Isotope Dilution as a Calibration Method for Solid Sampling Electrothermal Vaporization Inductively Coupled Plasma Mass Spectrometry FRANK VANHAECKE, SYLVIE BOONEN, LUC MOENS AND RICHARD DAMS L aboratory of Analytical Chemistry, Ghent University, Institute for Nuclear Sciences, Proeftuinstraat 86, B-9000 Ghent, Belgium The present paper reports on the use of isotope dilution as a continuous improvements in instrumentation LA-ICP-MS has method of calibration for solid sampling ETV-ICP-MS.The become a well established solid sampling technique, predompossibilities and limitations of this calibration strategy were inantly used for geological applications,12–16 but also for the evaluated by determining the Cd or Se content in solid CRMs analysis of metals,17–19 glasses20,21 and plastics22. of different origin. It was shown that since isotope ratios are In addition to LA-ICP-MS,ETV-ICP-MS also offers interesonly slightly affected or not affected at all by (i) matrix ting possibilities for the direct determination of trace and ultraeffects, (ii) signal drift and instrument instability and trace elements in solid samples.Both Voellkopf et al.23 and (iii) variations in the vaporization and/or transport efficiency, Gre�goire et al.24 have reported on the analysis of solid samples isotope dilution allows accurate analyses to be carried out using ETV-ICP-MS, but both research groups preferred (mean deviation between solid sampling ETV-ICP-MS results the analysis of slurries rather than of dry solid samples.and certified values <10%). The precision attainable is Wang et al.25 explored the feasibility of ‘real’ solid sampling, determined by the sample homogeneity and is hence but encountered some difficulties that hampered the practical comparable to that obtained using other calibration techniques, use of solid sampling ETV-ICP-MS. Argentine and Barnes,26 such as (i) external calibration with either a solid standard or however, successfully used ETV-ICP-MS for the determination an aqueous standard solution or (ii) standard additions.An of non-volatile impurities in semiconductor-grade organometimportant advantage of isotope dilution over the allic materials and process chemicals. Finally, earlier work aforementioned calibration techniques for solid sampling ETV- carried out in our laboratory, concerning the accurate determi- ICP-MS, however, is that the use of an elemental internal nation of As and Se in CRMs of both plant and environmental standard is no longer required.For some materials, accurate origin using solid sampling ETV-ICP-MS, has been reported analytical results could not be obtained as at least one of the in earlier publications.27–29 isotopes involved was observed to be subject to spectral As for all solid sampling techniques, also for solid sampling interference. The use of several parameters allowing spectral ETV-ICP-MS, accurate calibration is not obvious.Wang interferences to be detected is discussed. Finally, solid et al.25 concluded that in order to obtain optimum accuracy, sampling ETV-ICP-MS was used for the determination of the external calibration with a CRM with a composition as similar Cd content in tobacco as a ‘real-life’ sample and the results as possible to that of the sample should be used. The problem obtained using isotope dilution and single standard addition for of calibration in solid sampling ETV-ICP-MS, however, has calibration were compared with one another and with the been studied extensively in our laboratory and it has been result obtained (after taking the sample into solution) using demonstrated that different strategies allow accurate analyt- pneumatic nebulization ICP-MS.ical results.27 Keywords: Inductively coupled plasma mass spectrometry; The application of a CRM with a similar matrix composition electrothermal vaporization; solid sampling; isotope dilution ; and analyte content as a solid standard indeed permitted calibration excellent results to be obtained, although only provided that a suitable elemental internal standard (added to both the samples and the standard) was used.Although the results Since its commercial introduction in 1983, ICP-MS has aroused obtained for As and Se in solid materials of different origin great interest, and during the past decade it has proven its using Sb as an internal standard27,29 clearly illustrate the utility for the determination of trace and ultra-trace elements potential of this calibration