Use of the AT,’ Signal as a Diagnostic Tool in Solid Sampling Electrothermal Vaporization Inductively Coupled Plasma Mass Spectrometry Journal of Analytical Atomic Spectrometry FRANK VANHAECKE,* GABOR GALBACS,~ SYLVIE BOONEN LUC MOENS AND RICHARD DAMS Laboratory of Analytical Chemistry Ghent University Institute for Nuclear Sciences Proeftuinstraat 86 B-9000 Ghent Belgium The utility of the Ar2+ signal (at mass-to-charge ratio rn/z=80) as a diagnostic tool in solid sampling electrothermal vaporization inductively coupled plasma mass spectrometry (ETV-ICP-MS) is reported. Simultaneous monitoring of the argon dimer signal with the signal(s) of the analyte element(s) indicated that non-spectral interferences caused by matrix components co-volatilizing with the analyte element(s) can strongly affect the analyte signal profiles in solid sampling ETV-ICP-MS of samples of biological or environmental origin.This observation led to a more profound understanding of why for a given matrix the signal profiles strongly differ from one element to another and why for a given element the signal profile is seen to be strongly dependent on the matrix. These matrix effects were also observed to cause a curvature in the sample mass response curves (analyte signal intensity as a function of sample mass). It is shown that at least in some instances the use of the Ar2+ signal as an internal standard allows (i) this non-linearity to be corrected for and (ii) accurate analysis results to be obtained. Finally it is demonstrated that simultaneous registration of the argon dimer and the analyte signal(s) is useful during optimization of ashing and vaporization temperatures.Keywords Argon dimer (Ar2+ ); solid sampling; electrothermal vaporization; inductively coupled plasma mass spectrometry; matrix efects Since inductively coupled plasma mass spectrometry (ICP-MS) combines low detection limits with multi-element capabilities and a high sample throughput it is a highly desirable technique for trace element determination in a variety of matrices. Originally ICP-MS was mainly intended for the analysis of aqueous samples but at present there is an increasing interest in the direct analysis of solid samples (solid sampling). Direct analysis of solid materials is of course important for materials that cannot or only with great difficulty be brought into solution.Moreover in general solid sampling limits the neces- sary amount of often laborious and time-consuming sample pre-treatment leading to a reduced risk of contamination or analyte losses and finally as samples are analysed without dilution lower limits of detection are also to be expected. Next to laser ablation ICP-MS which has evolved into a well established solid sampling technique over the past 10 electrothermal vaporization (ETV) ICP-MS also offers pos- sibilities for the direct determination of trace and ultratrace elements in solid samples. * Senior research assistant of the Belgian National Fund for Scientific Research. t Present address Attila J6zsef University Department of Inorganic and Analytical Chemistry D6m Ter 7-P.O.Box 440 H-6720 Szeged Hungary. Both Voellkopf et al.’ and GrCgoire et aL6 reported on the analysis of solid samples using ETV-ICP-MS but both research groups preferred to analyse slurries rather than dry solid samples. Wang et aL7 explored the feasibility of ‘real’ solid sampling ETV-ICP-MS and highlighted some difficulties hampering routine use of this technique at present. They concluded that in order to obtain optimum accuracy external calibration using a solid certified reference material (CRM) with a composition as similar as possible to that of the sample should be used. Argentine and Barnes8 used ETV-ICP-MS for the direct determination of volatile and non-volatile impurities in semiconductor-grade organometallic materials and process chemicals.The technique showed great potential particularly for the non-volatile impurities since complete separation of the (volatile) matrix from the analytes could be obtained by application of a suitable temperature programme. Quantification of the latter impurities was performed using aqueous standards and the results obtained compared within a factor of two with the results obtained using electrothermal vaporization inductively coupled plasma atomic emission spec- trometry (ETV-ICP-AES) and ICP-AES after decomposition. Since for the volatile impurities the ETV-ICP-MS results were degraded owing to the occurrence of severe non-spectral interferences caused by co-volatilization of matrix and analyte ETV-ICP-MS was assessed to be inferior to ETV-ICP-AES for the determination of volatile components. In our laboratory however trace amounts of As and Se (both giving rise to volatile components evaporating from the sample at tempera- tures below SOOOC) were successfully determined in CRMs of plant and environmental Although also for these matrices important non-spectral interferences could be estab- lished several calibration methods were demonstrated to offer possibilities for accurate quantifi~ation.