Determination of Ag, Pb and Sn in aqua regia extracts from sediments by electrothermal atomic absorption spectrometry using Ru as a permanent modiÆer Jose� Bento Borba da Silva,a Ma�rcia Andreia Mesquita da Silva,b Adilson Jose� Curtius*b and Bernhard Welzb aDepto. de Quý�mica da Universidade Estadual de Maringa�, 87020-900 Maringa�, P.R., Brazil bDepto. de Quý�mica da Universidade Federal de Santa Catarina, 88040-900 Floriano�polis, S.C., Brazil. E-mail: curtius@qmc.ufsc.br Received 5th July 1999, Accepted 1st September 1999 Ruthenium, deposited on a L'vov platform, is proposed as a permanent modiÆer for the determination of Ag, Pb and Sn in aqua regia extracts from sediments by electrothermal atomic absorption spectrometry.The coating process is simple: a solution containing Ru is pipetted repeatedly on to the platform inserted in a graphite tube and is submitted to a temperature program. In a 50% v/v aqua regia solution, high pyrolysis temperatures could be used: 1200 �C for Ag and Pb, and 1500 �C for Sn.At these temperatures, similar characteristic masses to those found for a nitric acid medium, using a Pd±Mg modiÆer, were obtained, showing that the high concentration of chloride does not interfere with the determination. In the aqua regia medium, the permanent modiÆer is much superior in comparison with Pd or PdzMg, modiÆers applied as a solution, which could not stabilize the analytes satisfactorily. Very long tube lifetimes, around 1700 cycles, were obtained for Pb and Sn in this medium.Three sediment reference materials were partially dissolved using a mixture of aqua regia and hydrogen peroxide in a microwave oven. The results for Ag and Pb were in agreement with the recommended values, demonstrating the efÆciency of the extraction. However, for Sn, the precison was less satisfactory, indicating that the extraction may be less efÆcient and reproducible for this analyte. Other advantages of the permanent Ru modiÆer are the low blanks due to in situ cleaning of the modiÆer and the shorter analysis time in comparison with the modiÆers in solution.Introduction The total analyte concentration in a solid environmental sample, such as a sediment or a soil, can only be determined after a fusion with sodium carbonate1 or lithium metaborate, 2±4 or after an acid digestion in the presence of hydroØuoric acid and frequently also perchloric acid.5 Such digestion procedures are relatively time consuming and not without problems for the following analysis. Hence, instead of a total digestion, it is usual in environmental analysis to perform an extraction with boiling aqua regia.This is justiÆed on the assumption that heavy metals that are so strongly bound to the mineral matrix that they do not go into solution with such a digestion will also not be taken up by plants and cannot be dissolved by water and bacteria. In the meantime, reference materials have been produced for which, in addition to the certiÆed total concentration, information is also provided on the fraction that can be extracted in aqua regia.Aqua regia must be considered a difÆcult matrix for electrothermal atomic absorption spectrometry (ETAAS) because, due to its high chloride content, it may cause analyte loss during the pyrolysis stage, gas phase interferences in the atomization stage, as well as pronounced corrosion of graphite tubes, so that all aspects of the stabilized temperature platform furnace (STPF) concept6 have to be considered, particularly the use of appropriate chemical modiÆers.Chemical modiÆers have been used routinely in the determination of a great number of analytes by ETAAS. Generally, the use of modiÆers allows high pyrolysis temperatures, reducing or eliminating volatilization and vapor phase interferences and minimizing background signals.7 Solutions containing Pd with or without Mg have been employed for a wide range of elements, especially for