摘要:

Observations on the Use of Ion Spray Mass Spectroscopy for Elemental Speciation in Aqueous Solutions Without Recourse to Chromatography† BARRY L. SHARPa, AZLI B. SULAIMANb , KAREN A. TAYLORa AND BRIAN N. GREENc aDepartment of Chemistry, L oughborough University, L oughborough, UK L E11 3TU bDepartment of Chemistry, Faculty of Science, Universiti T eknologi Malaysia, KB 791, 80990 Johor Bahru, Johor,Malaysia cMicromass UK L td., 3 T udor Road, Altrincham, Cheshire, UK WA14 5RZ This paper describes preliminary experiments that are selected is achieved at the expense of losing molecular information, with the limited aim of assessing whether ion spray (IS) mass particularly where organic ligands are involved.The study of spectrometry, in its current state of development, is a viable such ligands has traditionally been addressed by non-analytical technique for practical speciation without recourse to the use chemists, using, for example, conventional forms of mass of chromatography.Further, it attempts to demonstrate that spectrometry. The challenge must be to bring the two the ion spray interface, in contrast to electrospray ion approaches to bear simultaneously so that quantitative and generation, can handle, without modification, wholly aqueous qualitative information on both the inorganic and organic media as well as mixed solvents and organic solvents. The components can be obtained simultaneously. instrumentation employed for this study was an unmodified, Perversely, many problems in elemental speciation derive commercially available, triple quadrupole mass spectrometer.from the use of chromatography. Identification of species is by Thus, experiments are described which demonstrate the retention behaviour only; therefore, it is dependent on a priori capability of IS to provide speciation information on, in this information and the availability of suitable standards, and case, Ni in aqueous and methanolic media with and without there is no definitive identification of species.Only limited the presence of the chelating ligand, EDTA. Additional information is provided about the nature of the ligand and its experiments are reported for aluminium in combination with bonding to the analyte. In many cases, where the complexes citric acid. The results confirm previous work and show that are thermodynamically and kinetically labile, they will tend to the source and interface conditions can be tailored to provide dissociate on the column, precluding the determination of either molecular information or to reveal the presence of stability constants.In order to achieve ecient separation, elements as their elemental ions. Aqueous solutions are column packings and solvent systems have to be tailored to handled successfully by the IS interface without modification specific forms of the analyte so that a complete analyte budget and reveal the expected speciation patterns.However, several cannot be obtained in a single run. Analyte retention, inecient problems were encountered: the spectra, even for single separation, dilution, contamination and incompatibility of component solutions, are complex and highly sensitive to solvent systems with the source add to the problems. operating conditions; new species can be generated in the Much of the work published to date in the atomic spec- ionisation–extraction process and conditions have to be trometry literature has sought to demonstrate the viability of tailored to ions of a particular polarity; elemental sensitivity is electrospray (ES) mass spectrometry as a technique for the limited; and care has to be taken to avoid sample carry over.direct speciation of elements in solution without recourse to Given the complexity of the spectra, it is concluded that real chromatography. The purpose of this work is to try to assess samples will require prior separation by chromatographic or how far this is true and what constraints might exist in using electrophoretic techniques and that MS–MS instrumentation is the technique for practical speciation, particularly in aqueous of considerable value for confirming the identification of solutions.The experiments are of a preliminary nature, but individual species. have been selected to answer these specific questions. Keywords: Ion spray; electrospray; elemental speciation; The use of the electrospray phenomenon as a means of MS–MS; atomic spectrometry producing ions for mass spectrometry has been pioneered by Fenn and co-workers.1–5 It enables the direct transfer of ions, already present in solution, to the gas phase at atmospheric Elemental speciation has been one of the principal growth pressure.This process is often referred to as ion evaporation. areas in analytical atomic spectrometry. This reflects the Kebarle and co-workers,6 in modelling the equivalent electrical recognition that in the environmental and biological sciences, circuit of the electrospray process, characterised the ion evapor- the transport, pool dynamics and toxicology of the elements ation as an electrophoretic eect.Thus, electrospray was is dependent on their chemical speciation. Significant progress considered to be a special case of an electrolysis cell. Ions, has been made in elemental speciation by the direct coupling migrating from the surface of the evaporating drop, carry of separation techniques, notably the various forms of HPLC, charge through the gas phase (as opposed to the liquid phase) to powerful elemental detectors such as ICP-MS.The strength to the counter electrode. The excess of the complimentary of such hyphenated techniques is that they provide low limits charge that accumulates in the liquid inside the capillary can, of detection, high accuracy and precision and almost unamonce the accumulated potential exceeds ~2 V, bring about biguous identification of elemental associations.However, this oxidation or reduction reactions that enable electrons to be released to, or withdrawn from, the external circuit. † Submitted for the Guest Editor Issue on Electrospray Techniques, May 1997. The onset of the electrospray eect is critically dependent Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 (603–609) 603on the solvent properties, particularly the surface tension and except that the charge reduction to the singly charged ion the conductivity.6–9 The production of ions depends on two takes place in the gas phase.22 processes operating in tandem.Firstly, accumulation of charge This paper attempts to assess the potential of IS for practical must distort the liquid surface, leading to the formation of the speciation by considering simple metal–ligand systems in aque- Taylor cone and ultimately electrohydrodynamic nebulisation; ous only solvent and employing modern, but unmodified, this is opposed by the surface tension.Droplet separation equipment of the type used for organic mass spectrometry. results in charge separation and requires transport of charge through the body of the liquid; this is facilitated by conductive EXPERIMENTAL solutions. Ion evaporation occurs as a result of increasing charge density on the desolvating droplets. An adverse eect Ion Spray and Mass Spectrometer of the solvent is that it does introduce a degree of species The instrument used was a Micromass (Altrincham, Cheshire, selectivity into the ion production process.Although ES pro- UK) Quattro II tandem mass spectrometer equipped with a duces solvated ions into the gas phase, with subsequent nebuliser assisted electrospray (ion spray) source. The ion de-clustering, disruption of the solvation sphere is inevitable spray source (see Fig. 1) employs a 127 mm id (229 mm od) and therefore solvation energy influences the eciency of ion stainless steel capillary tube concentric within a 330 mm id production.tube of the same material. The end of the centre capillary High surface tension, low conductivity solvents, such as pure protrudes 0.5 mm beyond the end of the outer tube and a flow water, can be sprayed by appropriate capillary tip geometry10 of N2 passes in the annulus between the tubes to assist the or raising the capillary voltage;6 however, the onset of coronal nebulisation process.Potentials up to 4 kV can be applied to discharge sets an upper limit. Henion and co-workers,11–13 the inner capillary, which was fed with liquid from a Harvard following earlier work,14 largely circumvented these problems (South Natick, Massachusetts, USA) Model 22 syringe pump. by separating the particle generation and ion evaporation The spray from the needle is directed into a chamber into stages. This was achieved by configuring the electrospray which a drying flow of N2 is introduced.Aerosol and drying capillary as the central tube in a pneumatic nebuliser, gas exit the chamber through a biased (typically 0.5 kV) employing N2 as the nebuliser gas. The technique, referred to stainlesssteel counter electrode (the ‘Pepperpot’) which consists as ion spray (IS), is very much more tolerant of solvent of a circular array of holes cut at intersecting angles to the properties and liquid flow rates and allows the production of transmission axis.The spacing between the capillary tip and ions at lower capillary voltages than would otherwise be counter electrode is typically 10 mm. possible. Ion spray is the technique employed in this paper. De-clustering is controlled by a number of factors, including Electrospray and ion spray belong to the family of soft the capillary voltage, drying gas flow and temperature, but ionisation techniques that allow large and relatively fragile most significant are the pressure in the sampler/skimmer space molecules to be introduced intact into a mass spectrometer.and the de-clustering potential. In this instrument, the However, the non-thermal mechanism of ion production can, de-clustering potential (referred to as the cone voltage) is from appropriate molecules, produce multiply charged species. applied between the sampler cone and the grounded skimmer This complicates spectral interpretation for low mass ions, but cone. However, the skimmer lens is biased typically 5 V above has the considerable advantage of bringing many high relative the sampler cone and therefore the majority of fragmentation molecular mass ions into the mass-to-charge ratio range of occurs in the high field region between the skimmer lens and low cost quadrupole mass filters.15 Thus electrospray and ion the sampler cone.spray have found wide application in the study of biomolecules Ions are introduced into the instrument through a conven- such as proteins, peptides and nucleic acids16,17 and non- tional two stage dierentially pumped sampler (orifice 200 mm) covalent complexes.18,19 Other applications have included inor- and skimmer (orifice 1.5 mm) interface.The mass spectrometer ganic species,20,21 inorganic complexes,22 metalloporphoryns,23 (see Fig. 2) comprises 4 stages: an rf only hexapole to collect, anticancer drugs,24 phosphopeptides25 and pesticides.26 focus and thermalise ions from the source; an analyser quadru- The studies reported in the atomic spectrometry literature pole and associated detector; a second hexapole (to which can (with the exception of the excellent paper by Corr and Larsen27) be added a dc bias potential), which acts as a collision cell; have largely concentrated on analyte solvent interactions and and finally, a second quadrupole also equipped with its own have, in all cases, employed methanol to aid the ES–IS process.detector. The detectors used in this instrument employ photo- This it does by promoting the formation of a stable Taylor multipliers to detect light produced from secondary electron cone (through reduced surface tension) and enabling stable impact on a phosphor. The electrons are producedby accelerat- nebulisation at lower voltages. It also accelerates subsequent ing analyte ions onto an Al target.droplet evaporation. This undoubtedly leads to the production of a more uniform aerosol in which many droplets can reach the critical charge density for ion evaporation on similar time scales.However, for non-covalently bound analytes, any change in the solvent must change the speciation. This is not to say that the major analyte species in the original sample will not be represented in the spectrum, but new competitive equilibria will be established whose influence is uncertain. Previous work28 has shown that ES–IS sources produce in the gas phase metal ion–solvent clusters and these tend to dominate the spectra under low energy conditions.Raising both the number and energy of the collisions that the analyte species experience before mass analysis results in a gradual de-clustering until a point where bare metal ions can be observed in the spectrum. De-clustering follows the typical hydrolysis series that would be observed for species in solution, e.g., Fig. 1 Nebuliser assisted electrospray (ion spray) source, counter electrode and vacuum interface for the Quattro II mass spectrometer. M(H2O)nm+�M(H2O)n-1(OH)(m-1)++H+ et seq. 604 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12Fig. 2 Schematic diagram of the Quattro II mass spectrometer. The instrument can be operated in a variety of modes. spectrum obtained under very soft conditions with the declustering voltage set at 35 V. The spectra are dominated by Conventional MS experiments are carried out using only the first quadrupole and detector. MS–MS experiments are per- solvent clusters and none of the major peaks exhibit a plus 2 peak of the right intensity that would indicate the presence of formed either by fixing the first quadrupole transmission and scanning the second (product ion scan) or fixing the second Ni.The series of peaks m/z 57, 75, 93 and 115 occurred in many of the spectra with varying relative intensity. They and scanning the first (precursor ion scan). A further option allows simultaneous scanning of both quadrupoles (neutral appear to be water clusters, although the base peak at m/z 57 remains unidentified.The peak at m/z 115 may be the sodium loss scanning) to detect the loss of neutral fragments. Mass scanning up to m/z 4000 is available. adduct of a protonated cluster occurring at m/z 93. Raising the de-clustering voltage to 150 V yielded the spectrum in The operating parameters, unless otherwise stated, are those shown in Table 1. Data acquisition and presentation was Fig. 3(b), which shows that the Ni is now present almost entirely as elemental ions. The peak at 108 was observed in accomplished using VG MassLynks software. many spectra and corresponds to Na2NO3+; it is usually accompanied by a smaller peak at m/z 193, which is the adduct Chemicals Na2NO3 (NaNO3 ); higher adducts have also been observed. Looking at the smaller peaks in the spectrum obtained at a Working solutions were prepared by dilution of appropriate cone voltage of 35 V, e.g., at m/z 179, some of them do have standards using ultra-pure water (18 MV) from a laboratorythe correct isotope ratio for Ni.To confirm this an MS–MS reagent grade water system (Elga, Maxima, High Wycombe, experiment was performed in which the second quadrupole Buckinghamshire, UK). The Ni and Al solutions (as nitrates) was set to m/z 58 and the first scanned (precursorion spectrum). were prepared by dilution of commercial standards (Spectrosol, Ar at 3.5×10-3 mbar was used as the collision gas and the Merck Ltd., Poole, Dorset, UK).EDTA solutions were precollision energy was 70 V. The resulting spectrum is shown in pared from the diammonium salt (Puriss, Fluka, Gillingham, Fig. 3(c) and shows a series of peaks in the range m/z 138–212. Dorset, UK) and those of citric acid from anhydrous citric At first, allocation of these was uncertain, but it transpired acid (Microselect, Fluka, Gillingham, Dorset, UK). that the instrument had been extensively used with acetonitrile solvent and the source was undoubtedly contaminated.This RESULTS AND DISCUSSION is illustrated by the peak at m/z 42, which is CH4CN+, the protonated form. Thus, the spectra fall into two series of which Metal Ion and Metal–Ligand Solutions Under Varying NiNO3+ is the common component. The following assign- Conditions ments are made: Nickel nitrate in aqueous solution m/z 138—58NiNO3H2O; 15mdash;58NiNO3(H2O)2; The initial experiments were carried out with unbuered m/z 174—58NiNO3(H2O)3 aqueous solutions of 10-3 M Ni(NO3)2.Fig. 3(a) shows the m/z 161—58NiNO3CH3CN; 202—58NiNO3(CH3CN)2 m/z 179—58NiNO3CH3CN(H2O); 197— Table 1 Instrumental conditions 58NiNO3CH3CN(H2O)2 Source and interface— and for each of these the corresponding peaks at m/z+2 due Electrospray capillary 2.5–3.0 kV Flow rate 5 ml min-1 to 60Ni are also present in the original spectrum [Fig. 3(a)]. Gas (N2 ) flow rate 0.25 l min-1 The spectrum presented in Fig. 3(a) is scaled to the strongest Drying gas (N2) or curtain flow rate 1.67 l min-1 peak and therefore some of the components listed above are Gas temperature 65 °C very weak as displayed. The assignment for m/z 197 is tentative because the 1975199 ratio is not correct for Ni. No assignment Sampler–skimmer cone pressure— is made for the peak at m/z 212. To confirm the presence of 0.8 mbar, increased to 1.2 mbar for the Al–citric acid study adducts due to acetonitrile, a neutral loss scan was performed Collision cell— in which both quadrupoles were scanned with a mass dierence Collision energy 70 V of 41.The spectrum is shown in Fig. 3(d) and confirms Gas pressure 1.5–5.0 10-3 mbar particularly the association of m/z 161 and 179, and to a lesser Collision gas Ar extent 202 with CH3CN. However, the principal peaks are at Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 605Fig. 3 (a), Positive ion spray spectrum of 10-3 M solution of unbuered Ni(NO3)2.Cone voltage 35 V. (b), Positive ion spray spectrum of 10-3 M solution of unbuered Ni(NO3 )2. Cone voltage 150 V. (c), Positive precursor ion MS–MS spectrum for m/z 58 of 10-3 M solution of unbuered Ni(NO3)2. Cone voltage 35 V, collision energy 70 V. (d) Neutral loss positive ion MS–MS spectrum for mass 41 of 10-3 M solution of unbuered Ni(NO3)2. Cone voltage 35 V, collision energy 70 V. (e), Positive product ion spectrum of m/z 179 of 10-3 M solution of unbuered Ni(NO3)2.Cone voltage 35 V, collision energy 70 V. ( f ), Positive product ion spectrum of m/z 181 of 10-3 M solution of unbuered Ni(NO3)2. Cone voltage 35 V, collision energy 70 V. (g), Positive ion spray spectrum of 10-3 M solution of unbuered Ni(NO3)2 in 50% MeOH–H2O solution. Cone voltage 35 V. (h), Positive ion spray spectrum of 10-3 M solution of unbuered Ni(NO3)2 in 50% MeOH–H2O solution. Cone voltage 150 V. 115/117 and 64. The last is almost certainly the sodium adduct When a high de-clustering potential was employed, the spectrum was almost identical with that obtained with water only with acetonitrile, sodium being ubiquitous in many of the spectra.The peaks at m/z 115/117 are most interesting and solvent, as might be expected, as the speciation information is lost. Since no adjustment was made to the instrument when appear to be associated with Ni. A possible assignment is 58/60NiOCH3CN+. Finally, for completeness, the product ion changing solvents, it does demonstrate the capability of the IS source in overcoming the problems of solvent compatibility.spectra of m/z 179/181 are shown in Figs. 3(e) and ( f ), clearly illustrating the association of these peaks with 58Ni and 60Ni, respectively. Note, however, that under the conditions used, the oxide peaks (m/z 74/76) are dominant. Nickel nitrate–EDTA system The next step was to look at the eect of adding a chelating Nickel nitrate in methanolic solution ligand and in the first instance EDTA was chosen, which is One of the motivations for carrying out this work was to known to form very stable complexes (Kf=4.2×1018 ) with Ni, ascertain whether it is possible to work with water without even at low pH values.The pressure in the sampler/skimmer the addition of methanol. Comparative spectra for the lower cone region was slightly higher for these experiments, 1.2 as and higher de-clustering potential are shown for a 50% compared with 0.8 mbar previously, which would lead to less methanol–water mixture in Figs. 3(g) and (h). Not surprisingly, tendency to fragmentation because the collision energy is the speciation, as indicated for the low de-clustering voltage reduced. Figs. 4(a) and (c) show the positive and negative ion spectra, are quite dierent. The solvent peaks at m/z 57 and spectra, respectively, of 10-4 M EDTA. The first point of note 61 remain, but the dominant peaks associated with Ni now is the simplicity of the negative ion spectra compared with appear at: those obtained in positive ion mode.The most readily identifi- able species in the positive spectrum are [EDTA+H]+ and m/z 152/154—58/60NiNO3MeOH; 184/186— 58/60NiNO3(MeOH)2; [EDTA+Na]+ at m/z 293 and 315, respectively. Both have m/z+1 ions of about 10% of the base peak intensity, which is m/z 216/218—58/60NiNO3(MeOH)3 606 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12consistent with a 10 carbon molecule such as EDTA.Reference ation of H2O would yield m/z 99, but further work would be necessary to have confidence in such assignments. to the literature on protein MS shows that Na readily adducts across adjacent carbonyl oxygens in the peptide link29,30 and it appears that protons and Na are behaving similarly with Aluminium/aluminium–citric acid system EDTA. The point to note here is that new species have been formed that are not representative of those found in Similar observations were made when citric acid was chosen as the ligand.Again, the pressure in the sampler/skimmer cone the solution phase. Contrast this with the negative ion spectra, where predictably [EDTA-H]- at m/z 291 and region was increased slightly to 1.2 mbar, which provided softer conditions for the weaker citrate complexes. Figs. 5(a) [EDTA-2H]2- at m/z 145 are present. Even here, however, new species are created with a dominant and unidentified peak and (b) show the positive and negative ion spectra for 10-4 M citric acid (C6O7H8), respectively. The base peak in the positive occurring at m/z 344.The eect of adding equimolar Ni2+ to the EDTA is shown, spectrumcorresponds to the Na adduct at m/z 215. A secondary peak at m/z 216 of about 6% of the base peak intensity occurs, for positive ion mode, in Fig. 4(b) and for negative ion mode in Fig. 4(d ). In the case of the positive ion spectra, the presence as would be expected for a molecule containing 6 carbons.The negative ion spectrum is very simple, being dominated by of the divalent metal completely removes the less stable adducts to yield [58/60Ni(EDTA–H)]+ at m/z 349 and 351. The fragmen- [(C6O7H8–H)]- at m/z 191 and [(C6O7H8–2H)]2- at m/z 95. Addition of 10-4 M Al3+ as nitrate yields the positive ion tation pattern is also changed with a reduction in the relative magnitude of m/z 163, but an enhancement of m/z 257. Looking spectra shown in Figs. 5(c) and (d). As might be expected in these acidic conditions, complex formation is not particularly at the negative ion spectra, the free ligand at m/z 291 and the unidentified peak at m/z 344 have disappeared and the spec- favoured and the protonated ligand (m/z 193) becomes the dominant peak in the spectrum. The most likely Al–citric acid trum is dominated by [58/60Ni(EDTA–3H)]- at m/z 347 and 349. Once again the fragmentation pattern is influenced by the associations are found at m/z 253 [C6O7H8Al(OH)2]+ and 298 [C6O7H8Al(NO3)(OH)]+.The Al, as might be expected, metal ion with the appearance of peaks at m/z 196 and 244. It might be expected that the dominant species in solution would occurs mainly as weakly hydrolysed products, the series m/z 115 [Al(OH)2(H2O)3 ]+, m/z 133 [Al(OH)2(H2O)4]+, m/z 142 be Ni(EDTA)2- (m/z 173/174), but under the conditions employed this does not appear. [Al(NO3)(OH)(H2O)2]+, m/z 160 [Al(NO3)(OH)(H2O)3]+ and m/z 178 [Al(NO3)(OH)(H2O)4]+ being just apparent in To confirm the presence of Ni in the peaks assigned to the Ni–EDTA complexes, MS–MS experiments were carried out the spectrum, particularly at the lower cone voltage.in which product ion spectra were obtained for m/z 349 and 351. The spectra are shown in Figs. 4(e) and ( f ), respectively. CONCLUSIONS The bare metal ions are clearly present and further any pattern that repeats in the second spectrum with a two mass unit IS–MS implemented on a modern instrumental platform is, as demonstrated here, capable of providing a wealth of chemical separation will almost certainly contain Ni.For example, m/z 117 could correspond to [58Ni(CH2COOH)]+ and an elimin- information. Anyone who has studied solution chemistry Fig. 4 (a), Positive ion spray spectrum of 10-4 M solution of EDTA as the diammonium salt. Cone voltage 25 V. (b), Positive ion spray spectrum of 10-4 M solution of EDTA as the diammonium salt with the addition of 10-4 M Ni(NO3 )2.Cone voltage 25 V. (c), Negative ion spray spectrum of 10-4 M solution of EDTA as the diammonium salt. Cone voltage 25 V. (d), Negative ion spray spectrum of 10-4 M solution of EDTA as the diammonium salt with the addition of 10-4 M Ni(NO3)2. Cone voltage 25 V. (e), Positive product ion spectrum of m/z 349 of 10-4 M solution of EDTA as the diammonium salt with the addition of 10-4 M Ni(NO3)2. Cone voltage 35 V, collision energy 70 V. ( f ), Positive product ion spectrum of m/z 351 of 10-4 M solution of EDTA as the diammonium salt with the addition of 10-4 M Ni(NO3)2.Cone voltage 35 V, collision energy 70 V. Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 607Fig. 5 (a), Positive ion spray spectrum of 10-4 M solution of citric acid. Cone voltage 20 V. (b), Negative ion spray spectrum of 10-4 M solution of citric acid. Cone voltage 20 V. (c), Positive ion spray spectrum of 10-4 M solution of citric acid with the addition of 10-4 M Al(NO3)3 .Cone voltage 20 V. (d), Positive ion spray spectrum of 10-4 M solution of citric acid with the addition of 10-4 M Al(NO3)3 . Cone voltage 10 V. cannot fail to be excited by seeing the species, for which one and the ones which most closely reflect the expected speciation are obtained when, as might be anticipated, the source polarity can write sensible chemical equations, appearing in the mass spectra. However, one must draw back from this evangelical matches that of the solution ion.Thus, in the work presented here, where positively charged metals are complexed with zeal and ask whether, if faced by a requirement for routine quantitative analysis on complex samples, could the technique organic ligands with -COOH and -OH functional groups (representative of a very large range of naturally occurring be a viable option. In its present state of development, the answer is clearly no. ligands), the negative ion mode provides the simplest and most useful information.The experiments described above, though of a preliminary nature, enable some general conclusions to be drawn. The limits of detection for the ES–IS technique are generally in the nM region for organic molecules, but the detection limits for low relative molecular mass inorganic species do not appear Potential Advantages of Electrospray–Ion Spray to be so low. In our work, detection limits for metal containing species were in the mM range, which for Ni would correspond The source is relatively cheap and simple to build.Conditions to a limit of about 60 ppb. in the source and interface can be varied to preserve molecular Problems were encountered in this work with contamination, information or to provide elemental information. Because of for example from acetonitrile and sodium. Although this can multiple charging, high relative molecular mass compounds, obviously be overcome, it is to be expected that very low flow e.g., metalloproteins, can be studied.A wide variety of flow techniques will not be as robust with respect to sample rates, from nl min-1 to ml min-1, can be accommodated by carryover as conventional atomic spectrometric techniques the IS technique. Electrospray alone is limited to flow rates in operating on millilitre per minute flows. This must impact on the ml min-1 range. A further advantage of IS compared with expectations of using ES–IS for high throughput speciation ES, at least as implemented on the Quattro instrument, is that studies.the solvent system can be changed from a purely aqueous to Perhaps the major conclusion that can be drawn from this an organic medium with no change in the instrumental work is that, in its current state of development, ES–IS is parameters and with entirely predictable consequences for the complementary to, but not a replacement for, existing tech- speciation and the resultant spectrum.niques. Real samples almost certainly require a prior separation by chromatographic or electrophoresis techniques to enable Disadvantages of Electrospray/Ion Spray simplification and interpretation of the ES–IS spectra. Even with prior separation, there are considerable benefits in The principal disadvantage of the technique is that the spectra working with MS–MS instruments which aid in identification are very complex, even for single component systems (multiply and provide additional information on molecular structure.charged species may also be present). Furthermore, the Unfortunately, this does rather obviate some of the benefits of observed spectra are extremely sensitive to the experimental using a relatively cheap source. The principal application for conditions. This is at once a major benefit and a serious ES–IS, as demonstrated in the recent papers of Corr and drawback. Provided that the sample contains only a few species co-workers,22,28 is likely to be in confirming the identity of of known origin, conditions can be tailored to provide elemenseparated species prior to their quantification by existing tal or molecular information.However, extrapolating to a real techniques. sample with many unknown components, the spectra would be dicult if not impossible to interpret. Changing the experimental conditions under these circumstances would not necessarily yield useful information, other than in the bare metal REFERENCES ion mode, where the speciation has been lost.In this respect the technique is far removed from the ideal of a transparent 1 Yamashita, M., and Fenn, J. D., J. Phys. Chem., 1984, 88, 4451. 2 Yamashita, M., and Fenn, J. D., J. Phys. Chem., 1984, 88, 4671. transducer that provides unperturbed information on the 3 Whitehouse, C. M., Dreyer, R. N., Yamashita, M., and Fenn, sample. J. D., Anal. Chem., 1985, 57, 675. A further diculty is that new species are created in the gas 4 Wong, S.F., Meng, C. K., and Fenn, J. D., J. Phys. Chem., 1988, phase that do not necessarily appear in solution. There are 92, 546. clearly diculties in trying to study at one time species that 5 Fenn, J. D., Mann, M., Meng, C. K., and Wong, S. F., Mass Spectrom. Rev., 1990, 9, 37. exist in solution in dierent polarities. The simplest spectra 608 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 126 Ikonomou, M. G., Blades, A. T., and Kebarle, P., Anal.Chem., 20 Siu, K. W. M., Gardner, G. J., and Berman, S. S., Anal. Chem., 1991, 63, 1989. 1989, 61, 2320. 7 Smith, D. P. H., IEEE T rans. Ind. Appl., 1986, IA-22, 527. 21 Corr, J. J., and Anacleto, J. F., Anal. Chem., 1996, 68, 2155. 8 Tang, L., and Kebarle, P., Anal. Chem., 1991, 63, 2709. 22 Blades, A. T., Jayaweera, P., Ikonomou, M. G., and Kebarle, P., 9 Agnes, G. R., and Horlick, G., Appl. Spectrosc., 1994, 48, 649. J. Chem. Phys., 1990, 92, 5900. 10 Chowdhury, S. K., and Chait, B. T., Anal. Chem., 1991, 63, 1660. 23 Van Berkel, G. J., McLuskey, S. A., and Glish, G. L., Anal. Chem., 11 Bruins, A. P., Covey, T. R., and Henion, J. D., Anal. Chem., 1987, 1992, 64, 1586. 59, 2642. 24 Poon, G. K., Bisset, G. M. F., and Mistry, P., J. Am. Soc. Mass. 12 Covey, T. R., Bonner, R. F., Shushan, B. I., and Henion, J. D., Spectrom., 1993, 4, 588. Rapid Commun. Mass Spectrom., 1988, 2, 249. 25 Huddlestone, M. J., Annan, R. S., Bean, M. F., and Carr, S. A., 13 Huang, E. C., and Henion, J. D., J. Am. Soc. Mass Spectrom., J. Am. Soc. Mass. Spectrom., 1993, 4, 710. 1990, 1, 158. 26 Lin, H. Y., and Voyksner, R. D., Anal. Chem., 1993, 65, 451. 14 Mach, L. L., Kralik, P., Rheude, A., and Dole, M., J. Chem. Phys., 27 Corr, J. J., and Larsen, E. H., J. Anal. At. Spectrom., 1996, 11, 1215. 1970, 52, 4977. 28 Stewart, I. I., and Horlick, G., J. Anal. At. Spectrom., 1996, 15 Hofstadler, S. A., Bakhtiar, R., and Smith, R. D., J. Chem. Educ., 11, 1203. 1996, 73, A82. 29 Tang, X. J., Ens, W., Standing, K. G., and Westmore, J. B., Anal. 16 Smith, R. D., Loo, J. A., Edmonds, C. G., Barinaga, C. J., and Chem., 1988, 60, 1791. Udseth, H. R., Anal. Chem., 1990, 62, 882. 30 Grese, R. P., Cerny, R. L., and Gross, M. L., J. Am. Chem. Soc., 17 Burlingame, A. L., Baillie, T. A., and Russel, D. H., Anal. Chem., 1989, 111, 2835. 1992, 64, 467. 18 Ganem, B., Li, Y. T., and Henion, J. D., J. Am. Chem. Soc., 1991, Paper 7/01420A 113, 6294. Received February 28, 1997 19 Schwartz, B. L., Light-Wahl, K. J., and Smith, R. D., J. Am. Soc. Mass. Spectrom., 1994, 5, 201. Accepted April 17, 1997 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 609