strategy, the latter also shows in a variety of matrices.In its standard configuration, ICP-MS important drawbacks. A consequence of this approach is that is mainly intended for the analysis of aqueous samples, solid sampling ETV-ICP-MS cannot be considered to be an although at present there is an increasing interest in the direct independent method, as one has to rely on certified values, analysis of solid samples.based on results obtained by other analytical techniques. Direct analysis of solid samples is of course of great impor- Moreover, the uncertainty of a certified analyte content is tance for materials that cannot or only with great difficulty be always much larger than the uncertainty of the concentration brought into solution. Moreover, in general, solid sampling in an aqueous standard solution. In addition, some information limits the necessary amount of often laborious and time- on the sample matrix is required and, finally, suitable reference consuming sample pre-treatment, leading to a reduced risk of materials are not available for all sample types.contamination and/or analyte losses, and as samples are Although when using an appropriate elemental internal analysed without dilution, also lower LOD are to be expected. standard, even external calibration with aqueous standard The possibilities of LA for the introduction of solid samples solutions was observed to provide acceptable results,27 single in ICP-MS have been reported extensively in the literature.1–5 standard addition was considered to be the most straight- Next to the general advantages of solid sampling mentioned forward and practicable method.27,29 Since matrix-induced above, LA-ICP-MS also offers the possibility for both lateral signal suppression was established to be strongly dependent and in-depth profiling of solid samples of different origin.6–11 LA-ICP-MS is somewhat expensive, but as a result of on the sample mass,27,29,30 accurate results could only be Journal of Analytical Atomic Spectrometry, February 1997, Vol. 12 (125–130) 125Table 1 Operating conditions for the ETV system and the ICP mass obtained with the latter calibration method when using an spectrometer appropriate internal standard or taking the sample masses as close together as is practically possible.The latter approach is ET V system— not only fairly wearisome and time-consuming, but the blank Type SM-30, Gru�n Analytische Mess- is also less accurately corrected for and there is no correction Systeme for signal drift, instrument instability and fluctuations in the Temperature program Multi-step temperature program vaporization and transport processes, such that the first consisting of: approach is to be preferred. 1. ‘Drying’ step (30 s at#120 °C) 2. Boosting step (1 s), during which It is obvious from the above-mentioned results that the use the power applied is 25% of the of an elemental internal standard was observed to be advanta- maximum power, allowing the geous (standard addition methods) or even imperative (external heating rate of the furnace to be calibration methods), depending on the calibration method increased at moderate temperatures used.However, the selection of a suitable elemental internal in order to obtain a stable ashing standard is not obvious.Most importantly, an appropriate temperature rapidly 3. Ashing step (30 s at#200–250 °C) internal standard should show an analogous (furnace) chemis- 4. Intermediate step (12 s at the try as the analyte element(s) and should of course only be ashing temperature), to switch the present at negligible levels in the samples under consideration.27 valve to the &lsquosuring’ position Although only of secondary importance with solid sampling and allow the plasma to stabilize ETV-ICP-MS, in general, internal standardization for correc- 5.Vaporization step (15 s) tion of matrix-induced signal suppression or enhancement and 6. Intermediate step, to switch the valve to the ‘venting’ position (after for improving the precision has been observed to be most the end of the measurement) efficient if the mass number of the internal standard is chosen 7. Cleaning step (2×3 s at#2700 °C) fairly close to that of the analyte element(s).30–32 Since a suitable internal standard should fulfil all three conditions ICP mass spectrometer— simultaneously, it is clear that selection of such an internal Type Perkin-Elmer SCIEX ELAN 5000 standard sometimes poses an unsurmountable problem.Rf