~ The use of an internal standard was observed to be advantageous (standard additions methods) to imperative (external calibration methods) depending on the calibration method used.The present paper reports on the results of a more funda- mental study in which the Ar,’ signal was systematically monitored and used as a diagnostic tool in solid sampling ETV-ICP-MS.Application of the AT,+ signal intensity as an indication of plasma loading effects and non-spectral inter- ferences was introduced for pneumatic nebulization ICP-MS by Beauchemin et a1.12 Recently the potential of this approach was confirmed by Chen and Houk,” who extended its appli- cation range by also using a variety of other polyatomic ions. GrCgoire et aL6 used the argon dimer signal to monitor matrix effects in slurry sampling ETV-ICP-MS allowing them to select an appropriate calibration strategy for the determination of trace metals in samples containing substantial matrix components. In the present study it is shown that systematic monitoring Journal of Analytical Atomic Spectrometry December 1995 Vol. 10 1047of the Ar,+ signal in solid sampling ETV-ICP-MS allows a more profound understanding of the signal profiles (signal intensity as a function of time) observed.Registration of the Ar2+ signal was also shown to be useful during optimization of ashing and vaporization temperatures. Finally it is demon- strated that at least in some instances the Ar2+ signal can be used as an internal standard correcting for matrix-induced signal suppression effects such that the necessity of adding an appropriate elemental standardg can be avoided. EXPERIMENTAL Instrumentation The ETV system used is a commercially available graphite furnace of the boat-in-tube type (SM-30 Grun Analytische Mess-Systeme Ehringhausen Germany). Although this device was originally designed for solid sampling Zeeman atomic absorption spectrometry (AAS) some simple modifications (described el~ewhere’~) sufficed to make it compatible for use with both ICP-AES and ICP-MS.Sample holders (‘boats’) can be easily and reproducibly loaded into the cylindrical furnace with the aid of a pair of tweezers sliding on a rail which is rigidly mounted in front of the furnace. After loading the sample one end of the furnace is closed using a shutter kept in position by a catchspring. During operation an Ar flow the flow rate of which is controlled by a mass flow con- troller (Model 5876 Brooks Instruments Veenendaal The Netherlands) is swept through the furnace transporting vapor- ized ‘sample particles’ into the central channel of the ICP. The multi-step temperature programme of the furnace (Table 1) is controlled by a computer program developed in-house while Table 1 spectrometer Operating conditions for the ETV system and the ICP mass (a) ETVsystem- Type SM-30 Grun Analytische Mess- Temperature programme the multi-step temperature Systeme programme consisted of (1) (2) boosting step (1 s) ‘drying’ step (30 s at about 120 “C) during which the power applied is twice as high as for the ashing step allowing the heating rate of the furnace to be increased at moderate temperatures and hence i i stable ashing temperature to be obtained rapidly intermediate step (11 s at the ashing temperature) to switch the valve to the ‘measuring’ position and allow the plasma to stabilize (3) ashing step (30 s) (4) (5) vaporization step (15 s) ( 6 ) intermediate step (7) to switch the valve to the ‘venting’ position cleaning step (2 times 3 s at about 2700 “C) (b) ICP mass spectrometer- Type Rf power Sampling depth Carrier gas flow rate Auxiliary gas flow rate Plasma gas flow rate Lens voltages Sampling cone Skimmer cone Perkin-Elmer SCIEX-Elan 5000 10 mm 0.3-1.2 1 min-’ 1.0 1 min-’ 12 I min-’ Tuned using pneumatic nebulization-no further tuning required when switching to solid sampling ETV-ICP-MS 1000-1300 W P -58.1 