ISSN:0267-9477    

DOI:10.1039/a701420a

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

At-line Determination of Mercury in Process Streams Using Atomic Fluorescence Spectrometry† NOEL K. BRAHMAa , WARREN T. CORNSa , PETER B. STOCKWELLa, LES EBDONb AND E. HYWEL EVANS*b aP.S. Analytical L td, Arthur House, Unit 3 Crayfields Industrial Estate, Orpington, Kent, UK BR5 3HP bUniversity of Plymouth, Department of Environmental Sciences, Drake Circus, Plymouth, Devon, UK PL 4 8AA Atomic fluorescence spectrometry (AFS) has been coupled It can be seen from eqn. 2 that Hg0 is capable of reducing with vapour generation sample introduction for the at-line Hg2+ to Hg22+, as the electrode potential favours the reverse determination of mercury in concentrated sulfuric acid reaction.This principle is sometimes used to remove mercury production streams. Matrix and heat eects were overcome by from the roast gas in the contact process. However, it should the use of discrete sample injection techniques. A sample also be noted that any removal of Hg2+ by precipitation of collection and pre-treatment manifold is described in insoluble compounds, or Hg0 by outgassing, will cause the combination with permanganate oxidation and SnCl2 reduction reaction to proceed in the forward direction.chemistries for coupling with a fully automated instrument. Although mercury removal rates of greater than 95% are Continuous monitoring of single grab samples of concentrated obtained with such processes, some of the mercury is still sulfuric acid over periods of 24 and 15 h resulted in precisions transferred with the roast gas to the adsorber towers and is of 5.8 and 8.9% RSD and mean mercury concentrations of present in the final product acid.During this process it is 1.11 and 1.08 mg g-1, respectively, compared with 1.2 and unavoidable that some mercury compounds (Hg0, Hg2Cl2, 0.95 mg g-1 obtained by a laboratory based method. The HgCl2) are also transferred with the wet roast gas to the application of the instrument in on-site trials at a sulfuric acid reaction chamber.plant is described. The mercury concentration in sulfuric acid The proportions of mercury species in the roast gas after was monitored continuously over a period of 10 d without scrubbing are unknown. One would assume that mercury(0) intervention and showed excellent agreement with results of and mercury(I) would be oxidised to mercury(II) during the laboratory analyses performed over the same period. production of sulfuric acid. If the mercury is present as Hg0, HgI and HgII species it will be necessary to oxidise all the HgI Keywords: Mercury cold vapour; sulfuric acid; atomic and Hg0 to HgII in order to achieve a total mercury result.fluorescence spectrometry; at-line determination ; sample When the oxidant is in excess, complete oxidation to mer- pre-treatment cury(II ) species is possible. O-line, this oxidation would normally be performed using an oxidant such as potassium Sulfuric acid has many industrial applications ranging from permanganate, potassium persulfate or bromine.After oxi- fertiliser production to food processing. Consequently, users dation, the resulting HgII species can readily be reduced using of acid have set stringentquality criteria specifically for mercury an acidic solution of tin(II) chloride to give a total mercury which need to be met by their acid suppliers. At present acid result. process streams at smelting plants are monitored manually The toxic eects of mercury compounds on environmental several times a day for mercury, requiring that a sample be systems have been well established over the last few decades, collected and analysed in the laboratory.This is laborious, resulting in many techniques for their detection. Originally, prone to contamination and time consuming. It can often take colorimetric and spectrophotometric methods, which were up to a day before the mercury concentration in the acid is based on mercury ions forming chelates with dithizone and known.During this delay period several hundred tonnes of dinaphthylthiocarbazone, were used.1 These methods are acid may have been produced which do not meet the requi- troublesome and suer from contamination and poor sensi- site criteria. tivity. Today, the most commonly used method for determining The presence of mercury in sulfuric acid is due to the mercury is cold vapour atomic absorption spectrometry association of mercury(II) sulfide (cinnabar) with the sulfide (CVAAS), a method originally developed in 1963 by Poluetov bearing ores.When the sulfide ore is roasted in air at 600 °C and Vitkun.2 It was not until 1968 that Hatch and Ott3 refined the cinnabar decomposes to produce mercury vapour and and popularised this method; since then many modifications sulfur dioxide: have been made to this principle. The atomic fluorescence technique is particularly good for HgS+O2�Hg0+SO2 (1) the determination of mercury as it is an atomic vapour at The subsequent roast gas can then be passed over activated room temperature and undergoes resonance fluorescence. carbon filter beds to remove the mercury vapour.In some West,4 in 1974, showed theoretically that cold vapour atomic production plants the roasted gas is purged through a solution fluorescence spectrometry (CVAFS) should be more sensitive of mercury(II) chloride to form mercury(I) chloride, which and produce considerably less spectral interference from non- precipitates, thus removing the majority of the mercury from specific absorption compared with the corresponding AAS the gas.For the disproportionation equilibrium: technique. Thompson and Reynolds5 were the first to use the cold Hg22+<Hg0+Hg2+ EG=-0.131 V (2) vapour technique for the determination of mercury in 1975. They modified a conventional atomic absorption spectrometer. More recently Godden and Stockwell6 developed a non- † Presented at the Eighth Biennial National Atomic Spectroscopy Symposium (BNASS), Norwich, UK, July 17–19, 1996.dispersive fluorescence spectrometer specifically for mercury Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 (611–616) 611Table 1 Instrumental conditions for PSA 10.223 at-line mercury determination. Their system was based on a conventional analyser molecular fluorescence spectrometer coupled to a vapour generation process. The detector arrangement of Godden and Carrier gas 300 ml min-1 argon Stockwell6 was simple in design and rugged in construction Sheath gas 300 ml min-1 argon and therefore an ideal starting point for the translation of a Dryer gas 2500 ml min-1 air Autozero On laboratory instrument into a process analyser. Timer On There are several potential problems which are likely to be Timer interval 15.00 min encountered during the determination of mercury in concen- Filter factor (smoothing) 4 trated sulfuric acid.The first is the diculty of obtaining a Runs 1 continuous sample for analysis which is representative of the Fit type Least squares straight line main stream. This has to be delivered to the instrument with Gain 1000×4.0 Delay 120 s the minimum of delay, without significant change in composi- Measure 150 s tion or physical state, and at a level of cleanliness with which Memory 120 s the analyser can cope. Loop size 300 ml The viscosity, corrosiveness and temperature of the acid Sample flow rate 0.7 ml min-1 must also be taken into account when designing such a system Reductant flow rate 0.7 ml min-1 because hot sulfuric acid is extremely oxidising.The instrument must be capable of at-line digestion, automatic calibration and sensitive detection, as well as full data management. Ideally Sequence of events the system should also have some sort of error monitoring With the standard instrument set up shown in Fig. 1 the system.The long term dependability and stability required for stream selection device was fitted with four valves, any one of process analysers is not usually encountered in laboratory which could be selected at a given time. The sample/standard techniques, so an accurate and reliable analytical instrument was pumped to a six-port discrete injection valve where a fixed monitoring a high-throughput process can be a valuable asset. volume of the sample was introduced to the mixing nifold. An inaccurate, unreliable instrument can be a menace which At the mixing manifold chemical pre-treatment was performed can lead to losses through lower eciency or, worse still, whereby the sample was mixed first with an oxidant to convert through plant shutdown.all the available mercury into the divalent state, Hg2+, then A bulk chemical manufacturer was approached to investigate the mercury was reduced to the gaseous phase, Hg0, by the possibility of configuring an at-line system for the determireduction with tin(II) chloride.The mixing coils were 0.8 mm nation of mercury in sulfuric acid. The manufacturer currently id and 200 and 150 cm long for the oxidation and reduction, analyses its acid stream three times a day manually, using respectively, resulting in respective residence times of 86 and CVAAS. 65 s. The gaseous mercury was stripped from solution using The sulfuric acid sample is 96% m/m H2SO4, has a density the gas–liquid separator in the presence of a carrier gas, and of 1.821 g cm-3, has an SO2 dissolved gas content of approxithen passed through a hygroscopic membrane dryer (Nafion, mately 0.1% m/v and is free from particulate matter.The Perma-Pure Products, Farmingdale, NJ, USA). The dried temperature of the acid at the point of sampling is around sample gas was then introduced to a fluorescence detector for 55°C. The acid contains approximately 10 mg l-1 of Fe (from determination. The detector returned a transient signal which the stainless steel tubing used in the piping) and less than was forwarded to the computer in BCD code, and then 1 mg l-1 of As, Cd, Cr, Cu, Mg, Mn, Ni, Pb and Zn.transmitted via a DAC output to the process control room through the system’s Touchstone software. EXPERIMENTAL Instrumentation Modes of operation Basic features of at-line system With an analysis program such as that shown in Table 2, the instrument can be configured into almost any analysis An at-line system was specifically designed to determine mersequence.The at-line instrument has two modes of operation. cury in liquid samples with full process and data control via The first is when the instrument performs one calibration daily an IBM compatible computer. The software allows the operand then analyses the sample repeatedly over the day against ator to fully customise the instrument to suit the application; that single calibration, which is a feasible mode of operation automatic calibration and analysis protocols can be fully given the excellent stability of the system, as demonstrated programmed within the software.The unit also employs an later. The second mode of operation is when the instrument error monitoring system which checks for the flow of sample, continually re-calibrates itself between samples. reagents and gases, as well as for leaks within the system. Should any of the reagents or gases fail to be present an alarm will sound and the instrument will shut down.The main Reagents components of the system are a stream selection valve, a peristaltic pump, a six-port switching valve, a gas–liquid separ- Standards were prepared using a mercury(II) stock solution (SpectrosoL, Merck, Poole, UK) in both 10% (v/v) and ator, an atomic fluorescence spectrometer and a computer fitted with a 4–20 mA DAC output. The instrumental con- concentrated sulfuric acid (AnalaR, Merck). The reductant was prepared from tin(II) chloride dihydrate (AR, Aldrich, ditions typically used during at-line process analysis are shown in Table 1.Gillingham, UK), a 4% m/v solution was prepared in 30% v/v HCl (AR, Fisons, Loughborough, UK). The reductant was purged with argon (Pureshield, BOC, Guildford, UK) for Sample collection approximately 20 min in order to remove any residual mercury vapour. The oxidant was either a 0.1 M potassium bromate– The sample was collected either via a fast loop principle from the flowing process stream or by sub-sampling a constantly potassium bromide solution (ConvoL, Merck) or a 5.0% m/v solution of potassium permanganate (SpectrosoL low in Hg, replenished over-flow vessel.Hence, a fresh representative sample of the process stream was delivered to the instrument Merck) in de-ionised water (Option 3, Elga, High Wycombe, UK). with minimum delay and no change in composition. 612 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12Fig. 1 Schematic layout of the at-line atomic fluorescence instrument for the analysis of sulfuric acid. Table 3 Instrument settings for PSA 10.023 Merlin detector and PSA Table 2 Analysis program used for laboratory trials using the bromination chemistry 10.004 vapour generator used for laboratory analysis Carrier gas 300 ml min-1 argon Line Valve Tag Sheath gas 300 ml min-1 argon 1 B Std 1, Blank 10% (v/v) H2SO4 Sample/blank flow rate 8.0 ml min-1 2 C Std 2, 10 mg ml-1 Hg Reductant flow rate 4.0 ml min-1 3 D Std 3, 50 mg ml-1 Hg Amplification 1×7.5=7.5 (1–10 000) 4 E Sample ??? mg ml-1 (Sample, 10% v/v) Dryer gas 2500 ml min-1 air 5 E Sample ??? mg ml-1 (Sample, 10% v/v) Filter factor (smoothing) 32 6 E Sample ??? mg ml-1 (Sample, 10% v/v) Delay 10 s 7 E Sample ??? mg ml-1 (Sample, 10% v/v) Rise 30 s 8 A (o ) Time-loop back to line 1 Measure 30 s Memory 60 s Method Development RESULTS AND DISCUSSION To determine the ecacy of various oxidation methods the Method Development following sample pre-treatments were performed.(a) To analyse The results for the oxidised and un-oxidised sample are shown the sample at dierent dilutions, without any oxidation stage, in Table 4. These results suggest that around 50% of the using the standard additions approach. Omitting the oxidation mercury in the sample is likely to be present as mercury(I ). stage produces a Hg0 and HgII result only because HgI will The 1% v/v dilution gave exactly the same value as that of not be readily reduced to Hg0. (b) To analyse the sample using the 10% v/v solution (10.5 mg ml-1 Hg), suggesting that the the proposed two oxidation methods, again with the standard sulfuric acid did not cause a matrix interference in the mercury additions approach (at 1 and 10% dilutions).Oxidising the determination. sample will produce a total mercury result. The bromination pre-oxidation resulted in a mercury con- The acid sample was first analysed without any oxidation centration of 17.6 mg ml-1 Hg, which was 92% of the manufac- at a dilution of 1 and 10% v/v.Mercury standards were prepared in the same matrix at 20, 40 and 60 ng ml-1 Hg. Next the sample was analysed using potassium bromate– Table 4 Results for the comparison of oxidation sample pre- bromide as the oxidant. To each acid-cleaned 100 ml calibrated treatment methods flask, 2 ml of 0.1 M potassium bromate–bromide solution was Hg concentration in added.Standard additions was again employed at the same concentrated H2SO4/ level as before (20, 40 and 60 ng ml-1 Hg). Each flask was mg ml-1 then diluted to the mark with either a 1 or 10% solution of the acid sample. The flasks were then left to stand for approxi- Oxidation pre- Manufacturer’s mately 20 min. Prior to analysis each flask was pre-reduced treatment method This work value Recovery (%) with a few drops of hydroxylamine hydrochloride to convert 10% sample, no 10.4±0.04 19.0±1.0 54 the excess bromine to bromide.The samples were then analysed oxidation in the same manner as the un-oxidised sample.Instrumental 1% sample, no 10.5±0.05 19.0±1.0 55 oxidation conditions are shown in Tables 2 and 3. 17.6±0.06 19.0±1.0 92 10% sample, Lastly, the samples and standards were prepared by the KBr–KBrO3 oxidation same method as for bromination, except that the bromate– 10% sample, KMnO4 18.1±0.03 19.0±1.0 95 bromide was replaced with 5.0 ml of 5.0% m/v KMnO4 oxidaton solution as the oxidant.Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 613turer’s value (Table 4). This was probably due to the oxidation O-line Site Trial of HgI to HgII, or to the removal of other interfering species Having configured the system as described above, and using by oxidation. The potassium permanganate pre-oxidation permanganate as the pre-oxidant, it was installed at a zinc resulted in higher recoveries than those obtained without smelting plant.Two separate production streams of sulfuric oxidation, with a mercury concentration of 18.1 mg ml-1 acid required monitoring for mercury, namely the final product (Table 4) being obtained, which was similar to that obtained ‘white’ acid and the un-bleached ‘black’ acid, the black colour by the bromination technique. being caused by carbon in the sulfur dioxide gas used to The results of the o-line investigations indicated that it was produce the acid.necessary to pre-oxidise the sulfuric acid sample in order to obtain close to 100% recovery for Hg. It is probable that preoxidation converted the HgI species to the Hg2+ ion, which is Continuous recalibration mode readily reducible with an acidic solution of tin(II) chloride. It A single sample of concentrated ‘white’ sulfuric acid was would appear that around 50% of the mercury present in the obtained from the production line and analysed continuously sulfuric acid sample was present as a HgI salt, most probably for 15 h (36 analyses) using the at-line system operated in Hg2SO4 .continuous recalibration mode, as shown in Fig. 3. This sample The next stage was to transfer the chemistries and the sample was also analysed using laboratory-based CVAAS for compari- to the PSA 10.223 at-line mercury analyser for further son. Mean results are shown in Table 5 and excellent agreement development. was obtained with the laboratory result.The standardisation procedure was also very stable, with the 0.543 and 1.358 mg Laboratory Trial The at-line monitoring system was initially tested in the laboratory using a grab sample of concentrated sulfuric acid from a zinc smelting plant which was diluted to 10% v/v. During this trial the tests were performed using an acidified solution of potassium bromate–bromide as the oxidant. Fig. 2 shows the results of continual re-calibration and analysis. This resulted in a relative standard deviation (RSD) of 1.36% for the sample and less than 8.0% for the standards over the test period of 36 h.In three instances the sample peak height was greater than that of the top standard. The greatest variation appeared to be in the blank solution which was 10% v/v sulfuric acid, which may have been due to carry-over from the Fig. 3 Results of o-line site trial for the determination of Hg in ‘white’ sulfuric acid, showing the stability of the instrument when using top standard.the continuous re-calibration mode. It had previously been necessary to add hydroxylamine hydrochloride after the pre-oxidation stage in order to reduce Table 5 Comparison of continuous mercury analysis with batch mode excess oxidant; however, this was also eliminated by increasing analysis for o-line site trial the concentration of the tin(II) chloride solution from 2 to 4% so that it was in excess. Sample dilution was further reduced Hg concentration/mg g-1 by using the oxidant stream as the sample carrier stream.These changes also improved the S/N. The bromination oxi- Continuous AFS Laboratory CVAAS dation gave rise to a noisier signal than the permanganate Continuous recalibration 1.08±0.1 (n=36) 0.95 oxidation, so it was decided to use the latter method for the Daily calibration 1.11±0.06 (n=77) 1.2 site trials. Fig. 2 Graph showing the variation in peak height % in the continuous re-calibration mode. 614 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12g-1 Hg standards giving RSDs of 3.2 and 7.3%, respectively. The instrument was calibrated ten times over the analysis period. Daily calibration mode The stability of the system was also tested by calibrating the unit once at the start of the day, then analysing a single sample of concentrated ‘black’ sulfuric acid repeatedly for 24 h. Results of the continuous analysis are shown in Fig. 4 and illustrate the excellent stability of the unit in this mode of operation.Mean results are shown in Table 5, and show that the sample concentration varied less than 6% over 24 h and 77 repeat analyses. Also, excellent agreement between the continuous and laboratory systems was obtained. At-line Site Trial The system was now set up at-line, i.e., connected to the production stream for continuous monitoring of the concentration of Hg in both ‘white’ and ‘black’ acid production streams. The instrument was calibrated over a 50 min period, then five determinations of Hg concentration were made at 9 min intervals, after which the cycle was repeated.Results for Hg concentration over a 10 d period are shown in Fig. 5. It is evident from the results that the concentration of Hg in both ‘white’ and ‘black’ acid streams followed a similar trend, with several peaks in concentration occurring. The regions of blank Fig. 5 Results of at-line site trial with continuous monitoring of the data between runs 67–70 and 60–62, for the ‘white’ and ‘black’ sulfuric acid production stream showing the variation in the concen- acids, respectively, represent periods of plant shut-down.tration of Hg in: (a) ‘white’ acid; (b) ‘black’ acid over a 10 d period. During the 10 d period only eight laboratory analyses were performed for each of the production streams. These results are shown in Table 6 alongside the results obtained with the at-line monitoring system, and excellent agreement was Table 6 Comparison of data for Hg concentration in ‘white’ sulfuric obtained.During the monitoring period the Hg concentration acid, obtained with the at-line continuous monitoring system and by o-line laboratory batch analyses in the product acid rose above the 1 mg ml-1 action limit during the peaks shown in Fig. 5. These peaks could be Hg concentration/mg g-1 identified much earlier using the at-line monitoring system, allowing the acid to be diverted to a low grade holding tank Sampling time† At-line analysis O-line laboratory analysis until the Hg concentration dropped back below the action 21/09/96 1.25 1.3 level.This illustrates two of the major advantages of continuous 22/09/96 0.64 0.6 at-line monitoring. Firstly, it is possible to identify the trend 23/09/96 0.61 0.6 in Hg concentration, thereby allowing action to be taken much 24/09/96 0.58 0.6 earlier than if only single grab samples were analysed once 25/09/96 0.75 0.8 26/09/96 0.70 0.8 every day; and secondly, results were obtained in real time 27/09/96 0.56 0.5 rather than after the 5 h delay typical of laboratory analyses. 28/09/96 0.64 0.6 † O-line laboratory samples were taken at 6.30 am and at-line CONCLUSIONS samples at either 6.20 am or 6.40 am.The at-line mercury detection system was successfully con- figured for the determination of mercury in concentrated sulfuric acid. The o-line chemistries were successfully applied at-line.A robust oxidation method was developed to cater for all forms of mercury which may be present in the acid. Good agreement was obtained using two dierent oxidation techniques. The permanganate technique was chosen over the bromination oxidation as it requires fewer reagents and minimises instrument signal noise. The system components proved resistant to attack from concentrated acids. The standards and reagents were proven to be eective and stable for more than 10 d. The reagent flow rates were adjusted for minimal consumption so that the operator would only need to prepare the reagents once a week, thus minimising operator attention. The sample collection manifold also proved eective for the delivery of a fresh representative sample of hot sulfuric acid to the instrument. The extreme weather conditions experienced Fig. 4 Results of o-line site trial for the determination of Hg in ‘black’ sulfuric acid using the daily calibration mode. (-12 °C) did not appear to aect the instrument’s performance Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 615in any way. Very good agreement with laboratory results REFERENCES utilising CVAAS was obtained from the at-line AFS system. 1 Lindstedt, G., Analyst, 1970, 95, 264. The system was proved eective for the at-line determination 2 Poluetov, N. S., and Vitkun, R. V., Zh. Anal. Khim., 1963, 18, 33. of mercury in concentrated sulfuric acid process streams, giving 3 Hatch, W. R., and Ott, W. L., Anal. Chem., 1968, 40, 2085. the process operators a greater degree of control of their 4 West, C. D., Anal. Chem., 1974, 46, 797. production process. 5 Thompson, K. C., and Reynolds, G. D., Analyst, 1975, 96, 771. 6 Godden. R. G., and Stockwell, P. B., J. Anal. At. Spectrom., 1989, 4, 301. Thanks are given to PS Analytical and the University of Paper 6/08550D Plymouth for the funding of this research through a Teaching Received December 23, 1996 Company Scheme. Thanks are also given to Budelco for provision of a trial sample. Accepted February 10, 1997 616 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12