power 1300 W As a result, more recently efforts have been made to circum- Sampling depth 10 mm vent the necessity of using such an elemental internal standard. Aerosol carrier gas flow rate 0.750 1 min-1 In an earlier publication,33 it was demonstrated that at least Intermediate gas flow rate 1.0 1 min-1 Outer gas flow rate 12 1 min-1 in some cases (determination of As in a solid CRM of plant Lens voltages Tuned using pneumatic nebulization; origin), the argon dimer (Ar2+) could be used as an internal no further tuning required when standard.Since isotope dilution (ID) as a calibration method switching to solid sampling ETV- only involves determination of an isotope ratio in the sample, ICP-MS the tracer and a mixture of both, matrix effects, signal drift Sampling cone Nickel; 1.0 mm orifice diameter and instrument instability and fluctuations in the vaporization Skimmer cone Nickel; 0.75 mm orifice diameter and transport processes should not have an adverse effect on the results obtained,34 and application of an elemental internal standard would hence no longer be required.The feasibility of Germany) via a 10 mm id silicone rubber tubing. In order to the latter calibration strategy for solid sampling ETV-ICP-MS (i) reduce the amount of deposition of evaporated sample and was investigated and the results obtained are reported in the furnace material on the torch, the interface and the lens stack present paper.and (ii ) avoid degradation of the interface pump oil, a threeway valve was used to vent vapours generated during the drying, ashing and cleaning steps. Operating conditions are EXPERIMENTAL summarized in Table 1. Instrumentation Measurements The ETV system used is a commercially available graphite furnace of the boat-in-tube type (SM-30, Gru�n Analytische Measurement parameters are summarized in Table 2.Fast Mess-Systeme, Ehringhausen, Germany). Although this device hopping between the nuclides monitored is necessary to obtain was originally designed for solid sampling Zeeman-effect AAS, a representative image of the corresponding signal profiles. On some simple modifications (described elsewhere35) sufficed to the other hand, an efficient use of the total measuring time make it compatible for use with both ICP-AES and ICP-MS.requires a high ratio of actual measuring time to mass spec- Graphite sample holders (‘boats’) can be easily and reproduci- trometer settling time and hence, relatively large dwell times. bly loaded into the cylindrical graphite furnace with the aid of As a compromise, 30 ms was used as the dwell time per a pair of tweezers, sliding on a rail, which is rigidly mounted measuring point.36 The three-way valve was switched manually in front of the furnace.After loading the sample, one end of to the ‘measuring position’ 12 s before the start of the vaporiz- the furnace is closed using a shutter kept in position by a ation stage, while the measurement itself was started 2 s before catch-spring. During operation, a flow of argon, the flow rate the beginning of the vaporization stage (Table 1). Each determi- of which is controlled by a mass-flow controller (Model 5876, nation consisted of three measurements of the blank (empty Brooks Instruments, Veenendaal, The Netherlands) is swept through the furnace, transporting the sample aerosol, formed Table 2 Measurement parameters by condensation of the vaporized sample, into the central channel of the ICP.The multi-step temperature program of Dwell time 30 ms the furnace (Table 1) is controlled by a computer program Scanning mode Peak hop transient Sweeps per reading 1 developed in-house, while the temperature can be monitored Readings per replicate 800 divided by the number of nuclides moni- using an optical pyrometer (PY20, Gru�n Optik, Ehringhausen, tored Germany), specially designed for use with this type of ETV Points per spectral peak 1 system.The ETV system was coupled to a Perkin-Elmer Total measurement time ca. 30 s SCIEX ELAN 5000 ICP mass spectrometer (U� berlingen, 126 Journal of Analytical Atomic Spectrometry, February 1997, Vol. 12boat) and five measurements of the isotope ratio under con- accurately the number of atoms present of both isotopes involved) of the tracer obtained as described above was sideration in the sample, the tracer and the mixtures (consisting of a given amount of sample, to which an appropriate amount determined by inverse ID (the tracer was considered as the sampleand