V B +9.7 V s2 -9.0 v El +4.0 V Nickel 1.0 mm orifice diameter Nickel 0.75 mm orifice diameter the temperature can be monitored using an optical pyrometer especially designed for use with this type of ETV system (PY20 Grun Optik Ehringhausen Germany).The ETV system was coupled to a Perkin-Elmer SCIEX-Elan 5000 ICP mass spec- trometer via 10mm id silicone rubber tubing.In order to (i) reduce the amount of deposition of vaporized sample material on the torch the interface and the lens stack and (ii) avoid degradation of the interface pump oil a three-way valve was used to vent vapours generated during drying ashing and cleaning steps. Operating conditions are summarized in Table 1. Measurements Measurement parameters are summarized in Table 2. In order to obtain a compromise between fast hopping between the nuclides monitored (to obtain a representative image of the corresponding signal profiles) and an efficient use of the total measuring time (high ratio of actual measuring time to mass spectrometer settling time) 30ms was used as the dwell time per measuring point.” Although a quadrupole filter can only transmit ions of one mass-to-charge ratio at each moment during the acquisition and the voltages applied to the quadru- pole rods have to be changed in order to hop between the nuclides of interest during one measurement the term ‘simul- taneous monitoring’ is used throughout the text in order to indicate that both the analyte signal(s) and the Ar,+ signal were recorded during the same firing of the furnace.In order to protect the electron multiplier against excessively high count rates the OmniRange device was used for the monitoring of both the Ar2+ signal intensity and analyte signal intensities giving rise to excessively high count rates. The OmniRange device can selectively and reproducibly reduce the sensitivity of the mass spectrometer by varying the ion transmission efficiency at the exact time a given mass-to-charge ratio is being measured.The OmniRange setting for the argon dimer [and if necessary for the analyte(s)] was always selected such that the Ar2-‘ signal intensity was more or less comparable to the intensity of the analyte signal in the matrix under consider- ation. The three-way valve was switched manually to the ‘measuring position’ 11 s before the start of the vaporization stage while the measurement itself was started 1 s before the beginning of the vaporization stage (Table 1). The total measurement time exceeded the duration of the transient signals monitored. Integration over a more limited time interval could be carried out using Microsoft Excel 5.0.Samples All the solid materials investigated during this study are CRMs of biological or environmental origin Mussel Tissue (BCR CRM 278) Olive Leaves (BCR CRM 062) Rye Grass (BCR CRM 281) Sea Lettuce (BCR CRM 279) and Tomato Leaves (NIST SRM 1573) of biological origin and Light Sandy Soil (BCR CRM 142) and River Sediment (BCR CRM 320) of environmental origin. RESULTS AND DISCUSSION Fig. 1 shows the AT,’ signal profiles during the vaporization step for (a) an empty sample holder for (b) a sample holder Table 2 Measurement parameters Dwell time 30 ms Scanning mode Peak hop transient Readings per replicate Points per spectral peak 1 Total measurement time ~ 3 0 s Sweeps per reading 1 800/number of nuclides monitored 1048 Journal of Analytical Atomic Spectrometry December 1995 Vol.10Fig. 1 8oAr2+ signal intensity as a function of time during the vaporization stage of the multi-step temperature programme for 0 (empty sample holder) 0.527 and 2.030mg of solid sample of Sea Lettuce (BCR CRM 279) containing about 0.5mg of Sea Lettuce and (c) containing about 2 mg of Sea Lettuce. The marked difference between the profiles under consideration indicates that also in solid sam- pling ETV-ICP-MS the argon dimer can be used as an indication of matrix effects. Both for the empty sample holder and for the solid samples the Ar2+ signal profile shows a strong suppression at the beginning of the vaporization. The magnitude of this suppression was seen (i) not to be affected by the sample mass (Fig. 1) and (ii) to increase with an increasing difference between the ashing and the vaporization temperature.Hence it is believed that since the Ar2+ signal intensity strongly depends on the carrier gas flow rate (Fig. 2) this suppression is to be ascribed to sudden changes in the carrier gas flow rate caused by rapid heating of the furnace. Fig. 1 