ISSN:0267-9477    

DOI:10.1039/a608550d

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Use of a Segmented Array Charge Coupled Device Detector for Continuum Source Atomic Absorption Spectrometry With Graphite Furnace Atomization JAMES M. HARNLY*a , CLARE M. M. SMITH†a, DESMOND N. WICHEMSa , JUAN C. IVALDI‡b, PETER L. LUNDBERGb AND BERNARD RADZIUKc aUSDA, ARS, BHNRC, Food Composition L aboratory, Building 161, BARC-East, Beltsville, MD 20705, USA bT he Perkin-Elmer Corporation, 761Main Avenue, Norwalk, CT 06589, USA cBodenseewerk Perkin-Elmer GmbH, Postbox 101164, D-88647 U� berlingen, Germany A commercially available echelle spectrometer with a line background correction, (2) computation of wavelength segmented array charge coupled detector (SCD) was used with integrated absorbance (a more fundamental absorbance a xenon arc lamp and graphite furnace atomizer for measurement independent of the source linewidth, the entrance continuum source atomic absorption spectrometry (CS-AAS).slit-width, and the width of the pixels),1,2 (3) detection limits Approximately 67% of the spectral wavelengths corresponding equivalent to or better than those for conventional, line source to the resonance transitions used for routine AAS AAS (LS-AAS),1,2 (4) construction of calibration graphs that determinations were available on the SCD.As many as eight can cover almost any desired analytical range,3 and (5) the elements were determined simultaneously with a read ability to accommodate variations in the position and width frequency of 50 Hz for each array.The high luminosity of the of the analytical profile, thus permitting the use of large echelle and the high quantum eciency SCD provided entrance slit-widths1 and pressurized atomization.4,5 photoelectron levels that ranged from equivalent to 7 times More sophisticated solid-state arrays, such as charge coupled higher than those previously measured by CS-AAS using a devices (CCDs),6 have the potential of further expanding the linear photodiode array (LPDA) detector.The low read noise capabilities of CS-AAS by lowering the detection limits and of the SCD resulted in the absorbance measurements being oering multiwavelength (multi-element) detection. The advanlimited by the photon shot noise of the continuum source. tage of CCDs over LPDAs arises from fundamental dierences Detection limits were obtained that ranged from equivalent to in their design and performance. The LPDA used in previous a factor of 3 better than those previously obtained for CS-AAS studies1–5 was composed of a row of pixels 25 mm wide and and from a factor of 2 worse to a factor of 10 better than 2500 mm high, while CCDs are usually composed of square (or those for conventional, line source AAS.Sensitivities, as close to square) pixels, 5–25 mm on a side, arranged in large, determined by intrinsic mass (mass necessary for an two-dimensional arrays. The read noise (the random noise absorbance of 0.0044 pm s), were similar to those measured component associated with the measurement of the charge on previously with an LPDA.The high resolution of the echelle an individual pixel) for the LPDA used previously was 2800 e- allowed detailed inspection of the spectra surrounding the (manufacturer’s specification), while that for a quiet CCD is wavelength of the elements determined. Data were displayed typically less than 50 e-. using contour absorbance plots. Molecular peaks were The low read noise of a CCD means that photon shot noise observed within the spectral window of the sub-arrays for As will be the limiting noise source at all but the lowest intensities.( 193.7 nm) and Se (196.0 nm). These peaks were spectrally This is the opposite of the results obtained previously with the and temporally resolved from the analyte peaks and LPDA, where read noise was dominant over most of the disappeared in the presence of a Pd chemical modifier. A low measurement range (1.25×108 e-).The first advantage of a sensitivity Pd line was identified that was 15 pm from the Se shot noise limit (over a read noise limit) is illustrated in line. The Pd and Se peaks were resolved using a spectral Table 1. The intensities in column 3 were obtained experimen- bandwidth of 3 pm per pixel. tally with an LPDA.1 The expected shot noise (square root of Keywords: Segmented array; charge coupled device; continuum the intensity) shown in column 4 ranges from 4.6 to 1.1 times source; atomic absorption spectrometry; graphite furnace less than the 2800 e- read noise of the LPDA for the five atomization elements shown.In reality, the 2800 e- read noise specified by the manufacturer of the LPDA was never achieved (3800 e- was routinely measured) and the ratios should be even greater. Use of a linear photodiode array (LPDA) as a detector for The fact that the shot noise is less than the read noise dictates continuum source atomic absorption spectrometry (CS-AAS) that the signal-to-noise ratios (S/N) and the detection limits has resulted in improved performance characteristics.1–5 for CCDs will be better than those of the LPDA.The CCD Placing a short array (256 pixels each 25 mm wide) in the focal advantage increases as the intensity decreases. Thus, the use plane of a high resolution echelle spectrometer allows a small of CCDs will be most advantageous in the far UV where the wavelength interval (typically less than 1 nm) immediately continuum source is weakest.around the analytical line of interest to be continuously The second advantage of the shot noise limit is that the S/N monitored. This optical arrangement permits (1) accurate o- is independent of the spectral bandpass. With the LPDA, the best detection limits were achieved with the largest possible † Present address: Department of Chemistry, University College entrance slit-width (500 mm). With a CCD, improved detection Cork, Cork, Ireland.limits (as discussed above) can be obtained with a small ‡ Present address: Cetac Technologies, Inc., 5600 South 42nd Street, Omaha, NE 68107, USA. entrance slit width and high resolution. Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 (617–627) 617Table 1 Advantage of shot noise over read noise Element Wavelength/nm Intensity*/e- Shot noise/ÓIntensity Shot noise advantage† As 193.7 384000 620 4.5 Se 196.0 370000 608 4.6 Pb 217.0 1150 000 1072 2.6 Cd 228.8 3 110 000 1760 1.6 Ni 232.0 3230 000 1797 1.6 Mn 279.5 7 200 000 2683 1.1 * 20 ms integration, 20 A lamp current, and 500×500 mm entrance slit.†=2800/ÓIntensity. The availability of large CCD arrays enables large regions approximately 2 ms with a read noise of 15 e-. Finally, the echelle design provides a luminosity higher than that of any of the two-dimensional spectrum of an echelle spectrometer to be covered simultaneously. From these large arrays, small sub- of the other available echelles and a maximum resolution of 3 pm per pixel.sets of pixels can be isolated for computing absorbances for the various wavelengths of interest. Thus, CCDs and echelle In this study, a prototype CS-AAS instrument was constructed from an Optima/SCD, a continuum source, and a spectrometers are inherently compatible with multi-element detection. Unfortunately, there are also some disadvantages: graphite furnace atomizer. This instrument was characterized for source transmission, read noise, analytical sensitivity, detec- low quantum eciency in the UV and a lengthy read time of the CCD.A grid of electrodes overlaying the array surface is tion limits, and spectral resolution. Multi-element measurements were made for up to eight elements at a time and results necessary for the collection and transfer of the charges generated by photons striking the pixels. This grid limits the are reported for ten elements. Measured values were compared with those obtained previously using a Spectraspan echelle penetration of UV photons into the active detector region.Whereas an LPDA has a quantum eciency of approximately (Spectraspan V, Fisons Instruments, Beverly, MA, USA) with an LPDA detector (Spectraspan/LPDA). 30% at 200 nm, CCDs have little response below 400 nm. With a phosphor coating (which absorbs UV photons and fluoresces over a band ovisible wavelengths), quantum eciencies of 10–25% can be attained at 200 nm.A much more elegant technique, thinning the arrays and illuminating EXPERIMENTAL them from the back side, provides higher quantum eciencies Instrumentation (40–50%), but is less widely available and is more expensive.6 The size of the large CCD arrays necessary to cover the The instrument consisted of a 300 W Xe arc lamp (Cermax LX300UV, ILC Technology, Sunnyvale, CA, USA), a echelle spectrum is problematic for graphite furnace atomization. The transient nature of the furnace signal dictates a HGA-500 graphite furnace and power supply (Perkin-Elmer, Norwalk, CT, USA) with a graphite platform and an AS-40 minimum absorbance computation frequency of 50 Hz.A 1000×1000 pixel array requires 100 ms to read at a 10 MHz autosampler (Perkin-Elmer), and an Optima 3000 ICP optical emission spectrometer (Perkin-Elmer).7,8 The ICP is optically data acquisition frequency; an array read frequency of 10 Hz. Each pixel must be read, and the charge cleared, even though coupled to the entrance slit of the echelle by two mirrors; the collimating mirror produces a collimated beam from a point only a small fraction of the pixels possess useful information.Ideally, the array should be read as quickly as possible since source within the ICP and the focusing mirror focuses the collimated beam onto the entrance slit. For this study, the ICP the integration and read processes must be kept separate to avoid blurring the image.For example, if the array is allowed and the collimating mirror were removed and replaced by the furnace and the Xe lamp. The Xe lamp was placed close to to integrate the source intensity for 16 ms, it must be read in 4 ms (a 250 Hz read frequency) to allow an absorbance compu- the furnace, separated only by a shield with an iris diaphragm. The shield prevented the lamp radiation from melting the non- tation frequency of 50 Hz. Thus, the read frequency for a 1000×1000 pixel array would be about 25 times too slow.metal components of the furnace and was positioned so that the iris allowed a collimated beam of light, approximately Faster read times can only be achieved by reducing array sizes, and wavelength coverage, or by increasing the number of read 6 mm in diameter, to pass into the furnace. This beam was then focused onto the entrance slit of the echelle by the ports for the array. At this time, none of the large, twodimensional CCDs available for spectroscopic applications focusing mirror.The vertical and side-to-side positions of the lamp were optimized for maximum transmitted intensity. appears to be well suited for multi-element CS-AAS with graphite furnace atomization. The specifications of the Optima/SCD are shown in Table 2 along with the specifications of the Spectraspan/LPDA instru- Recently, a solid-state detector was introduced which appeared to remedy some of the problems encountered in ment which was used previously for CS-AAS.1–5 Table 3 lists the spectral bandwidths (SBWs) for each of the instruments at adapting large CCD arrays to multi-element CS-AAS.This detector, consisting of a series of one-dimensional CCDs a series of wavelengths for several entrance slit-widths. The Optima/SCD was used in the high, normal and low resolution imbedded in a monolithic chip,7 has been described as a segmented CCD array (SCD) and was designed in conjunction modes with entrance slit-widths of 31, 62 and 185 mm, respectively.The CCD arrays could also be operated in the high or with an echelle spectrometer8 for use as a detection system for a commercially available emission spectrometer (Optima 3000, low resolution mode with the 12.5 mmwide pixels used independently or in pairs (a 25 mm wide pixel). To eliminate confusion, Perkin-Elmer, Norwalk, CT, USA) with an inductively coupled plasma (ICP) source. The Optima echelle and SCD must be the widths of the entrance slit and pixels will be specified in the text.It should be noted that the Optima employs a 2.5 viewed as an integral unit since the positioning of the CCDs in the chip is dependent on the optical dispersion. The times demagnification of the image onto the SCD. Thus, a 25 mm pixel matches a 62 mm wide entrance slit and a 12.5 mm Optima/SCD has several features attractive to CS-AAS. The one-dimensional CCD arrays do not require an overlying grid wide pixel matches a 31 mm wide entrance slit.By comparison, the Spectraspan/LPDA was routinely operated with entrance of electrodes and, as a result, provide a quantum eciency of 60% at 200 nm. In addition, each array can be read in slits of 500 mm to optimize the S/N. The SBWs obtained with 618 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12Table 4 Availability of ETAAS wavelengths on Optima Table 2 Comparison of spectrometer/detection systems Spectraspan/LPDA Optima/SCD Element Primary wavelength(s) Available Not available Spectrometer characteristics— Ag 328.1 × Al 309.3 × F-number 13 6.7 Focal length/nm 750 500 As 193.7 × Au 242.8 × Grating: width (h×w)/mm2 46×96 80×160 grooves/mm-1 79 79 Ba 553.6 × Bi 233.1 × blaze angle (°) 63.4 63.4 Entrance slit (h×w)/mm2 500×500 250×31.2 Cd 228.8 × Co 240.7 × 250×62.5 250×185 Cr 357.9 × Cu 324.7 × Spectral luminosity/mm2 nm-1 0.0132 0.0286 Fe 248.3 × In 325.6 × Reciprocal linear dispersion/nm mm-1 0.0670 0.100 Mn 279.5 × Mo 313.3 × Spectral bandwidth/pm 33.6 3.1* 6.2† Ni 232.0 × Pb 217.0 × 18.1‡ Luminosity/mm2 0.00044 0.00009* 283.3 × Pd 247.6 × 0.00018† 0.00054‡ Pt 265.9 × Sb 217.6 × Detector characteristics — Se 196.0 × Si 251.6 × Format 1×256 1×16 1×32 Sn 244.6 × 286.3 × Size (h×w)/mm2 2500×25 170×25 170×12.5 Ti 364.3 × V 318.4 × Read noise/e- 2800 15 Quantum eciency Tl 276.8 × Zn 213.9 × (at 200 nm) (%) 25.7 61.6 Saturation charge/e- 1.25×108 0.98×106 * 31 mm entrance slit-width.7 Table 5 Elements determined simultaneously † 62 mm entrance slit-width.7 ‡ 185 mm entrance slit-width.7 Suite 1 Suite 2 (10 ms integration) (1 ms integration) Table 3 Spectral bandwidth of spectrometers (pm) Wavelength/ Wavelength/ Array Element nm Array Element nm Spectraspan Optima 0 Se 196.0* 52 Ni 231.6† Element 25 mm 500 mm 31mm 62mm 185 mm 8 As 193.7 103 Mn 279.5 15 Se 196.0* 107 Tl 276.8 As (193.7) 1.84 32.5 3.03 6.06 17.9 35 Sb 217.6 120 Pb 283.3 Se (196.0) 1.87 32.9 3.07 6.14 18.1 39 Zn 213.9 Zn (213.9) 2.06 35.9 3.35 6.70 19.8 42 Pb 217.0 Pb (217.0) 2.10 36.4 3.40 6.79 20.1 58 Bi 223.1 Cd (228.8) 2.23 38.4 3.58 7.16 21.2 63 Cd 228.8 Ni (232.0) 2.26 38.9 3.63 7.26 21.5 Mn (279.5) 2.80 46.9 4.37 8.75 25.9 * The Se 196.0 nm line was available on opposite ends of adjacent Pb (283.3) 2.85 47.6 4.43 8.87 26.2 orders.† This line was chosen erroneously instead of array 71 at the Ni 232.0 nm line. entrance slit-widths of 500 mm (the largest available width) and 25 mm (the smallest available width) are also shown in Table 3.at a frequency of 50 Hz (once every 20 ms). Data were acquired for two dierent suites of elements. For the first suite, intensities Wavelength Availability for eight elements with wavelengths between 194 and 229 nm The Optima/SCD was designed for the detection of emitted (Table 5) were integrated over a 10 ms interval. Following the intensities from a high temperature ICP. Obviously, the most integration period, 8 ms (#1 ms per array) were required to suitable wavelengths for the ICP, which are found on the SCD, read the eight arrays.This allowed 2 ms before the start of the will not always be the best wavelengths for AAS, which uses next data acquisition cycle. For the second suite, intensities for resonance transitions. Table 4 shows that one-third of the four elements with wavelengths between 232 and 283 nm preferred wavelengths for 28 elements commonly determined (Table 2) were integrated over a 1 ms interval.A total of 4 ms by AAS are not available on the Optima/SCD. For half of the was required to read the four arrays, allowing 15 ms before elements listed as ‘Not available’, secondary lines could be the next data acquisition cycle. The graphite furnace atomizfound on the SCD with sensitivities within a factor of 4 of the ation program (Table 6) was the same for both suites of primary lines. Only Co, Fe,Mo and V could not be determined elements. by CS-AAS using the preset wavelengths of the SCD.The raw data consisted of the intensities for each pixel for each array acquired 50 times a second. Thus, for eight arrays of 16 pixels each, a single atomization of 5 s yielded 32000 Data Acquisition and Storage total intensities. The 32000 intensities were stored in a file on disk. At a later date, the files were converted to Excel The start of the data acquisition program was triggered manually when the furnace power supply reached the start of (Microsoft, Redmond, WA, USA) files.All data processing was performed using either Excel or a program written in Labview the atomization cycle. The data acquisition program was designed to read each array for the specified suite of elements (National Instruments, Austin, TX, USA). Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 619Table 6 Graphite furnace atomization program* Step Ramp time/s Hold time/s Temperature/°C Dry 20 10 250 Char 5 5 300 Atomize 0 5 2500 Clean out 1 5 2700 * 20 ml samples sizes used in all cases.Absorbance Contour Plots The contour plots for absorbance were created using Excel from the 16×250 (pixel×time) intensity matrix obtained for each element. The matrix size was first reduced by discarding the data for the first and last pixel, neither of which were meaningful, and by averaging successive sets of ten intensities Fig. 1 Variance versus intensity in analog-to-digital converter units with respect to time.The result was a 14×25 intensity matrix (ADUs). Linear regression for first-order equation gave a slope of 0.0864 and an intercept of 1.679. This translates into a gain of 11.6 e- (I14,25) where each intensity represented the average over 0.2 s. per ADU and a read noise of 15.0 e-. The first intensity for each pixel was used as a reference to compute absorbances for each of the 25 time segments: The photoelectron charge measured in each pixel is analogous to the time integrated cathodic current of a photomul- An,m=log AIn,1 In,mB for m=1–25 (1) tiplier tube.The measured photoelectron charge is determined by the source intensity, the spectral luminosity (light trans- where n is the pixel number and m is the time in units of 0.2 s. mitted per spectral bandwidth) of the echelle, the quantum This was repeated for each pixel (n=2–15). To remove the eciency of the detector, and the integration time.In this flicker noise of the continuum source, corrected absorbances, study, the comparison of the measured photoelectron charge A¾n,m, were computed by subtracting the absorbance of a of the Optima/SCD with that of the Spectraspan/LPDA was selected reference pixel in each scan: facilitated by the use of the same 300W Xe arc lamp and A¾n,m=An,m-ARef,m for n=2–15 (2) graphite furnace atomizer with both instruments. In each case, the graphite furnace served to isolate a 6 mm diameter beam This was repeated for each time segment (m=2–25).of nominally collimated radiation from the lamp (see under Experimental). The transmitted beam was imaged onto the Standards entrance slit of the spectrometer by a lens (Spectraspan) or mirror (Optima) mounted at a distance of one focal length Standard solutions were prepared by dilution of 1000 mg ml-1 from the slit. For the Spectraspan/LPDA, the ratio of the focal stock solutions of As, Bi, Cd, Mn, Pb, Sb, Se, Tl and Zn and length of the lens (80 mm) to the 6 mm diameter of the beam 5000 mg ml-1 of Pd supplied by Inorganic Ventures yielded an F-number of 13, which was equal to that of the (Lakewood, NJ, USA).Sub-boiling distilled nitric acid was spectrometer, i.e., the source image filled the collimating mirror obtained from Seastar Chemicals (Seattle, WA, USA). SRM of the echelle. For the Optima/SCD, the mirror was located 1643c AcidifiedWater was obtained from the National Institute 102 mm from the entrance slit providing an F-number of 16.9, of Standards and Technology (Gaithersburg, MD, USA).which underfilled the collimating mirror of the echelle, which had an F-number of 6.7. Underfilling the mirror did not result in any lost intensity from the source. RESULTS AND DISCUSSION Comparing the specifications of the two instruments in Measured Intensity and Read Noise Table 2, it can be seen that the spectral luminosity (at 200 nm) and the quantum eciency of the Optima/SCD are both The Optima/SCD is expected to provide improved detection limits for CS-AAS for those elements determined at wave- slightly more than twice that of the Spectraspan/LPDA (ratios of 2.2 and 2.4, respectively).Hence, at 200 nm and for the same lengths below 250 nm (Table 1) because of the low read noise and high quantum eciency of the CCDs and the increased entrance slit-width, the intensity measured by each pixel of the SCD is 5.3 times that measured by each pixel of the LPDA.luminosity of the echelle. Initial experiments were conducted to verify the read noise level and to compare the photoelectron The combined spectral luminosity-quantum eciency ratio is based on the assumption that the optical transmission factor charges generated in the pixels of the Optima/SCD with those previously measured for the Spectraspan/LPDA. (a factor accounting for the reflection and transmission eciencies of all the optical components) is the same for both The read noise and the detector gain of the SCD were determined for each array (As, 193.7 nm) by plotting variance instruments and that the entrance slit-heights for the Optima and Spectraspan are 250 and 500 mm, respectively.The ratio versus photoelectron charge measured in analog-to-digital converter units (ADUs) as shown in Fig. 1. The general equation of 5.3 is relatively constant between 200 and 400 nm since the spectral luminosity, which is inversely proportional to wave- for the relationship is: length, will vary similarly for both spectrometers and since the s2=S/G+NR2/G2 (3) ratios of the quantum eciencies for the two detectors vary from 1.8 to 2.6 over this region.where s2 is the variance (in ADUs), S is the photoelectron charge (in ADUs), G is the detector gain (e- per ADU) and Experimentally determined intensities for the two instruments are shown in Table 7. The measured intensities in the NR is the read noise (e-).9 Application of linear regression to the data in Fig. 1 yielded the solid line with a slope of 0.0864 first two columns (e- per pixel) were determined with a 20 A lamp current and integration times of 20 and 10 ms for the and an intercept of 1.679. Using eqn. (3), these values yielded a gain of 11.6 e- per ADU and a read noise of 15.0 e-, which Spectraspan/LPDA and Optima/SCD, respectively (see under Experimental). In the next two columns, e- per nm is com- are in good agreement with the manufacturers specifications of 12.0 e- per ADU and 15 e-, respectively. puted for a common integration time (20 ms) per unit of 620 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12Table 7 Intensity comparison e- per pixel (×10-3) e-per nm (×10-3) Spectrospan/LPDA Optima/SCD Spectrospan/LPDA* Optima/SCD† Ratio of e- per nm Element (mm) A Optima/SCD Spectrospan/LPDAB A20 A‡ 20 ms§ 500×500B A20 A‡ 10 ms§ 250×62¶B A20 A‡ 20 ms§ 500dB A20 A‡ 20 ms§ 250d B As (193.7) 384 208 11.8 68.6 5.8 Se (196.0) 370 185 11.2 60.6 5.4 Zn (213.9) 1230 482 34.3 144 4.2 Pb (217.0) 1150 342 31.6 100 3.2 Cd (228.8) 3110 290 82.2 81.5 0.99 Ni (232.0) 3230 1940 83.1 536 6.4 Mn (279.5) 7200 4780 153 1090 7.1 Pb (283.3) 15 800 9720 333 2190 6.6 * (e- per pixel)/Dl, where Dl is acquired from Table 3 for a 500 mm slit-width.† (e- per pixel)×(20 ms/10 ms)/Dl, where Dl is acquired from Table 3 for a 62.5 mm slit-width. ‡ Lamp current. § Integration time.¶ Entrance aperture, height (mm)×width (m). d Entrance aperture, height (mm). spectral bandwidth. The ratio of e- per nm for the two instruments can be compared with the ratio of spectral luminosity- quantum eciency from the previous paragraph. The ratios of the adjusted intensities ranged from 0.99 to 7.1. Without Cd (228.8 nm), the average ratio was 5.5, slightly higher than the value of 5.3 computed in the previous paragraph. The range of ratios, from 3.2 to 7.1, was most likely due to subtle optical dierences between the two instruments.The reason for the low ratio for Cd (0.99) is not readily obvious. Since both echelles have the same blaze angle and number of grooves per millimetre, they have identical distributions of wavelengths on the orders. A careful examination of the positioning of the detectors in the focal plane shows that, for both echelles, the location of the Cd wavelength is the same with respect to the order center; far to the low wavelength side.Data for the other elements suggest that there Fig. 2 Intensity versus pixel number for atomization of 20 ml of 100 ng ml-1 of As (2000 pg). A, 10th scan 0.2 s after the start of is no significant dierence in the transmission eciency of the atomization (prior to appearance of the analyte) and B, 110th scan two echelles. The only explanation that is left is that there is 2.2 s after the start of atomization (in the middle of the analytical peak).a problem with the Cd array on this particular Optima. Absorbance Calculation can be used as the reference pixels provided that they accurately The computation of time and wavelength integrated represent the transmitted intensity of the source. For example, absorbance (Al,t ) has been previously described.1 Briefly, this using pixels 8 and 9 as the analytical pixels and pixels 2–5 is accomplished by computing a wavelength integrated and 12–15 as the reference pixels, absorbance is computed as: absorbance (Al) for each scan (or read) of the array and then summing these absorbances for the time interval covering the Al=DlP .9 j=8 log AIR IjB (4) duration of the analyte absorption signal.Al is computed by selecting reference pixels to both sides of the analytical absorption profile, using the average reference intensity to compute where DlP is the spectral width of a single pixel and IR is the an absorbance for each of the analytical pixels covering the average reference pixel intensity computed as: absorption profile, and summing the absorbances.With a high resolution instrument, the calculation of Al is changed in two important ways; far fewer pixels are used in the calculation IR= .5 n=2 In+ .15 n=12 In 8 (5) and the structure (absorption or emission by non-analytical species) in the spectrum around the analytical line is much In the second step, Al,t was computed by summing the better defined (compared with the lower resolution of the wavelength integrated absorbance from each scan: Spectraspan with a 500 mm entrance slit-width).Both of these factors make the selection of pixels to be used in the absorbance Al,t=Dt .250 i=1 Al,i (6) calculation critical. Fig. 2 shows the intensity versus pixel number for the 110th scan (with 50 scans per second, this scan shows the absorption where Dt is the time interval between absorbance calculations and Al,i is the wavelength integrated absorbance of the ith profile 2.20 s after the start of atomization) for the atomization of 20 ml of 100 ng ml-1 (2000 pg) of As.It can be seen that scan. In eqn. (6), i=250 corresponds to data acquisition for 5 s at 50 Hz. After normalization by DlP and Dt, Al,t has units four pixels (numbers 7–10) cover the absorption profile at this concentration. Any or all of the remaining non-sample pixels of picometres×seconds (pm s). Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 621Intrinsic Mass from both sides of the absorption profile. Similarly, it can be shown that the baseline noise for the shot noise limited The intrinsic mass is the mass necessary to provide an case is: absorbance of 0.0044 pm s. In theory, intrinsic mass is dependent only on the design and atomization eciency of the sAl= 0.434DlpÓn+1 ÓI (8) graphite furnace and is independent of all other instrument parameters.2 Hence, with a common furnace and the same Since both n and I are linearly dependent on the slit-width, temperature program, the intrinsic masses should be constant eqn.(7) shows that the absorbance noise will decrease as the despite dierences in the spectrometer and detector entrance slit is widened for the read noise limited case while characteristics. eqn. (8) shows that the absorbance noise is approximately In this study, the same furnace was employed with the constant for the shot noise limited case. Since wavelength Optima/SCD that had previously been used with the integrated absorbance is independent of the entrance slit-width Spectraspan/LPDA.The entrance apertures used for routine (see preceding section) the S/N for the noise limited case is operation of the two instruments, however, were dramatically also independent of the slit-width. Thus, a small slit-width will dierent. The Spectraspan/LPDA used a 500 mm wide slit, give the same S/N and oer much better resolution. Far fewer 25 mm wide pixels, had a spectral bandwidth of 32–48 pm sample pixels (ns) will be necessary and the number and (between 193 and 283 nm, Table 3), and used 32 pixels to position of the reference pixels (nR) will be dependent on the cover the absorption profile and determine a wavelength spectral structure surrounding the absorption profile.Under integrated absorbance. The Optima/SCD used a 62 mm these conditions, the assumption that the number of sample entrance slit-width, 25 mm wide pixels, had a spectral bandwith pixels and reference pixels is equal is no longer reasonable.of 6–9 pm (Table 3), and used only two or three pixels to Eqn. (8) can then be expressed as: compute the wavelength integrated absorbance. Thus, intrinsic mass,2 based on normalization of the wavelength integrated sAl= 0.434ÓnS+nS/nR ÓI (9) absorbance by the spectral width of the pixels, is critical to an accurate comparison of the sensitivities of the two instruments. It can be seen that the absorbance noise can be minimized by Table 8 compares the intrinsic masses of the Optima/SCD keeping the number of sample pixels as small as possible and and the Spectraspan/LPDA.In general, the agreement between the number of reference pixels large compared with the number the two instruments was very reasonable. The intrinsic masses of sample pixels [eqn. (9) and Table 9]. for the Optima/SCD ranged from 109 to 167% of the values Table 10 shows the experimentally determined absorbance of the Spectraspan/LPDA.The consistently high bias of the noise for the Spectraspan/LPDA and the Optima/SCD and intrinsic mass, or lower sensitivity, suggests that either more the ratio of the noises for four elements falling between 193 pixels should have been used to measure absorbance in the and 229 nm. The ratio of the absorbance noises (in units of wings of the profile or the furnace performance was not the pm s) ranges from 0.5 to 3.1. These values were obtained with same for the Optima/SCD experiments.Selection of the necesthe conditions shown in Table 7. The Spectraspan/LPDA sary number of pixels to cover the absorption profile was employed an entrance aperture 500×500 mm, a lamp current easier with the Spectraspan/LPDA since the large entrance of 20 A, an integration time of 20 ms, and n=nS=nR=20. The slit-width produced images on the array that had the rectangu- Optima/SCD employed an aperture 250×62 mm, a 20 A lamp lar shape of the entrance slit function.With the Optima/SCD, the images at low concentration are approximately Gaussian and it is more dicult to decide where the absorption signal Table 9 Noise dependence on number of sample and reference pixels has ended. In Table 8, all the reported intrinsic masses were [evaluation of ÓnS+(nS/nR) from eqn. (2)] computed using four or more analytical pixels. Number of sample pixels (nS) Number of reference pixels (nR) 1 2 3 4 5 6 Absorbance Noise 1 1.4 2.0 2.4 2.8 3.2 3.5 It was previously shown1,4 that the baseline noise for wave- 2 1.2 1.7 2.1 2.4 2.7 3.0 length integrated absorbance (sAl) can be computed for the 3 1.2 1.6 2.0 2.3 2.6 2.8 read noise limited case from the number of pixels needed to 4 1.1 1.6 1.9 2.2 2.5 2.7 cover the absorbance profile (n), the read noise of the LPDA 5 1.1 1.5 1.9 2.2 2.4 2.7 (sR), and the average intensity reaching each pixel (I ): 6 1.1 1.5 1.9 2.2 2.4 2.6 7 1.1 1.5 1.8 2.1 2.4 2.6 8 1.1 1.5 1.8 2.1 2.4 2.6 sAl= 0.434DlpsRÓn+1 I (7) 9 1.1 1.5 1.8 2.1 2.4 2.6 10 1.0 1.5 1.8 2.1 2.3 2.5 where it is assumed that a total of n reference pixels is selected Table 10 Experimentally determined noise levels Table 8 Intrinsic mass (mi)* Baseline absorbance noise/pm s mi/pg Element Wavelength/nm Spectraspan/LPDA Optima/SCD Wavelength/ Spectraspan/ Optima/ Improvement Element nm LPDA* SCD† LPDA/SCD As 193.7 9.7 10.5 Se 196.0 11 12 As 193.7 0.026 0.0084 3.1 Se 196.0 0.019 0.0092 2.1 Pb 217.0 1.9 3.0 Sb 217.6 — 10 Pb 217.0 0.010 0.0078 1.3 Cd 228.8 0.0039 0.0084 0.5 Cd 228.8 0.27 0.29 Tl 276.8 — 11 Mn 279.5 0.86 1.4 * 20 ms integration, 20 A lamp current, 500×500 mm entrance slit, and nS=nR=20.Pb 283.3 3.9 6.7 † 10 ms integration, 20 A lamp current, 250×62 mm entrance slit, and nS=2, nR=6. * Mass necessary to give an absorbance of 0.0044 pm s. 622 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12current, a 10 s integration time, nS=2, and nR=6.If the bandwidth significantly smaller than that used for LS-AAS with a deuterium arc lamp for background correction, dramati- Optima/SCD used a 20 ms integration time, the measured intensities would increase by a factor of 2, the absorbance cally broadened the absorption profile and made it dicult to observe any structure around the analytical line. As a result, noise would decrease by a factor of Ó2, and the improvements shown in Table 10 would increase by a factor of Ó2.It can be it was generally assumed that the structure was non-existent or insignificant. seen that the use of the low noise CCD should result primarily in an improvement in the detection limits in the far UV. As a first approximation, this assumption is reasonable; the relatively low temperature of the graphite furnace results in a fairly simple absorption spectrum. With complex sample Detection Limits matrices and some chemical modifiers, however, this assumption is not necessarily valid.10,11 With the inverse Zeeman As shown in the previous two sections, the absorbance signal eect, broad band background interferences are corrected at and noise are both highly dependent on the number of sample the analytical line and magnetically susceptible spectral inter- pixels chosen to cover the absorption profile.The plots in ferences are minimized by making reference measurements Fig. 3 correspond to the use of pixel 8, pixels 8 and 9, pixels within several picometres of the line center. With the 7–9 and pixels 7–10 as the sample pixels for the atomization Optima/SCD, the spectrum can be visually inspected for of 20 ml of 10 ngml-1 (200 pg) of As.In each case, pixels 2–5 interferences with a resolution of 3–4 pm. An examination of and 12–15 were used as the reference pixels. The computed the spectra around some of the elements determined in this absorbances and the detection limits for all four cases are study will be presented in a following section.listed in Table 11. Temporally, the absorbance was integrated Table 12 shows the Optima/SCD detection limits for ten from 1.1 to 3.0 s. It can be seen that the best detection limit elements whose wavelengths fall between 193 and 283 nm and was obtained with only one sample pixel, although the range compares them with those for the Spectraspan/LPDA and a of detection limits was fairly small. As more sample pixels were state-of-the-art, conventional, commercially available spec- added, the absorbance increased but so did the absorbance trometer.12,13 In this study, all the reported detection limits noise, as predicted in Table 9.Fig. 3also shows the absorbances were computed using two analytical pixels. The ratios of the computed using pixel 6 and pixel 11 as the sample pixels. It is experimental detection limits (for the Spectraspan/LPDA and obvious that, at this concentration, pixels 6 and 11 do not lie the Optima/SCD) show reasonable agreement with the ratios on the absorption profile.obtained for the absorbance noise (Table 10). In general, the Pixels 2–5 and 12–15 can be used to determine the average detection limits for the Optima/SCD range from a factor of 2 reference intensity, provided that they accurately represent the worse to a factor of 10 better than those for the conventional transmitted intensity of the source. Any extraneous (nonspectrometer. The exception is Cd, which suers from low light analyte) absorption or emission at these wavelengths will bias transmission (Table 7) on the Optima.It must be remembered the absorbance calculation. Such a bias will most likely change that the conventional detection limits were obtained in the with time given the transient nature of furnace atomization. single-element mode and are, on average, a factor of 2 worse With the Spectraspan/LPDA, an entrance slit-width of 500 mm when four elements are determined simultaneously.was used (a spectral bandwidth of 32 pm for As at 193.7 nm) to increase the radiation reaching each pixel and to improve the S/N. The large entrance slit, while still providing a spectral S/N versus Resolution The resolution and S/N were investigated for instrument entrance slit-widths of 31, 62 and 185 mm and pixel widths of 12.5 and 25 mm. In Fig. 4, the emission profiles for a Se (196.0 nm) electrodeless discharge lamp (EDL) are shown for the three slit-widths with a pixel width of 12.5 mm. Using the data in Table 2, the theoretical resolution of the Optima at this wavelength is 3.1, 6.1 and 18.1 pm.Barnard et al.7 reported experimentally determined resolutions of 5.9 and 8.2 pm for slit-widths of 31 and 62 mm, respectively (with a pixel width of 12.5 mm). Resolutions of 5.7, 7.4 and 17.2 pm were estimated Table 12 Detection limits Detection limit/pg Wavelength/ Spectraspan Optima Fig. 3 Absorbance versus time for atomization of 20 ml of 10 ngml-1 Element nm LPDA* SCD† Conventional‡ of As (200 pg).Absorbances computed as shown in eqns. (1) and (2) with pixels 2–5 and 12–15 used as the reference pixels and A, pixel 8; As 193.7 28 12 6 B, pixels 8 and 9; C, pixels 7–9; D, pixels 7–10; E, pixel 6; and F, pixel Se 196.0 54 16 9 11 as the sample pixels. Zn 213.9 — 0.07 0.1 Pb 217.0 6 4 8 Sb 217.6 — 8 15 Table 11 Signal and detection limit versus number of sample pixels* Bi 223.1 — 5 6 Cd 228.8 0.4 0.7 0.1 Number of Analytical Absorbance Detection Tl 276.8 — 1 9 reference pixels pixels Absorbance noise limit/pg Mn 279.5 0.5 0.2 0.6 Pb 283.3 0.9 0.4 4 1 8 0.0413 0.0010 11 2 8, 9 0.0506 0.0016 12 3 7–9 0.0563 0.0018 15 * 20 ms integration, 20 A lamp current, 500×500 mm entrance slit, and nS=nR=20, with Perkin-Elmer HGA-500 furnace. 4 7–10 0.0586 0.0020 15 † 10 ms integration, 20 A lamp current, 250×62.5 mm entrance slit, and nS=2, nR=6, with Perkin-Elmer HGA-500 furnace.* The number of reference pixels was held constant at 8; pixels 2–5 and 12–15. ‡ Perkin-Elmer SIMAA 6000 with THGA furnace.12,13 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 623Table 13. The S/N were obtained using 25 mm wide pixels for the two largest slit-widths and 12.5 mm wide pixels for the 31 mm slit. As discussed in the previous section, the S/N is, in theory, independent of the entrance slit-width since the wavelength integrated absorbance and absorbance noise remain constant.This theory, however, assumes that first, the spectral bandwidth of the spectrometer is larger than the half-width of the absorption profile, and second, the detection intervals (pixels) are small compared with the spectral bandwidth. The first assumption is not valid for measurements with the Optima with a 31 mm entrance slit. For Se, the spectral bandwidth of 2.9 pm at 196.0 nm is approaching the 1.0 pm half-width of the absorption profile.Thus, it is not surprising that the 62 mm slit provides a better S/N than the 31 mm slit. This is consistent with the theoretical model presented by O’Haver.14 The S/N should approach a constant as the spectral Fig. 4 Intensity versus pixel for Se electrodeless discharge lamp run at a power level of 6 W with a pixel width of 12.5 mm and entrance slit-widths of A, 31.2; B, 62.5; and C, 185 mm. Table 13 Normalized S/N versus entrance slit-width Entrance slit-width from Fig. 5. The experimentally determined values for the 31 and 62 mm wide pixels from this study are undoubtedly biased Element (nm) 31 mm 62mm 185 mm high because the spectral widths of the pixels (3.0 pm for the As (193.7) 0.57 1.0 0.86 12.5 mm pixel and 6.1 pm for the 25 mm pixel) are similar to Se (196.0) 0.51 1.0 0.78 that of the spectral bandwidth and there was no mathematical Pb (217.0) 0.46 1.0 0.88 deconvolution. Cd (228.8) 0.50 1.0 0.95 The S/N for the three entrance slit-widths are shown in Pixel (a) (c) (d) (b) 14 12 10 8 6 4 2 14 12 10 8 6 4 2 1.0 2.0 3.0 4.0 5.0 1.0 2.0 3.0 4.0 5.0 Time /s Fig. 5 Contour plots of absorbance as a function of time (0.2 s intervals) and wavelength (6.1 pm intervals) for the atomization of As (193.7 nm) at 2700 °C: (a), 20 ml of 10 ngml-1 As; (b), 20 ml of 100 ng ml-1 As; (c), 20 ml of 100 ng ml-1 As in 250 mg ml-1 Pd; and (d), 1643c Acidified Water (certified value, 82.1±1.2 ng ml-1 As). 624 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12bandwidth is increased. In every case, the S/N for the 185 mm variation around an absorbance of zero with only the analyte and occasional interferents providing well defined absorbance slit-width is close to that of the 62 mm entrance slit and much better than that of the 31 mm slit. The slightly worse S/N for peaks. Fig. 5 shows absorbance as a function of wavelength and the 185 mm slit suggests either the presence of a flicker noise component, which becomes more significant as the entrance time for the atomization of 0.2 and 2 ng of As (193.7 nm) in 0.5% nitric acid [Fig. 5 (a) and (b), respectively], 2 ng of As slit-width is increased, or the use of a non-optimum number of pixels to integrate the absorbance, i.e., a slightly better S/N with 5 mg of Pd [Fig. 5 (c)], and SRM 1643c Acidified Water [Fig. 5 (d), certified value of 82.1±1.2 ng ml-1 of As, or 1.6 ng would have been obtained with one fewer pixels. The second assumption is also not valid, since the spectral of As in 20 ml].Fig. 5 (a) was obtained from the same data used in Fig. 3. The absorbance of the As can be clearly seen widths of the 12.5 and 25 mm pixels are equal to the spectral bandwidth for the 31 and 62 mm slits, respectively. As a result, on pixels 8 and 9 between 1.4 and 2.6 s. A non-analyte peak can be seen on pixel 11 between 0.6 and 1.2 s. The results in resolution of the absorption profile is lost and, as discussed earlier, the estimate of the spectral bandwidth is biased high.Fig. 5 (a) and (b) were obtained on a dierent day than those in Fig. 5 (c) and (d). A slight shift in the center of the As peak In addition, integration in the intensity domain, rather than in the absorbance domain, can cause non-linearity in the can be seen between days; the analyte peak, centered on pixel absorbance calculation and a deterioration of the S/N. This is oset by the relationship in eqn. (9) which shows that the best S/N is obtained with the fewest sample pixels.Obviously, the S/N is a complex function of the spectral bandwidth, the pixel width, and width of the absorption profile. High Resolution Spectra The Optima/SCD provided a 4000 point intensity matrix, 16 pixels×250 time elements (reads every 0.02 s for 5 s), for each element. This intensity matrix was then reduced to a 14×25 absorbance matrix by dropping the extreme pixels and by averaging successive sets of ten absorbances to provide a temporal resolution of 0.2 s (see under Experimental).Fig. 5 shows four of the 14×25 absorbance matrices presented as Fig. 7 Absorbance versus time for the atomization of 20 ml of contour plots. The flicker component of the continuum source 100 ng ml-1 Se in 250 mg ml-1 Pd. Pixels 7 and 8 were used as the has been removed in Fig. 5 by using pixel 13 as a reference analytical pixels for the Se trace and pixels 5 and 6 were used as the analytical pixels for the Pd trace [see Fig. 6 (d)]. (see under Experimental). As expected, the plots show minor Pixel (a) (c) (d) (b) 14 12 10 8 6 4 2 14 12 10 8 6 4 2 1.0 2.0 3.0 4.0 5.0 1.0 2.0 3.0 4.0 5.0 Time /s Fig. 6 Contour plots of absorbance as a function of time (0.2 s intervals) and wavelength (6.1 pm intervals) for the atomization of Se (196.0 nm) at 2700°C: (a), 20 ml of 100 ng ml-1 Se; (b), SRM 1643c. Acidified Water (certified value, 12.7±0.7 ng ml-1 Se); (c), 20 ml of 250 mg ml-1 Pd; and (d), 20 ml of 100 ng ml-1 Se in 250 mg ml-1 Pd.Color key as Fig. 5. Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 6258 in Fig. 5 (a) and (b), is centered between pixels 7 and 8 in reference intensities at higher analyte concentrations. A careful search of various wavelength tables failed to reveal any elemen- Fig. 5 (c) and (d). A similar shift can also be observed for the non-analyte peak. This shift arose from a slight change in the tal absorption peaks 18 pm from the As line at 193.7 nm, which suggested that the non-analyte peak was due to a entrance optics.The non-analyte peaks are spatially and temporally resolved molecular absorption process. This possibility was further supported by the disappearance of the peak when the same from the analyte peak. It can be seen that any of the pixels not on the analyte peak could be used as reference pixels for mixed standard solution was used with the Pd chemical modifier.There was no non-analyte peak in the Pd matrix the time interval when the analyte signal was present, i.e., the selection of pixels 2–5 and 12–15 in previous sections is blank (not shown). Fig. 6 shows absorbance as a function of wavelength and justified. Care must be taken, however, in selecting reference pixels so that the intensities are not biased by the broadening time for the atomization of 2 ng of Se (196.0 nm) in 0.5% nitric acid [Fig. 6 (a)], SRM 1643c Acidified Water [Fig. 6 (b), of the profile at higher concentrations [compare Fig. 5 (a) and (b)]. In this case, pixels 2–5 and 12–15 are also suitable certified value of 12.7±0.7 ng ml-1 of Se or 0.254 ng of Se in 20 ml], the Pd blank [Fig. 6 (c)], and 2 ng of Se in 5 mg of Pd [Fig. 6 (d)]. A non-analyte peak can be seen in Fig. 6 (a) and (b) (between pixels 2 and 5, from 1.0 to 1.4 s), which is spectrally and temporally resolved from the analyte peak (between pixels 6 and 9, from 1.6 to 2.8 s).The non-analyte peaks in Fig. 6 (a) and (b) disappear with a Pd chemical modifier [Fig. 6 (c) and (d)]. Another non-analyte peak is observed, however, between pixels 5 and 6, running from 3.2 s to the end of the atomization cycle. In Fig. 6 (d), it can be seen that this non-analyte peak lies immediately adjacent to the analyte peak (between pixels 6 and 8, from 3.2 to 3.6 s) but is not resolved with 62 mm slit and 25 mm pixel widths. If an absorbance is computed using pixels 5 and 6 for the analytical intensity, then the top trace in Fig. 7 is obtained. Fig. 8 Intensity versus pixel number for the atomization of Se using Using pixels 7 and 8 for the analytical intensity, the lower a 31 mm wide entrance slit with 12.5 mm wide pixels. The top scan, for trace is obtained. The timing and the duration of the non- Se, was obtained 2.0 s after the start of atomization of 20 ml of analyte peak suggests that it is associated with the volatilization 100 ng ml-1 Se [see Fig. 6 (a)]. The bottom two scans, for Se and Pd of the Pd. Wavelength tables published by the National and for Pd alone, were obtained at 2.2 and 2.8 s, respectively, after the Institute of Standards and Technology15 show the presence of start of the atomization of 20 ml of 100 ng ml-1 Se in 250 mg ml-1 Pd [see Fig. 6 (d)]. a weak Pd line at 196.011 nm (compared with the Se line at Pixel (a) (c) (d) (b) 14 12 10 8 6 4 2 14 12 10 8 6 4 2 1.0 2.0 3.0 4.0 5.0 1.0 2.0 3.0 4.0 5.0 Time /s Fig. 9 Contour plots of absorbance as a function of time and wavelength for (a), 20 ml of 10 ngml-1 Sb (217.6 nm); (b), 20 ml of 10ngml-1 Pb (217.0 nm); (c), 20 ml of 10 ngml-1 Tl (276.8 nm); and (d), 20 ml of 10 ngml-1 Zn (213.9 nm). The time intervals were 0.2 s and the wavelengths were 6.8, 6.8, 8.8 and 6.7 pm in (a), (b), (c) and (d), respectively. Color key as Fig. 5. 626 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12196.026 nm). Using 31.2 mm slit and 12.5 mm pixel widths, the albeit very expensive, detector for CS-AAS.Improved detection intensity plots shown in Fig. 8 were obtained. The resolution limits were obtained for most of the elements tested between of 3.1 pm per pixel allows the Pd and Se peaks to be resolved 190 and 280 nm. The ability to inspect the absorption spectra (middle trace). The identities of the peaks were easily verified in the region around the analytical wavelengths oers the by the atomization of a Se standard in 0.5% nitric acid (top potential for exploring spectral interferences for AAS.Although trace) and a Pd blank (bottom trace). The peak centers were the wavelength selection for the SCD oered only about 67% between 4 and 6 pixels apart, between 12.4 and 18.6 pm, in of the elements routinely determined by AAS, the results very reasonable agreement with the tabulated dierence of obtained here demonstrate that an SCD-type detector, when 15 pm. combined with an echelle spectrometer, could provide true As and Se were the only elements of those listed in Table 5 simultaneous multi-element capability for CS-AAS with graph- for which non-analyte peaks were observed within the spectral ite furnace atomization.window of the CCD arrays. Fig. 9 shows plots obtained for 0.2 ng of Sb (217.6 nm), Pb (217.0 nm), Tl (276.8 nm) and Zn (213.9 nm). It can be seen that the baseline absorbance falls REFERENCES fairly consistently within an absorbance range of ±0.001 with random values falling within the limits of ±0.003.Only the 1 Harnly, J. M., J. Anal. At. Spectrom., 1993, 8, 317. analytical peaks exceeded the range of ±0.003. In addition, 2 Smith, C. M. M., and Harnly, J. M., Spectrochim. Acta, Part B, 1994, 49, 387. no non-analyte peaks were observed at these wavelengths with 3 Harnly, J. M., Smith, C. M. M., and Radziuk, B., Spectrochim. a Pd modifier. Acta, Part B, 1996, 51, 1055. In general, the lack of spectral structure in Figs. 5, 6 and 9 4 Smith, C. M. M., and Harnly, J. M., J. Anal. At. Spectrom., 1995, supports the general assumption that the relatively low tem- 10, 187. peratures (<2700 °C) of the graphite furnace produce reason- 5 Harnly, J. M., Spectrochim. Acta, Part B, 1993, 48, 909. ably simple spectra. In every case, the analyte and non-analyte 6 Sims, G. R., in Charge-T ransfer Devices in Spectroscopy, ed. Sweedler, J. V., Ratzla, K. L., and Denton, M. B., VCH, New peaks could be temporally and/or spatially resolved. The use York, 1994, pp. 9–58. of chemical modifiers, as shown here, can have advantages 7 Barnard, T. W., Crockett, M. I., Ivaldi, J. C., and Lundberg, P. L., and disadvantages. The non-analyte spectral features observed Anal. Chem., 1993, 65, 1225. for As and Se in nitric acid disappear with the use of the Pd 8 Barnard, T. W., Crockett, M. I., Ivaldi, J. C., Lundberg, P. L., modifier. The modifier itself, however, produced a significant Yates, D. A., Levine, P. A., and Sauer, D. J., Anal. Chem., 1993, peak very close to the Se peak. Chemical modifiers would 65, 1231. appear to be problematic because they are used in concen- 9 Mortara, L., and Fowler, A., Proc. SPIE—Int. Soc. Opt. Eng., 1981, 290, 28. trations that far exceed the concentration of the analyte. 10 Massman, H., T alanta, 1982, 29, 1051. Undoubtedly, some chemical modifiers will be more trouble- 11 Heitmann, U., Schutz, M., Becker-Ross, H., and Florek, S., some than others. With the exception of Se (196.0 nm), the Pd Fresenius’ J. Anal. Chem., 1996, in the press. modifier provided no spectral structure in the wavelength 12 The Perkin-Elmer Corporation, SIMAA 6000 Atomic Absorption regions examined in this study. The Mg(NO3 )2–NH4H2PO4 Spectrometer, Part Number B050-4270/8.94, 1994. modifier, however, has been shown to provide complex struc- 13 The Perkin-Elmer Corporation, personal communication. ture at the Cd wavelength (228.8 nm).11 Further study is 14 O’Haver, T. C., Anal. Chem., 1991, 63, 164. 15 National Institute of Standards and Technology, Circular 488, necessary to explore the presence of spectral interferences and Section 4 Ultraviolet Multiplet T able (Z=1 to 64) and Section 5 the suitability of specific chemical modifiers. Ultraviolet Multiplet T able (Z=72 to 88). CONCLUSION Paper 6/08440K Received December 17, 1996 The high luminosity, quantum eciency, and resolution and the low read noise of the Optima/SCD make it a powerful, Accepted February 14, 1997 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 627