the standard of natural composition was considered of tracer was added), respectively. In some instances, the Omnirange device was used.This option permits a selective as the ‘tracer’) using pneumatic nebulization ICP-MS. Finally, for use with solid sampling ETV-ICP-MS, the tracers obtained and reproducible reduction of the sensitivity of the mass spectrometer by varying the ion transmission efficiency at the were diluted to an appropriate concentration level, depending on the analyte concentration in the sample. exact time that a given mass-to-charge ratio is being measured and hence allows measurement at higher concentration levels. As reported on in an earlier publication,27 identical behaviour of an analyte in the sample and in the standard can only be guaranteed if the standard solution (in this instance, the Samples tracer) is inserted into the graphite sample holder (10 ml using a micropipette) and dried under an IR lamp before loading of In order to evaluate the accuracy of the results obtainable using ID as a calibration technique in solid sampling ETV- the solid sample (typically 1–2 mg).In all cases, the amount of tracer added to the solid sample was selected such that the ICP-MS, efforts were made to determine Cd or Se in a number of CRMs of biological or environmental origin. For the isotope ratio for the mixture was close to the average of the corresponding isotope ratios in the sample and the tracer. determination of Cd, Aquatic Plant (BCR CRM 060), Light Sandy Soil (BCR CRM 142) and Sewage Sludge of Industrial Finally, for the determination of Cd in the tobacco sample solutions obtained as described above, an external calibration Origin (BCR CRM 146) were selected as samples.For Se, Estuarine Sediment (BCR CRM 277), Sea Lettuce (BCR CRM solution was prepared by appropriate dilution of a commercially available Cd standard solution (1 g l-1); In was added 279) and Wheat Flour (NIST SRM 1567A) were used. No sample pre-treatment except for homogenization (by shaking) as an internal standard.and weighing was required before analysis. The moisture content of these materials was determined according to the Calculations ‘drying instructions’, given in the certificates, and correspond- All calculations were carried out according to the formulae ingly corrected for. presented by Longerich.37 In solid sampling ETV-ICP-MS, After having demonstrated the possibilities and limitations calculation of the standard deviation of the results obtained of ID as a calibration method in solid sampling ETV-ICP-MS, is, however, somewhat more complicated than for aqueous Cd was determined in Johnson heavy cigarette-tobacco as a solutions, since for each replicate measurement of a mixture ‘real-life’ sample. In order to evaluate the accuracy of this of (solid) sample and trace amount of sample taken was analysis, tobacco samples were also taken into solution and of course slightly different.However, since the uncertainty due the solutions obtained were analysed using pneumatic nebuliz- to inhomogeneity of the samples clearly exceeded other contri- ation ICP-MS with external calibration as a calibration tech- butions (e.g., the uncertainty in the isotope ratio measurement nique.To #1.2 g of tobacco, 5 ml of 14 mol l-1 HNO3 and for the sample and the tracer), the analyte concentration was 1 ml of 10 mol l-1 HCl (both purified by sub-boiling distil- calculated for each replicate measurement of a mixture of lation) were added and the mixtures obtained were subjected sample and tracer, respectively, and the standard deviation for to a step-wise temperature program (15 min at 80°C, 15 min the n individual results was calculated subsequently and was at 110 °C and 2×90 min at 240 °C) in a high-pressure asher considered as the standard deviation on the final result.