shows that for an empty sample holder the Ar2+ signal intensity continuously increases and finally reaches a more or less stable value. When analysing a liquid standard solution e.g. 5-30 pl of a 200 pg I-' Cd standard solution the Ar2+ profile was observed to be identical with that of an empty sample holder. For a solid sample however the Ar2+ signal profile shows important alterations. Since these alterations become more important with increasing sample mass and since the over-all shape of the Ar2+ signal profile during vaporization strongly depends on the matrix under consideration there is little doubt that also in solid sampling ETV-ICP-MS the Ar2+ signal can be used as an indication of non-spectral interferences caused by volatilizing matrix components and hence as an indication of the actual sensitivity of the instrument at each moment during vaporization.This suppression of the * 0.25 0.5 0.75 1 1.25 u - Carrier gas flow ratd min-' Fig. 2 'OArz+ signal intensity monitored as a function of the carrier gas flow rate at three rf powers during the vaporization stage of the multi-step temperature programme for samples of about 0.7 mg of River Sediment (BCR CRM 320) (vaporization temperatures 1400 "C) Ar2+ signal intensity could for example be caused by a cooling effect on the plasma caused by the introduction of (possibly macromolecular) matrix components leading to a decreased probability of Ar2+ formation and/or ionization.Evaporation of gaseous products from the matrix during vaporization could possibly also give rise to Ar2+ signal suppression owing to (i) a dilution of the carrier gas (ii) a change in the position of the Ar,+ maximum density zone in the plasma leading to an alteration of the extraction efficiency and/or (iii) a widening of the central channel of the ICP. Simultaneous monitoring of the argon dimer and the signal from analyte elements allowed a better understanding of the signal profiles observed to be obtained. During previous appli- c a t i o n ~ ~ .~ ~ it was observed that for a given material the signal profiles observed strongly depend on the analyte element and that for a given analyte element the signal profile strongly depends on the matrix. The results of simultaneous monitoring of the argon dimer with As Cd and Co respectively are shown for Olive Leaves (BCR CRM 062) in Fig. 3(a)-(c). A closer look at the analyte profiles concerned indicates that in each instance the actual shape of the profile is determined by both the analyte element (appearance time and duration of signal) and the matrix (suppression of the analyte signal by co-volatilizing matrix components). For As for example the first dip in the Ar2+ profile coincides with a dip in the 75As+ profile and it is clear that also the remainder of the 75As+ profile is strongly affected by the variation in the Ar2+ intensity (or more correctly the actual sensitivity of the mass spec- trometer).It is important to stress that the 75As+ signal profile is not perturbated by the presence of 40Ar35C1+ since no significant alteration in the shape of the signal profile is observed on addition of about 100 pg of C1 per mg of sample (about a 140-fold excess in comparison with the original C1 0 2.5 5 7.5 10 12.5 15 Timds 5 Fig. 3 "Ar,+ and M+ signal intensity as a function of time during the vaporization stage of the multi-step temperature programme for solid samples of Olive Leaves (BCR CRM 062) where M+ is (a) 75As+ (b) l1'Cd+ and (c) 59C0+ Journal of Analytical Atomic Spectrometry December 1995 Vol. 10 1049content 0.7 mg g- ').Similar signal profiles were also observed for e.g. Hg Sb and Se. For Cd on the other hand [Fig. 3(b)] condition that sample masses were chosen as close as possible to one an~ther.'?'~ - - the appearance time and the signal duration are such that the ll1Cd+ profile exactly fits under a peak in the AT2+ profile and hence apparently does not suffer from signal suppression. Co finally is less volatile than As and Cd and hence its signal is only detected later after the start of the vaporization than for more volatile elements. Since during the presence of the Co+ signal the Arz+ signal only shows a continuous increase in the signal intensity also for Co no dips in the profile are observed. It is also interesting that after about 5 s the sensitivity of the instrument gradually increases (as indicated by the continu- ously increasing argon dimer intensity) such that the tailing of the As and Co peaks (apparently) becomes more prominent.A more realistic representation of the signal profile can however be obtained by plotting the ratio of the signal intensity of the analyte element to the signal intensity of the argon dimer as is illustrated in Fig. 4 for Co. Fig. 4 clearly indicates that about 13 s after the start of the vaporization the signal ratio (59C0+ 80Ar2+) has already decreased to about 5% of its maximum value such that in this instance the acquisition duration and integration range can be limited to for example 15 s. Since it was demonstrated in Fig. 1 that the matrix effects encountered become more important with increasing sample mass it is clear that these matrix effects also cause curvature of the sample mass response curves (analyte signal intensity as a function of sample mass).It was demonstrated in a previous paper" that for As and Se this curvature could be corrected for by using Sb as an internal standard on condition that an aliquot of (liquid) Sb standard was brought into the sample holder and dried under an infrared lamp before solid sample loading. The use of Sb as an internal standard therefore corrected for matrix effects and allowed As and Se to be determined accurately in CRMs of both biological and environ- mental origin either by external calibration methods using liquid as well as solid standards (CRMs with similar matrix composition and analyte content) or by standard additions methods.If no internal standard was used accurate quantifi- cation was only possible using standard additions and on ' 140 120 0.8 + 2 2 0.4 a?. 0 In 0.2 n h c .- c 100 2 E 80 - c p f Q) 60 .E 20 0 - 0 2.5 5 7.5 10 12.5 15' Tirnds Fig. 4 Signal ratio (59C0+ "Ar,+) and 59C0+ signal intensity as a function of time during the vaporization stage of the multi-step temperature programme for about 0.7mg of solid sample of Olive Leaves (BCR CRM 062). In order to facilitate the evaluation of the importance of the memory effect the horizontal line indicates 5% of the maximum value for the (59C0+ *'Ar,+) signal ratio In general selection of a suitable internal standard is not self-evident. Most importantly an appropriate internal stan- dard should show an analogous (furnace) chemistry and should of course only be present at negligible levels in the samples under consideration.Although only of secondary importance with solid sampling ETV-ICP-MS in general internal stan- dardization for correction for matrix-induced signal suppres- sion or enhancement and for improving the precision has been observed to be most efficient if the mass number of the internal standard was chosen to be relatively close to that of the analyte element(^).'^-'' Since a suitable internal standard should fulfil all three conditions simultaneously it is clear that selection of such an internal standard sometimes poses an insurmountable problem. It was found however that also Ar2+ shows potential as an internal standard as is illustrated in Fig.5 in which the 7 5 A ~ + signal intensity and the ratio (75A~+ Ar2+) calculated by summation of the individual signal ratios for every reading are plotted as a function of the sample mass for Sea Lettuce. Since the signal ratio (75A~+ Ar2+) increases linearly with increasing sample mass it is clear that application of the Ar2+ signal as an internal standard allows the curvature of the 75As+ sample mass response curve to be corrected for accurately . Table 3 presents the results obtained for the quantification of As (by single standard addition) in Sea Lettuce (BCR CRM 279) using (i) Sb9 and (ii) Ar2+ as an internal standard. The results obtained were not deteriorated by the presence of C1 as even on addition of amounts of C1 exceeding the original content of the sample no 40Ar35C1+ interference on the 7 5 A ~ + signal could be established.' All the results were obtained under the same experimental conditions (e.g.calibration by single standard addition five sub-samples per replicate meas- uring sequence) discussed in detail el~ewhere.~ Comparison of these results with one another and with the certified value clearly demonstrates that in this instance Ar,+ corrects as successfully as Sb for the matrix-induced suppression observed such that its application as an internal standard allows accurate 14 I h .r C 12 .- i 6 a c .G 4 0 0.5 1 1.5 2 2.5 3 Sample mass/mg 300 .