ISSN:0267-9477    

DOI:10.1039/a608440k

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Rapid and Quantitative Microwaveassisted Recovery of Methylmercury From Standard Reference Sediments CHUN MAO TSENG, ALBERTO DE DIEGO†, FABIENNE M. MARTIN AND OLIVIER F. X. DONARD* L aboratoire de Chimie Bio-Inorganique et Environnement, EP CNRS 132, Universite� de Pau, He�lioparc, 64000, Pau, France. E-mail: Olivier.Donard@univ-pau.f r A simple and rapid method has been developed to determine cury (HgII)6 but they do not exceed 1.5% of the total mercury present in sediments.1 The development of analytical methods methylmercury in sediments. The procedure is based on the quantitative microwave-assisted leaching of methylmercury for the determination of low levels of MeHg+ in sediments is still dicult at present, due to problems related to sample from sediments with an acidic extractant.Sample preparation is achieved in an open focused microwave field by heating handling (e.g., sampling, storage and sample preparation techniques) and analytical quality control (e.g., comparison of about 1 g of dry sediment suspended into 10 ml of 2 mol l-1 HNO3 during 3–4 min at a power of 40–60 W.The extracted dierent methods, use of certified reference materials and interlaboratory studies).7 To improve the quality of MeHg+ mercury compounds are ethylated, cryogenically trapped in a chromatographic phase, successively eluted and detected in an analysis and validate dierent techniques, 18 expert laboratories in the European Union took part in an intercomparison electrothermally heated quartz furnace by CV AAS.The method has been validated by the analysis of the certified exercise organised by the Community Bureau of Reference (BCR) in 1995.8 IAEA-356 and CRM 580 reference materials and one sediment sample, BCR S19, from an intercomparison study Both the sample preparation step and the determination of MeHg+ in sediments remain a problem. The method most within the framework of a certification exercise sponsored by the Community Bureau of Reference (BCR) of the European widely employed for the extraction and separation of MeHg+ in sediments is still the classic Westo�o� technique.9,10 Acidic Commission.The detection limit of the procedure is 0.5 ng of methylmercury per gram of dry sediment. After analysis of the leaching,11–14 alkaline digestion15–19 or steam distillation,17–21 followedby one or two separation steps, e.g., solvent extraction, sediments mentioned above (methylmercury ranging from 5 to 80 ng g-1 ), a mean methylmercury recovery of 97% and an ion exchange, distillation or aqueous derivatization (hydride generation or ethylation), are other isolation methods used RSD of 7% were obtained.Under optimum conditions, sample throughput is restricted by the instrumental analysis time to separate MeHg+ from HgII. Quantification can then be undertaken by electron capture, atomic fluorescence and (about 20 min per sample), rather than by the sample preparation step.In addition, the methylmercury extraction atomic emission or absorption spectrometry. The sample preparation methods mentioned above are not only laborious and eciency of four dierent acids, nitric, hydrochloric, sulfuric and acetic acid, has been investigated for comparison. Results time-consuming, but also lack sucient eciency and reliability.19,22,23 For example, HCl leaching at room tempera- on the behaviour and stability of methylmercury in a microwave field are also provided.ture17 does not quantitatively release methylmercury compounds from sediment samples. Alternatively, both alkaline Keywords: Mercury speciation ; microwave-assisted extraction; digestion by 25% KOH in methanol and distillation quantitat- sediments ; ethylation; cryogenic trapping; gas ively release MeHg+ from sediments, but it takes 1 to 6 h to chromatography; atomic absorption spectroscopy get complete recoveries.15–21 The eciency of the microwave-assisted extraction technique Mercury speciation has been a field of continuous concern for sample preparation in environmental applications has been over along period.Such interest is mainly due to the toxicologi- evaluated for dierent matrices, e.g., soils, sediments and cal impact, ecological problems and biogeochemical cycling of biological tissues, since it was first applied in 1975.24 This mercury, involving distribution, accumulation, transformation technique has been confirmed as one of the best methods for and transport pathways in the natural environment.1,2 Mercury mineralization25–27 and selective leaching28–30 of analyte comcompounds are discharged into the aquatic environment from pounds.It has also been verified as an appropriate tool for various anthropogenic and natural sources and are mainly rapid preparation of solid samples for organometallic speciremoved through suspended solids to be deposited in sedi- ation analysis.31 Recently, an open low-power focused microments.Sediments, one of the most significant parts of the wave system has been successfully applied to organotin mercury biogeochemical cycle, play an important role as the speciation analysis.32–34 In addition, further development of ultimate sink of mercury species.3,4 Mercury species accumu- the technique has led to flow-through sample extraction lated in sediments, however, could be transformed and released schemes for continuous on-line determination and it has into the aquatic environment under specific conditions.2 Hence, facilitated automated methods.35,36 Nevertheless, careful analysis of sediments is a prerequisite for the frequent monitor- optimisation of the conditions of the microwave extraction ing of mercury levels in water.5 procedure is required in terms of stability of the target com- The determination of total mercury is not sucient for pounds under a microwave field, prior to speciation analysis.understanding the toxicological impact and pathway of mer- Essential parameters, such as the extraction medium, power cury species in the environment. Methylmercury (MeHg+) applied and exposure time, must be fully optimised.32–34 compounds are considerably more toxic than inorganic mer- In this article, a simple, rapid and accurate protocol for sample preparation and determination of MeHg+ in sediments is presented. Microwave-assisted acid leaching of the sample † On leave from the Department of Analytical Chemistry, University of the Basque Country, 644 P.K., 48080 Bilbao, Spain. provides fast, reproducible and reliable results after analysis of Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 (629–635) 629the extract by an automated on-line ethylation–cryogenic controlled by the flowmeter, was allowed to purge and deliver the volatile mercury species to the cryogenic trap via two trapping–GC–quartz furnace AAS (Et–CT–GC–QFAAS) system.electric Teflon valves. During the derivatization step, the cryogenic trap was immersed in a liquid N2 cold bath (-196 °C), lifted by the pneumatic pump through N2 gas. EXPERIMENTAL Volatile mercury species were trapped in 2.5 g of Chromosorb W-HP (60–80 mesh) coated with 10% SP-2100 (Supelco, St Instrumentation Germain en Lage, France) and packed in a silanized U-shaped Methylmercury extractions were undertaken in a 50 ml round- (45 cm length, 5 mm id) Pyrex column.In the desorption step, bottomed open vessel (150 mm length, 35 mm id) with a the trapped species were sequentially eluted on the basis of borosilicate glass condenser, using a Microdigest A301 their boiling points by gradual heating of the column [wrapped (2450 MHz, maximum power 200W) microwave digestor with 1 mm diameter Nichrome wire (1.75 V×m-1)] by means (Prolabo, Fontanay-sous-bois, France), equipped with a TX32 of an electrical power supply set at 30 V.To prevent H2O programmer, which allows the applied energy to be selected condensation, the PTFE tubing transfer lines between (i) the from 10 to 200 W in increments of 10 W. The exposure time, reaction vessel and the cold trap and (ii) the cold trap and the up to 99 min, can be set in steps of 1 min. The focused single- quartz furnace were also wrapped with Nichrome wire and mode m digestor is an open system32–34 which repro- heated by an adjustable voltage. All the connections between ducibly delivers energy by a magnetron and intensively focuses components were made with Teflon tubing and Omnifit PTFE it onto the sample by means of a wave-guide.The temperature variable-bore connectors. The atomisation of mercury com- of the sample can be continuously measured in real time by a pounds was achieved in a T-shaped quartz furnace (light path Megal 500 thermometer (Prolabo) connected to a micro- length 20 cm, 1 cm id), electrothermally heated at 800 °C by computer, so that the determination of the temperature profile an MHS-20 controller (Perkin-Elmer, Norwalk, USA) and of the sample solution during microwave irradiation is also they were detected by an atomic absorption spectrophotometer possible.(Model 5100, Perkin-Elmer). The AAS instrument operated at Determination of the recovered mercury species was per- 253.7 nm, with a 0.7 nm slit width. A hollow cathode lamp formed by an automated on-line hyphenated system: on-line was used as excitation source.Signal acquisition as well as all automation of the ethylation reaction, cryogenic trapping and the steps in the analysis were controlled through the electronic chromatographic separation of the ethylated mercury species panel, which was programmed using BORWIN chromato- and their detection by quartz furnace AAS (Fig. 1). It was graphic software. controlled by a computer equipped with BORWIN software (JMBS Developments, Lyon, France).The set-up consisted of a 200 ml reagent flask, a peristaltic pump (Ismatec S. A., Reagents Glahbrugg-Zu�rich, Switzerland), two- and three-way electric Teflon valves, a 250 ml reaction and purge vessel, PTFE Analytical grade chemicals (Merck, Darmstadt, Germany; Strem, Newburyport, USA) and Milli-Q water (Millipore, transfer lines of 3 mm id, a cryogenic trap, a pneumatic pump (Joucomatic S. A., Rueil-Malmaison, France), two flowmeters, Milford, MA, USA) were used throughout unless otherwise stated.An approximately 0.01% m/v solution of sodium an electric control panel and a quartz furnace. The mercury species were ethylated in the 250 ml reaction vessel, in which tetraethylborate (NaBEt4) was prepared daily in a glove bag (Bioblock Scientific) filled with N2 by dissolving the reagent 50 ml of sample solution reacted with 10 ml of 0.01% m/v NaBEt4 solution, quantitatively pumped by the peristaltic in water. NaBEt4 solution was always kept in ice and darkness after preparation and throughout the analysis.A 2 mol l-1 pump, using Tygon pumping tubing (2.0 mm id) (Bioblock Scientific, Illkirch, France). The solution was continuously solution of sodium acetate buer was prepared by dissolving 82 g of sodium acetate in a final volume of 500 ml of water stirred by a magnetic Teflon bar. Helium carrier gas, its rate Fig. 1 Schematic diagram of the automated on-line hyphenated system for mercury speciation analysis. 1, Reagent flask; 2, reaction and purge vessel; 3, cryogenic trap; 4, quartz furnace; 5, control panel; 6, flowmeters. 630 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12containing 57.2 ml of glacial acetic acid. The acid solutions, HNO3 , HCl (2 mol l-1) and H2SO4 (1 mol l-1), were prepared by diluting with water the appropriate volume of the corresponding concentrated acids to a final volume of 500 ml. Standard Solutions and Certified Reference Sediments A standard stock solution of 1000 mg ml-1 of HgII was prepared by dissolving mercury(II) chloride in 1% HNO3, and that of 1000 mg ml-1 of methylmercury by dissolving methylmercury chloride in methanol.All stock solutions were stored in a refrigerator and protected against light. Working standard solutions were prepared by appropriate dilution in water of the stock solutions and they were stored one week at most. IAEA-356 and CRM 580 reference sediments are certified Fig. 2 Temperature profile in the chromatographic column during for methylmercury by, respectively, the International Atomic the desorption step: voltage, 30 V.Typical chromatogram correspond- Energy Agency (IAEA) and the Community Bureau of ing to 5 ng of MeHg+ and 5 ng of HgII in Milli-Q water obtained by Reference (BCR) of the European Communities. BCR S19 is Et–CT–GC–QFAAS (absorbance in arbitrary units). Peak at 2.75 min: a candidate reference material which was recently used in an MeHgEt; peak at 3.10 min: HgEt2.intercomparison exercise organised by the BCR.8 Hg0 is assumed to come from decomposition of inorganic Analytical Operating Conditions mercury in a direct reduction process. A typical chromatogram corresponding to a mixture of 5 ng of HgII and 5 ng of MeHg+ The analytical performance of the determination technique dissolved in water is also provided in Fig. 2. The linear was investigated using mercury(II) chloride and methylmercury calibration range extends from 0.5 to 20 ng as Hg for both chloride standards in Milli-Q water.Optimum operating con- MeHg+ and HgII, with a sample volume of 50 ml. The RSD ditions of the system are listed in Table 1. The analytes react (n=10) for 5 ng as Hg is less than 10% for each mercury in the following way with NaBEt4 to give the corresponding species. The detection limits, calculated as 3 times the standard volatile mercury species:37 deviation of ten blank measurements, are 50 pg for labile HgII HgII+2 NaBEt4�HgEt2+2 Na++2 BEt3 (1) and MeHg+, respectively.MeHg++NaBEt4�MeHgEt+Na++BEt3 (2) Analytical Procedure ‘BEt3’ represents an unstable compound which reacts with air and water. Heating of the column by an electrical power A sample of approximately 1 g of homogenised dry sediment supply set at 30 V resulted in the temperature profile displayed and 10 ml of acid solution were placed in an extraction tube in Fig. 2. A voltage of 30 V was selected in order to get good and exposed to microwave irradiation at 60W for 3 min.After peak resolution and minimise the analysis time. As a result, microwave irradiation, the sample solution was transferred ethylated mercury species were sequentially eluted from the into a 15 ml tube and centrifuged at 5000 rpm for 5 min. The column by thermal desorption in order of increasing molecular supernatant was poured into a 22 ml Pyrex vial with a Teflon weight with the following retention times: MeHgEt, 2.75±0.04 cap (Supelco) and stored in a refrigerator until analysis. For min; HgEt2, 3.10±0.05 min.When the concentration of inor- the determination of mercury species, an aliquot of 1 ml of the ganic mercury in the sample is very high (as in the case of extract was added to 50 ml of Milli-Q water in a 250 ml sediments) a peak for Hg0 is also observed at 1.36±0.03 min. reaction vessel and the pH of the solution was adjusted to 4 by adding acetate buer. Mercury species were then determined Table 1 Optimum parameters of the automated Et–CT–GC–QFAAS by means of the Et–CT–GC–QFAAS automated on-line system for mercury speciation analysis system.The three-point standard additions method was always used in the determination step, in order to avoid possible Ethylation— matrix interferences. Blanks were run after each triplicate Derivatization 10 ml of 0.01% m/v NaBEt4 analysis to check for possible memory eects. Solution pH #4, 0.5 ml of 2 mol l-1 acetic/ acetate buer In the experiments to check the stability of MeHg+ in a Reaction time 3 min microwave field and to optimise the microwave extraction procedure, the sample, e.g., 10 ml of acidic solution spiked Cryogenic trapping— GC column U-shaped glass tube, 45 cm length, with 25 ng of MeHg+ or the real sediment with the extractant, 5 mm id respectively, was exposed to microwave irradiation for 1–8 GC phase 10% SP-2100 on Chromosorb min at a power ranging from 10 to 160 W.The procedure for W-HP 60/80 mesh size MeHg+ determination was the same as that described above. Carrier gas Helium (99.995%) Pre-cooling duration 1 min Purging duration 10 min RESULTS AND DISCUSSION Purging flow rate 150 ml min-1 The behaviour of methylmercury in a microwave field is Desorption— Sng gas Helium (99.995%) generally aected by the nature of the extraction media, due Stripping flow rate 150 ml min-1 to dierences in the ability of the solvent to absorb and Desorption voltage 30 V propagate microwave energy.38,39 Monitoring of temperature Data acquisition— during the microwave exposure may give clues to understand- Instrument Perkin-Elmer AAS 5100 ing the mechanisms involved in the process.The temperature Wavelength 253.7 nm profiles corresponding to five dierent solvents, 100% Quartz furnace temperature 800 °C CH3COOH, 2 mol l-1 HCl, HNO3 , 1 mol l-1 H2SO4 and Acquisition duration 5 min Milli-Q water, exposed to microwave irradiation (Fig. 3), Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 631Fig. 3 Temperature profile of dierent solvents under microwave Fig. 4 Recovery of methylmercury from the IAEA-356, BCR S19 and irradiation: power 60 W; volume 10 ml. CRM 580 sediments using dierent acidic extractants: power, 60 W; irradiation time, 3 min; volume, 10 ml. Error bars represent standard deviation of three measurements. reveal a rapid increase of temperature in the initial stage to reach a plateau after 3 min.The maximum temperatures achieved are 10–20°C above the normal boiling points in all yielded lower average recoveries, e.g., about 85% and 55%, the cases. This superheating eect, which has been reported respectively. Lower MeHg+ recoveries obtained after 1 mol before,40 may be responsible for the high eciency of the l-1 H2SO4 leaching might be the result of: (i) coprecipitation microwave-assisted extraction technique. of methylmercury with sulfate precipitates, such as calcium Preliminary investigations were carried out to confirm the sulfate;45 (ii) adsorption to fine particles of organic matter; (iii) stability of MeHg+ in dierent solvent media under microwave complexation with organic ligands, such as lipids and proteins irradiation.Ten millilitres of the acidic solutions mentioned (thiol groups); or (iv) incomplete recovery from sediments. It above were successively spiked with 25 ng of MeHg+ and is worth mentioning that both the inorganic mercury and the exposed to 60W of power during varying heating times.The methylmercury content of the H2SO4 extract decreased mark- MeHg+ content of the corresponding leachates was then edly with storage time, even when it was stored in a refrigerator. determined. For 1 mol l-1 H2SO4 and 100% CH3COOH solvents, averaged MeHg+ recoveries of only 80–90% were Matrix Interferences in Sediment Extracts obtained after 8 min of irradiation. For the rest of the studied conditions, quantitative recovery of MeHg+ was always The analytical procedure described in the experimental section achieved.MeHg+ losses are probably due to evaporation of allows direct determination of MeHg+ in sediments. extractant during vigorous heating. The use of a reflux con- Nevertheless, the most appropriate acid solvent must be selec- denser in extraction procedures based on the open microwave ted in order to minimise matrix eects in the determination digestion of the sample at low power during long heating times step.Interferences in the analysis of the leachates were quant- is, therefore, highly recommended, in order to avoid possible ified by comparing the dierence between the analytical signal losses by evaporation of the extractant and target analytes. corresponding to spiked (5 ng of MeHg+) and non-spiked sediment extracts of each acid solution. Results in Fig. 5 are normalised to the signal obtained for a solution of 5 ng of Evaluation of Acid Extractants for Quantitative MeHg+ MeHg+ in HNO3 exposed to microwave irradiation.The best Recovery from Reference Sediments and most consistent sensitivity, not seemingly aected by Nitric, hydrochloric, sulfuric and acetic acids have commonly matrix eects, was obtained from the nitric acid leachate. been used in the extraction of organomercury compounds Matrix eects of dierent levels seem to interfere, however, in from sediments.11–14,41,42 In this study, their methylmercury the analysis of the HCl, H2SO4 and CH3COOH leachates.In extraction eciency from three materials, IAEA-356, BCR S19 the case of HCl, the low yield obtained in the reagent blank and CRM 580, which are anthropogenically contaminated (spiked and unspiked HCl solution, not leachate), 80%, is samples of dierent origin, have been investigated. The MeHg+ possibly due to chlorine interference.15 In the presence of content of the last two sediments is 10–20 times higher than that of the first one.To avoid the potential cleavage of the mercury–carbon bond or demethylation processes by concentrated strong acids,17,43,44 2 mol l-1 HNO3, 2 mol l-1 HCl, 1 mol l-1H2SO4 ,and 100% CH3COOH have been considered. The stability of MeHg+ in those extractants after microwave digestion at 60W power for up to 6 min was also proved, as stated before. Results obtained after the procedure described in the experimental section show that, after 3 min heating at 60 W, the recovery of methylmercury varies significantly as a function of the extractant used (see Fig. 4). The best recoveries were obtained with HNO3 and HCl solutions, resulting in values which lie within the confidence intervals of the certified MeHg+ concentration for each sediment. Insucient recovery of methylmercury from sediments after a HCl leaching at room Fig. 5 Dierence between the analytical signals corresponding to temperature have been reported elsewhere.17 This phenomenon spiked (5 ng of MeHg+) and non-spiked sediment extracts: power, is not observed when the microwave extraction technique is 60 W; irradiation time, 3 min; volume, 10 ml.‘Reagent blank’ stands applied. The eciency of microwave extraction is related to for simple acidic solutions, not for leachates. Dierence between signals the intensive reaction between the sample matrix and the acid normalised to the signal obtained for a solution of 5 ng of MeHg+ extractant, due to the focused microwave energy and to the in 2 mol l-1 HNO3 exposed to microwave irradiation.Error bars represent standard deviation of three measurements. superheating eect of the extractants. CH3COOH and H2SO4 632 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12sediment matrices, on the contrary, an averaged yield of 99% is no risk of MeHg+ loss during microwave irradiation. BCR S19 and CRM 580 were used in this study. MeHg+ recovery was obtained.No interference seems to aect MeHg+ determination in sulfuric standard solution (reagent blank), but poor from the BCR S19 reference sediment at dierent conditions of power applied and irradiation time is given in Fig. 7. Similar yields are obtained when analysing the H2SO4 leachates corresponding to each one of the sediments. The lowest yields results were obtained for the CRM 580 material. High recoveries (>85%) are achieved in all the cases.For both materials, correspond to CH3COOH extracts, e.g., 45.3% and 12.6% for BCR S19 and IAEA-356, respectively. It is also worth men- exposure time required to recover 100% of MeHg+ at 20, 40, 60 and 80W are 4, 3, 2 and 1 min, respectively. Microwave- tioning that serious matrix interference, resulting in yields of only 49.4 and 31.1% for BCR S19 and IAEA-356, respectively, assisted leaching with HNO3 is, therefore, confirmed as a rapid and ecient sample preparation method for mercury speciation was also observed in alkaline sediment extracts, after alkaline leaching of the reference materialswith 25%KOH in methanol, analysis in sediments, in which the MeHg+ moiety is completely preserved.Microwave irradiation of the sample during a finding which has already been described by Horvat et al.17 After comparison of the four acidic solutions studied, extrac- 3 min at 60 W, using 2 mol l-1 HNO3 as extractant, is recommended as the optimum procedure for quantitative MeHg+ tion with 2 mol l-1 HNO3 is recommended for organomercuric speciation analysis of sediments by microwave-assisted leach- recovery from sediments.ing, prior to detection by Et–CT–GC–QFAAS. Analytical Figures of Merit Optimisation of Microwave Leaching of Sediments with The whole analytical procedure proposed for methylmercury 2 mol l-1 HNO3 for MeHg+ Analysis determination in sediments is presented schematically in Fig. 8. The performance of this procedure has been evaluated critically In a first step, the stability of 25 ng of MeHg+ spiked into 10 ml of 2 mol l-1 HNO3 solution during microwave by the analysis of the BCR S19 and CRM 580 materials.The three-point standard additions method was always used in the irradiation at various microwave power settings and exposure times was investigated. Results are shown in Fig. 6(a). The determination step, in order to avoid possible matrix interferences. An RSD of 7% was obtained after six independent signal for MeHg+ decreases with increasing power applied and exposure time. For example, lower signals for MeHg+ replicates.The detection limit was 0.5 ng of MeHg+ per g of dry sediment, with a linearity range from 0.5 to 20 ng of were systematically observed after 4 min at a heating power of 160 W. Losses should not be attributed to decomposition MeHg+. The slopes of the calibration curves corresponding to HNO3 leachates and Milli-Q water spiked with MeHg+ by microwave attack, but to evaporation during extreme heating for long periods of time.A correlation between loss of did not dier significantly, showing no matrix eect in nitric acid medium. IAEA-356 reference material was also analysed MeHg+ signal and loss of solution volume after 50% of extractant loss is shown in Fig. 6(b). Exposure times of 8, 7, 5 following the procedure in Fig. 8. The results corresponding to the three sediments, presented in Table 2, are in good and 3 min assure 100% yields of MeHg+ at 40, 80, 120 and 160 W, respectively. Therefore, optimisation of MeHg+ extrac- agreement with the certified values.Fig. 9 shows typical chromatograms corresponding to each one of the sediments analysed. tion eciency by 2 mol l-1 HNO3 in real sediments was undertaken in the range 20–80W and 1–4 min, in which there A higher content of inorganic mercury in CRM 580 led to the appearance of a little peak corresponding to Hg0. Fig. 7 Recovery of methylmercury from BCR S19 reference sediment at dierent combinations of microwave power and irradiation time: extractant, 10 ml of 2 mol l-1 HNO3.Recoveries normalised to that of a sediment leachate exposed to microwave irradiation at 60 W for 3 min. Table 2 Certified and determined MeHg+ values for several certified reference sediments Concentration of MeHg+/ng g-1* Sediment Determined† Certified Fig. 6 (a) Stability of MeHg+ and (b) direct correlation between loss BCR S19 51.9±5.1 53.1±8.6 CRM 580 79.6±3.0 75.4±5.0 of MeHg+ signal and solution volume, after exposure to microwave field at dierent combinations of microwave power and irradiation IAEA-356 5.49±0.72 5.87±0.41 time: 10 ml of 2 mol l-1 HNO3 spiked with 25 ng of MeHg+.MeHg+ content normalised to that of a solution unexposed to microwave *Calculated for dry mass. †Six independent experiments. irradiation. Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 633time, instead of by the sample preparation step, as is the case for other analogous methods reported in the literature.CONCLUSIONS A simple, fast, precise and accurate method for the determination of methylmercury in sediments is presented. After microwave-assisted extraction with nitric acid, the methylmercury concentration is directly determined by an automated on-line Et–CT–GC–QFAAS hyphenated system. After careful evaluation of the extraction eciency, matrix interference and stability of methylmercury in a microwave field for several acid solutions, nitric acid was chosen as the most appropriate acidic extractant.It assures high quality methylmercury determination in dierent sediment matrices. Microwave-assisted acid leaching of methylmercury from sediments was investigated by a matrix approach to optimise the extraction eciency. The results from the analysis of three standard reference materials demonstrate the simplicity, eciency, reproducibility and accuracy of the proposed procedure.It can not only significantly reduce the sample preparation time, but also Fig. 8 Schematic flow diagram of the proposed procedure for the analysis of MeHg+ in sediments. enhance the recovery and preserve the organospecies, compared with conventional extraction methods, distillation, ultrasonic assisted leaching and supercritical fluid extraction. The open focused microwave-assisted extraction method developed in our laboratory has also been applied to a wide variety of environmental matrices and organometallic species.32–34 The authors thank Mr.D. Mathe� (Prolabo) for allowing the use of a prototype microwave digestor, R. Lobinski, J. Szpunar and V. O. Schmitt for fruitful contributions to the microwaveassisted speciation-related research, and K. J. Lamble for correcting the article. C. M. Tseng acknowledges the Taiwan Government for his PhD grant. A. de Diego is grateful to the Spanish Government for his post-doctoral fellowship.REFERENCES 1 Craig, P. J., in Organometallic Compounds in the Environment, Principles and Reactions, ed. Craig, P. J., Longman, Essex, 1986, pp. 65–101. 2 Moore, J. W., and Ramamoorthy, S., in Heavy Metals in Natural Waters, Applied Monitoring and Impact Assessment, eds. Moore, J. W., and Ramamoorthy, S., Springer-Verlag, New York, 1984, pp. 125–160. 3 Kuto, A., Akagi, H., Mortimer, D. C., and Miller, D. R., Nature (L ondon), 1977, 270, 419. 4 Rudd, J. W. M., and Turner, M. A., Can. J. Fish. Aquat. Sci., 1983, 40, 2218. 5 Miller, D. R., and Akagi, H., Ecotoxicol. Environ. Safety, 1979, 3, 36. 6 Hempel, M., Chau, Y. K., Dutka, B. J., McInnis, R., Kwan, K. K., Fig. 9 Typical chromatograms obtained by Et–CT–GC–QFAAS of and Liu, D., Analyst, 1995, 120, 721. 1 ml of nitric acid extract, after microwave-assisted extraction of about 7 Quevauviller, Ph., Maier, E. A., and Griepink, B., in Quality 1 g of IAEA-356, BCR S19 and CRM 580 sediments: extractant, 10 ml Assurance for Environmental Analysis, eds.Quevauviller, Ph., of 2 mol l-1 HNO3; power, 60 W; irradiation time, 3 min; absorbance Maier, E. A., and Griepink, B., Elsevier Science, Amsterdam, in arbitrary units. 1995, pp. 1–25. 8 Quevauviller, Ph., Fortunati, G. U., Filipelli, M., Baldi, F., Bianchi, M., and Muntau, H., Appl. Organomet. Chem., 1996, 10, 537. The proposed analytical procedure reduces markedly the 9 Westo�o� , G., Acta. Chem.Scand., 1967, 20, 1790. time needed for sample preparation. While conventional sol- 10 Westo�o� , G.. Acta. Chem. Scand., 1968, 22, 2277. vent extraction methods12,17,18 take about 4 h, microwave- 11 May, K., Stoeppler, M., and Reisinger, K., T oxicol. Environ. assisted techniques only need about 5 min for quantitative Chem., 1987, 13, 153. recovery of the target analytes. Extraction times of about 90 12 Hempel, M., Hintelmann, H., and Wilken, R.-D., Analyst, 1992, 117, 669. and 30 min are, respectively, required when distillation17,21 or 13 Hintelmann, H., and Wilken, R.-D., Appl.Organomet. Chem., supercritical fluid extraction procedures46 are employed. In 1993, 7, 173. addition, keeping the number of analytical steps to a minimum 14 Quevauviller, P., Donard, O. F. X., Wasserman, J. C., Martin, considerably reduces the sources of analytical errors. F. M., and Schneider, J., Appl. Organomet. Chem., 1992, 6, 221. Centrifuged sediment extracts can be directly injected into the 15 Bloom, N.S.. Can. J. Fish. Aquat. Sci., 1989, 46, 1131. Et–CT–GC–QFAAS hyphenated system without further 16 Liang, L., Horvat, M., Cernichiari, E., Gelein, B., and Balogh, S., T alanta, 1996, 43, 1883. manipulation. Sample throughput is controlled by analysis 634 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 1217 Horvat, M., Bloom, N. S., and Liang, L., Anal. Chim. Acta, 1993, 34 Szpunar, J., Schmitt, V. O., Lobinski, R., and Monod, J.-L., 281, 135.J. Anal. At. Spectrom., 1996, 11, 193. 18 Lee, Y. H., Munthe, J., and Iverfeldt, A° ., Appl. Organomet. Chem., 35 Tsalev, D. L., Sperling, M., and Welz, B., An2, 117, 1729. 1994, 8, 659. 36 Welz, B., Tsalev, D. L., and Sperling, M., Anal. Chim. Acta, 1992, 19 Horvat, M., Mandic, V., Liang, L., Bloom, N. S., Padberg, S., 261, 91. Lee, L.-H., Hintelmann, H., and Benoit, J., Appl. Organomet. 37 Rapsomanikis, S., and Craig, P. J., Anal. Chim.Acta, 1991, Chem., 1994, 8, 533. 248, 563. 20 Floyd, M., and Sommers, L. E., Anal. L ett., 1975, 8, 525. 38 Jocelyn, J. R. P., and Be�langer, J. M. R., T rends Anal. Chem., 21 Horvat, M., May, K., Stoeppler, M., and Byrne, A. R., Appl. 1994, 13, 176. Organomet. Chem., 1988, 2, 515. 39 Zlotorzynski, A., Crit. Rev. Anal. Chem., 1995, 25, 43. 22 Cappon, C. J., and Smith, J. C., Anal. Chem., 1977, 49, 365. 40 Baghurst, D. R., and Mingos, D. M. P., J. Chem. Soc., Chem. 23 Cela, R., Lorenzo, R. A., Mejuto, M. C., Bollain, M. H., Casais, Commun., 1992, 674. M. C., Botana, A., Rubi, E., and Medina, M. I., Mikrochim. Acta, 41 Longbottom, J. E., Dressman, R. C., and Lichtenberg, J. C., 1992, 109, 111. J. Assoc. O. Anal. Chem., 1973, 56, 1297. 24 Abu-Samra, A., Morris, J. S., and Koirtyohann, S. R., Anal. 42 Horvat, M., Byrne, A. R., and May, K., T alanta, 1990, 37, 207. Chem., 1975, 47, 1475. 43 Tseng, C. M., de Diego, A., Martin, F., Amouroux, D., and 25 Nadkarni, R. A., Anal. Chem., 1984, 56, 2233. Donard, O. F. X., J. Anal. At. Spectrom., in the press. 26 Fischer, L. B., Anal. Chem., 1986, 58, 261. 44 Puk, R., and Weber, J. H., Anal. Chim. Acta, 1994, 292, 175. 27 Hocquellet, P., and Candillier, M.-P., Analyst, 1991, 116, 505. 45 Hoening, M., and Guns, M. F., in Quality Assurance for 28 Lopez-Avila, V., Young, R., and Berkert, W. F., Anal. Chem., Environmental Analysis, eds. Quevauviller, Ph., Maier, E. A., and 1994, 66, 1097. Griepink, B., Elsevier Science, Amsterdam, 1995, pp. 63–88. 29 Lopez-Avila, V., Benedicto, J., Charan, C., and Young, R., Environ. 46 Emteborg, H., Bjorklund, E., Odman, F., Karlsson, L., Sci. T echnol., 1995, 29, 2709. Mathiasson, L., Frech, W., and Baxter, D. C., Analyst, 1996, 30 Stout, S. J., Dacunha, A. R., and Allardice, D. G., Anal. Chem., 121, 19. 1996, 68, 653. 31 Szpunar, J., Schmitt, V. O., Donard, O. F. X., and Lobinski, R., T rends Anal. Chem., 1996, 15, 181. Paper 7/00832E 32 Donard, O. F. X., Lalere, B., Martin, F., and Lobinski, R., Anal. Received February 5, 1997 Chem., 1995, 67, 4250. AcceptedMarch 13, 1997 33 Lale�re, B., Szpunar, J., Budzinski, H., Garrigues, P., and Donard, O. F. X., Analyst, 1995, 120, 2665. Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12