the more volatile elements,8 and are recommended in the software of the instruments and in the manuals from the instrument manufacturers.9 The most common way of applying the modiÆer is by pipetting its solution together with or after the sample and calibration solutions.More recently, the modiÆer has been applied as a metal deposit on the graphite tube surface or on the L'vov platform, acting as a ``permanent'' modiÆer, making possible between 50 and more than 1000 atomization cycles10,11 without repeating the treatment of the tube or platform.Platinum group metals (PGMs), such as Pd, Pt, Ir, Ru and Rh, as well as carbide-forming elements, such as Zr, Nb, Ta and W, have been used as permanent modiÆers for the determination of volatile elements by ETAAS.7,11±15 In this sense, Ir and Ru, due to their high melting-points, should allow higher pyrolysis temperatures than Pd.14 On the other hand, Tsalev et al.16 did not Ænd a pronounced correlation between the maximum pyrolysis temperature and the melting-point of the PGMs using Pd, Rh and Ru chlorides in the determination of 18 analytes.The fact that the carbide-forming elements and the PGMs have a similar behavior as chemical modiÆers has also been attributed to their catalytic action.17 The high boiling-points of Ir and Ru, as compared with that of Pd, allows their use as permanent modiÆers, while Pd is lost at relatively low temperatures (1800 �C), and cannot be employed for this purpose.18 In spite of the wide use of PGMs as chemical modiÆers, Pd alone or mixed with Mg as solutions, and Ir or Rh as permanent modiÆers, have been preferred while Ru has rarely been used.Sturgeon et al.19 employed Ru, Pt, Pd and Rh to sequester and concentrate the hydrides of Bi, Se, As, Sn and Sb in a graphite tube. Palladium showed the best performance, leading to higher signal appearance temperatures, which was attributed J.Anal. At. Spectrom., 1999, 14, 1737±1742 1737 This Journal is # The Royal Society of Chemistry 1999to its afÆnity for H2. Only for Se was the same signal appearance temperature obtained for Pd and Ru. Mixtures of PdzPtzRhzRu were also used as thermal stabilizers in the hydride collection.20 Using PGM chlorides as modiÆers for 18 analytes, Tsalev et al.16 observed that the maximum pyrolysis temperature decreased in the order RuwRhwPd. In another study, using different metals as modiÆers, their efÆciency diminished in the order PdwRuwCewAgwPt.21 Cai and McDonald22 investigated Ru as a potential chemical modiÆer for Pb and proposed the formation of an intermetallic compound PbRu2 as the stabilizing mechanism.In this work, the determination of Ag, Pb and Sn by ETAAS after partial dissolution with aqua regia was investigated. High pyrolysis temperatures are desirable to eliminate chlorides from the matrix effectively, without losing the analytes.This situation seems to be very challenging to test Ru, deposited on a L'vov platform, as a permanent modiÆer. Experimental Apparatus An Aanalyst 100 atomic absorption spectrometer (Perkin- Elmer, Norwalk, CT, USA), equipped with an HGA-800 graphite tube atomizer, an AS-72 autosampler and a deuterium-arc background corrector, was operated under the conditions recommended by the manufacturer. Integrated absorbance (peak area) was used exclusively for signal evaluation.The hollow cathode lamp for Ag (Hitachi, Mitorika, Ibaraki, Japan) was operated at 15 mA and the hollow cathode lamps for Pb and Sn (Perkin-Elmer) at 10 and 25 mA, respectively. The volume pipetted into the graphite tube was 20 mL for the test sample and calibration solutions. The volume of the chemical modiÆer when added in solution was 10 mL. Argon, 99.996% (White Martins, Saƒo Paulo, S.P., Brazil), was used as sheath gas. Pyrolytic graphite coated graphite tubes (Perkin-Elmer, Part No.B010-9322) with a total pyrolytic graphite platform (Perkin-Elmer, Part No. B010- 9324) were