(HPA, Ku� rner, Rosenheim, Germany). Since a fine white precipitate was observed in the digestion vessels, HF was added and the sample was subjected a second time to the RESULTS AND DISCUSSION multi-step temperature program.The solutions obtained were Preliminary Study diluted to 25 ml and In was added as an internal standard to correct for matrix effects, signal drift and instrument instability. During a preliminary study, the precision of the isotope ratio measurements and the mass discrimination observed using ETV-ICP-MS were compared with the corresponding values Standards obtained using pneumatic nebulization ICP-MS.For series of five successive measurements of a 100 mg l-1 For the determination of Cd, the tracer available for these experiments was enriched in 110Cd (93.63%) and was purchased Cd standard solution with pneumatic nebulization ICP-MS, the average RSD for the 110Cd5113Cd isotope ratio was from Campro Scientific (Veenendaal, The Netherlands). About 5 mg of this material, consisting of CdO, were taken into observed to be 0.20%, while the mass discrimination factor K (defined as the ratio of the ‘measured 110Cd5113Cd isotope solution by adding #2 ml of 14 mol l-1 HNO3 (purified by sub-boiling distillation).A stock solution (5 mg l-1) was ratio to the true value’) was established to be 0.94. For a series of five successive ETV-ICP-MS measurements of aliquots of obtained by dilution with Millipore Milli-Q water to a volume of 1 l. For the determination of Se, a standard enriched in 82Se 10 ml of a 200 mg l-1 Cd tracer solution, the RSD for the isotope ratio was observed to vary from 0.77 to 2.2%, with an (Se, 92.2% in 82Se) from Euriso-top (Saint Aubin, France) was used.Approximately 10 mg of the metallic Se standard were average value of 1.5%. Also for ETV-ICP-MS, the average mass discrimination factor was observed to be 0.94. taken into solution using 30 ml 14 mol l-1 HNO3 and subsequently diluted to 1 l using Millipore Milli-Q water. In order For a 100 mg l-1 Se standard solution, the 82Se577Se isotope ratio could be measured with an RSD of 0.75% using pneu- to allow accurate determination (the signal intensity for both isotopes involved significantly exceeding the blank level) of matic nebulization ICP-MS (n=5).With pneumatic nebulization ICP-MS, the mass discrimination factor (defined as the the isotope ratios (110Cd5113Cd and 77Se582Se) in the tracers using ETV-ICP-MS, these standards were ‘diluted’ with a ratio of the ‘measured 82Se577Se isotope ratio to the true value’) was established to be 1.14.For ETV-ICP-MS on the standard of natural composition. For Cd, the tracer finally used was obtained by mixing ‘enriched Cd’ with ‘natural Cd’ other hand, the RSD for five successive measurements of a 10 ml aliquot of a 100 mg l-1 Se tracer solution varied from in a 151 (m/m) ratio, giving a 110Cd5113Cd ratio of 6.37 instead of 195.1, while for Se, two portions of ‘enriched Se’ were mixed 0.93 to 1.8%, with an average value of 1.4%.The average value for the mass discrimination factor was found to be with three portions of ‘natural Se’ (m/m) giving an 82Se577Se ratio of 8.65 instead of 144.3. The concentration (or more 1.16. By means of a t-test (95% confidence level), it was Journal of Analytical Atomic Spectrometry, February 1997, Vol. 12 127demonstrated that this value did not differ significantly from instrument instability and (iii ) variations in the vaporization and/or transport efficiency,34 it could be expected that ID as the value obtained using pneumatic nebulization ICP-MS.The precision for the isotope ratio measurements attainable a calibration technique allows accurate analysis with solid sampling ETV-ICP-MS, while the application of an elemental using ETV-ICP-MS is clearly somewhat poorer than that obtained using pneumatic nebulization ICP-MS. This is prob- internal standard is no longer required. Of course, a prerequisite for accurate analysis by ID is the ably to be attributed to the transient nature of the signals involved.The precision of the isotope ratios obtained, however, availability of two ‘free’ isotopes, i.e., two isotopes which are not spectrally interfered by isobaric M+ ions, doubly charged were assessed to be sufficiently good to warrant an investigation of the possibilities of using ID as a calibration method in solid ions, oxide or other polyatomic ions. Therefore, in all cases, it should be carefully checked that this condition is fulfilled.This sampling ETV-ICP-MS. is especially the case in solid sampling ETV-ICP-MS, since as a result of the ‘nature’ of the technique (e.g., dry plasma Use of Parameters Allowing Spectral Interferences to be conditions, introduction of relatively large amounts of C into Detected the ICP), unforeseen spectral interferences can occur. Several ‘tools’ can, however, be used to check fulfilment of All relevant figures and analytical results obtained for the this condition.Firstly, the experimentally determined Cd or CRMs investigated are summarized in Tables 3–5. In order to Se isotope ratio can be compared with the