ISSN:0267-9477    

DOI:10.1039/a700832e

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Continuous Hydride Generation System for the Determination of Trace Amounts of Bismuth in Metallurgical Materials by Atomic Absorption Spectrometry Using an On-line Stripping-type Generator/ Gas–Liquid Separator SOLANGE CADORE* AND NIVALDO BACCAN Universidade Estadual de Campinas, Instituto de Quý�mica, Caixa Postal 6154, 13.083–970 Campinas, SP, Brazil A method for the determination of trace amounts of bismuth chemist with an excellent tool to detect and quantify hydrideusing flow injection and atomic absorption spectrometry with forming elements, owing to its high sensitivity and hydride generation was developed. The introduction of 50 ml of simplicity.13–16 sample and tetrahydroborate solution into the HCl and The concept of the technique of hydride generation was aqueous carriers in a merging zones manifold allows the developed by Holak17 and is based on the reaction between formation of bismuthine, which is separated from the liquid the acidified sample and a reducing agent, which forms the phase in an on-line stripping-type generator/gas–liquid volatile hydride.Since then, eorts have continually been made separator. The calibration graph is linear from 0.1 to to improve the reproducibility and sensitivity, introducing 100 ng ml-1 Bi with a detection limit (3s) of 320 pg ml-1 Bi tetrahydroborate as reductor18 and quartz atomization tubes.19 (corresponding to 16 pg Bi). The relative standard deviation In order to increase the analytical signal it is recommended for 20 replicates varies from 10% for 0.1 ng ml-1 Bi to 1.9% that the hydride should be generated and transferred to the for 100 ng ml-1 Bi, with an injection frequency of up to 150 atomizer as quickly as possible, diminishing its dilution by the samples h-1.NiII, CoII, AgI, HgII, SeIV and SbIII interfere, but carrier gas. The concept of flow injection analysis (FIA)20 is they can be masked with a thiourea–KI solution. The prescribed as the method of choice to overcome this problem applicability of the proposed method to metallurgical samples of dilution and increase the sample throughput.Low sample was demonstrated by the analysis of certified reference volumes are involved and miniaturization, which reduces the materials. contact time between the reagents, minimizes interference eects.21 The combination of bismuthine (BiH3) generation Keywords: Bismuth determination; metallurgical materials ; with atomic absorption spectrometry has been the subject of flow injection hydride generation; atomic absorption a number of investigations with automated methods,21–25 in spectrometry dierent materials.A° stro�m21 used an FIA system in which a bismuth sample of Quality control of industrial products demands a continuous 700 ml was injected into a continuous flowing stream of HCl. development and improvement of new analytical chemical Under the best conditions for the system, a detection limit of methods.The presence of elements at trace levels requires 0.08 ng ml-1 Bi was obtained. Chan et al.22 described an instrumental methods which in turn might be automated. In automated method in which the sample reacted with a masking particular, the presence of hydride-forming elements in steel reagent and the resulting mixture was introduced into an HCl may impart deleterious changes to the physical properties of stream. The hydride, once generated, was separated and trans- this material, as desirable or undesirable eects. Among these ported to the atomizer by an argon flow, after passing through elements, bismuth is important because its addition to metallur- an H2SO4 impinger in order to eliminate excess of water gical materials can aect their quality, positively or negatively, vapour, resulting in a detection limit of 20 ng g-1 Bi.Chan depending on its concentration and the composition of the and Hon23 utilized a flow injection (FI) system with the material.1,2 It has been reported that the addition of bismuth introduction of 200 ml of sample solution, obtaining a detection to Al–Mg alloys improves their corrosion resistance due to limit of 0.17 ng of bismuth using calibration by standard the protective action of Bi2Mg3 compounds in its structure.3 additions.The determination of bismuth and other elements As well as avoiding the formation of graphite nodules, bismuth forming volatile covalent hydrides was described by Schmidt also promotes iron carbide stabilization during the solidifi- et al.;24 according to these workers, the sensitivity of the cation process, when added to iron or steels.4 On the other automated system is comparable to, and, in most cases, better hand, even small amounts of this element may produce a than the manual technique and is considered superior with decrease in hot ductility, workability and cause the rupture of respect to reproducibility and ease of operation. A detection alloys5 and steels.6 limit of 0.24 ng ml-1 Bi was obtained with a high sample flow Among the methods that have been used for bismuth rate (50–100 ml min-1).Yamamoto et al.25 found that gas determination are those based on spectrophotometry7,8 and segmentation with N2 in FIA is an eective method to prevent polarography,9 but they lack sensitivity or selectivity. Flame sample zone broadening during mixing before gas generation. atomic absorption spectrometry10 shows selectivity but the When 0.5 ml of sample was used, Bi, As, Se, Sb and Te were sensitivity is not sucient for metallurgical analysis.Graphite determined with detection limits of 0.04–0.3 ng. furnace atomization has good sensitivity,5,11,12 but occasionally The eciency of the separation of the generated hydride matrix elements other than the analyte may be volatilized from the liquid phase is important in order to ensure that the together, causing interferences.The hydride generation method, coupled with atomic absorption, has provided the analytical analyte is transferred to the atomizer. This is provided by a Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 (637–642) 637Fig. 1 Schematic diagram of the hydride generation system. L= Injection volume; RC=reaction coil; PP=peristaltic pump; AAS= atomic absorption spectrometer. gas–liquid separator and the most common type is a U-tube;21,26 however, sometimes it is dicult to obtain a constant drain of the liquid phase and problems such as dilution of the generated hydride, which causes a low signal and low reproducibility, might occur.Many other designs have been described in the literature, making modifications to improve the hydride separation. A packed U-tube,27 a cooled U-tube,28 porous membranes29 and a porous tube30 separator have all been proposed, having smaller dead volumes than Utube separators. Despite the higher sensitivities claimed, most of these separators show the best performance with a high acid concentration, producing a large amount of H2 when in contact with the tetrahydroborate solution.A very convenient method to remove the generated hydride from the aqueous Fig. 2 Stripping-type gas–liquid reactor–separator used in the con- medium was described by Schmidt et al.24 as a typical forced tinuous-flow hydride generation system. outlet separator made with a medium porosity Bu�chner fritted glass funnel with a stopper inserted in the upper end.Three (Micronal, Model B332 II). Samples and reagents were aspir- tubes were inserted through the stopper to accommodate the ated through Tygon tubes (Technicon) with appropriate flow sample input and gaseous hydride, argon and waste removal. rates for each solution. Sample loops, reaction coils and The argon used to purge the hydride was introduced through transmission lines were prepared using Teflon tubing (CPL, the lower end of the funnel.This system works with large 0.8 mm id). For the introduction of solutions into the system, volumes on-line. Brindle and co-workers31,32 described another a proportional injector made from acrylic was used.33 For the forced outlet gas–liquid separator based on a glass frit for the separation of the formed hydride from the liquid phase a determination of arsenic and antimony using DC In stripping-type gas–liquid reactor–separator was utilized, both designs, sample and reducing agent solutions are continuinstead of a U-shaped25 separator.The separator used consists ously introduced into the generator/separator chamber. An of three parts made of acrylic and is shown in Fig. 2. The hole argon flow of 0.4–0.5 l min-1 was introduced into the generator through which the mixture (gaseous hydride+liquid fraction) through the glass frit to remove the hydrides from solution. enters the separator and those through which the liquid Another flow of argon (2 l min-1) was introduced into the fraction leaves were at the same height, allowing the presence vessel from the top. The high flow of argon sweeps the hydrides of a small liquid residue (the residence volume is about 200 ml) into the plasma through the outlet part of the device via a U-tube buer tank, partly filled with water.This buer tank is inside the separator. N2 introduction at the bottom strips the necessary in this design in order to moderate any pressure hydride through a sintered glass frit to the atomizer, using a fluctuation in the hydride transportation line to the plasma.Teflon tube (10 cm×3 mm), in a uniform and constant way. This paper describes a simple stripping-type gas–liquid This enhances the separation eciency which is related to the separator that eciently enhances the separation and transpor- sensitivity. tation of BiH3 to the quartz atomization cell. This eciency After separation, the hydride was transported with a conis reflected in a greater sensitivity and excellent reproducibility, trolled N2 flow rate to an electrothermal cell, constructed from due to the low residence volume (about 200 ml) inside the a 170×8 mm id quartz tube, fused at the centre with a chamber.Furthermore, small sample and reagent volumes 100×2 mm id quartz tube to form a T-shaped atomizer. The (50 ml) are used in a merging-zones configuration of the FI cell was wound with 2 m of Ni–Cr wire (0.45 mm diameter; manifold.The N2 flow, used as purge gas, was 120 ml min-1, 8.3 V m-1) and isolated with an asbestos strip, except for the causing an ecient mixing and removal of the bismuth hydride 2 cm extremities of the tube. This allowed more stable signals with subsequent transference to the atomization cell. The because of the elimination of ignition of the H2 at the end of relevant parameters were optimized taking advantage of this the tube.34 In order to obtain higher temperatures and to stripping-type generator/gas–liquid separator, increasing the minimize heat changes, the cell was surrounded with a two eciency of the separation of BiH3 from the liquid phase.part aluminium tube, with a 2 mm thick tube-wall, fixed The interference eect of a number of ions was evaluated as together under pressure, as shown in Fig. 3. Before use, the well as the use of masking agents. The method developed was cell was washed with a 1+9 HF solution and then treated successfully applied to the determination of bismuth in metal- with 5% m/v dichlorodimethylsilane or trimethylchlorosilane lurgical materials. in toluene for 2 h and dried for 1 h at 90–100 °C.The silane solution reacts with hydroxyl groups, eliminating active sites on the glass surface, allowing the best sensitivity and repeat- EXPERIMENTAL ability of the desired reaction.35 This treatment was repeated Apparatus after about 500 injections, and could be repeated as often as necessary.The BiH3 atomization temperature was obtained The FI system is shown in Fig. 1 and consists of a mergingzones manifold33 with a six-channel peristaltic pump with a Varivolt regulator, which is connected to the Ni–Cr 638 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12pared with those obtained with graphite furnace atomization, according to the procedure described by Gladney.36 RESULTS AND DISCUSSION Using the merging zones manifold, equal volumes of sample and reducing agent are introduced into separate carrier streams, which have the same flow rate in order to ensure perfect homogenization of the solutions.The generation of BiH3 occurs in a reaction coil which is located after the confluence Fig. 3 Heated quartz tube used as the electrothermal atomizer for on-line atomization of BiH3. point of the system. As the hydride is generated in an acidic medium the carrier for the sample is an HCl solution while de-ionized water was chosen for the tetrahydroborate carrier, because of the formation of H2 in the presence of acid.When the tetrahydroborate reacts with the acidified sample solution the following reactions take place: BH4-+3H2O+H+�H3BO3+8H Bi3++6H �BiH3+3H+ The hydride is separated from the liquid phase in a gas– liquid separator and then swept to the atomizer where, according to Welz and Melcher,37 the atomization in the quartz tube Fig. 4 Dependence of the absorbance on the reaction coil length for is due to collisions between hydrogen radicals: dierent analyte concentrations. CHCl=1.0 mol l-1 ; CNaBH4 =1.0%; Tat=900 °C; N2 flow rate=110 ml min-1.BiH3+H � BiH2+H2 BiH2+H � BiH+H2 coil of the cell. The electrothermal cell was aligned in the optical path of an atomic absorption spectrometer (Varian, BiH+H�Bi+H2 Model Gemini AA 12/1475) equipped with a deuterium background corrector. During all the optimization steps, the background corrector was kept on in order to check for the existence of significant background.We did not observe any Optimization of FI Hydride Generation System relevant background that needed to be corrected. The bismuth Reaction coil (reaction time) and injection volume hollow cathode lamp was operated at 8 mA with a wavelength of 223.1 nm and a slit-width of 0.2 nm. The signals obtained The BiH3 is formed when the sample reacts with a tetrahydrowere registered by an Epson LX-800 printer. borate solution; hence, the time for which the solutions are in contact was evaluated.Dierent lengths (10, 15, 30, 50 and 75 cm) of reaction coil were prepared and tested while also Reagents varying the injection volume (30, 50, 75, 100 and 150 ml) for All reagents used were of analytical-reagent grade. De-ionized three dierent concentrations of bismuth (40, 80 and water was used throughout. 120 ng ml-1). For each reaction coil the absorbance signal The bismuth standard solution was prepared from metallic increases with the increase in injected volumes.Fig. 4 shows bismuth (J. T. Baker, 99.99%) treated with concentrated HCl the eect of the reaction coil in the determination of bismuth and HNO3 and finally diluted with 1.0 mol l-1 HCl. Appropriate dilution was made from this solution, whenever necessary, with 1.0 mol l-1 HCl. Solutions of sodium tetrahydroborate were prepared by dissolving NaBH4 powder (Merck) in 0.05 mol l-1 KOH and were stored in plastic bottles, under refrigeration.The solution is stable for about 2 weeks, with no loss of the observed absorption signal. Solutions of interferent ions and masking agents were prepared by the dissolution of appropriate salts in acid or de-ionized water. Fig. 5 Eect of the amount of NaBH4 used on the hydride generation with dierent analyte concentrations. Vinj=50 ml; RC=15 cm; N2 flow Samples rate=120 ml min-1; Tat=900°C; carrier flow rate (HCl and H2O)= 1.9 ml min-1.T in alloy. A 0.1 g amount of certified alloy was dissolved in 3 ml of concentrated HNO3 and heated. The filtered solution Table 1 Peak area for the injection of 4 ng Bi. Reactor–separator was treated with concentrated HCl and heated. The sample height: 6 cm; 1.0 mol l-1 HCl; 1.0% NaBH4 in 0.05 mol l-1 KOH; N2 was diluted to 50.0 ml with 1.0 mol l-1 HCl. at 120 ml min-1 Steels. A 0.1 g amount of material was treated with 3 ml of Injection volume/ Bismuth concentration/ Peak area/ concentrated HNO3. After evaporation, concentrated HCl was ml ngml-1 As added and the solution was heated.The sample was diluted 30 133.3 3.849 to 10.0 ml with 1.0 mol l-1 HCl. 50 80.0 3.901 Bronze/Brass. Appropriate amounts of material were treated 75 53.3 4.058 with concentrated HNO3 and HCl and heated. After filtration, 100 40.0 3.846 the solutions werwith 1.0 mol l-1 HCl. The results 150 26.7 3.665 obtained with hydride generation measurements were com- Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 639for a sample and reducing agent volume of 50 ml. The absorbance is slightly better for a 15 cm coil, decreasing when the reaction time increases, due to sample dilution. In addition, the instability of the BiH3 should be considered. As has previously been reported, bismuth hydride is unstable and thermal decomposition38 might occur, even at ambient temperature. 21 Fujita and Tanaka38 concluded that bismuth hydride is very unstable, decomposing slowly, even at a temperature of 25°C. They included a kinetic factor to be considered, inducing the decomposition of the hydride.Fig. 7 Influence of the height of the separator on the absorbance of Furthermore, they suggested that hydride generation was dierent analyte concentrations. Analytical conditions as in Fig. 5. relatively rapid, but the product hydride was thermally unstable CHCl=1.0 mol l-1. and should be removed from the solution as quickly as possible to produce the maximum absorbance value.This is a condition 0.5 to 5.0 mol l-1; hence, control of the acidity during sample that was used in our approach with the stripping-type gas– preparation is not a critical step for bismuth determination. liquid separator. Thus, considering the sensitivity and repro- However, this is valid only for HCl medium. If there is a ducibility, a reaction coil of 15 cm is recommended for the mixture of HCl with another acid, this becomes a serious determination of bismuth.With respect to injection volume, problem, considering that the absorbance signal decreases with the reproducibility and the frequency decrease when volumes an increase in the concentration of the second acid. Replacing greater than 100 ml are injected. To choose this parameter, HCl with HNO3, H2SO4, H3PO4 or HClO4 leads to a lower however, it is important to consider the eciency of the sensitivity owing to interactions between the analyte and hydride separation from the liquid phase.Injecting dierent NO3-, SO42- or ClO4- or even the formation of less reactive volumes of sample with variable concentration but with the species between H3PO4 and BH4-. same final mass of bismuth, the peak area was calculated. Table 1 shows that volumes between 30 and 100 ml can be injected without significant change in the area. In this instance, Eect of the reactor–separator height the stripping-type gas–liquid separator is ecient for this The proposed stripping-type reactor–separator consists of three system and a volume of 50 ml was used.separate parts. The intermediate part can have dierent heights, which might aect the final result because it increases the Eect of sodium tetrahydroborate concentration distance to the atomizer. Three dierent heights were tested: 4, 6 and 10 cm. An increase in the length leads to a decrease The purpose of the tetrahydroborate is to provide hydrogen in the absorbance signal (Fig. 7), reflecting the dilution of the radicals to react with bismuth in order to generate the hydride. hydride. The best eciency was obtained with the intermediate As can be seen in Fig. 5, the absorbance increases with the part at 4 cm; however, it was also noticed that some water concentration of BH4- up to 2.0% m/v, for all the bismuth vapour condensed in the Teflon tube wall used to connect the concentrations tested. A concentration of 1.0% m/v was separator to the atomizer, causing less reproducibility of the chosen, considering that higher concentrations of NaBH4 signals.This cannot be controlled by the use of drying agents generate an excess of H2 , which leads to low reproducibility such as H2SO4 or CaCl2. Thus, the 6 cm intermediate part due to the dilution of BiH3. was found to be the most appropriate and was used during the remainder of this work. Eect of acid concentration Using the same concentration of HCl for dilution of the Eect of carrier gas flow rate bismuth solution and as its carrier, it was observed (Fig. 6) Flow rates of N2 less than 90 ml min-1 showed lack of that the generation of BiH3 is virtually independent of the reproducibility and large, poorly shaped signals. Above acidity above 0.5 mol l-1, as has previously been reported 130 ml min-1, a higher injection frequency associated with by A° stro�m21 and Yamamoto et al.25 It is important to emphasdilutionof the BiH3 was observed and, consequently, a decrease ize that the presence of the acidic medium is essential to in the signals.For this work, flow rates between 110 and generate the hydride. When the sample is prepared with 120 ml min-1 were chosen. de-ionized water instead of HCl, the absorbance signal can decrease up to five times, showing that the presence of H+ is needed to liberate the H2 that will react with the Bi3+ ion. On Eect of reagent flow rate the other hand, no change in absorbance is observed when the When the merging zones manifold is used the flow rate of both sample carrier is kept constant (1.0 mol l-1) and the concencarrier reagents must be the same.It was observed that up to tration of the acid used to prepare the sample changes from a flow rate of 2.8 ml min-1 the absorbance signal increases Fig. 6 Eect of the HCl concentration on the eciency of hydride generation with dierent analyte concentrations. CNaBH4 =1.0% in Fig. 8 Bismuth absorbance as a function of the atomization tempera- 0.05 mol l-1 KOH; Vinj=50 ml; carrier flow rate (HCl and H2O)= 1.9 ml min-1. ture with dierent analyte concentrations. Conditions as in Fig. 5. 640 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12Fig. 11 Interferences of some cations on the formation of BiH3 in the Fig. 9 Calibration graph for bismuth using the on-line hydride gener- aqueous phase. CBi=40 ng ml-1. ation system. MnII, MoVI, TiIV, VV and ZnII did not aect the absorbance and above this value the sample and reducing agent do not signal even when they were present in a 5000- or 10000-fold have sucient time to react before reaching the reactor– excess and were not considered as liquid phase interferents.separator. Considering the injection frequency and the sensi- AsIII, AsV and PbII did not interfere in the gaseous phase in a tivity, a flow rate of 1.9 ml min-1 was considered suitable for 10000-fold excess, nor did SbV and SnIV in a 5000-fold excess.bismuth determination. However, a significant reduction in the bismuth signal was caused by small amounts of CoII, CuII, NiII, WVI, HgII, SbIII Eect of the temperature of atomization and AgI. The eect of the interferents in the liquid and gaseous phases is shown in Figs. 11 and 12, respectively. The compe- Once separated, the hydride is quickly transported to the tition for the reducing agent can explain the interference in atomization cell and no critical eects of the changes in the the liquid phase as suggested by Smith.40 In the gaseous phase, atomization temperature were observed.It was verified that the competition for both the tetrahydroborate and the hydro- the atomization temperature is not critical above 850 °C, as gen radicals inside the atomizer39 seems to be the cause of the can be seen in Fig. 8, and 900 °C is suggested as a working absorbance signal reduction. temperature. This value is consistent with those reported In order to overcome these interferences, several reagents by A° stro�m,21 Chan et al.22 and Crock,28 with a similar profile were tested as masking agents.Considering that there is a to the plot of absorbance versus quartz cell temperature. competition for the reducing agent it is advisable to introduce a second reductor species into the system. To maintain the Analytical Performance of the Method simplicity of the FI manifold, the studied reagents were dissolved in 1.0 mol l-1 HCl and used as the sample carrier.The calibration graph with typical analytical signals (Figs. 9 When the sample and the reductor are kept in contact, the and 10) is linear from 0.1 to 100 ng ml-1 of bismuth. The masking agent is already acting over the interferent species, detection limit, calculated as three times the standard devon thus reducing the competition for the tetrahydroborate. As of the blank signal, was 320 pg ml-1, which corresponds to thiourea, L-cysteine and thiosemicarbazide have a disulfide 16 pg of bismuth. The relative standard deviation, for 20 group that can easily be reduced,41 they were investigated as replicate determinations, varies from 10% for 0.1 ng ml-1 Bi masking agents.The results showed that a 0.2% thiourea to 1.9% for 100 ng ml-1 Bi. The injection frequency, under the solution and a mixture of thiourea (0.2%)–L-cysteine (1%) optimized conditions, was 120–150 samples h-1. were ecient in preventing interferences if only one interferent species was present in the sample.Ascorbic acid, citric acid, Interferences hydroxylamine and potassium iodide were also studied but were not suciently ecient to overcome all the interferent According to Dedina,39 the interferences shown by the hydride species. The use of a thiourea (0.2%)–KI (10%) solution generation technique can be classified into two groups: the allowed the recovery of the bismuth signal in the presence of liquid phase interferences, where hydride formation occurs, one or more foreign ions.Table 2 shows the recovery of the and the gaseous phase interferences, occurring either during analytical signal for bismuth spiked with dierent types of hydride transport or in the atomizer. In order to identify interferents in the presence of the thiourea–KI solution. For species that could interfere with the determination of bismuth samples with a high content of tin, the Sn2+ was separated by in the liquid phase, 18 foreign ions were studied.Five species precipitation as metastannic acid, which is formed in an HNO3 (Sn, Sb, As, Se, Pb) which generate hydrides, and are classified medium, and filtered prior to injection. as potential gaseous interferents, were also studied. Arsenic, tin and antimony were investigated in dierent valency states. In addition, mercury was also investigated because it can be Applications reduced to Hg0 when it is in contact with tetrahydroborate.A The accuracy of the method was examined by analysing several species was considered as a potential interferent if the dierence NIST SRMs. The determination of bismuth was also performed between the absorbance for bismuth and that for bismuth in the presence of a particular interferent was higher than 10%. Under these conditions, AlIII, CaII, CdII, CrIII, FeII, FeIII, MgII, Fig. 10 Typical analytical signals showing the reproducibility of the Fig. 12 Influence of some elements occurring in the gaseous phase injections. (The dotted line represents the monitoring of the background.) on the absorbance of the analyte. CBi=40 ng ml-1. Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 641Table 2 Recovery of the analytical signal of Bi in the presence of interferent species under the masking eect of thiourea (0.2%).KI (10%) Sample Bi5interferent ratio Analytical signal recovery for Bi (%) Bi5NiII5CoII5CuII5AgI 15500055000510005200 94 Bi5SnIV 155000 100 Bi5SbIII 15500 100 151000 90 Bi5SeIV 1510 100 15100 91 Bi5NiII5CoII5CuII5AgI5SeIV5SbIII5SnIV 15100051000510005100525550051000 93 Table 3 Bismuth determination in samples of metallurgical interest 4 Aborn, R.H., Bull. Bismuth Inst., 1975, 7, 1. 5 Headridge, J. B., and Thompson, R., Anal. Chim. Acta, 1978, 102, 33. Amount of Bi (%) 6 Zhou, N., Frech, W., and Lundberg, E., Anal. Chim. Acta, 1983, 153, 23. Sample* This work ETAAS Certified value 7 Marczenko, Z., Spectrophotometric Determination of Elements, SRM 54D 0.044¡¾0.001 . 0.044¡¾0.005 Ellis Horwood, Chichester, 1976. SRM 361 0.00036¡¾0.00001 . 0.0004¢Ó 8 Beinrohr, E., and Hofbauerova¢¥, H.,Mikrochim. Acta, 1989, II, 119. SRM 362 0.0018¡¾0.0001 . 0.002¢Ó 9 Rooney, R. C., Analyst, 1976, 101, 749. SRM 364 0.00096¡¾0.00002 . 0.0009¢Ó 10 Barnett, W. B., and McLaughlim, E. A., Jr., Anal. Chim. Acta, Brass 0.0011¡¾0.0002 0.0011¢Ô 1975, 80, 285. Bronze de Can.o¢¥ n 0.0059¡¾0.0001 0.0067¢Ô 11 Andrews, D.G., and Headridge, J. B., Analyst, 1977, 102, 436. Fluorescent bronze 0.00075¡¾0.00001 ND¡× 12 Headridge, J. B., and Smith, D. R., T alanta, 1972, 19, 833. 13 Drinkwater, J. E., Analyst, 1976, 101, 672. * SRM 54D: Sn (88.5%); Sb (7.04%); Cu (3.62%). SRM 361: 14 Hon, P. K., Lau, O. N., Chung, W. C., and Wong, M. C., Anal. Ni (2.0%); Cr (0.69%); Mn (0.66%); Mo (0.19%). SRM 362: Chim. Acta, 1980, 115, 355. Mn (1.04%); Cu (0.50%); Ni (0.59%); Co (0.30%); Cr (0.30%). 15 Vanloo, B., Dams, J., and Hoste, J., Anal. Chim. Acta, 1983, SRM 364: Mn (0.25%); Cu (0.24%); Ni (0.14%); Mo (0.45%); 151, 391. Co (0.15%); Ti (0.24%). Brass: Cu (50.90%); Zn (20.40%); 16 Welz, B., and Melcher, M., Spectrochim. Acta, PartB, 1981, 36, 439. Pb (0.2%); Sn (0.6%); P (0.0.1%); As (0.0.1%); Sb (0.0.1%); 17 Holak, W., Anal. Chem., 1969, 41, 1712. Ni (<0.25%); Fe (0.1%). Bronze de Can.o¢¥ n: Cu (60.97%); 18 Braman, R. S., Justen, L.L., and Foreback, C. C., Anal. Chem., Sn (10.35%);Zn (<2%). Fluorescent bronze: Cu (#90%);Sn (#10%); 1972, 44, 2195. Zn (<2%); P (0.2.1.5%). 19 Chu, R. C., Barron, G. P., and Baumgarner, P. A. W., Anal. ¢Ó Value for reference only, not certified. Chem., 1972, 44, 1476. ¢Ô Standard deviation within 10%. 20 Ru¡Æz¢§ ic¢§ka, J., and Hansen, E. H., Flow Injection Analysis, Wiley, ¡× ND: Not detected. New York, 2nd edn., 1988. 21 A¡Æ stro¡§m, O., Anal. Chem., 1982, 54, 190. 22 Chan, C. Y., Baig, M. W. A., and Pitts, A. E., Anal. Chim. Acta, in dierent metallurgical samples. The results obtained 1979, 111, 169. (Table 3) were compared with certified values or graphite 23 Chan, W.-F., and Hon, K.-P., Analyst, 1990, 115, 567. furnace measurements. The close agreement between the results 24 Schmidt, F. J., Royer, J. L., and Muir, S. M., Anal. L ett., 1975, obtained by the proposed method and the certified values 8, 123. shows the good accuracy of the method.This was also con- 25 Yamamoto, M., Makoto, Y., and Yamamoto, Y., Anal. Chem., firmed by the analysis of steels, bronze and brass, the results 1985, 57, 1382. 26 Vijan, P. N., and Wood, G. R., At. Absorpt. Newsl., 1974, 13, 33. of which are included in Table 3. 27 Pierce, F. D., Lamoreaux, T. C., and Fraser, K. S., Appl. Spectrosc., 1976, 30, 38. CONCLUSIONS 28 Crock, J. G., Anal. L ett., 1986, 19, 1367. 29 Pacey, G. E., Strata, M. R., and Gord, J. R., Anal. Chem., 1986, Bismuth determination using hydride generation atomic 58, 502. absorption spectrometry coupled with an FI system was found 30 Yamamoto, M., Takada, K., Kumamaru, T., Yasuda, T., and Yokoyama, S., Anal. Chem., 1987, 59, 2446. to be simple, with low reagent consumption, relatively inter- 31 Brindle, I. D., Alarabi, H., Karshman, S., Le, X., Zheng, S., and ference-free and sensitive. This sensitivity combined with signal Chen, H., Analyst, 1992, 117, 407. reproducibility is mainly due to the stripping-type gas.liquid 32 Chen, H., Brindle, I. D., and Zheng, S., Analyst, 1992, 117, 1603. separator because of the low residence volume of the liquid 33 Bergamin Fo, H., Zagatto, E. A. G., Krug, F. J., and Reis, B. F., phase therein, which leads to better separation eciency. Anal. Chim. Acta, 1978, 101, 17. The results obtained for metallurgical samples showed good 34 Thompson, K. C., and Thomerson, D. R., Analyst, 1974, 99, 595. 35 Parisis, N. E., and Heyndrickx, A., Analyst, 1986, 111, 281. accuracy and precision. 36 Gladney, E. S., At. Absorpt. Newsl., 1977, 16, 114. 37 Welz, B., and Melcher, M., Analyst, 1983, 108, 213. The authors thank Carol H. Collins for assistance with the 38 Fujita, K., and Tanaka, T., T alanta, 1986, 33, 203. English in this manuscript. 39 Dedina, J., Anal. Chem., 1982, 54, 2097. 40 Smith, A. E., Analyst, 1975, 100, 300. 41 Boampong, C., Brindle, I. D., Le, X., Pidwerbesky, L., and REFERENCES Ponzoni, C. M. C., Anal. Chem., 1988, 60, 1185. 1 Fleming, H. D., and Ide, R. G., Anal. Chim. Acta, 1976, 83, 67. 2 Yamamoto, M., Yamamoto, Y., and Yamashige, T., Analyst, 1984, Paper 6/06553H 109, 1461. Received September 24, 1996 3 Baba, Y., Hagiwara, M., and Hamada, J., Bull. Bismuth Inst., 1974, 4, 1. Accepted February 18, 1997 642 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12

