Non-spectroscopic suppression of zinc in ICP-MS in a candidate biological reference material (IAEA 392 Algae)† M. J. Campbell* and A. To� rve�nyi International Atomic Energy Agency, Agency’s Laboratories at Seibersdorf and Vienna, A-2224 Seibersdorf, Austria. E-mail: M.Campbell@IAEA.Org Received 1st March 1999, Accepted 30th April 1999 ICP-MS was used for the simultaneous determination of nine analytes in a candidate reference material, IAEA 392 Algae (Scenedesmus obliquus), following microwave digestion with nitric acid and dilution to volume with de-ionised water.For eight analytes, the results showed reasonable agreement with consensus mean results obtained from a world-wide intercomparison exercise on this material. However, the value obtained for Zn was 55.4% of the consensus mean result (69.6 versus 125.6 mg g-1, respectively). Analysis of the same digest solutions by flame AAS gave results which were in good agreement with the consensus mean (127.1 versus 125.6 mg g-1), indicating that all the Zn was in solution.When the method of standard additions was employed, the result was overestimated by #17% by ICP-MS. A 1+9 aqueous dilution of the digests gave an acceptable result for Zn (122.0 mg g-1). Evidence is presented to demonstrate that the underestimation of Zn by external calibration ICP-MS is due to a non-spectroscopic suppression of the Zn in the digest solution caused by nitric acid. accordance with the latest international guidelines.2–4 Introduction Nevertheless, the consensus mean value often lies close to the Microwave digestion in sealed Teflon containers using nitric mean value derived from an expert group and represents a acid followed by ICP-MS is probably one of the most versatile robust estimate of the ‘true’ value.methods available to the analytical chemist for the analysis of It was decided that IAEA 392 would be certified, for up to biological materials. It has the advantages of higher analytical 17 analytes, in accordance with the recommendations of an throughput and reduced contamination risk over conventional expert group5 on the basis of results obtained from four invited hot-plate digestion methods.It has been widely applied to a laboratories and the IAEA’s Seibersdorf laboratory. This large range of materials and (with some minor exceptions) is approach was intended to identify the practical limits of the characterised by accurate and precise results.This method was application of current guidelines relating to quality measures, recently used by this laboratory in the certification campaign uncertainty quantification and traceability for the certification of an IAEA candidate reference material (IAEA 392 Algae); of the trace element contents of a natural matrix reference however, the apparent Zn recovery was seriously underesti- material. The techniques used were neutron activation analysis mated by ICP-MS giving only 55.4% of the consensus mean (NAA), X-ray methods, electrochemical methods, thermal value, derived from a preliminary statistical evaluation of a ionisation mass spectrometry, AAS, ICP-OES and ICP-MS.world-wide intercomparison exercise on this material.1 If the The results of the certification campaign are currently being digests were diluted 1+9 in de-ionised water, an improved evaluated and, if appropriate, certified values for the trace recovery was obtained.The usually robust method of standard element content of the material will be published later this additions failed to compensate completely for the eVect in the year. Consequently, it should be stressed at this juncture that undiluted digests, resulting in an overestimation of the result. the values cited in this paper for the trace element contents of Evidence is presented to demonstrate that the underestimation this material do not constitute reference values. of Zn by external calibration ICP-MS is due to a non- The results for Zn from the preliminary statistical evaluation spectroscopic suppression of the Zn in the digest solution of the 1996 world-wide intercomparison exercise on IAEA 392 caused by nitric acid.obtained from NAA, AAS and ICP-MS are presented in Fig. 1 which shows the classical ‘S’-shape format with the horizontal part of the ‘S’ indicating values that agree with the consensus Intercomparison study mean. The 95% confidence interval is very narrow due to the The Analytical Quality Control Service (AQCS) of the large number of laboratories which contributed to the data set International Atomic Energy Agency (IAEA) organised a (126 accepted out of 142 submitted laboratory means).From world-wide intercomparison exercise to determine trace Fig. 1, it appears that the consensus mean is slightly overelements in an algae material [IAEA 392 Algae (Scenedesmus estimated with respect to two of the three techniques illustrated. obliquus 208)] during 1996.A wide range of analytes (38 in The means and 95% confidence intervals for the individual total ) were determined with varying degrees of precision and techniques are: AAS, 129.9±4.3; ICP-MS, 120.7±4.5; and accuracy.1 The IAEA is currently reviewing its policy relating NAA, 122.1±5.1 mg g-1 Zn, respectively. The overall consensus to such exercises which are no longer suYcient to characterise mean is 125.6±2.5 mg g-1 Zn. On the basis of these results, it a matrix as a reference material of high metrological value in is apparent that the Zn value obtained for IAEA 392 by ICP-MS in-house (69.6±7.8 mg g-1), during the certification study mentioned above, is a significant underestimate of the †Presented at the 1999 European Winter Conference on Plasma Spectrochemistry, Pau, France, January 10–15, 1999.