used. The platform was pre-treated with Ru, in a similar way to that described previously forapplying 40 mL of a 500 mg mL21 Ru solution on to the platform and submitting the tube to the temperature program given in Table 1. This procedure was repeated 25 times in order to obtain a deposit of 500 mg of Ru, as a permanent modiÆer.This temperature program also served to remove volatile contaminants, and hence to ensure low blank values in the Ænal analysis.11,23 The absolute blank values for Ru as well as for PdzMg, when applied in solution, and also for Ru when applied as a permanent modiÆer, are shown in Table 2, demonstrating the puriÆcation effect of the permanent modiÆers. The graphite furnace temperature program for each analyte was optimized and is shown in Table 3. A microwave oven, MLS-1200 MEGA (Milestone, Sorisole, B.G., Italy), was used to dissolve the samples. Reagents and solutions All chemicals used were of analytical-reagent grade, unless otherwise speciÆed.Water was de-ionized in a Milli-Q system (Millipore, Bedford, MA, USA). Hydrochloric acid (Merck, Darmstadt, Germany, No. 334) and nitric acid (Carlo Erba, Milan, Italy, No. 408015) were further puriÆed by sub-boiling distillation in a quartz still (Ku» rner Analysentechnik, Rosenheim, Germany). Hydrogen peroxide, 30% v/v (Merck, No. 507016), was used as supplied. The following 1000 mg mL21 stock solutions were used: ruthenium (Fluka, Buchs, Switzerland, No. 84033) in 1 mol L21 hydrochloric acid; silver (Fluka, No. 85137) in 0.5 mol L21 nitric acid; lead (Spex, Edison, NJ, USA, No. PLK10-Pb) in 0.3 mol L21 hydrochloric acid; and tin (Spex, No. PLK10-Sn) in 1 mol L21 hydrochloric acid. For the modiÆer in solution the following 10.0°0.2 g L21 stock solutions were used: magnesium nitrate solution, modiÆer for graphite furnace AAS (Merck, No.B593213 431); and palladium nitrate solution, modiÆer for graphite furnace AAS (Merck, No. B936989 710). The calibration solutions for Ag, 1±5 mg L21, for Pb, 10± 50 mg L21, and for Sn, 20±80 mg L21, were obtained by dilution with 0.2% v/v nitric acid. The modiÆers in solution were also diluted with 0.2% v/v nitric acid. Samples Three certiÆed reference materials were analysed, one from the National Research Council of Canada (Ottawa, Canada): Marine Sediment (MESS-2), and two from Canadian CertiÆed Reference Material Project (Ottawa, Canada): Stream Sediment (STSD-2) and Lake Sediment (LKSD-3).To each aliquot of about 250 mg of the material, weighed directly in the PTFE Øask of the microwave system, 5 mL of aqua regia and 1 mL of hydrogen peroxide were added. A four-step program was used in the microwave oven: 5 min at 250 W, 5 min at 400 W, 5 min at 650 W and 5 min at 250 W. The volume of the Ænal solution was made up to 10 mL with aqua regia used to wash the Øask.A solid residue remained on the bottom of the calibrated Øask. These solutions were diluted 1z19 for Pb, 1z19 for Ag in LKDS-3 and in STDS-2 and 1z9 for Sn in STDS-2 with 50% v/v aqua regia. The other solutions were not further diluted. An aliquot of the supernatant was transferred into the autosampler cup. Before use, all glassware was kept in an Extran solution (Merck) for 12 h and in an ultrasonic water-bath for 30 min and was then rinsed with Milli-Q water.Then, it was kept in a 20% aqua regia solution for at least 48 h and washed four times with Milli-Q water. The PTFE Øasks of the microwave system and the autosampler cups were kept in a warm 50% v/v nitric acid solution for 4 h and were then washed several times with Milli-Q water. Results and discussion Pyrolysis temperature Fig. 1(a)±(c) shows the pyrolysis curves for Pb, Ag and Sn, respectively, in 50% v/v aqua regia solution using A–Ru as a permanent modiÆer; B–PdzMg added in solution as a modiÆer; C–no modiÆer; and D–Pd alone in solution as a modiÆer. For all three elements the Ru-treated platforms provide the best stabilization of the analyte up to fairly high Table 1 Temperature program for the Ru coating of the L'vov platform Step Temperature/�C Ramp/s Hold/s Ar Øow rate/mL min21 1 90 5 15 250 2 140 5 15 250 3 1000 10 10 250 4 2000 0 5 0 5 20 1 10 250 Table 2 Blank values obtained for different modiÆers, applied directly in 0.2% v/v nitric acid (20 mL of modiÆers), and as permanent modiÆers.