1.5 67+3 1040 f 50 1.1 kO.1 (0.17)+ 219k9 * Precision expressed as the standard deviation (n = 5). Value for T1 is not given by NIES.The value was obtained by Seiji Nakamura with ID SIMS. 70 Journal of Analytical Atomic Spectrometry January 1996 VoE. 11Table 4 Determination of atmospheric trace metal concentrations at Hidaka (Japan) by ID ICP-MS multi-element analysis Found*/ng m-3 Run No. 1 2 3 4 5 6 7 8 9 10 11 12 Period November 1989 December 1989 January 1990 February 1990 March 1990 April 1990 May 1990 June 1990 July 1990 August 1990 September 1990 October 1990 Aerosol/pg rnp3 5.4 4.6 4.3 8.1 8.2 10.2 11.6 9.8 7.6 6.9 5.0 6.1 Ni 0.32 0.31 0.37 0.79 0.46 0.51 0.64 0.51 0.20 0.26 0.35 0.25 c u 0.52 0.82 0.72 1.6 1.7 1.3 1.4 1.7 1.8 I .2 0.89 1 .o Zn 6.2 5.9 5.4 9.2 9.0 9.3 9.6 8.8 3.9 3.8 3.3 4.4 Cd 0.09 1 0.07 1 0.056 0.1 1 0.11 0.12 0.12 0.1 1 0.050 0.046 0.040 0.053 T1 0.020 0.02 1 0.018 0.026 0.025 0.020 0.035 0.030 0.017 0.01 1 0.041 0.012 Pb 4.5 4.1 2.1 4.8 6.2 5.7 7.0 6.2 3.0 2.3 2.4 2.5 *Blank value is subtracted.recovery was 96+2% whereas the recoveries of the matrix elements were less than 2% as shown in Table 1. As the organic matter in the aqueous phase after the back-extraction procedure is destroyed completely by heating with a few drops of concentrated HC10,,9 the precision of the ID ICP-MS analysis was less than 3% relative standard deviation (n = 5). Table 2 shows the absolute blanks obtained by the present method not including the contribution from the filter. As the filters used in this work contained relatively high amounts of Ni Cu and Zn the total procedural blanks including the filter were measured separately by the ID ICP-MS method.Typical values of these are also shown in Table 2. The analytical results for the NIES No. 8 CRM Vehicle Exhaust Particulates summarized in Table 3 resulted from five independent determinations. The accuracy of the method is evident from a comparison of the results with the certified values. The airborne particulate samples collected from Hidaka (Japan) were analysed by the proposed ID ICP-MS method. The analytical results obtained after total decomposition and extraction with dithizone followed by back-extraction with HNO are shown in Table 4. The results indicate that baseline levels of atmospheric trace elements can be determined with good precision. Detection limits ranged from 0.15 pg m-3 for Cd to 30pgm-3 for Zn based on three times the standard deviation of the blanks.Finally it is interesting to compare the values obtained during the periods February-May and July-October; the values obtained during February-May were higher than those obtained during July-October for all the elements measured with the exception of TI. The authors thank J. W. McLaren (National Research Council of Canada Ottawa) for helpful comments during the prep- aration of this manuscript. REFERENCES Duce R. A. Hoffman G. L. and Zoller W. H. Science 1975 187 59. Mukai H. Ambe Y. and Morita M. J. Anal. At. Spectrom. 1990 5 75. Dick A. L. Geochim. Cosmochim. Acta 1991 55 1827. Berg T. Royset O. and Steinnes E. Atmos. Environ. 1993 27A 2435. Akatsuka K. Hoshi S. McLaren J. W. and Berman S. S. Bunseki Kagaku 1994 43 61. McLaren J. W. Beauchemin D. and Berman S. S. Anal. Chem. 1987 59 610. McLaren J. W. Mykytiuk A. P. Willie S. N. and Berman S. S. Anal. Chem. 1985 57 2907. Akatsuka K. McLaren J. W. Lam J. W. and Berman S. S. J. Anal. At. Spectrom. 1992 7 889. Sandell E. B. Colorimetric Determination of Traces of Metals Interscience New York 2nd edn. 1950 pp. 90-100. Paper 51041 240 Received June 26 1995 Accepted October 27 1995 Journal of Analytical Atomic Spectrometry January 1996 Vol. 11 71