ISSN:0267-9477    

DOI:10.1039/a606553h

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Flow Injection for Automation in Atomic Spectrometry† Invited Lecture J. L. BURGUERA* AND M. BURGUERA IVAIQUIM (Venezuelan Andean Institute for Chemical Research), Faculty of Sciences, University of L os Andes, P.O. Box 542, Me� rida 5101-A, Venezuela The use of FI methods to extend the capabilities of atomic Flow injection, which was initially based on the sequential insertion of a discrete liquid sample into an unsegmented spectrometric detectors is discussed with emphasis on continuously flowing stream of a suitable liquid, was originally developments in automation.Topics such as direct sample designed to automate serial assays. Today, an FI system, owing introduction and on-line dilution, calibration, dissolution, to its modular character, provides the most flexible way to liquid–liquid extraction, precipitation reactions, ion-exchange automate the number of operations necessary in a chemical preconcentration, generation of volatile species and microwave assay.FI operations are very flexible and versatile, since flows sample pre-treatment are used to illustrate the versatility of can be mixed, trapped, reversed, split, recombined and sampled, coupling FI with either FAAS, ETAAS, ICP-AES, ICP-MS while contact times with selected sections of reagents or sensing and/or AFS. Finally a brief overview of the advances in this surfaces can be precisely controlled. The usefulness of integrat- research field is presented. ing FI with AS was recognized very early in the development Keywords: Flow injection ; automated analysis; atomic of FI systems.9–13 The first paper which used the words ‘flow spectrometry injection’ in the title and which described the presentation of samples to AAS appeared in 1979,9 while reports of FI combinations with ICP-AES were made by three groups.11–14 The combination of FI techniques with AAS15–24 and ICP- As in other scientific, technical and social fields, the reduction AES9,18,19,21,25,26 has been reviewed since 1985.In 1989 and of human participation is one of the primary goals of today’s 1995, the first two books which treated the FI–AS literature analytical chemistry.1 Automation can be considered as the selectively were published,27,28 and in 1997 a third book, which automatically controlled operation of an apparatus, process treats in depth and exclusively the fundamentals and appli- or system by mechanical or electronic devices which makes cations of the field, will appear.29 A fully automated FI system possible the partial or complete replacement of human organs requires: (1) the introduction of standard solutions and sample of observation, eort and decision.According to this definition, solutions with dierent analyte concentrations in an automatic the automation of a chemical analysis can be implemented by fashion with the aid of an autosampler and/or the use of the replacement of human intervention in at least one of the electrically actuated injection valves; (2) accurate control for major operations of the analytical process, viz., the sampling the automatic functioning of peristaltic pumps; (3) straightforstage, pre-treatment stage, measurement, data acquisition or ward automation of the transport–reaction system in order to data treatment.1 If one considers the analytical process to reduce human participation to the desired levels; and (4) a consist of three essential stages (preliminary operations, micro-computer furnished with a passive interface for data measurement and transducing of the analytical signal and data acquisition and treatment and an active interface for concollection and processing), the first stage, namely preliminary trolling operational modules, and governed by appropriate operations (sampling and sample preservation, dissolution/ software.30 Undoubtedly, any FI–AS system can be considered desegregation, separation, analytical reaction and transport to to have to some extent a certain degree of automation, mainly the detector), has so far scarcely been automated despite the of the preliminary operations of the analytical process.31 great interest in reducing human participation in such time- However, in this respect much research has largely been consuming and tedious activities, which are also the source of focused on the development of FI as an eective sampling major errors and are occasionally even hazardous.2 interface with a variety of possible configurations with dierent The developments in atomic spectrometric (AS) techniques optical or electrochemical instruments typically used in routine provide some of the most sophisticated and elegant methods laboratories.Therefore, it is not our aim to present in this for the detection and determination of minor, trace and paper an exhaustive review of the large variety of alternatives ultratrace metals in clinical, environmental, agronomic and previously reported in the field, but to present a rational industrial samples.3 Just as there have been extremely rapid overview of the areas most strongly influenced by advances in automation.developments in AS methods, in particular ETAAS, ICP-AES and, more recently, ICP-MS, there has been a parallel development and expansion in the use of FI4–7 techniques. Features FLAME ATOMIC ABSORPTION such as high precision, high sampling rates, low carryover, SPECTROMETRY calibration techniques, high degree of flexibility, microsample manipulation, ability to perform rapid on-line pre-treatment Zagatto et al.9 initiated the FI–AAS approach with a semichemistry and easy automation have all provided the stimulus automatic FI procedure for the determination of calcium, magnesium and potassium by AAS.In their procedure, a for the development of FI–AS based methodology.3,8 manual double proportional injector32 permitted the utilization of sample and reagent merging zones.The sample dilution and reagent addition processes were performed with two stages † Presented at the Fourth Rio Symposium on Atomic Spectrometry, Buenos Aires and Iguazu�, Argentina, November 24–30, 1996. requiring no human intervention. Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 (643–651) 643Wolf and Stewart10 described the implementation of an FI Bysouth and Tyson54 described a flow manifold that automatically diluted solutions presented to a flame atomic absorp- system adapted for AAS with a remarkable degree of automation.An automatic sampler moved to withdraw either tion spectrometer while maintaining a constant flow to the nebulizer. The manifold is based on a fixed-speed pump, sample (to fill the injection valve loop), carrier solution or loop washing solution. The digital output of the integrator, together with a computer-controlled stream-switching valve and a computer-controlled injection valve configured for which gave a printed and electronic digital output of the peak areas, was stored on magnetic tape cassettes and the data were stream switching.The computer is also used for data acquisition and handling. In addition to the production of solutions processed by use of a programmable calculator. They reported excellent precision (and hence low detection limits) obtainable for a conventional calibration procedure, the system is also used for a calibration procedure whereby the stock standard for zinc and copper based on the improvement in nebulizer performance achieved when the flow rate of the carrier is solutionis automatically diluted by afactor until its absorbance matches that of the sample. controlled by a suitable pump rather than by the oxidant flow rate.Also, Stewart et al.33 reported a microprocessor system Ara�ujo and Lima55 described three FI systems for the automatic determination of sodium, potassium, lithium, mag- to control automatically a sampler, a pair of solenoid gas valves, a fraction collector and the appropriate detector for nesium, zinc, copper and iron in certain biological fluids by AAS.A microcomputer-controlled autosampler coupled to either standard, titration or dilution systems. The sampler has two states, rinse and sample, which are activated by a continu- an injection was used to handle the sample. The setups were designed to allow a split-stream process for ous voltage at either of two locations on a microswitch which is mechanically linked to the motor output.Sample tray determinations requiring high sample dilution. An automatic on-line FI dilution system was developed by advance is automatic with each load-to-rinse transition. The pair of solenoid valves are activated individually to switch the Fang et al.49 based on the injection of low- and submicroliter sample volumes and their dispersion in a mixing coil.Precise sample insertion valve between its load and inject positions. The valve control circuit uses a full wave bridge, as direct control of sample uptake and injection was possible in capillary tubing using computer-controlled stepper motor-driven current solenoid valves are used. The drop-counting feature of the fraction collector is used as the control point for the peristaltic and small-bore Neoprene pump tubes. Sperling et al.56 made an attempt to improve the limits of microcomputer.Nebulizer performance was also considered by Yoza et al.,34 who described a manifold which included a detection for Zn, Cu and Ca by practical evaluation of dedicated signal processing. The signal evaluation mode, the pre-coil (to damp the shock during sample injection) and water- and air-compensation method (to reduce the eect of selection of the peak window and the degree of smoothing applied were investigated. In the peak-height mode, for a wide the flow/aspiration rate ratio).The enhancing eect of organic solvents was utilized by Fukamachi and Ishibashi,35 who range of carrier flow rates the FI technique permits detection limits comparable to those calculated for conventional steady- injected aqueous solutions of a number of metals into a carrier stream of an immiscible solvent (either n-butyl acetate or state evaluation to be obtained, even if the sample volume injected is several times smaller than that conventionally isobutyl methyl ketone) propelled solely by the ‘suction’ of the nebulizer.Some other advantages of using FI in conjunction aspirated. In contrast, the detection limits for integrated signals are considerably worse for FI than steady-state evaluation at with AAS were also pointed out by some of the pioneers of such a combination, e.g., the on-line implementation of the carrier flow rates similar to the uptake rate of conventional aspiration. However, detection limits similar to or better than splitting of the stream for the simultaneous determination of four elements,36 solvent extraction,37 standard addition those for conventional nebulization may be obtained by FI if the ensemble summation procedure for non-integrated FI method38,39 and calibration methods.40,41 Simonsen et al.42 described a completely automated FI–AAS signals is used with at least five replicates at optimum carrier flow rates.This relationship could be improved by reducing procedure for the determination of zinc in human serum.The instrumentation was a simple construction controlled by a the carrier flow rate but the eect was dierent for dierent elements and not in conformity with other aspects of FI such microcomputer with a program written in BASIC to operate an autosampler, a single three-way valve and integration. The as sample throughput. For integrated signals the signal evaluation window should be adjusted to the peak width or the FI system was pumpless, using the negative pressure created by the nebulizer, and the sampling was performed by means signal should be treated with a smoothing routine in order to achieve optimum signal-to-noise ratios.of the valve. No pre-treatment of the sample was necessary and, although it was possible to make peak height measure- A fully automated FI–AAS microtechnique for the determination of metals in whole blood, plasma and blood cells was ments, peak area measurements were used in the procedure in order to carry out quantification by use of an aqueous graph.described by Burguera and Burguera.57 In the FI configuration used, the whole blood sample and saline solution are intro- Most of the early applications of FI–FAAS used FI simply as a sample introduction system.43 Nearly all the applications duced in parallel into the carrier stream of water by means of the valve injector. After the sample and the saline solution described used the limited dispersion FI mode so that the sample handling system could be considered as an automated ‘discrete have passed through dierent lengths of tubing they mix downstream while the sample plug is mixed and dispersed into nebulization’44 or ‘micro-sampling nebulization’45 accessory.However,it should be noted that the sample introduction devices3 the saline solution. The cycle program was fixed to determine zinc in (i) whole blood, by directing the mixed sample–saline were manually actuated, so that the automation of the sample introduction process was only apparently complete.solution directly toward the instrument nebulizer, (ii) blood plasma, by the on-line filtration of the whole blood sample One of the drawbacks of FAAS is its limited linear working range. Thus, depending on which tasks are to be executed before its introduction into the instrument, and (iii) blood cells, by the iterative change of the flow direction of the carrier prior to detection, FI systems may be programmed to dilute or to preconcentrate the analyte so that its concentration falls stream through the filter in order to push the retained blood cell samples toward the instrument nebulizer.within the range of the detector, avoiding tedious multiple dilution steps. The combination of FI techniques with AAS FI procedures based on liquid–liquid extraction processes require the development of the capability to mix and separate has brought about a significant enhancement of the performance of the latter and has given rise to various on-line reproducibly two immiscible liquids.4 In their original matrix, constituents to be detected are transferred on-line from their possibilities, including introduction with a high dissolved solid content,46–48 microsampling,49 zone sampling50 and original matrix to an entirely new and dierent matrix.The merits, theory and applications of such matrix modifications preconcentration and separation processes.28,50–53 644 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12have been discussed.4,58,59 Earlier studies of the utility of the spectrometer. This ‘filterless’ procedure improved the eciency compared with those procedures using on-line filter liquid–liquid extraction in FI with spectrophotometric detection systems were simultaneously carried out by Karlberg and devices, owing to minimization of dispersion and the automation of this preconcentration approach.Similar manifolds Thelander60 and Bergamin et al.61 Since then, a large number of semi-automated and automated FI procedures with AS were described by Sperling et al.,56 the analyte elements (Cd, Co and Ni) being preconcentrated and separated from the detection have been discussed in several monographs.35,59–74 These basically described the on-line enhancement of sensitivity bulk of the biological matrix by on-line coprecipitation without filtration. achieved by using organic solvent carrier streams, extraction of metal ions in aqueous samples into the organic phase, which Ion-exchange preconcentration has proved to be an eective means for sensitivity enhancement of FAAS58–89 and also a led into the loop of an injector situated in an integrated feed system of a spectrometer.In particular, liquid–liquid FI extrac- means of removing interferences.90–92 This approach has undergone rapid automation levels since the first paper by Olsen tions can be easily adapted for the simultaneous determination of several compounds.The species can be non-selectively et al.93 was published, in which a confluenced buer improved mixing and countercurrent elution improved the column oper- extracted into an organic phase, which is detected by a multichannel detector, such as a spectrophotometer and AAS ation. The earlier on-line exchange column preconcentration systems94,95 were rapidly supplemented by more selective pre- instrument.75 Alternatively, the partition of the extractable and non-extractable species between both immiscible phases is concentration systems, which include time-based sample loading, 96 countercurrent elution of the column,97–100 the use of detected with two detectors, which are either in series or in parallel, or by a single, fast reading computer-controlled ‘on- multi-port sample introduction devices for accommodating microcolumns in a loop96,101,102 and the use of air-flow seg- tube’ photometric detection system with millisecond time resolution. 70 As far as AS applications are concerned, one of the ments between the sample and eluent103 or the use of air-flow transportation to limit dispersion between the dierent zones main design criteria is that the overall procedure produces an increase in sensitivity, and thus the flow rate of the sample with the aim of improving the enrichment factors or decreasing the eluate volume. Major contributions to the research and carrier stream should be greater (by as large a factor as possible) than that of the extracted stream.If this flow is to be development of manifolds for on-line enrichment in FAAS were made by the research groups of Fang96–98,101,102 and introduced into an FAAS instrument at, say, 4.0 ml min-1, the practical limitations of most sample handling manifolds will Welz.99 In 1993, Fang58 published a book which provided a systematic treatment of basic principles, practical guidelines keep the preconcentration factor to less than 4.18 This limitation may be overcome by relocating the injection valve so and useful hints for those wishing to upgrade their analytical methods by separation and preconcentration.A reversed-phase that it serves as an interface between the extraction manifold and the spectrometer introduction system,76,77 with increases extraction column method has recently been employed in an on-linepreconcentration technique.104 Two microcolumns were in sensitivity of between 15- and 20-fold in comparison with the direct nebulization of aqueous solutions.77 A similar system linkedinto the manifold for gold preconcentration with reverse- flow elution to limit eluent dispersion.A computer program was described by Fang et al.,78 who obtained a 60-fold increase in sensitivity by using an air carrier stream. was developed to control simultaneously the FI and the simplex optimization of experimental conditions.A fully auto- Valca�rcel and co-workers have demonstrated that precipitation reactions as separation tools can be used to good eect mated FI–AAS system for the determination of trace amounts of copper in waters with on-line preconcentration in an ion- in FI–AAS systems.79–83 Such techniques can be used for the direct or indirect assay of various ions, with or without exchange column was recently described by Burguera et al.105 The manifold incorporated a microcolumn with a time-based filtration and with or without subsequent dissolution of the precipitate.The usual operation of a continuous precipitation injector and it was arranged in such a way that the flow of the sample enters through the narrower end of the column system involves introducing the sample containing the analyte into a carrier solution including the precipitating reagent and, as the solenoid valve is switched to the alternate position, the analyte elution occurs from the broader to the narrower (usually a cation).The mixing of the two results in the formation of a precipitate that is retained on a suitable filter. end of the column. On-line generation of volatile metal species is well suited to This is located prior to the measuring instrument and allows passage of the filtrate, which proceeds on to the detector. Thus FI–AS systems. Such chemistries yield highly sensitive and selective assays, and the capability of FI systems for control the analyte is determined from the dierence between the initial concentration of the precipitating reagent in the carrier of the reaction parameters has made this combination powerful.Much of the first published work in the area of FI–vapor and in the filtrate (non-dissolution method). The eciency of a continuous precipitation system is influenced directly by the generation (FIVG) was concerned with using the on-line mixing of the sample with the reductant solution in an acidic injected sample volume and the geometric characteristics of the precipitation coil.The major drawbacks of on-line precipi- medium in order to minimize the dead volumes, sample size and reagent consumption.106–109 The semi-automation of tation systems are the cleaning of the filter after a number of analytical determinations and the adsorption of impurities sample introduction in continuously flowing streams of reagents partially fulfilled this goal. However, the eects of both at the surface or within the precipitate,83 which makes this approach dicult for full automation of the preliminary chemical interferences were substantially minimized because the use of short tubular reactors favored the optimization of operations.These drawbacks were later overcome by Valca�rcel and Gallego80 and Peterson et al.84 by the conversion of the experimental conditions and the kinetic discrimination for the main reaction over competing interfering reactions. The soluble species into soluble compounds in which the precipitate is led directly to the spectrometer.These continuous precipi- selective determination of dierent species was also possible by changing the acidity of the sample solution.107 Later, tation systems with precipitate dissolution require no filter cleaning. The system was automated by means of an injection considerable interest was primarily shown in the reduction of the volume in gas–liquid separation,110–113 the implementation valve, a dual-pump set-up and an electric timer synchronizing the stop and start of both pumps.Fang et al.85 automated a of on-line conditions for the generation of volatile species, 114–119 on-line ion-exchange separation hydride gener- coprecipitation–dissolution procedure without filtration for the determination of lead in the presence of high concentrations ation,110,120,121 and on-line selective determination of chemical species.51 However, a number of researchers are now actively of iron by FAAS.A knotted reactor permitted the coprecipitation –dissolution process in a three-line FI manifold which pursuing the theme of full automation of the FIVG manifold so as to improve the performance of the data acquisition and used two pumps and an injector actuated by the computer of Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 645treatment. Negretti de Bratter et al.122 developed a computer sensitivity for mercury determination, an amalgamation unit was used.program which made it possible to define the user’s own analytical procedures. With the aid of the implemented com- Gludenis and Tyson145 designed a double FI system for the FAAS determination of trace metals in slurried cocoa powder puter language, it was possible to operate valves by commands using a laboratory-made interface with the computer, and to on-line digested under stopped-flow/high pressure conditions in a resistively heated thermal oven.A two-stage, dual col- create reports for printer output or for other programs, including speadsheets. If the data display was of poor quality, the umn with a three-way valve assembly allowed controlled de-pressurization of the system and gas–liquid separation. The program oered several methods of data smoothing and filtering. Their FIVG manifold permitted the automation of a thermal oven was mounted on a ceramic stand-o above the box housing the power supply and control electronics.The microtechnique for selenium determination in human body fluids, such as human serum and breast milk. Since the first tubing was wrapped in vertical coils around the resistive heating elements contained within the oven cavity. The autommercially available accessory for FIVG–AAS has been marketed,123 a considerable variety of developments are con- mation of this kind of system is limited because the seven valves must be operated to achieve a high degree of matrix tinuously being reported, e.g., the use of on-line ion-exchange separation to modify the sample matrix and the optimization decomposition, to achieve reliable de-pressurization of mineralized samples and to obtain ecient gas–liquid separation of the HG conditions124,125 for a reliable determination of tin in canned food,117 and the on-line addition of reagents for CV before sample introduction into the spectrometer.The impetus behind the development of the earlier on-line generation to determine mercury in urine.126 The use of an electronically controlled time-based device constitutes a multi- microwave sample mineralization systems was later focused mainlyon the analysis of solid samplesand on the minimization function accessory suitable for the versatile on-line determination of tin in some biological tissues by FI–HGAAS.127 The of the problems encountered with the production of gases (bubbles) during the decomposition, which caused serious FI configuration allows the insertion and on-line mixing of variable volumes of sample, acid and reductant with reduced disruption in the detection system.146,147 The methodologies described are similar and consist mainly of a carrier stream, dead volumes and minimal manipulation of the samples. Cobo Ferna�ndez et al.128 described semi-automatic FI and usually aqueous, digestion on-line in PTFE tubing, a cooling device and a direct interface with AAS for analysis.143,144,148–151 continuous-flow systems for the determination of SeIV and SeVI in sea-water samples by HGAAS with on-line pre-reduction of A report by Burguera et al.152 described a fully automated computer-monitored manifold for the determination of metals SeVI to SeIV .To reduce SeVI quantitatively to SeIV, the sample is acidified with concentrated HCl and the sample coil is in whole blood by on-line, microwave-assisted mineralization and FI–AAS. The samples were drawn and at the same time heated to 140 °C; HG is performed in an ice-bath to prevent decomposition of NaBH4.pumped directly from the patient’s forearm vein to a time injector, which was automatically controlled to bring the Electrolytic dissolution with FI, which was first proposed by Bergamin et al.129 to determine soluble aluminium in steels, sample–reagent mixture into the carrier stream. In this way, it was shown for the first time that in vivo sample uptake and was later coupled to an FAAS detector by Yuan et al.130 One of the alloy sample surfaces was finely polished with a lathe on-line sample pre-treatment reduce sample handling and make complete automation of the analysis possible.Later before the analysis. This polished surface was placed on a silicone-rubber slot and gently fixed with an adjustable clamp, publications153,154 illustrated the rapid advancement in technology for the microwave system. Automated on-line providing a sealed electrolysis cell.The electrolysis current was adjusted with a laboratory-made amperostatic supply and FI–HGAASsystems have been developed for the determination of total arsenic in urine and geothermal waters and of inorganic together with the peristaltic pump was switched on or o by the FI quartz timer. The dissolved sample solution was arsenic and organic arsenic metabolites after ion-exchange separation from urine samples. The sequential separation of delivered directly to the detector by the FI pump.As traditional digestion procedures involving hot-plate treat- inorganic (AsIII and AsV ) and organic (monomethylarsonic acid and dimethylarsinic acid) arsenic species allowed its determi- ment of samples are often lengthy, with the disadvantages of possible losses due to volatilization, degradation, etc., resulting nation by continuous HG by mixing downstream the acidified euents (or digest for total arsenic) with pre-reducing and in poor reproducibility, alternative energy sources are currently receiving a great deal of attention.131 In addition to ultrasound, reductant solutions.One of the research aspects which is most active at present electricity and UV radiation,132–135 microwaves have been increasingly utilized in the last few years, originally with is the use of on-line liquid–liquid extraction155 and microwave chemistry for speciation studies.154–159 Welz et al.159 described domestic microwave ovens and later with focused microwave systems and closed-bomb systems.28,136–139 The successful use a merging zone FI manifold which made possible on-line microwave-assisted acid digestion, followed by the pre- of microwave mineralization for the dissolution of dierent kinds of samples has prompted the investigation and develop- reduction of AsV to AsIII and its determination in biological materials by HGAAS.The atomic absorption spectrometer, ment of total automation of the dissolution process using robotics140 and on-line digestion systems.142 The first FI on-line the FI system and the autosampler were controlled by a computer which also calculated the absorbance and integrated microwave digestion system for AAS was that reported by Burguera et al.142 for the determination of metals in blood absorbance values.The eciency of acid digestion is increased by pressure which is built up in-line by a flow restrictor. serum, plasma or whole blood. The procedure involved the synchronous merging of reagent and sample introduced Lo�pez-Gonza�lvez et al.160,161 described HPLC coupled with a post-column semi-automatic on-line microwave-assisted min- through a double-loop injector and its mixing in a Pyrex coil where the decomposition takes place while being subjected to eralization HGAAS system to determine arsenite, arsenate, diethylarsinate, monomethylarsonate, arsenobetaine and microwave irradiation.Tsalev et al.143,144 designed and evaluated a system for an arsenocholine in drinking water, sewage and harbor sea-water, synthetic fish extract and sediment extract160 and urine.161 automated on-line microwave oven digester for use with CV (for mercury) and HG (for As, Bi and Sn)–AAS.Urine and Pitts et al.162 described a totally automated FI approach employing on-line microwave energy for inorganic selenium environmental water samples were mixed with an appropriate reagent and loaded on an autosampler; all measurements, speciation by HG–quartz furnace AAS.The sample is introduced into the system via the first peristaltic pump and then calibration and data processing were performed automatically by means of a personal computer, a microwave oven program- pumped through a heating coil inside the microwave unit. The heated solution is then passed into a reaction coil, and thence mer and a computer plus a printer. In order to increase the 646 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12to a cooling coil, contained in an ice-bath. Following the Pd-treated stabilized temperature platform furnace prior to its cooling step, the sample is passed to a mixing valve, where it atomization. joins a stream of sodium tetrahydroborate solution, which is An ecient FI system with an on-line anion-exchange resin pumped by a second peristaltic pump and on to the gas–liquid column incorporated into a time-based injector was described separator. The hydrogen selenide is purged in a stream of for the removal of sulfur anion spectral interference for the argon from the separator to the gas flame-heated quartz atom determination of manganese by ETAAS.173 The interference cell.The sample is first analyzed for selenium(IV), then micro- eect was made possible by retaining the interferent species on wave energy is used to reduce selenium(VI) and provide a total the resin column and eluting the analyte with nitric acid. The selenium measurement.The autosampler, the dual-channel analyte plug was introduced into the atomizer with a sampling peristaltic pump and valve switching were computer-controlled, arm assembly by means of positive disent with air while the software also oered peak height and peak area through a time-based solenoid valve. The entire system was measurement facilities. Cobo-Ferna�ndez et al.163 presented a controlled by a computer, independent of the spectrometer. semi-automated and on-line system consisting of an anion- Two of the pioneers of the FI–microwave process, Burguera exchange chromatographic column, microwave-induced ther- and Burguera,174 described for the first time the design and mal oxidation of trimethylselenium in the presence of persulfate operation of an on-line automated microwave-assisted min- and microwave-induced thermal reduction of SeVI to SeIV in eralization and FI–ETAAS system for the determination of HCl medium, followed by HG and AAS detection for the lead in biological materials.Slurried solid sample or blood determination of trimethylselenium, SeIV and SeVI in tap water. was mixed with an acidic mixture in the merging-zone mode Recently, Burguera et al.156 developed a complete automatic and mineralized downstream in a PTFE coil located inside and an on-line system for the selective determination of SeIV the microwave oven. The gases formed during the mineraliz- and SeVI in citrus fruit juices and geothermal waters by HGAAS ation process were then removed on-line by diusion through with microwave-aided heating pre-reduction of SeVI to SeIV.a porous PTFE membrane located in a gas diusion unit. In The samples and the pre-reductant solutions (4 mol l-1 HCl order to improve the removal of generated vapors, an ice trap and 12 mol l-1 HCl for SeVI) which circulated in a closed-flow was located after the liquid–gas separator. An injection valve circuit were injected by means of a time-based injector.164 This was actuated to introduce the digest into a capillary collector mixture was displaced by a carrier solution of 1% v/v HCl tube of a sampling arm assembly, from which a defined volume through a PTFE coil located inside the focused microwave of the mineralized sample was then dispensed into the graphite oven and mixed downstream with a borohydride solution to tube by an air flow via a time-based solenoid injector.The generate the hydride.The main advantage of the method is entire system was controlled by a computer, independent of that the selective determination of SeIV and SeVI is performed the spectrometer. The spectrometer autosampler used for the in a closed system with minimum manipulation, exposure to introduction of the chemical modifier was pre-programmed to the environment, sample waste and operator attention. The synchronize with the operation of the flow system. Further injector and the microwave oven were operated by an electronic studies were subsequently reported.175–177 The system was used timer.Here also, an ice trap was used to lower the temperature also for the determination of