(n~3) ModiÆer Blank signal/s Ag Pb Sn 15 mg Pdz10 mg Mg 0.007°0.002 0.029°0.005 0.021°0.005 10 mg Ru 0.155°0.008 0.056°0.013 0.006°0.002 500 mg Ru permanent 0.001°0.001 0.001°0.001 0.001°0.002 1738 J. Anal. At. Spectrom., 1999, 14, 1737±1742pyrolysis temperatures. The only element that exhibits a fairly strong dependence of the integrated absorbance signal on the pyrolysis temperature is Sn.This is most likely due to the low sublimation point of SnCl2 of 650 �C, which coincides satisfactorily with the beginning of the drop of the pyrolysis curve at 600 �C. This loss mechanism has previously been proposed by Rayson and Holcombe,24 who also pointed to the high thermal stability of this gaseous molecule. The performance of PdzMg as a modiÆer, added in solution, is clearly inferior to that of the Ru-treated platform in the presence of this high chloride matrix. This may be explained through the stabilization mechanism of this modiÆer, which was investigated in detail by Ortner et al.25 The Pd penetrates to a depth of 10 mm into the pyrolytic graphite surface, and is activated by covalent bonding to the graphite lattice (intercalation compound).This ``activated Pd'' then forms a covalent bond with the analyte element during the drying stage.25 Although this mechanism was established only for As, it is most likely valid also for other elements.A penetration of Pd into the graphite structure has been described earlier by Majidi and Robertson,26 and it does undoubtedly play the key role in the stabilization of analytes by Pd. Other proposed mechanisms, such as a retention of the analyte in particles on the surface at temperatures w800 �C, are not possible from considerations on the diffusion of analytes in the corresponding matrices.25 Likewise, the formation of intermetallic compounds between the analyte and Pd does not occur due to the extreme concentration ratio of Pd : analyte w1000 : 1.The interference of chloride could then be explained as a competition with the Pd for the formation of intercalation compounds with graphite, i.e. less Pd can be intercalated in the presence of high chloride concentrations. This intercalation of chloride under similar conditions has been clearly demonstrated for total pyrolytic graphite tubes27 and is likely to occur here as well.Palladium alone, added in solution, although proposed by other groups as a modiÆer,28,29 is signiÆcantly less efÆcient than PdzMg in the high chloride matrix for all three elements investigated here. This is according to expectation as the magnesium nitrate can contribute to the stabilization of the analyte in at least three different ways. Firstly, the magnesium chloride which is undoubtedly formed under these high chloride conditions, hydrolyzes with the formation of MgO(s) and HCl(g).30 Secondly, the MgO can at least in part prevent analyte molecules, such as chlorides, from being lost in the pyrolysis stage by imbedding them in an oxidizing atmosphere.31 Thirdly, several workers19,32±34 reported the formation of a Pd±M±O bond (where M is the analyte) in the presence of the PdzMg modiÆer, a bond that is most likely not formed in a high chloride matrix in the absence of magnesium nitrate. For Tl it could be shown that ``reduced Pd'', i.e.Pd that was pipetted onto the platform Ærst and pyrolyzed at 1000 �C before the sample solution was introduced, could prevent a chloride interference effectively, but not Pd added in solution.31 This further supports the above-discussed competition between