value expected on evaluate these results thoroughly, several parameters were used the basis of the natural isotopic abundances of the isotopes as indicators for spectral interferences. These ‘tools’ will be involved. The difference between these values should be in introduced in the following paragraphs and thereafter, the agreement with a mass bias in favour of the higher mass ion.results obtained will be systematically discussed. In order to determine if the magnitude of the difference between As has already been discussed at the beginning of this paper, the experimentally determined isotope ratio and the true value when using external calibration or standard additions the can still be considered to be solely the result of this mass selection of a suitable (elemental) internal standard in solid discrimination effect, a mass discrimination factor (defined as sampling ETV-ICP-MS can sometimes pose an unsurmountthe ratio of the ‘measured isotope ratio to the true value’) can able problem. Since isotope ratios are not affected at all, or only slightly affected by (i ) matrix effects, (ii ) signal drift and be calculated for both the sample (Ks) and the tracer (Kt).A Table 3 Results for the determination of the 110Cd5113Cd and 82Se577Se isotope ratios in the solid samples investigated 110Cd5113Cd ratio (uncorrected for mass discrimination effects)— Material investigated 110Cd:113Cd* RSD (%) Sewage Sludge Industrial Origin (BCR CRM 146) 0.943 0.45 Aquatic Plant (BCR CRM 060) 1.013 3.8 Light Sandy Soil (BCR CRM 142) 1.034 2.9 Tobacco 1.005 0.78 82Se577Se ratio (uncorrected for mass discrimination effects)— Material investigated 82Se577Se† RSD (%) Wheat Flour (NIST SRM 1567A) 1.34 0.79 Estuarine Sediment (BCR CRM 277) 1.20 0.64 Sea Lettuce (BCR CRM 279) 1.58 12 *Value expected on the basis of the natural isotopic abundances of the isotopes involved, 1.022.†Value expected on the basis othe natural isotopic abundances of the isotopes involved, 1.144. Table 4 Mass discrimination factors obtained for the sample (Ks) and the tracer (Kt) and ratio Ks/Kt observed during the determination of Cd or Se using solid sampling ETV-ICP-MS CRM Ks* Kt Ks/Kt 110Cd5113Cd— Sewage Sludge Industrial Origin (BCR CRM 146) 0.92 0.92 1.00 Aquatic Plant (BCR CRM 060) 0.99 0.94 1.05 Light Sandy Soil (BCR CRM 142) 1.01 0.90 1.12 Tobacco 0.98 0.94 1.04 82Se:77Se— Wheat Flour (NIST SRM 1567A) 1.17 1.17 1.00 Estuarine Sediment (BCR CRM 277) 1.05 1.08 0.97 Sea Lettuce (BCR CRM 279) 1.38 1.24 1.11 *Values obtained for Ks obtained using pneumatic nebulization ICP-MS using a Cd or a Se standard solution, 0.94 and 1.16, respectively. Table 5 Solid sampling ETV-ICP-MS results (mg g-1) for the Cd or Se content in the materials analysed; calibration by ID Material investigated This work, mean±95% CL Certified value±95% CL Cd— Sewage Sludge Industrial Origin (BCR CRM 146) 80±10 77.7±2.6 Aquatic Plant (BCR CRM 060) 2.184±0.079 2.20±0.10 Se— Wheat Flour (NIST SRM 1567A) 1.204±0.031 1.102±0.088 Estuarine Sediment (BCR CRM 277) 1.664±0.093 2.04±0.18 128 Journal of Analytical Atomic Spectrometry, February 1997, Vol. 12substantial difference between the mass discrimination factor abundances of the isotopes involved by #38% (Table 3).(ii) This would lead to a mass discrimination factor for the for the sample and the tracer, respectively, indicates the presence of a spectral interference. Finally, a large RSD for sample of 1.38 and although for this determination the mass discrimination factor for the tracer was also seen to be fairly the experimentally observed isotope ratio can also be indicative of the presence of a spectral overlap. The latter is especially high (Table 4), the discrimination factor for the sample exceeded that for the tracer by >10%.Finally, (iii ) also the expected in the case of overlap with C-containing polyatomic ions, since the amount of C introduced into the plasma can large RSD of the 82Se577Se isotope ratio results is indicative of the occurrence of a spectral interference. Hence, no further differ strongly from one firing of the furnace to another. It should be noted that the first two ‘tools’ imply that it is attempts were made to obtain an analysis result.assumed that the analyte in the sample is of natural isotopic composition. However, this is also the case if external cali- Determination of Cd in Tobacco bration or standard additions are used and only one nuclide is monitored. Of course, care