titanium dioxide in soaps with of the hot euent and therefore to help to control flow some modifications in order to allow the confinement of the disturbances. sample plug within carbon tetrachloride plugs, thus minimizing the sample dispersion.175 In order to confine the sample plug, ELECTROTHERMAL ATOMIC ABSORPTION the analyte must be insoluble in the organic phase, and this SPECTROMETRY must also be immiscible with water and transparent to microwave energy.A similar automated system to that developed Historically, FI on-line sorbent extraction for ETAAS systems, for in vivo sample uptake and on-line measurement of metal pioneered by the research group of Fang et al.165 and the species by microwave-assisted mineralization and FI–FAAS152 Department of Applied Research of Perkin-Elmer in was later coupled with ETAAS for the determination of cobalt U� berlingen, was extended in its applications to the removal of in whole blood.176 Determination of iron and zinc in adipose some matrix components166 and the on-line purification of the tissue was possible in a totally closed system by the alternative complexing agents167 with increasing degrees of automation.46 exposure of the sampling unit in the microwave-irradiated A commercially available FI accessory for atomic spectrometry zone of a focused microwave oven.177 Subsequently, additional was used as the sample handling system, performing preconcen- flows of Triton X-100 and chemical modifier were introduced tration and matrix separation.168 The rotation speed of the to avoid the detrimental accumulation of solids on the wall of two peristaltic pumps, their stop and go intervals, the actuation the tubing and to minimize matrix interference eects, of the injection valve and the interaction with the furnace respectively.autosampler arm were programmed and controlled automatically by a computer.168 Interfacing the FI system to the graphite furnace was achieved simply by connecting the transfer INDUCTIVELY COUPLED PLASMA ATOMIC capillary of the FI system to the sample introduction capillary EMISSION SPECTROMETRY of the autosampler arm and disconnecting the dispenser pump The last three decades have witnessed extensive developments of the autosampler.An alternative approach to using a graphite in plasma spectrochemical methods, in particular ICP-AES furnace as the atomization unit is the in situ deposition of and, more recently, ICP-MS. Much of this development has FI–HG species inside the pre-heated graphite furnace.169,170 been based on the measurement of steady-state signals after The in situ technique is advantageous in that interferences can pneumatic nebulization of liquid samples, although attention be reduced as a result of (i) increased sensitivity, permitting has been given to discrete sample introduction techniques such sample dilution to decrease interferent concentration, and (ii) as pulse nebulization,178 electrothermal vaporization179 and elimination of the problems associated with variable rates of FI,180,181 which are characterized by signals of a transient hydride formation and evolution.However, while the hydride nature. The earliest publications concerning sample introduc- generation and the atomization cycle are performed semition in ICP-AES were concerned with exploiting the advan- automatically, the quartz transfer tip,171 which connects the tages of the FI technique which derived from the use of small FI system with the graphite furnace, is fitted manually.Li volumes.182,183 et al.172 described a semi-automated FI–HG method using Ito et al.184 published the first application of FI to ETAAS detection. The hydride generation cycle was conducted on-line, while the analyte (Sn) was trapped ‘in situ’ on a ICP-AES.A line manual sample injector was attached to a Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 647conventional glass concentric nebulizer. Although the sample INDUCTIVELY COUPLED PLASMA MASS SPECTROMETRY was injected into an approximately 50 ml air bubble, formed by lifting the end of the carrier tubing above the water for a ICP-MS is a relatively recent technique for trace multi-element short time, in order to prevent its dilution, peak intensities and isotopic analysis210,211 and still has some limitations, e.g., were about 65% of those obtained by continuous nebulization.a highly saline sample can cause both spectral interferences In order to decrease the dependence on the skill of the operator and matrix interferences.212–214 The feasibility of using FI and in injecting the sample, a semi-automatic sample injection ICP-MS as a multi-element detector for reversed-phase LC system was subsequently developed.185 Once the sample had and high-concentration dissolved solid solutions was presented been injected into an injection tee, then nebulization, followed by Thompson and Houk215 and Coedo and Dorado.216 This by flushing of the sample pathway, and argon introduction is a particularly attractive approach for the separation and were performed automatically in sequence by means of an selective detection of a broad range of species, which still needs eight-way motor-driven liquid chromatographic rotary valve.further automation. Bloxham et al.217 described an FI–ICP-MS An automatic peakector-integrator was employed in each method with matrix removal of indirectly interfering species channel of a three-channel spectrometer for simultaneous such as sodium and chloride ions by ion-exchange chromat- multi-element analysis. Liversage et al.186 described a system ography for the determination of dierent metal species in for the semiautomatic determination of As by FI–HG–ICP- sea-water.Huang et al.218 described a continuous-flow in situ AES. The flow injection block was made from Perspex and HG–nebulizer sample introduction system coupled with consisted of a moveable inner block sandwiched between two ICP-MS for arsenic determination in sea-water and urine with other blocks, both of which were bolted to the Perspex base FI analysis. With this sample introduction system, the entire in an adjustable position.A merging zone method was used injected sample was nebulized. The nebulization process, in to inject the sample and sodium borohydride reductant. which the liquid is shattered into fine droplets in an Ar stream, The initial studies in the FI–ICP-AES field have concen- is a very eective way to purge AsH3 from the liquid, probably trated on using FI as a means for sample introduction and more so than bubbling Ar through a static reservoir of bulk analytical advantages such as high precision, high sampling liquid, as in a conventional gas–liquid separator.Almost all rates for microliter sample volumes, minimizing spectral inter- the arsenic is liberated from the droplets as AsH3 and then ferences by using novel calibration procedures5 and freedom goes to the plasma by means of a cross-flow pneumatic from nebulizer or injector tip blockage problems (for samples nebulizer with a spray chamber of the Scott type.L-Cysteine of high dissolved solids content) have been demonstra- was employed as the pre-reductant; with this reagent only mild ted.187–189 However, the detection limits in these systems were nitric acid conditions are required for HG. These combinations poorer than those obtained with conventional continuous reduced the ArCl+ molecular interference formed when HCl aspiration of samples. Therefore, numerous designs have been is included in the reaction medium.proposed and applied to enhance sensitivity,22 including liquid– On-line isotopic dilution analysis was carried out by mixing liquid extraction,190,191 ion exchange,7,93,192–196 liquid chroma- the sample and spiked solution with the FI system.219 The tography197 or modifications of the ICP introduction spike was loaded continuously in the reagent or selected valve interface.26,197 As the principal limitation of FI–ICP experi- loop and the samples were loaded individually through the ments is the low analyte transport eciency associated with sample valve.Only the sample volume consumed during the pneumatic nebulization, improved interfaces have been measurement was mixed with the spike, and the remainder of described with special reference to the nebulizer design26,197 in the sample remained intact. With a change in the FI operating order to make FI–ICP detection limits comparable to those parameters, on-line sample and isotope dilution can be per- achieved with conventional continuous sample aspiration.formedsimultaneously with various dilution rates. The merging Matrix eects have been minimized by the on-line dilution of zone modes were time-based programmable by selecting the samples (in terms of the eects of sample injection volume and injection time and the pump motor speed. carrier stream flow rate),198,199 using a relatively high rf Crain and Kiely220 reduced the liquid waste by 52% when power,198 controlling the evolution of hydrogen,186,200 the FI–direct injection nebulization was used in place of continu- on-line electrolytic dissolution of steels and other iron ous pneumatic nebulization for the analysis of hazardous alloys,201–203 on-line microwave mineralization204 or by apply- and/or radioactive solutions by ICP-MS.The instrumental ing the generalized standard addition method.205,206 limits of detection obtained with both nebulization approaches A sophisticated computer-controlled system was described were comparable for Cd, Pb and U.However, some statistically by Martin et al.207 that allowed intelligent sample handling, significant dierences were observed only in the case of Ni, based on a Fiatron SHS-3000 FI instrument. This system and were attributed to the choice of nebulizer. automatically performed the analysis of undiluted samples, Unfortunately, the current status of FI–ICP-MS software carrying out up to 200-fold dilutions if necessary, selected the makes complete automation of such systems problematic.Not optimum calibration curve for each element and spiked the only is satisfactory support for transient signals lacking but samples with standards, thus permitting standard addition also the hardware and software which would make interfacing studies. It was reported that the analysis of 76 samples of a less daunting task. widely varying concentrations could be carried out totally automatically, with no operator intervention, using the system described coupled to a 36-channel ARL 3400 ICP spectrometer.ATOMIC FLUORESCENCE SPECTROMETRY An FI system has been combined with a laboratory robot and ICP-AES for automated sample preparation and analysis Pioneering work in the field of FI–AFS detection was reported by Morita et al.221 with the on-line cold vapor generation of of lubricating oils.208 The substantial improvement in precision was related to several factors associated with FI, such as a mercury.The sample was injected into an acid carrier and merged with the tin (II) chloride reductant solution downstream relatively constant solution flow rate (regardless of sample viscosity), a reduced build-up of carbon on the torch injector at a confluence point. After passing through a mixing coil, the mercury vapor generated was separated from the liquid phase tip and improved plasma stability. A computer-intelligent automated sample FI system was reported to perform on-line by passing the stream through a specially designed cylindertype gas–liquid separator.Bloxham et al.222 investigated further sample dilution, without operator knowledge or intervention, so that measurements were made in appropriate analyte the possibility of incorporating an on-line bromide–bromate oxidation step for the complete oxidation of humic-bound and concentration ranges.209 648 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 1213 Jacintho, A. O., Zagatto, A. G., Bergamin Fo., H., Krug, F. J., organic mercury species in marine biological reference mate- Resi, B. F., Bruns, R. E., and Kowalski, B. R., Anal. Chim. Acta, rials and sea-water. A heated reaction coil was incorporated 1981, 130, 243. in the FI manifold to increase the conversion of organic 14 Broekaert, J. A. C., and Leis, F., Anal. Chim. Acta, 1979, 109, 73. mercury into inorganic mercury(II) chloride. However, the 15 Tyson, J. F., T rends Anal.Chem., 1985, 110, 419. automation of FI–AFS systems has been limited because the 16 Ruzicka, J., Fresenius’ Z. Anal. Chem., 1986, 324, 745. methodology has been addressed to improvements of the flow 17 Fang, Z., Xu, S., Wang, X., and Zhang, S., Anal. Chim. Acta, 1986, 179, 325. process in order to increase the sensitivity for the determination 18 Tyson, J. F., Colloquium Atomspektrometrische Spurennalytik, of the analyte. Bondenseewerk Perkin-Elmer, U� berlingen, 1991, p. 175. 19 Tyson, J. F., Adv. At. Spectrosc., 1992, 1, 161. FURTHER DEVELOPMENTS 20 Fang, Z., Xu, S., and Tao, G., J. Anal. At. Spectrom., 1996, 11, 1. 21 Burguera, J. L., and Burguera, M., J. T race Elem. Electrolytes FI–AS instrumentation can potentially be used for the in situ Health Dis., 1993, 7, 9. on-line evaluation of chemical water quality parameters.223 22 Tyson, J. F., Anal. Chim. Acta, 1990, 234, 3. The development of chemical microsensors in conjunction with 23 Tyson, J.F., Analyst, 1985, 110, 419. FI–AS would make it possible to create miniaturized analysis 24 Tyson, F., Spectrochim. Acta Rev., 1991, 14, 169. 25 Christian, G. D., and Ruzicka, J., Spectrochim. Acta, Part B, systems for liquid handling and sensing, including the use of 1987, 42, 157. micropumps and valves, in micromachines with silicon struc- 26 LaFreniere, K. E., Rice, G. W., and Fassel, V. A., Spectrochim. tures.224 However, the main limitation of this approach is the Acta, Part B, 1985, 40, 1495.deployment of this type of instrumentation in remote locations, 27 Flow Injection Atomic Spectroscopy, ed. Burguera, J. L., Marcel which is restricted by the power requirements and lack of Dekker, New York, 1989. portability of the AS detectors. Also, the microcomputer 28 Fang, Z., Flow Injection Atomic Absorption Spectrometry, Wiley, New York, 1995. control and data acquisition system can be a significant part 29 Flow Analysis Atomic Spectrometric Detection, ed.Sanz Medel, of the total cost and power requirements of an in situ moni- A., Elsevier, Amsterdam, 1997, in the press. tor.225 Biosensors can advantageously be used in FI,226 where 30 Luque de Castro, M. D., and Valca�rcel, M., in Automation in their performance in many cases can be augmented. In other the L aboratory, ed. Hurst, W. J., VCH, New York, 1995, instances, depending on the physical processes and chemical pp. 35–89. reactions involved in a particular assay, it might be beneficial 31 Zhi, Z.L., Rý�os, A., and Valca�rcel, M., Anal. Chim. Acta, 1994, to replace the compromise working conditions of a biosensor 293, 163. 32 Bergamin Fo., H., Reis, B. F., and Zagatto, E. A. G., Anal. Chim. with the optimum conditions readily feasible in an FI system, Acta, 1978, 97, 427. whereby the overall operations can be fully automated and 33 Stewart, K. K., Brown, J. F., and Golden, B. M., Anal. Chim. therefore simplified. Possible coupling of HPLC separation Acta, 1980, 114, 119.with o-line ETAAS or an on-line real-time atomic detector 34 Yoza, N., Aoyagi, Y., and Ohashi, S., Anal. Chim. Acta, 1979, (ICP-AES)227 may become a synergistic combination with 111, 163. increasing degrees of automation, despite the intrinsically 35 Fukamachi, K., and Ishibashi, N., Anal. Chim. Acta, 1980, discontinuous operation of ETAAS and the low detection 119, 383. 36 Basson, W. D., and Van Staden, J. F., Fresenius’ J.Anal. Chem., limits achievable for some elements using ICP-AES. Also, 1980, 302, 370. drop-based automated liquid–liquid extraction using a single, 37 Mindel, B. D., and Karlberg, B., L ab. Pract., 1981, 719. microliter-volume organic drop228 may be suitable for AS. The 38 Tyson, J. F., and Idris, A. B., Analyst, 1981, 106, 1125. development of pervaporation229,230 automatic modules 39 Tyson, J. F., and Idris, A. B., Analyst, 1984, 109, 23. coupled to AS instrumentation will undoubtedly be suited to 40 Tyson, J.F., Appleton, J. M. H., and Idris, A. B., Anal. Chim. the monitoring of complex samples. Acta, 1983, 145, 159. Notwithstanding the clear advantages inherent in the auto- 41 Tyson, J. F., and Bysouth, S. R., J. Anal. At. Spectrom., 1988, 3, 211. mation of laboratory processes, one should also be aware of 42 Simonsen, K. W., Nielsen, B., Jensen, A., and Andersen, J. R., the risks involved: (1) the operator is to a greater or lesser J. Anal.At. Spectrom., 1986, 1, 453. extent detached from the analytical process, which results in 43 Riley, C., Rocks, B. F., and Sherwood, R. A., T alanta, 1984, the loss of sometimes valuable information and calls for more 31, 879. frequent controls; (2) some laboratory principals may over- 44 Uchida, T., Kojima, I., and Iida, C., Anal. Chim. Acta, 1980, estimate the potential of automation and may therefore be less 116, 205. critical in evaluating its eectiveness; and (3) the chemist may 45 Fry, R.C., Northway, S. J., and Denton, M. B., Anal. Chem., 1978, 50, 1719. lose sight of the prevalent role played by chemistry.131 46 Fang, Z., Welz, B., and Schlemmer, G., J. Anal. At. Spectrom., 1989, 4, 91. REFERENCES 47 Erler, W., Portala, F., and Schulze, H., Analysis of Samples with High Dissolved Solids Content Using Flow Injection Flame 1 Valca�rcel, M., and Luque de Castro, M. D., Automatic Methods AAS, No. 1.4-E (2930/5.89), Bodenseewerk Perkin-Elmer, of Analysis, Elsevier, Amsterdam, 1988. U�berlingen, 1989. 2 Valca�rcel, M., Quý�m. Anal., 1990, 9, 215. 48 Zhou, N., Frech, W., and Lindberg, E., Anal. Chim. Acta, 1983, 3 Burguera, J. L., and Burguera, M., J. Anal. At. Spectrom., 1995, 153, 33. 10, 473. 49 Fang, Z., Welz, B., and Sperling, M., Anal. Chem., 1993, 65, 1682. 4 Ruzicka, J., and Hansen, E. H., Flow Injection Analysis, Wiley, 50 Reis, R. F., Jacintho, A. O., Mortatti, J., Krug, F. J., Zagatto, New York, 2nd edn., 1988.E. A. G., Bergamin Fo., H., and Pessenda, L. C. R., Anal. Chim. 5 Karlberg, B., and Pacey, G. F., Flow Injection Analysis: A Acta, 1981, 123, 221. Practical Guide, Elsevier, Amsterdam, 1989. 51 Burguera, M., Burguera, J. L., Brunetto, M. R., De la Guardia, 6 Valca�rcel, M., and Luque de Castro, M. D., Flow Injection M., and Salvador, A., Anal. Chim. Acta, 1991, 161, 105. Analysis, Principles and Applications, Ellis Horwood, 52 Borja, R., De la Guardia, M., Salvador, A., Burguera, J.L., and Chichester, 1987. Burguera, M., Fresenius’ J. Anal. Chem., 1990, 338, 9. 7 McLeod, C. W., J. Anal. At. Spectrom., 1987, 2, 549. 53 Rondo�n, C., Burguera, M., Burguera, J. L., Brunetto, M. R., and 8 Tyson, J. F., Microchem. J., 1992, 45, 143. Carrero, P., J. T race Elem. Med. Biol., 1995, 9, 49. 9 Zagatto, E. A. G., Krug, F. J., Bergamin Fo., H., Jorgensen, 54 Bysouth, S. R., and Tyson, J. F., J. Anal. At. Spectrom., 1987, S. S., and Reis, B. F., Anal. Chim. Acta, 1979, 104, 279. 2, 217. 10 Wolf, W. R., and Stewart, K. K., Anal. Chem., 1979, 51, 1201. 55 Ara�ujo, A. N., and Lima, J. L. F. C., J. T race Elem. Electrolytes 11 Greenfield, S., Pure Appl. Chem., 1980, 52, 2509. 12 Greenfield, S., Ind. Res. Dev., 1981, 23, 140. Health Dis., 1989, 3, 97. Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 64956 Sperling, M., Koscielniak, P., and Welz, B., Anal. Chim. Acta, 101 Fang, Z., Ruzicka, J., and Hansen, E. H., Anal. Chim.Acta, 1984, 164, 23. 1992, 261, 115. 102 Fang, Z., Sperling, M., and Welz, B., Anal. Chim. Acta, 1992, 57 Burguera, J. L., and Burguera, M., L ab. Rob. Autom., 1991, 3, 119. 269, 9. 58 Fang, Z., Flow Injection Separation and Preconcentration, VCH, 103 Hirata, S., Honda, K., and Kumamaru, T., Anal. Chim. Acta, Weinheim, 1993. 1989, 221, 65. 59 Karlberg, B., and Pacey, G. E., Flow Injection Analysis. A 104 Di, P., and Davey, D. E., T alanta, 1994, 41, 565. Practical Guide, Elsevier, Amsterdam, 1989. 105 Burguera, J. L., Burguera, M., Carrero, P., Marcano, J., Rivas, 60 Karlberg, B., and Thelander, S., Anal. Chim. Acta, 1978, 98, 1. C., and Brunetto, M. R., J. Autom. Chem., 1995, 17, 25. 61 Bergamin Fo., H., Madeiros, J. X., Reis, B. F., and Zagatto, 106 Astro�m, O., Anal. Chem., 1982, 54, 190. E. A. G., Anal. Chim. Acta, 1978, 101, 9. 107 Yamamoto, M., Yasuda, M., and Yamamoto, Y., Anal. Chem., 62 Scokart, P. O., Meeus-Verdinne, K., and De Borger, R., Int.J. 1985, 57, 1382. Environ. Anal. Chem., 1987, 29, 305. 108 Chan, C. Y., Anal. Chem., 1985, 57, 1482. 63 Kumamaru, T., Nitta, Y., Matsuo, H., and Kimura, E., Bull. 109 Fang, Z., Xu, S., Wang, X., and Zhang, S., Anal. Chim. Acta, Chem. Soc. Jpn., 1987, 60, 1930. 1986, 179, 325. 64 Karlberg, B., Fresenius’ J. Anal. Chem., 1988, 329, 660. 110 Chen, H., Brindle, I. D., and Zheng, S., Analyst, 1992, 117, 1603. 65 Martý�nez-Jime�nez, P., Gallego, M., and Valca�rcel, M., Anal. 111 Carlos de Andrade, J., Pasquini, C., Baccan, W., and Van Loon, Chim.Acta, 1988, 215, 233. J. C., Spectrochim. Acta, Part B, 1983, 38, 1329. 66 Ba�ckstro�m, K., and Danielsson, L. G., Anal. Chem., 1988, 112 Pacey, G. E., Straka, M. R., and Gord, J. R., Anal. Chem., 1986, 60, 1354. 58, 502. 67 Kuba�n, V., Koma�rek, J., and Cajkova�, D., Chem. L isty, 1989, 113 Dedina, J., Anal. Chem., 1982, 54, 2097. 84, 376. 114 Petterson, J., Hansson, L., and Olin, A., T alanta, 1986, 33, 249. 68 Ba�ckstro�m, K., and Danielsson, L Erler, W., Schulze, H., and McIntosh, S., Perkin-Elmer 232, 301. T echnical Report, No. TSAA-12, Bodenseewerk Perkin-Elmer, 69 Ba�ckstro�m, K., and Danielsson, L. G., Mar. Chem., 1990, 28, 33. U� berlingen, 1992. 70 Kuba�n, V., Crit. Rev. Anal. Chem., 1991, 22, 477. 116 Yamamoto, M., Takada, K., Kumamuru, T., Yasuda, M., 71 Sweilah, J. A., and Cantwell, F. F., Anal. Chem., 1985, 57, 420. Yokoyama, S., and Yamamoto, Y., Anal.Chem., 1987, 59, 2446. 72 Tyson, J. F., Adeeyinwo, C. E., Appleton, J. M. H., Bysouth, 117 Barnes, R., and Wang, X., J. Anal. At. Spectrom., 1988, 3, 1083. S. R., Idris, S. R., and Sarkissian, L. L., Analyst, 1985, 110, 487. 118 Wang, X., and Barnes, R., J. Anal. At. Spectrom., 1988, 3, 1091. 73 Gallego, M., Silva, M., and Valca�rcel, M., Fresenius’ J. Anal. 119 Chan, W. F., and Hon, P. K., Analyst, 1990, 115, 567. Chem., 1986, 323, 50. 120 Fang, Z., Sun, L., Hansen, E.H., Olesen, J. E., and Henriksen, 74 Attiyat, A. S., J. Flow Injection Anal., 1987, 4, 26. L. M., T alanta, 1992, 39, 383. 75 Zagatto, E. A. G., Jacintho, A. O., Pessenda, L. C. R., Krug, 121 Tyson, J. F., Oey, S. G., Seare, N. J., Kibble, H. A. B., and F. J., Reis, B. F., and Bergamin Fo., H., Anal. Chim. Acta, 1981, Fellows, C., J. Anal. At. Spectrom., 1992, 7, 315. 125, 37. 122 Negretti de Bratter, V. E., Bratter, P., and Tomiak, A., J. T race 76 Nord, L., and Karlberg, B., Anal.Chim. Acta, 1981, 125, 199. Elem. Electrolytes Health Dis., 1990, 4, 41. 77 Nord, L., and Karlberg, B., Anal. Chim. Acta, 1983, 145, 151. 123 Welz, B., and Shubert-Jacobs, M., At. Spectrosc., 1991, 12, 91. 78 Fang, Z., Zhu, Z., Zhang, S., Xu, S., Gu, L., and Sun, L., Anal. 124 McIntosh, S., Li, Z., Carnrick, G. R., and Slavin, W., Spectrochim. Chim. Acta, 1988, 214, 41. Acta, Part B, 1992, 47, 701. 79 Valca�rcel, M., and Gallego, M., in Flow Injection Atomic 125 Welz, B., Schubert-Jacobs, M., and Guo, T., T alanta, 1992, Spectroscopy, ed.Burguera, J. L., Marcel Dekker, New York, 39, 1097. 1989, ch. 5, pp. 157–224. 126 Guo, T., and Baasner, J., Anal. Chim. Acta, 1993, 278, 189. 80 Valca�rcel, M., and Gallego, M., T rends Anal. Chem., 1989, 8, 34. 127 Burguera, M., Burguera, J. L., Rivas, C., Carrero, P., Brunetto, 81 Martý�nez-Jime�nez, P., Gallego, M., and Valca�rcel, M., Anal. M. R., and Gallignani, M., Anal. Chim. Acta, 1995, 308, 339.Chem., 1987, 59, 69. 128 Cobo Ferna�ndez, M. G., Palacios, M. A., and Ca�mara, C., Anal. 82 Martý�nez-Jimenez, P., Gallego, M., and Valca�rcel, M., Analyst, Chim. Acta, 1993, 283, 386. 1987, 112, 1233. 129 Bergamin Fo., H., Krug, F. J., Zagatto, E. A. G., Arruda, E. C., 83 Martinez-Jimenez, P., Gallego, M., and Valca�rcel, M., J. Anal. and Coutinho, C. A., Anal. Chim. Acta, 1986, 190, 177. At. Spectrom., 1987, 2, 211. 130 Yuan, D., Wang, X., Yang, P., and Hung, B., Anal.Chim. Acta, 84 Peterson, B. A., Fang, Z., Ruzicka, J., and Hansen E. H., Anal. 1991, 243, 65. Chim. Acta, 1986, 184, 165. 131 Bryce, D. W., Izquierdo, A., and Luque de Castro, M. D., 85 Fang, Z., Sperling, M., and Welz, B., J. Anal. At. Spectrom., 1991, Analyst, 1995, 120, 2171. 6, 301. 132 Bitsch, R., and Merck, E., L aborPraxis, 1994, 18, 76. 86 Atienza, J., Herrero, M. A., Maquieira, A., and Puchades, R., 133 Provan, G. J., Scobbie, L., and Chesson, A., J. Sci.Food Agric., Crit. Rev. Anal. Chem., 1992, 23, 1. 1994, 64, 63. 87 Bysouth, S. R., Tyson, J. F., and Stockwell, P. B., Analyst, 1990, 134 Goa, Q., Fan, C. L., and Wei, J. H., Fenxi Huaxue, 1994, 22, 418. 115, 571. 135 Sa�nchez, J., Garcý�a, S., and Milla�n, E., Analusis, 1994, 22, 222. 88 Hidalgo, M. M., Go�mez, M. M., and Palacios, M. A., Fresenius’ 136 Introduction to Microwave Sample Preparation. T heory and J. Anal. Chem., 1996, 354, 420. Practice, ed. Kingston, H. M., and Jessie, L.B., American 89 Yebra-Biurrun, M. C., Bermejo-Barrera, A., Bermejo-Barrera, Chemical Society, Washington, DC, 1988. M. P., and Barciela-Alonso, M. C., Anal. Chim. Acta, 1995, 137 De la Guardia, M., Salvador, M., Burguera, J. L., and Burguera, 303, 341. M., J. Flow Injection Anal., 1988, 5, 121. 90 Kamson, O. F., and Townshend, A., Anal. Chim. Acta, 1983, 138 Chakraborty, R., Das, A. K., Cervera, M. L., and De la Guardia, 155, 253. M., Fresenius’ J. Anal. Chem., 1996, 355, 99. 91 Burguera, J. L., Burguera, M., Rivas, C., Carrero, P., Gallignani, 139 Burguera, M., and Burguera, J. L., Quý�m. Anal., 1996, 15, 112. M., and Brunetto, M. R., J. Anal. At. Spectrom., 1995, 10, 479. 140 Norris, J. D., Preston, B., and Ross, L. M., Analyst, 1992, 117, 3. 92 Oey, S. G., Seare, N. J., Tyson, J. F., and Kibble, H. A. B., 141 Matousek de Abel de la Cruz, A., Burguera, J. L., Burguera, M., J. Anal. At. Spectrom., 1991, 6, 133. Wasin, S., and Rivas, C., Fresenius’ J.Anal. Chem., 1996, 354, 184. 93 Olsen, S., Pessenda, L. C. R., Ruzicka, J., and Hansen, E. H., 142 Burguera, M., Burguera, J. L., and Alarco�n, O. M., Anal. Chim. Analyst, 1983, 108, 905. Acta, 1986, 179, 351. 94 Zhang, Y., Riby, P., Cox, A. G., McLeod, C. W., Date, A. R., 143 Tsalev, D. L., Sperling, M., and Welz, B., Analyst, 1992, 117, 1729. and Cheung, Y. Y., Analyst, 1988, 113, 125. 144 Tsalev, D. L., Sperling, M., and Welz, B., Analyst, 1992, 117, 1735. 95 Valde�s-Hevia y Temprano, M.C., Pe�rez Parajo�n, J., Dý�az Garcý�a, 145 Gludenis, T. J., and Tyson, J. F., J. Anal. At. Spectrom., 1992, M. E., and Sanz-Medel, A., Analyst, 1991, 116, 1141. 7, 301. 96 Fang, Z., Xu, S., and Zhang, S., Anal. Chim. Acta, 1987, 200, 35. 146 Burguera, M., Burguera, J. L., and Alarco�n, O. M., Anal. Chim. 97 Fang, Z., and Welz, B., J. Anal. At. Spectrom., 1989, 4, 543. Acta, 1988, 214, 421. 98 Fang, Z., Guo, T., and Welz, B., T alanta, 1991, 38, 613. 147 Carbonell, V., De la Guardia, M., Salvador, A., Burguera, J. L., 99 Sperling, M., Xu, S., and Welz, B., Anal. Chem., 1992, 64, 3101. and Burguera, M., Anal. Chim. Acta, 1990, 238, 417. 100 Ma, R., Van Mol, W., and Adams, F., Anal. Chim. Acta, 1994, 148 Haswell, S. J., and Barclay, D., Analyst, 1992, 117, 117. 149 Carbonell, V., Morales-Rubio, A., Salvador, A., De la Guardia, 285, 33, 650 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12M., Burguera, J. L., and Burguera, M., J.Anal. At. Spectrom., 190 Siles Cordero, M. T., Vereda Alonso, E. I., Garcý�a de Torres, A., and Cano Pavo�n, J. M., J. Anal. At. Spectrom., 1996, 11, 107. 1992, 7, 1085. 191 Can�ada Rudner, P., Garcý�a de Torres, A., and Cano Pavo�n, 150 Welz, B., Tsalev, D. L., and Sperling, M., Anal. Chim. Acta, 1992, J. M., J. Anal. At. Spectrom., 1993, 8, 705. 261, 91. 192 Pereiro Garcý�a, M. R., Dý�az, M. E., and Sanz Medel, A., J. Anal. 151 Saraswati, R., Vetter, T.W., and Watters, R. L., Analyst, 1995, At. Spectrom., 1987, 2, 699. 120, 95. 193 Cox, A., McLeod, C., Miles, D. L., and Cook, J. M., J. Anal. At. 152 Burguera, J. L., Burguera, M., and Brunetto, M. R., At. Spectrom., 1987, 2, 553. Spectrosc., 1993, 14, 90. 194 Ebdon, L., Fisher, A. S., and Worsfold, P. J., J. Anal. At. 153 Burguera, J. L., Burguera, M., Rivas, C., Carrero, P., and Spectrom., 1994, 9, 611. Rondo�n, C., Quim. Anal., 1997, in the press. 195 Olsen, S., Pessenda, L.C. R., Ruzicka, J., and Hansen, E. H., 154 Burguera, J. L., Burguera, M., and Danet, A. F., Rev. Roum. Analyst, 1983, 108, 905. Chim., in the press. 196 Cresser, M. S., Ebdon, L. C., McLeod, C. W., and Burridge, 155 Taylor, M. J. C., and Van Staden, J. F., Anal. Chim. Acta, 1995, J. C., J. Anal. At. Spectrom., 1986, 1, 1R. 307, 1. 197 Lawrence, K. E., Rice, G. W., and Fassel, V. A., Anal. Chem., 156 Burguera, J. L., Carrero, P., Burguera, M., Rondo�n, C., Brunetto, 1984, 56, 289. M.R., and Gallignani, M., Spectrochim. Acta, Part B, 1996, 198 McLeod, C. W., Worsfold, P. J., and Cox, A. G., Analyst, 1984, 51, 1837. 109, 327. 157 Luque de Castro, M. D., Bryce, D. W., and Izquierdo, A., 199 Bussie`re, L., Dumont, J., and Hubert, J., Anal. Chim. Acta, 1989, T alanta, 1995, 42, 12l5. 224, 73. 158 Bryce, D. W., Izquierdo, A., and Luque de Castro, M. D., 200 Tioh, N. H., Israel, Y., and05. 159 Welz, B., He, Y., and Sperling, M., T alanta, 1993, 40, 1917. 201 Flock, J., and Ohls, K., Fresenius’ J. Anal. Chem., 1988, 331, 408. 160 Lo�pez-Gonza�lvez, M. A., Go�mez, M. M., Ca�mara, C., and 202 Bergamin, H. F., Krug, F. J., No�brega, J. A., Oliveira, P. V., Palacios, M. A., J. Anal. At. Spectrom., 1994, 9, 291. Reis, B. F., and Gine�, M. F., Anal. Chim. Acta, 1991, 245, 211. 161 Lo�pez-Gonza�lvez, M. A., Go�mez, M. M., Ca�mara, C., and 203 Yuang, D., Wang, X., Yang, P., and Hang, B., Anal.Chim. Acta, Palacios, M. A., Mikrochim. Acta, 1995, 120, 301. 1991, 251, 187. 162 Pitts, L., Worsfold, P. J., and Hill, S. J., Analyst, 1994, 119, 2785. 204 Bordera, L., Hernandis, V., and Canals, A., Fresenius’ J. Anal. 163 Cobo-Ferna�ndez, M. G., Palacios, M. A., Chakraborti, D., Chem., 1996, 355, 112. Quevauviller, P., and Ca�mara, C., Fresenius’ J. Anal. Chem., 205 Zagatto, E. A. G., Jacintho, A. O., Krug, F. J., Reis, B. F., Bruns, 1995, 351, 438.R. E., and Araujo, M. C. U., Anal. Chim. Acta, 1983, 145, 169. 206 Israel, Y., and Barnes, R. M., Anal. Chem., 1984, 56, 1188. 164 Carrero, P., Burguera, J. L., Burguera, M., and Rivas, C., 207 Martin, J. M., Dobbins, J. T., and Ihrig, P. J., Quantification of T alanta, 1993, 40, 1967. Metals in L iquid Samples by Computer Intelligent Flow Injection 165 Fang, Z., Sperling, M., and Welz, B., J. Anal. At. Spectrom., 1990, Inductively Coupled Plasma Emission Spectrometry,R. J.Reynolds 5, 639. Tobacco Co. and R. D. Department, Winston-Salem, NC, 1993. 166 Sperling, M., Yin, X., and Welz, B., J. Anal. At. Spectrom., 1991, 208 Granchi, M. P., Biggersta, J. A., Hillard, L. J., and Grey, P., 6, 295. Spectrochim. Acta, Part B, 1987, 42,169. 167 Welz, B., Yin, X., and Sperling, M., Anal. Chim. Acta, 1992, 209 Ihrig, P. J., Dobbins, J. T., and Reynolds, R. J., paper presented 261, 477. at the 1986 Winter Conference on Plasma Spectrometry, Hawaii, 168 Welz, B., Sperling, M., and Sun, X., Fresenius’ J.Anal. Chem., 2nd–8th January, 1986, paper No. 44. 1993, 346, 550. 210 Houk, R. S., Fassel, V. A., Flesh, G. D., Svec, H. J., Gray, A. L., 169 An, Y., Willie, S. N., and Sturgeon, R. E., Spectrochim. Acta, and Taylor, C. E., Anal. Chem., 1980, 52, 2283. Part B, 1992, 47, 10403. 211 Houk, R. S., Anal. Chem., 1986, 58, 97A. 170 Erber, D., and Cammann, K., Analyst, 1995, 120, 2699. 212 Tan, S. H., and Horlick, G., J. Anal. At. Spectrom., 1987, 2, 745. 171 Yan, X. P., and Ni, Z. M., Anal. Chim. Acta, 1994, 291, 89. 213 Douglas, D. J., and Kerr, L. A., J. Anal. At. Spectrom., 1988, 3, 749. 172 Li, Z., McIntosh, S., Carnrick, G. R., and Slavin, W., Spectrochim. 214 Story, W. C., Caruso, J. A., Heitkemper, D., and Perkin, L., Acta Part B, 1992, 47, 701. J. Chromatogr., Sci., 1992, 30, 427. 173 Burguera, J. L., Burguera, M., Rivas, C., Carrero, P., Gallignani, 215 Thompson, J. J., and Houk, R. S., Anal. Chem., 1986, 58, 2541. M., and Brunetto, M. R., J. Anal. At. Spectrom., 1995, 10, 479. 216 Coedo, A. G., and Dorado, M. T., J. Anal. At. Spectrom., 1994, 174 Burguera, J. L., and Burguera, M., J. Anal. At. Spectrom., 1993, 9, 1111. 8, 235. 217 Bloxham, M. J., Hill, S. J., and Worsfold, P. J., J. Anal. At. 175 Burguera, M., and Burguera, J. L., L ab. Rob. Autom., 1993, 5, 1. Spectrom., 1994, 9, 935. 176 Burguera, M., Burguera, J. L., Rondo�n, C., Rivas, C., Carrero, 218 Huang, M. F., Jiang, S. J., and Hwang, C. J., J. Anal. At. P., Gallignani, M., and Brunetto, M. R., J. Anal. At. Spectrom., Spectrom., 1995, 10, 31. 1995, 10, 343. 219 Viczia�n, M., La�sztity, A., Wang, X., and Barnes, R. M., J. Anal. 177 Burguera, J. L., Burguera, M., Carrero, P., Rivas, C., Gallignani, At. Spectrom., 1990, 5, 125. M., and Brunetto, M. R., Anal. Chim. Acta, 1995, 308, 349. 220 Crain, J. S., and Kiely, J. T., J. Anal. At. Spectrom., 1996, 11, 5251. 178 Cresser, M. S., Prog. Anal. At. Spectrosc., 1981, 4, 219. 221 Morita, H., Kimoto, T., and Shimomura, S., Anal. L ett., 1983, 179 Barnes, R. M., and Fodor, P., Spectrochim. Acta, Part B, 1983, 16, 1187. 38, 1191. 222 Bloxham, M. J., Hill, S. J., and Worsfold, P. J., J. Anal. At. Spectrom., 1996, 11, 511. 180 Ito, T., Kawaguchi, H., and Mizuike, A., Bunseki Kagaku, 1980, 223 Alexander, P. W., Di Benedetto, L. T., Dimitrakopoulos, T., 29, 332. Hibbert, D. B., Ngila, J. C., Sequeira, M., and Shiels, D., T alanta, 181 Marshall, J., Haswell, S. J., and Hill, S. J., J. Anal. At. Spectrom., 1996, 43, 915. 1987, 2, 79R. 224 Van der Schoot, B. H., Jeanneret, S., Van der Berg, A., and 182 Greenfield, S., and Smith, P. B., Anal. Chim. Acta, 1972, 59, 341. Rooij, N. F., Anal. Methods Instrum., 1993, 1, 1. 183 Sebatiani, E., Ohls, K., and Riemer, G., Fresenius’ Z. Anal. 225 Andrew, K. N., Blundell, N. J., Price, D., and Worsfold, P. J., Chem., 1973, 264, 105. Anal. Chem., 1994, 66, 916A. 184 Ito, T., Kawaguchi, H., and Mizuike, A., Bunseki Kagaku, 1980, 226 Hansen, E. H., Talanta, 1994, 41, 939. 29, 332. 227 Sanz-Medel, A., Analyst, 1995, 120, 799. 185 Ito, T., Nakagawa, E., Kawaguchi, H., and Mizuike, A., 228 Liu, H., and Dasgupta, P. K., Anal. Chem., 1996, 68, 1817. Mikrochim. Acta, 1982, I, 423. 229 Prinzing, V., Ogbomo, I., Lehn, C., and Schmidt, H. L., Sens. 186 Liversage, R., Van Loon, J. C., and De Andrade, J. C., Anal. Actuators, B, 1990, 1, 542. Chim. Acta, 1984, 161, 275. 230 Papaefstathiou, I., Tena, M. D., and Luque de Castro, M. D., 187 Jacintho, A. O., Zagatto, E. A. G., Bergamin Fo., H., Krug, F. J., Anal. Chim. Acta, 1995, 308, 246. Reis, B. F., Bruns, R. E., and Kowalski, B. R., Anal. Chim. Acta, 1981, 130, 243. Paper 6/07974A 188 Greenfield, S., Spectrochim. Acta, Part B, 1983, 38, 95. Received November 25, 1996 189 Alexander, P. W., Finlayson, R. J., Smythe, L. E., and Thalib, A., Analyst, 1982, 107, 1335. Accepted January 3, 1997 Journal of Analytical Atomic Spectr