chloride and Pd for intercalation as the most likely mechanism of interference and the low stabilizing power d modiÆers under these conditions. The use of reduced Pd as a modiÆer was not investigated in this work as it was considered too time consuming, particularly in comparison with the use of a permanent modiÆer.We have no explanation for the somewhat higher sensitivity without modiÆer in comparison with Pd alone as the modiÆer for Ag, and in part also for Pb. In essence, this demonstrates that Pd, added in solution, has very little stabilizing power for these elements under the conditions of this experiment. In other words, Pd cannot be intercalated into the graphite lattice under these conditions or the ``activated Pd'' cannot form a stable covalent bond with the analyte, whereas this is apparently possible, at least in part, in the presence of magnesium nitrate.It is unlikely that the condensation of matrix vapor,35,36 and the trapping of the analyte on such clusters is responsible for the observed effect, as in that situation, the interference should be more pronounced in the presence of Mg in addition to Pd, compared with Pd alone, which is not the case.For Sn, in the absence of a modiÆer, only a very small integrated absorbance signal of about 0.05 s could be measured at low pyrolysis temperatures, which disappeared completely at Table 3 Temperature program for the determination of Ag, Pb and Sn using the Ru-treated platform Step Temperature/�C Ramp/s Hold/s Ar Øow rate/mL min21 1 90 5 10 250 2 140 5 20 250 3 1200 (Ag,Pb); 1500 (Sn) 10 30 250 4a 1800 (Ag, Pb); 2300 (Sn) 0 5 0 5 2650 1 5 250 6 20 1 10 250 aReading in this step.Fig. 1 Pyrolysis temperature curves for (a) 1 ng Pb; (b) 0.1 ng Ag; and (c) 1.6 ng Sn, all in 50% v/v aqua regia. A: Ru-coated platform; B: 15 mg Pdz10 mg Mg as modiÆer in solution; C: without modiÆer; and D: J. Anal. At. Spectrom., 1999, 14, 1737±1742 1739w800 �C. This observation is in agreement with the work of Brown and Styris,37 who investigated the atomization of SnCl2 by mass spectrometry and observed only SnO(g) and SnCl2(g), but no free Sn(g).These workers concluded that atomic Sn, which can be detected by AAS, is only formed on the graphite surface. Tube lifetime Fig. 2(a) and (b) shows the average integrated absorbance and relative standard deviation (RSD; n~20) for Pb and Sn, respectively, in 50% v/v aqua regia solution over the lifetime of a tube with a platform treated with Ru as described under Experimental. No long-term measurements were made for Ag, as a similar behaviour can be anticipated, because it is certainly not the analyte, but the aggressive matrix and the atomization and clean-out temperatures that determine the tube lifetime.38 Most striking is the very long lifetime of the tubes in the presence of this high acid concentration without the need for any recoating. The Pb experiment was terminated after 1700 atomization cycles, as the integrated absorbance started to drop signiÆcantly.A visual inspection showed that the platform was still in good condition, but some corrosion appeared at the inner and outer tube walls.The tube from the Sn experiment broke after about 1750 atomization cycles. For comparison, Rohr et al.,39 in an extensive lifetime test for graphite tubes in the presence of various matrices, found the most severe corrosion and tube breakage after only 240 atomization cycles for a 6 mol L21 HCl matrix. The signal stability for Sn was excellent over the entire lifetime of the tube without any signiÆcant drift in sensitivity, and for the majority of the individual measurement intervals of 20 atomization cycles each, the RSD was around 2.5%; only for a few of them–more pronounced towards the end of the tube lifetime–was the RSD between 5 and 10%.The integrated absorbance values for Pb appear to be slightly less consistent over the lifetime of the tube with an overall