‘true’ value. J. Anal. At. Spectrom., 1999, 14, 1313–1316 1313Fig. 1 Results from selected techniques used in the intercomparison of IAEA 392 Algae.Table 1 Comparison of results for the determination of Zn in various reference materials by ICP-MS Matrix This work/ 95% confidence Reference value/ 95% confidence mg g-1 interval mg g-1 interval Cabbage (IAEA 359) 23.2 7.0 38.9 — Algae (IAEA 392) 69.6 7.8 125.6 28 Oyster Tissue (NIST SRM 1566a) 531.2 11.9 830 57 Spinach (NIST SRM 1570a) 52.1 13.1 82 3.0 River Sediment (NIST SRM 1645) 1670 — 1720 170 then the vessels are sealed and tightened using a torque wrench Experimental prior to digestion.As a further confirmation of the adequacy All ICP-MS measurements were made using an Elan 6000 of the digestion procedure, a suite of eight other analytes was (Perkin-Elmer SCIEX, Concord, Ontario, Canada) and a determined in the algae material and the results, presented in Perkin-Elmer AAnalyst 800 (Perkin-Elmer, Norwalk, CT, Table 2, confirm that its performance was generally acceptable. USA) was used for the atomic absorption measurements.Both instruments were operated in their standard configurations. Standard additions For ICP-MS, data were accumulated in peak jump mode for In order to investigate the discrepancy observed between the the analytes of interest, with a 2 s integration per peak. The ICP-MS results for Zn in IAEA 392 and the consensus mean AAS measurements for Zn were performed using an air– value from the world-wide intercomparison, a variant of the acetylene flame at a wavelength of 213.9 nm with a slit-width technique of standard additions was used.Nine sub-samples of 0.7 nm with deuterium background correction. of algae (#100 mg: average 104.8 mg, RSD 3.5%) were taken In order to investigate whether or not the suppression of for digestion: three sub-samples were left unaltered, three had Zn observed in IAEA 392 for ICP-MS analysis was a matrixappropriate spike additions made prior to an overnight cold specific problem, a number of other biological matrices and a digestion stage in nitric acid and the last three were spiked river sediment sample were analysed for their Zn content following microwave digestion when the solutions were made following microwave digestion (Table 1). For the work up to volume (50 ml ).This enabled the influence of the spiking reported here, 100 mg of samplee digested and made up to a final volume of 50 ml (undiluted); in some cases, an Table 2 Results for the analysis of IAEA 392 Algae by ICP-MS aliquot of 1 ml was further diluted 1+9 in de-ionised water compared with the intercomparison consensus mean values (Millipore, Bedford, MA, USA; Milli-Q 18.5 MV).A cocktail of Sc, Y and In was added to all solutions for internal Element This work/ 95% confidence Nominal value/ 2sa standardisation purposes. The results are presented in Table 1. mg g-1 interval mg g-1 Chromium 4.32 0.15 4.37 3.2 Microwave digestion Cobalt 2.99 0.09 3.27 0.8 A simple, robust, microwave digestion program was developed Copper 19.1 1.19 22.4 15 Iron 468 16 491 156 to give quantitative recoveries for a range of analytes in Lead 0.67 0.10 0.55 0.3 various biological reference materials.In order to minimise Magnesium 1970 160 2314 740 polyatomic interferences, the only reagents used were 5 ml of Manganese 60.2 1.49 66.4 15.8 concentrated nitric acid (Merck, Darmstadt, Germany) and Nickel 0.55 0.13 0.64 0.48 de-ionised water (Millipore, Milli-Q).A Milestone microwave Zinc 69.6 7.8 125.6 28 oven (MLS 1200 mega, Milestone S.R.L., Sorisole, Italy) was aBased on the standard deviation of all (statistically) accepted results used. In this system, samples are placed in Teflon liners which for the intercomparison. are supported inside a resin sleeve, reagents are added and 1314 J. Anal. At. Spectrom., 1999, 14, 1313–1316chronology to be studied and gave a total of six results (three values each calculated against the pre- and post-spiking standard additions responses, respectively).The undiluted samples were also analysed by flame AAS. A typical result for both techniques is plotted in Fig. 2 and the results are given in Table 3. In an unrelated study, it had been noted that the apparent Zn recovery improved when the samples were diluted, so the solutions were diluted 1+9 and re-analysed at a later date (Fig. 3). Note, in Fig. 2 and 3, the y-axis is expressed as ‘equivalent concentration’ by dividing the signal