- C Fig. 5 75As+ signal intensity and (75A~+ "Ar,+) signal ratio as a function of the sample mass (solid sample) of Sea Lettuce (BCR CRM 279).The dashed line is only intended as an indication of the trend observed Table3 Results (pgg-') for the determination of As in Sea Lettuce (BCR CRM 279); calibration by single standard addition (two indepen- dent determinations for each internal standard used) Mean (SEM)* Mean (SEM)* Certified value using Sb as using Ar,+ as 3.09 0.20 3.18 (0.22) 2.96 (0.34) 3.24 (0.17) 3.31 (0.28) k 95 YO confidence limit internal standard internal standard * Standard error of the mean. 1050 Journal of Analytical Atomic Spectrometry December 1995 Vol. 10results to be obtained. Owing to the transient nature of the signals involved application of Sb as an internal standard for the determination of As and/or Se implies (i) integration of the analyte signal intensity and the Sb+ signal intensity separately and (ii) division of the integrated signal intensities obtained (75A~+ 12'Sb+ and/or 82Se+ 12'Sb+).For this approach it is mandatory that the internal standard (Sb) shows a completely analogous behaviour to the analytes (As and Se). When using the argon dimer as an internal standard the ratio (As' ATz+) is calculated for every reading individu- ally and the final 'signal ratio' is obtained by summation of all the individual ratios. This approach offers the advantage of real-time correction for non-spectral interferences and should be valid more universally since similarity of signal profiles is no longer requested. Finally it should be stressed that for example in order to evaluate the importance of the differences in mass number between an analyte element and the argon dimer it is advisable to check the utility of the Ar,+ signal as an internal standard for each analyte and each matrix under consideration e.g.by plotting the ratio (M' Ar2+) as a function of the sample mass. During the analysis of materials of environmental origin registration of the argon dimer has also proved to be useful during the optimization of the vaporization temperature. For some materials of environmental origin (e.g. soils or sludges) ashing at a moderate temperature (200-400°C) does not provide an effective removal of the matrix. For the determi- nation of relatively volatile elements (As Cd Se) it is therefore advisable to work at a moderate vaporization temperature in order to volatilize the latter analytes before the major matrix components.Fig. 6(a)-(c) shows the signal profiles for 'llCd+ and Ar2+ for River Sediment (BCR CRM 320) at about 1370 1550 and 1700 "C respectively. The Ar2+ signal profiles indi- .; 1.4 . . . . n...X a . . . I . . . . 1 L . L 5 0 2.5 5 7.5 10 12.5 15 17.5 20 0 2.5 5 7.5 10 12.5 15 17.5 20 v > v) 0 c .- c .- 0 2.5 5 7.5 10 12.5 15 17.5 20 Timds Fig. 6 '*Ar,+ and 'llCd+ signal intensity as a function of time during the vaporization stage of the multi-step temperature programme for solid samples (about 0.7 mg) of River Sediment (BCR CRM 320) at a vaporization temperature of (a) 1370 (b) 1550 and (c) 1700 "C cate an increasing amount of co-volatilization of the matrix on increasing the vaporization temperature from 1370 to 1700°C.Since weighing of the empty sample holders before and after each analysis of a sample of one of the environmental materials investigated (Light Sandy Soil or River Sediment) systematically indicated a significant loss of mass a (chemical) reaction between the solid sample material and the sample holder material (pyrolytically coated graphite) could possibly contribute to the matrix effect observed. No significant loss of mass was established for the sample holders when analysing materials of biological origin. It is important in order to limit (i) the amount of matrix-induced signal suppression (ii) the deposition of sample material on the torch the interface and the lens stack and (iii) the degradation of the interface pump oil that no excessive amount of matrix material co-volatilizes with the analyte(s) as is for example to a very large extent the case in Fig.6(c). Hence simultaneous registration of the analyte signal (as an indication of analyte vaporization) and the argon dimer (as an indication of the amount of co-volatilizing matrix material and/or the extent to which the chemical reaction between the sample and the graphite sample holder proceeds) is useful during the optimization of the vaporization temperature for matrices of environmental origin. Simultaneous registration of the argon dimer also provides an explanation of why in these instances the signal intensity does