drop in sensitivity of about 25%, and some irregularities around 400 and 1250 atomization cycles. It must be kept in mind, however, that these experiments were carried out over a period of about 2 weeks each, most of the time unattended, and with a repeated change of the measurement solution, so that these inconsistencies should not be over-interpreted.Similar to the Sn experiment, the RSD was mostly around 2.5%, and only in a few series were values between 5 and 10% obtained. Fig. 3 shows a series of atomization pulses for 1 ng Pb in 50% v/v aqua regia over the lifetime of the tube, using a Ru-treated platform.Over more than 1000 atomization cycles the signal shape is almost symmetrical, which is unusual for an endheated graphite tube atomizer, and is probably due to the high stabilizing power of Ru. Around 900 atomization cycles a small early peak appeared, which became more and more prominent with increasing tube age. This is an indication that some of the Pb is no longer stabilized to the same degree, most likely because the Ru layer breaks, and some Pb is atomized directly from the exposed graphite surface.It should be noted, however, that there was essentially no change in the integrated absorbance recorded between atomization cycle 900 and 1600, which means that no Pb is lost in the pyrolysis stage. Towards the end of the tube lifetime, the second peak started to shift to even later appearance, and a shoulder appeared where the original peak maximum used to be. This is a further indication of a breakdown of the Ru layer. Most likely parts of the layer are peeling off, and hence lose contact with the platform, resulting in a further delay of the Pb atomization. Rademeyer et al.,7 for an Ir-coated graphite tube, also observed the appearance of a second early peak for Pb after 220 Ærings, Fig. 2 Average integrated absorbance and RSDs of 20 consecutive measurements each for (a) 1 ng Pb and (b) 1.6 ng Sn, both in 50% v/v aqua regia, using a Ru-treated platform. Fig. 3 Atomization pulses for 1 ng Pb in 50% v/v aqua regia, using a Ru-treated platform, after 1, 900, 1200, 1400 and 1600 atomization cycles. 1740 J. Anal. At. Spectrom., 1999, 14, 1737±1742which they attributed to an atomization from a different surface. On inspection by scanning electron microscopy they found that there was no visible residue of Ir in the center of the tube where the sample was deposited, which had apparently migrated to the cooler ends of the tube. It should be stressed again that the lifetime of the tube and the long-term performance of the Ru permanent modiÆer found in this work are far better than previous results obtained in our laboratory before11 or published in the literature, particularly considering the matrix.The only comparable published lifetime of about 1500 atomization cycles for a W±Rh treated platform40 was obtained for Cd in 0.2% v/v HNO3, and the platform had to be recoated repeatedly after every 300±350 atomization cycles. In later work the same group reported a lifetime of only about 600 atomization cycles for Cd in Æsh slurry, also with one recoating after 300 cycles.41 Analytical application The Ægures of merit for the determination of Ag, Pb and Sn in 50% v/v aqua regia are shown in Table 4.The obtained characteristic masses (the analyte mass that produces an integrated absorbance of 0.0044 s) were similar to those reported in the literature8 and by the instrument manufacturer9 for the analytes in 0.02% v/v nitric acid, using PdzMg (15 mgz10 mg) in solution as modiÆer.As is also shown in Table 4, using a Ru-treated platform, the obtained characteristic masses in aqua regia medium (50% v/v) and in 0.2% v/v nitric acid were similar. Hence, the Ru-treated platform allows sufÆciently high pyrolysis temperatures to eliminate efÆciently the effect of the chloride resulting from the high concentration of aqua regia. The limits of detection (LOD), deÆned as the mass (absolute) or the concentration (relative) that gives an integrated absorbance equal to three times the standard deviation of ten measurements of a solution close to the blank, based on 20 mL, are also shown in Table 4.The results for the analysis of the certiÆed reference materials arewn in Table 5. Since the characteristic masses in the aqua regia medium, as already discussed, were similar to those in 0.2% v/v nitric acid, the calibration solutions were prepared in the latter medium. The concentrations obtained for Ag and Pb using aqua regia extraction agreed well with the certiÆed or recommended values for the total content, which is not so much a proof of the accuracy of the proposed method, but an indication of the efÆciency of the aqua regia extraction, at least for the samples investigated in this study.For these analytes, the RSDs were below 10%. For Sn, the agreement was also good; however, for this analyte, the RSDs were fairly high. It was observed that the absorption pulse for Sn in the sample extract was different from the absorption pulses for this analyte in the calibration solution in 0.2% v/v nitric acid and in 50% v/v aqua regia.While the former was noisy, and not symmetrical, the latter was smoother and more symmetrical. Also, the background signals were much higher in the extracts, probably surpassing the correction capability of the continuum source background corrector. It was observed that the differences in the absorption pulses and in the background signals were not due to the presence of the aqua regia in the extract, but to some concomitants of the samples.Conclusions The use of a Ru coating on a L'vov platform, as a permanent modiÆer for Ag, Pb and Sn, is advantageous in comparison with the use of Pd alone or mixed with Mg in solution. The coating procedure is simple and efÆcient and the modiÆer is cleaned in situ by applying the conditioning temperature program, resulting in much lower blank signals in comparison with the use of the cited modiÆers in solution.Since the modiÆer does not have to be pipetted into the tube, the analysis time is shorter. High pyrolysis temperatures could be used, allowing the determination of the analytes in an aqua regia medium using calibration solutions in dilute nitric acid without matrix-matching. Particularly outstanding was the lifetime of the Ru-treated platform in the aggressive medium of aqua regia, making this permanent modiÆer a potential candidate for the stabilization of other volatile elements, and particularly for a routine analysis of aqua regia extracts using ETAAS.The aqua regia was very efÆcient in extracting Ag and Pb from the sediment materials, but less efÆcient for Sn. Acknowledgements The authors are grateful to Conselho Nacional de Pesquisas e Desenvolvimento Tecnolo�gico (CNPq) and to CoordenacÀaƒo de AperfeicÀoamento de Pessoal de Ný�vel Superior (CAPES). M. A. Mesquita da Silva has a scholarship from CAPES and J.B. Borba da Silva and B. Welz from CNPq. We also thank Dr. C. Gre�goire of the Canadian Geological Survey for providing the certiÆed reference materials. We dedicate this paper to the memory of our colleague, the late Dr. Eduardo Stadler. Table 4 Figures of merit for the determination of the analytes in 50% v/v aqua regia, using the Ru-treated platform Characteristic mass/pg Limit of detection (k~3, n~10) Analyte Founda Foundb Literaturec Absolute/pg Relative/mg L21 Ag 1.8°0.2 1.9°0.1 1.5 10 0.5 Pb 9.5°1.2 10.0°0.5 10.0 10 0.5 Sn 10.0°0.1 10.0°0.5 10.0 35 1.7 aIn 50% v/v aqua regia medium, using the Ru-treated platform.bIn 0.2% nitric acid, using the Ru-treated platform. cIn 0.2% nitric acid using 15 mg Pdz10 mg Mg in solution, as modiÆer.9 Table 5 Concentrations of Ag, Pb and Sn in sediment reference materials (n~3) Sample Concentration in the sample/mg L21 Ag Pb Sn Found CertiÆed Found CertiÆed Found CertiÆed LKDS-3 3.06°0.01 2.8a 32°1.3 29a 2.3°1 3a MESS-2 0.18°0.002 0.18°0.02 22.7°0.1 21.9°1.2 1.9°0.5 2.27°0.42 STDS-2 0.58°0.01 0.5a 69°7.0 66a 4.8°1.3 5a aRecommended value.J. Anal. At. Spectrom., 1999, 14, 1737±1742 1741References 1 P. J. Potts, Handbook of Silicate Rock Analysis, Blackie, Glasgow, 1987. 