response for AAS and ICP-MS by the appropriate sensitivity factors such that both techniques can be represented on a single graph (Fig. 2). Influence of acid strength The influence of acid strength is illustrated in Fig. 4. A series of solutions were prepared containing fixed amounts of Sc, Y, In and Zn in 0, 1, 3, 5 and 10% v/v HNO3, respectively. This Fig. 3 Determination of Zn in IAEA 392 by standard additions at eVect was further studied by digesting IAEA 392 using a half 1+9 dilution by ICP-MS.and a fifth of the normal volume of nitric acid (5 ml ) and making the solutions up as before (Table 3). Results and discussion When low analyte recoveries are encountered, it is usually assumed that the discrepancy is due either to (i) incomplete dissolution or (ii) loss of the analyte either as a volatile component or to the walls of the reaction vessel. The results Fig. 4 Influence of nitric acid concentration on Sc, Y, In and Zn response.shown in Table 1 indicate that the combination of the digestion procedure used here, with quantification by ICP-MS, leads to a systematic underestimation of Zn in biological matrices. For NIST SRM 1645 River Sediment, an acceptable value was found, but in this case a modified method was used involving mixed acid attack and a much larger dilution (vide infra). For the biological matrices, the usual reasons for low recovery can be dismissed since AAS analysis of the same digest solutions Fig. 2 Determination of Zn in IAEA 392 by standard additions using ICP-MS and AAS. gave correct results, indicating that all the Zn was indeed in Table 3 Summary of standard additions results for Zn in IAEA 392 Algae before and after dilution compared with the results from aqueous calibration Regression coeYcient Slope Replicate/ Replicate 2/ Replicate 3/ mg g-1 mg g-1 mg g-1 Undiluted— Pre-spiked 0.9941 1130.4 147.81 145.56 146.13 Post-spiked 0.9980 1118.3 147.10 148.27 148.88 Versus aqueous — — 74.96 74.13 73.82 calibration Diluted 1+9— Pre-spiked 0.9988 203.6 124.78 126.98 121.29 Post-spiked 0.9953 205.0 119.94 122.15 116.60 Versus aqueous — — 93.6 91.0 95.3 calibrationa Versus aqueous 117.10 119.30 113.84 calibrationb aOriginal standard set also diluted 1+9, i.e., 0.1% HNO3.bCalculated from original calibration, i.e., 1% HNO3. J. Anal. At. Spectrom., 1999, 14, 1313–1316 1315solution (Fig. 2). Furthermore, the method used gives good analytical response of In. These eVects are consistent with an recoveries for eight other analytes on the basis of ICP-MS eVective reduction in the energy of the plasma in the central results (Table 2).Zn is underestimated by ICP-MS against an analyte channel. An eVect of reducing the nebuliser gas flow aqueous calibration, despite the use of internal standards, is to increase the residence time in the plasma which, to some giving roughly half of the expected concentration.extent, oVsets the reduction in plasma energy. This idea can The influence of the chronology of the Zn spike incorpor- be visualised using the ‘zone model’ of Vanhaecke et al.10 ation for standard additions was studied in this work since which relates the spatial occurrence of the maximum region Campbell et al.6 showed that the chronology of the addition of M+ density to operational parameters of the instrument of an isotopically enriched spike had a strong influence on the and the mass of the analyte.It can be calculated that the final result for the determination of Hg in a biological CRM. addition of nitrogen from nitric acid to the plasma is two However, the results of the pre- and post-digestion spiking orders of magnitude lower than that used by Wang et al.;9 study (Table 3) demonstrate that such an eVect does not play however, the influence of the time taken for evaporation and a role here. dissociation of the acid may play a role.Furthermore, the Standard additions (to independent digests) to the undiluted ionisation potential of Zn (9.4 V) is significantly higher than algae solution give better agreement between the ICP-MS those of the analytes used in their study and, therefore, would result and the consensus value for Zn, but lead to an overestim- be expected to be more sensitive to changes in the available ation by about 17% whereas 1+9 dilution of the digest energy of the plasma than easily ionised analytes.This could solution provides results which are in good agreement with explain why a correct result was obtained for NIST SRM the consensus value. If none of the nitric acid used for the 1645 River Sediment in this work since the total dilution factor digestion process was lost or consumed, the maximum acid (2500) is large enough to negate any eVect of nitrogen on Zn. content in the final solution would be 10%. The observation Indeed, the residual acid concentration was so low that the that simple dilution of the digest solutions gave higher apparent solutions were made up to 1% HNO3, to avoid the risk of Zn recoveries provided strong circumstantial evidence that loss of analytes from solution; consequently, the acid strength