has to be taken for elements for For the tobacco sample investigated, the 110Cd5113Cd signal which it is known that the isotopic composition could change ratio could be measured with an RSD of only 0.78% (Table 3).from one sample to another, such as Pb. A 4% difference could be established between the mass discrimination factors for the sample and the tracer, respectively (Table 4). This difference is comparable to or even somewhat Determination of Cd in CRMs smaller than that observed for Aquatic Plant (BCR CRM For Sewage Sludge of Industrial Origin (BCR CRM 146), an 060), and for the latter CRM an excellent result was obtained. excellent RSD (Table 3) was obtained for the 110Cd5113Cd Hence the Cd content was determined using solid sampling isotope ratio.Since the mass discrimination factors for the ETV-ICP-MS using both ID and single standard addition sample and the tracer were also in excellent agreement (nuclide monitored, 110Cd) for calibration and using pneumatic (Table 4), it could be assumed that both Cd isotopes involved nebulization ICP-MS, after taking the sample into solution, as were interference-free and hence, the result obtained is in described under Experimental.All results (corrected for the excellent agreement with the certified value (Table 5). Although moisture content) are summarized in Table 6. The result the RSD of the 110Cd5113Cd ratio was observed to be <1%, obtained using ID compares excellently with that obtained the 95% confidence limit of the final result was seen to increase using standard addition (both using solid sampling ETV- readily to >10%.This is to be attributed to a certain level of ICP-MS). The deviation between these solid sampling results inhomogeneity of the analyte element in the CRM under and the pneumatic nebulization ICP-MS result is somewhat investigation. This inhomogeneity is of course not corrected larger (#12%), although no significant difference could be for using ID and is the determining factor for the precision of established. the results obtained. Hence, the level of precision attainable is comparable to that obtained when using external calibration or standard addition as a means of calibration.27 For Aquatic CONCLUSIONS Plant on the other hand, the RSD obtained for the 110Cd5113Cd The results obtained for the determination of Cd or Se in ratio was significantly higher (Table 4) and there was a 5% CRMs of different origin indicate that ID allows accurate difference between the mass discrimination factors for the analyses to be carried out (the mean deviation between the sample and the tracer.Nevertheless, the result obtained for solid sampling ETV-ICP-MS results and corresponding certi- the Cd content in Aquatic Plant is in excellent agreement with fied values is<10%). In addition, this calibration strategy can the certified value. For Light Sandy Soil, however, no accurate be applied for all elements that show (at least) two interference- result could be obtained. This is probably to be attributed to free isotopes. The precision attainable is determined by the a spectral interference at m/z 110 since the experimentally sample homogeneity and is hence comparable to that obtained observed 110Cd5113Cd ratio (Table 3) exceeds the theoretical using other calibration techniques, such as external calibration value (which is in contradiction with a mass bias in favour of using either a solid standard or an aqueous standard solution the higher mass ion, generally observed in ICP-MS), while the or standard addition. The most important advantages of ID mass discrimination factor for the sample exceeds that for the over the forementioned calibration techniques for use with tracer by more than 10% (Table 4).solid sampling ETV-ICP-MS are: (i) the use of an elemental internal standard is no longer required and (ii) the possibility Determination of Se in CRMs of erroneous (biased) results is reduced as a result of the systematic check on the presence of spectral interferences.For the determination of Se in Wheat Flour (NIST SRM Since the availability of tracers is continuously increasing and 1567A), both the low RSD for 82Se577Se (Table 3) and the the price of these materials correspondingly decreasing, ID can excellent agreement between the mass discrimination factors be considered to be a viable alternative to other calibration for the sample and the tracer (Table 4), indicated the absence techniques, and its use could