2 A. A. Verbeek, M. C. Mitchell and A. M. Ure, Anal. Chim. Acta, 1982, 135, 215. 3 M. Bettinelli, Anal. Chim. Acta, 1983, 148, 193. 4 D. C. Bartenfelder and A.D. Karathanasis, Commun. Soil Sci. Plant Anal., 1988, 19, 471. 5 H. Agemian and E. Bedek, Anal. Chim. Acta, 1980, 119, 323. 6 W. Slavin, D. C. Manning and G. R. Carnrick, At. Spectrosc., 1981, 2, 137. 7 C. Rademeyer, B. Radziuk, N. Romanova, N. P. Skaugset, A. Skogstad and Y. Thomassen, J. Anal. At. Spectrom., 1995, 10, 739. 8 B. Welz, G. Schlemmer and J. R. Mudakavi, J. Anal. At. Spectrom., 1988, 3, 93. 9 Analytical Techniques for Graphite Furnace AAS, Bodenseewerk Perkin-Elmer, U» berlingen, 1984, Part No.B010-0180 (B332/E). 10 I. L. Shuttler, M. Feuerstein and G. Schlemmer, J. Anal. At. Spectrom., 1992, 7, 1299. 11 J. B. B. Silva, M. B. O. Giacomelli, I. G. Souza and A. J. Curtius, Microchem. J., 1998, 60, 249. 12 D. L. Tsalev, A. D'Ulivo, L. Lampugnani, M. Di Marco and R. Zamboni, J. Anal. At. Spectrom., 1995, 10, 1003. 13 E. Bulska and W. Jedral, J. Anal. At. Spectrom., 1995, 10, 49. 14 A. Volynsky and V. Krivan, J. Anal. At. Spectrom., 1997, 12, 333. 15 E. Bulska, K. Liebert-Ilkowska and A. Hulanicki, Spectrochim. Acta, Part B, 1998, 53, 1057. 16 D. L. Tsalev, V. I. Slaveykova and P. B. Mandjukov, Spectrochim. Acta Rev., 1990, 13, 225. 17 A. Volynsky, Spectrochim. Acta, Part B, 1996, 51, 1573. 18 B. Welz, G. Schlemmer and J. R. Mudakavi, J. Anal. At. Spectrom., 1992, 7, 499. 19 R. E. Sturgeon, S. N. Willie, G. I. Sproule, P. T. Robinson and S. S. Berman, Spectrochim. Acta, Part B, 1989, 44, 667. 20 K. Dahl, Y. Thomassen, I.Martinsen, B. Radziuk and B. Salbu, J. Anal. At. Spectrom., 1994, 9, 1. 21 M. Burguera and J. L. Burguera, J. Anal. At. Spectrom., 1993, 8, 229. 22 K. Cai and W. McDonald, Microchem. J., 1997, 57, 370. 23 D. Pozebon, V. Dresler and A. J. Curtius, J. Anal. At. Spectrom., 1998, 13, 7. 24 G. D. Rayson and J. A. Holcombe, Anal. Chim. Acta, 1982, 136, 249. 25 H. M. Ortner, U. Rohr, S. Weinbruch, G. Schlemmer and B. Welz, 2nd European Furnace Meeting, St. Petersburg, 1996, Book of Abstracts OIV/2. 26 V. Majidi and J. D. Robertson, Spectrochim. Acta, Part B, 1991, 46, 1723. 27 B. Welz, G. Schlemmer, H. M. Ortner and W. Birzer, Spectrochim. Acta, Part B, 1989, 44, 1125. 28 T. M. Rettberg and L. M. Beach, J. Anal. At. Spectrom., 1989, 4, 427. 29 P.-y. Yang, Z.-m. Ni, Z.-x. Zhuang, F.-c. Xu and A.-b. Jiang, J. Anal. At. Spectrom., 1992, 7, 515. 30 J. P. Byrne, C. L. Chakrabarti, D. C. Gre�goire, M. M. Lamoureux and T. Ly, J. Anal. At. Spectrom., 1992, 7, 371. 31 B. Welz, G. Schlemmer and J. R. Mudakavi, Anal. Chem., 1988, 60, 2567. 32 D. L. Styris, L. J. Prell and D. A. RedÆeld, Anal. Chem., 1991, 63, 503. 33 D. L. Styris, L. J. Prell, D. A. RedÆeld, J. A. Holcombe, D. A. Bass and V. Majidi, Anal. Chem., 1991, 63, 508. 34 Y.-p. Liao and A.-m. Li, J. Anal. At. Spectrom., 1993, 8, 633. 35 W. Frech, B. V. L'vov and N. P. Romanova, Spectrochim. Acta, Part B, 1992, 47, 1461. 36 B. V. L'vov and W. Frech, Spectrochim. Acta, Part B, 1993, 48, 425. 37 G. N. Brown and D. L. Styris, J. Anal. At. Spectrom., 1993, 8, 211. 38 B. Welz, G. Schlemmer, H. M. Ortner and W. Wegscheider, Prog. Anal. Specstrosc., 1989, 12, 111. 39 U. Rohr, H. M. Ortner, G. Schlemmer, S. Weinbruch and B. Welz, Spectrochim. Acta, Part B, 1999, 54, 699. 40 E. C. Lima, F. J. Krug and K. W. Jackson, Spectrochim. Acta, Part B, 1998, 53, 1791. 41 E. C. Lima, F. J. Krug, A. T. Ferreira and F. Barbosa, Jr. , J. Anal. At. Spectrom., 1999, 14, 269. Paper 9/905415D 1742 J. Anal. At. Spectrom., 1999, 14, 1737&plu