acid strength was involved in the suppression.In one digestion matched that of the external calibration. set, the acid content was reduced to 50 and 20% of its original A similar eVect of nitric acid concentration on sensitivity level and the results obtained (versus external aqueous stan- was observed for As in biological matrices11 which resulted in dards containing 1% v/v HNO3) were 63.4 and 93.3 mg g-1, erroneously high As values.Both Zn and As have relatively respectively, which is in agreement with the trend depicted in high first ionisation potentials (9.4 and 9.8 V, respectively) Fig. 4. However, for the latter digest, the solution was col- which may make them more sensitive to ‘spatial eVects’ within oured, indicating that the digestion process was not complete.the plasma. Confirmation that the eVect described in this The influence of acid strength was studied by monitoring the paper is not limited to ICP-MS was provided by Marichy sensitivity for a suite of analytes at constant concentration et al.,12 who reported a #15% decrease in optical emission (50 mg l-1) over a range of acid strengths. Clear evidence is intensities for Co II as the acid strength was increased from 0 shown (Fig. 4) which demonstrates significant non- to 1% HNO3; Co has a first ionisation potential of 7.9 V. spectroscopic suppression of Zn at nitric acid concentration From the work presented here, it is clear that Zn results levels where other analytes (Sc, Y and In) are largely unaVec- will only be reliable if the concentration of nitric acid is kept ted, rendering internal standardisation ineVective for Zn. AAS as low as possible and held constant.A similar eVect is results do not show evidence of any influence of nitric acid on predicted for analytes with high first ionisation potentials. Zn recovery from aqueous solutions and provide accurate results for the determination of Zn in IAEA 392 from the References same solutions. Stewart and Olesik7 studied the influence of nitric acid 1 Preliminary Statistical Evaluation of the Intercomparison Run concentration and nebuliser gas flows on aerosol transport IAEA 390, IAEA/AL/101 P, 1998, International Atomic Energy rates into an ICP and noted a sharp decrease in analyte signal Agency, Vienna, Austria.as the acid concentration increased from 0 to 2%, depending 2 Programme Performance Assessment External Reviewers’ Report on the Consultants’ Meeting on Analytical Quality Control Services, upon operating conditions. We observed this decline in signal IAEA/AL/114, 1998, International Atomic Energy Agency, between 0 and 2% HNO3 for Zn, but not for Sc, Y or In.Vienna, Austria. However, they report that the eVect is not so pronounced 3 Report of the Consultant’s Meeting on Traceability of IAEA-AQCS under ‘robust plasma conditions7’ which correspond to the Reference Materials to SI-Units, IAEA/AL/105, 1997, conditions we were using. In an earlier paper,8 they note that International Atomic Energy Agency, Vienna, Austria. analyte signal depression ( lasting from 5 to 25 min) may occur 4 ILAC Requirements for Accreditation of Certifiers of Reference as acid strength is increased, until the steady state is Materials (draft), 1997, International Standards Organisation, Geneva, Switzerland.re-established. This is attributed to competing evaporational 5 Report on the Consultants’ Meeting on Good Analytical Practice in processes between acid aerosol droplets and coalesced droplets the Analysis of IAEA Materials for Certification, 1996, in on the spray chamber walls and presumably aVects all analytes preparation. in a similar manner. In our work, the original digest solutions 6 M. J. Campbell, G. Vermeir, Ph. Quevauviller and R. Dams, are produced in a batch process and have similar acidity levels; J. Anal. At. Spectrom., 1992, 7, 617. they incorporate internal standards which should minimise 7 I. I. Stewart and J. W. Olesik, J. Anal. At. Spectrom., 1998, 13, 1249. this eVect, so it is unlikely that it is responsible for the Zn 8 I. I. Stewart and J. W. Olesik, J. Anal. At. Spectrom., 1998,13, 843. suppression observed. 9 J. Wang, E. H. Evans and J. Caruso, J. Anal. At. Spectrom., 1992, The observed behaviour of Zn may be due to the high local 7, 929. concentration of nitrogen atoms in the plasma (due to the 10 F. Vanhaecke, R. Dams and C. Vandecasteele, J. Anal. At. dissociation of the nitric acid). Wang et al.9 studied the eVects Spectrom., 1993, 8, 433. of introducing nitrogen into an argon ICP used for mass 11 M. J. Campbell, C. Demesmay and M. Olle�, J. Anal. At. spectrometry. They observed that when 3% of the nebuliser Spectrom., 1994, 9, 1379. 12 M. Marichy, M. Mermet and J. M. Mermet, Spectrochim. Acta, gas flow was made up of nitrogen, the central analyte channel Part B, 1990, 45, 1195. in the plasma broadened, the optimum M+ signal shifted to lower nebuliser gas flow rate, certain polyatomic species were significantly reduced and a slight decline was observed for the Paper 9/01639B 1316 J. Anal. At. Spectrom., 1999, 14, 1313–13