lead to an extension of the of important spectral interferences on the isotopes involved.application range of solid sampling ETV-ICP-MS. Hence, the result obtained compares favourably with the certified value (Table 5). Although for Estuarine Sediment F. V. is a Senior Research Assistant of the Belgian National (BCR CRM 277) an excellent RSD for the 82Se577Se ratio Fund for Scientific Research. could be obtained (Table 3), and only a 3% difference between the mass discrimination factor for the sample and that for the tracer (Table 4) could be established, the deviation between the solid sampling ETV-ICP-MS result and the certified value Table 6 Results (s) in mg g-1 for the determination of Cd in tobacco.is larger than for Wheat Flour (Table 5). This could possibly be attributed to a 40Ar37Cl+ interference on the 77Se+ signal, Solid sampling ETV-ICP-MS PN-ICP-MS explaining the lower Ks factor found (Table 4). Finally, for Sea ID, 1.39 (0.32)* External calibration, 1.23 (0.03)† Lettuce, several factors indicated the occurrence of a spectral Standard addition, 1.36 (0.19)* interference on 82Se. (i) The experimentally determined 82Se577Se ratio (uncorrected for mass discrimination effects) *n=5.†n=3. exceeds the value expected on the basis of the natural isotopic Journal of Analytical Atomic Spectrometry, February 1997, Vol. 12 12922 Marshall, J., Franks, J., Abell, I., and Tye, C., J. Anal. At. REFERENCES Spectrom., 1991, 6, 145. 1 Gray, A.L., Analyst, 1985, 110, 551. 23 Voellkopf, U., Paul, M., and Denoyer, E. R., Fresenius’ J. Anal. 2 Arrowsmith, P., Anal. Chem., 1987, 59, 1437. Chem., 1992, 342, 917. 3 Denoyer, E. R., Fredeen, K. J., and Hager, J. W., Anal. Chem., 24 Gre�goire, D. C., Miller-Ihli, N. J., and Sturgeon, R. E., J. Anal. 1991, 63, 445A. At. Spectrom., 1994, 9, 605. 4 van de Weijer, P., Baeten, W. L. M., Bekkers, M. H. J., and 25 Wang, J., Carey, J. M., and Caruso, J. A., Spectrochim. Acta, Part Vullings, P.J. M. G., J. Anal. At. Spectrom., 1992, 7, 599. B, 1994, 49, 193. 5 Paul, M., At. Spectrosc., 1994, 15, 21. 26 Argentine, M. D., and Barnes, R. M., J. Anal. At. Spectrom., 1994, 6 Imai, N., Anal. Chim. Acta, 1992, 269, 263. 9, 1371. 7 Pearce, N. J. G., Perkins, W. T., Abell, I., Duller, G. A. T., and 27 Vanhaecke, F., Boonen, S., Moens, L., and Dams, R., J. Anal. At. Fuge, R., J. Anal. At. Spectrom., 1992, 7, 53. Spectrom., 1995, 10, 81. 8 Chenery, S., and Cook, J. M., J. Anal. At. Spectrom., 1993, 8, 299. 28 Moens, L., Verrept, P., Boonen, S., Vanhaecke, F., and Dams, R., 9 Richner, P., and Evans, D., At. Spectrosc., 1993, 14, 157. Spectrochim. Acta, Part B, 1995, 50, 463. 10 Shibata, Y., Yoshinga, J., and Morita, M., Anal. Sci., 1993, 9, 129. 29 Boonen, S., Vanhaecke, F., Moens, L., and Dams, R., Spectrochim. 11 Ulens, K., Moens, L., Dams, R., Van Winckel, S., and Acta, Part B, 1996, 81, 271. Vandevelde, L., J. Anal. At. Spectrom., 1994, 9, 1243. 30 Thompson, J. J., and Houk, R. S., Appl. Spectrosc., 1987, 41, 801. 12 Broadhead, M., Broadhead, R., and Hager, J. W., At. Spectrosc., 31 Doherty, W., Spectrochim. Acta, Part B, 1989, 44, 263. 1990, 11, 205. 32 Vanhaecke, F., Vanhoe, H., Dams, R., and Vandecasteele, C., 13 Imai, N., Anal. Sci., 1990, 6, 389. T alanta, 1992, 39, 737. 14 Imai, N., Anal. Chim. Acta, 1990, 235, 381. 33 Vanhaecke, F., Galba�cs, G., Boonen, S., Moens, L., and Dams, R., 15 Perkins, W. T., Fuge, R., and Pearce, N. J. G., J. Anal. At. J. Anal. At. Spectrom., 1995, 10, 1047. Spectrom., 1991, 6, 445. 34 McLaren, J. W., Beauchemin, D., and Berman, S. S., Anal. Chem., 16 Perkins, W. T., Pearce, N. J. G., and Jeffries, T. E., Geochim. 1987, 59, 610. Cosmochim. Acta, 1993, 57, 475. 17 Yasuhara, H., Okano, T., and Matsumura, Y., Analyst, 1992, 35 Verrept, P., Dams, R., and Kurfu�rst, U., Fresenius’ J. Anal. Chem., 117, 395. 1993, 346, 1035. 18 Kogan, V. V., Hinds, M. W., and Ramendik, G. I., Spectrochim. 36 Denoyer, E. R., At. Spectrosc., 1994, 15, 7. Acta, Part B, 1994, 49, 333. 37 Longerich, H. P., At. Spectrosc., 1989, 10, 112. 19 Watling, R. J., Herbert, H. K., Delev, D., and Abell, I. D., Spectrochim. Acta, Part B, 1994, 49, 205. Paper 6/04133G 20 Imbert, J. L., and Telouk, P., Mikrochim. Acta, 1993, 110, 151. Received June 12, 1996 21 Moenke-Blankenburg, L., Schumann, T., Gu�nther, D., Kuss, H.-M., and Paul, M., J. Anal. At. Spectrom., 1992, 7, 251. Accepted August 7, 1996 130 Journal of Analytical Atomic Spectrometry, February 1997,