not reach a constant value starting from a given vaporization temperature (as is the case for the samples of biological origin in~estigated):~,~~ the signal intensity starts to decrease at excessively high vaporization temperatures as a result of matrix effects.Since for environmental samples an appropriate selection of the vaporization temperature can significantly reduce the amount of matrix components co-volatilizing with the ana- lyte(s) (e.g. As Cd or Se) and hence the extent to which these analyte signal intensities are suppressed careful optimization of the vaporization temperature also has a direct bearing on the curvature of the corresponding sample mass response curves (signal intensity as a function of the sample mass). As a result of this optimization the curvature of the sample mass response curves was observed to be much less pronounced for environmental than for biological matrices,16 since for the latter at least for the elements investigated no conditions could be found leading to a complete separation of volatiliz- ation of analyte(s) and matrix respectively.For biological matrices on the other hand optimization of the ashing temperature is important. An optimum ashing temperature provides an efficient removal of the matrix with- out analyte losses. Since the Ar,' signal profile gives an indication of matrix effects it is clear that the ashing tempera- ture also exerts an influence on the Ar2+ profile such that registration of the argon dimer gives an indication of the effectiveness of the ashing step. Simultaneous registration of the analyte signal (to ensure that no analyte is lost during ashing) and the argon dimer (to control the effectiveness of the ashing step) during vaporization following ashing at suc- cessively increasing ashing temperatures is therefore useful during optimization of the ashing temperature.CONCLUSIONS It is shown as has already been demonstrated for pneumatic nebulization ICP-MS by Beauchemin et aZ.12 and Chen and Houk13 and for slurry sampling ETV-ICP-MS by GrCgoire et u Z . ~ that also for solid sampling ETV-ICP-MS monitoring of the Ar2+ signal provides a direct indication of plasma loading effects and non-spectral interferences. Monitoring of the Ar2+ signal intensity not only allows differences in the shape of signal profiles between different analytes and/or matrices to be explained but is also of use for optimization of Journal of Analytical Atomic Spectrometry December 1995 VoZ.10 1051operational conditions. Finally since it was demonstrated that the Ar,’ signal may at least in some instances be used as an internal standard to correct for matrix-induced signal suppres- sion effects the necessity of adding an appropriate elemental internal standard (similar in furnace chemistry and only present at a negligible level in the sample) for accurate quantification may be circumvented. Further research is presently being carried out in order to evaluate the universality of this approach. REFERENCES Gray A. L. Analyst 1985 110 551. Arrowsmith P. Anal. Chem. 1987 59 1437. Denoyer E. R. Fredeen K. J. and Hager J. W. Anal. Chem. 1991 63 445. Ulens K. Moens L. Dams R. Van Winckel S. and Vandevelde L. J. Anal. At. Spectrom. 1994 9 1243. Voellkopf U. Paul M. and Denoyer E. R. Fresenius’ J. Anal. Chem. 1992 342 917. Grkgoire D. C. Miller-Ihli N. J. and Sturgeon R. E. J. Anal. At. Spectrom. 1994 9 605. Wang J. Carey J. M. and Caruso J. A. Spectrochim. Acta Part B 1994 49 193. 8 9 10 11 12 13 14 15 16 17 18 19 Argentine M. D. and Barnes R. M. J. Anal. At. Spectrom. 1994 9 1371. Vanhaecke F. Boonen S. Moens L. and Dams R. J. Anal. At. Spectrom. 1995 10 81. Boonen S. Vanhaecke F. Moens L. and Dams R. Spectrochirn. Acta Part B 1995 in the press. Moens L. Verrept P. Boonen S. Vanhaecke F. and Dams R. Spectrochim. Acta Part B 1995 50,463. Beauchemin D. McLaren J. W. and Berman S. S. Spectrochim. Acta Part B 1987 42 467. Chen X. and Houk R. S. J. Anal. At. Spectrom. 1995 10 in the press. Verrept P. Dams R. and Kurfurst U. Fresenius’ J. Anal. Chem. 1993 346 1035. Denoyer E. R. At. Spectrosc. 1994 15 7. Galbacs G. Vanhaecke F. Moens L. and Dams R. unpub- lished work. Thompson J. J. and Houk R. S. Appl. Spectrosc. 1987 41 801. Doherty W. Spectrochim. Acta Part B 1989 44 263. Vanhaecke F. Vanhoe H. Dams R. and Vandecasteele C. Talanta 1992 39 737. Paper 51040771 Received June 23 1995 Accepted August 21 1995 1052 Journal of Analytical Atomic Spectrometry December 1995 Vol. 10