ISSN:0267-9477    

DOI:10.1039/a607974a

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Application of Laser Sampling Microprobe Inductively Coupled Plasma Mass Spectrometry to the In Situ Trace Element Analysis of Selected Geological Materials ZHONGXING CHEN, WILL DOHERTY AND D. CONRAD GRE� GOIRE* Geological Survey of Canada, 601 Booth Street, Ottawa, Ontario, Canada K1S 5B6 A method is described for the direct solid sampling and minerals (e.g., tourmaline, wolframite, beryl, rutile, monazite, zircon, chromite and cassiterite), avoidance of volatile element analysis of geological materials by laser microprobe inductively coupled plasma mass spectrometry (LAM- loss (e.g., As, Se and B), reduction of some spectral interferences by polyatomic ions and provision for spatial analysis of small ICP-MS).An Nd5YAG laser operated in the Q-switched mode at a fundamental wavelength of 1064 nm was used. Both selected areas on the surface of the sample. In this study, trace element concentrations in apatite, monaz- the sample stage and the laser with its floor-mounted power supply were controlled by a personal computer in a mouse- ite, chromite, olivine, concretions and fused buttons of silicate rock powder were analysed in situ with a spatial resolution of driven Windows environment.The signal intensity was optimized to the maximum level by adjusting the nebulizer gas about 60–150 mm depending on the material. flow rate and ion optics using NIST 610 glass. Sample was pre-ablated for a few seconds to remove any surface EXPERIMENTAL contamination.NIST 610 or 612 silicate glass was used as a Instrumentation calibration standard and Ca, Ce, Si or Mn were used as an internal standard to correct for signal drift, dierences in A Perkin-Elmer SCIEX (Thornhill, Ontario, Canada) Model transport eciency and sampling yields for dierent geological 320 laser sampler was used in conjunction with a Perkinmaterials. Trace element concentrations in apatite, monazite, Elmer ELAN Model 5000a ICP-MS instrument.An Nd5YAG chromite, olivine, concretion and fused buttons of silicate rock laser was operated in the Q-switched mode with a fundamental powder were determined in situ with a spatial resolution of wavelength of 1064 nm. Both the sample stage and the laser 60–150 mm, depending on the materials analysed. Good with its floor–mounted power supply were controlled by a precision and accuracy and low solid limits of detection were personal computer in a mouse-driven Windows environment.obtained. Optimization of the plasma and mass spectrometer conditions was first accomplished using solution nebulization sample Keywords: L aser sampling; inductively coupled plasma mass introduction and aqueous standards. The signal intensity for spectrometry ; trace elements; geological materials laser sampling was optimized using NIST 610 glass as sample. For each analytical run, the nebulizer gas flow was adjusted As early as 1965, Ready1 studied laser vaporization processes to about 1.10 l min-1 of Ar, using NIST 610 glass containing and a review by Darke and Tyson2 summarized knowledge up about 450 mg g-1 of trace elements, so that (1) ThO+/Th+ to 1992 of the interaction of laser radiation with solids.This <0.5%, (2) sensitivity of La>40 counts s-1 mg-1 g and (3) interaction may cause both ablation and vaporization by sensitivity of Th>30 counts s-1 mg-1 g. processes that depend on both the characteristics of the laser beam and physical properties of the solid.The ability of Preparation of Samples and Standards focused laser radiation to volatilize virtually any material has provided the analytical chemist with a versatile technique of Four common geological sample types (rocks, mineral mounts, direct solid sampling for subsequent spectrometric analysis. thin sections and polished slabs) were laser sampled directly The history and developments up to the late 1980s of laser as received, without any sample preparation prior to analysis.sampling introduction systems were extensively reviewed by For rock powder samples, 0.2 g of powder sample and 0.4 g of Moenke-Blankenburg.3 lithium metaborate (LiBO2) were weighed into a platinum Laser sampling of solids has been combined with various crucible and thoroughly mixed. The mixture was then fused analytical techniques such as AAS,4 MIP-AES,5 DCP-AES,6 over a Meeker burner for 20 min and allowed to cool. The ICP-AES7,8 and ICP-MS.9–16 Gray9 first reported solid sample crucible was gently tapped and the silicate glass button was introduction by ICP-MS using lasers and thus introduced a retrieved.technique which has steadily gained in popularity, particularly NIST 610 or NIST 612 silicate glass reference material was for the direct analysis of geological materials. Sample introduc- used as a calibration standard. The NIST silicate glass button tion using lasers oers several advantages over conventional (as received) was mounted in epoxy and then the surface of pneumatic nebulization for the analysis of geological materials.the button was polished. Major and trace element concen- Many samples, particularly in geochemistry, are initially in trations of the two standards, determined by ICP-OES and solid form, and there are many advantages to a sample ICP-MS, are given in Table 1. introduction technique that does not require an intermediate sample dissolution step. These advantages include reduced Data Acquisition and Calibration reagent and labour costs, elimination of dilution errors, minimization of sample contamination and losses arising Data were acquired in the peak-hopping mode with a dwell time of 20 ms.The sample was pre-ablated for a few seconds from sample-handling steps, application to dissolution-resistant Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 (653–659) 653Table 1 Elemental composition (mg g-1) of NIST 610 and NIST 612 silicate glasses Constituent* NIST610 RSD (%) NIST612 RSD (%) SiO2 67.40% 70.0% TiO2 0.14% <0.02% Al2O3 1.80% 1.84% Fe2O3 0.11% <0.06% MnO 0.05% 0.01% MgO 0.08% <0.04% CaO 10.50% 11.00% Na2O 12.40% 12.80% K2O 0.06% <0.05% P2O5 0.08% <0.01% Total 93.2% 95.8% Ba 400 50 Be 400 34 Co 330 28 Cr 330 33 Cu 370 35 Ni 380 64 Sc 420 36 Sr 440 70 V 420 31 Zn 370 17 Ce 423 1.1 35 2.3 Dy 413 1.1 33 1.4 Ce 423 1.1 35 2.3 Dy 413 1.1 33 1.4 Er 420 0.0 35 2.3 Eu 420 0.0 33 2.9 Fig. 1 Time variation of the signals of selected elements in NIST 610 Gd 427 1.1 36 2.3 silicate standard. The intensity for each element was calculated as the Ho 420 0.0 35 2.3 mean countrate during laser sampling. La 410 0.0 33 2.5 Lu 410 0.0 34 2.4 Nd 417 1.1 33 2.5 Pr 420 0.0 35 2.3 Sm 423 1.1 35 3.6 Tb 420 0.0 35 2.3 Tm 427 1.1 35 2.3 Y 430 0.0 36 2.3 Yb 427 1.1 36 1.3 Ag 227 2.1 18 2.7 Bi 370 0.0 29 2.8 Cs 350 0.0 38 2.1 Ga 417 1.1 35 3.6 Hf 407 1.2 35 2.3 In 420 0.0 35 3.6 Mo 383 1.2 33 2.8 Nb 484 2.3 33 2.5 Fig. 2 Scanning electron micrograph of laser sampling craters of Pb 400 0.0 37 2.2 NIST 610, 70–80 mm in diameter. Rb 400 0.0 29 4.3 Ta 470 3.7 21 0.0 Th 423 1.1 34 2.4 be deconvoluted by mathematical means and in this study U 430 0.0 33 2.5 were corrected for by using internal standards. During a long Zr 433 1.1 38 2.1 run, the ICP-MS sensitivity can change and matrix eects can cause both signal enhancement and depression of analyte * Ba, Be, Co, Cr, Cu, Ni, Sc, Sr, V and Zn were determined by ICP-OES and other elements by solution nebulization ICP-MS.signals. Also, laser sampling yields can vary from one sample to another and from laser firing to firing because of dierent sample compositions and surface structures. These eects were to remove any surface contamination. Background levels for corrected for by frequent recalibration, in addition to internal each element were established by acquiring data for approxistandardization.The criterion used here for selecting elements mately 50 s prior to commencing laser sampling. Laser samas internal standards was as reported by Fryer et al.17 The pling for quantitative analysis was performed at 10 Hz. The concentrations of the elements chosen as internal standards laser was fired for 40 s to obtain steady-state signals (Fig. 1). were acquired using EPMA. A scanning electron micrograph Count rates were collected aexportable data files using the of a laser sampling crater for NIST silicate glass 610 is shown ELAN software.The data files were translated using a in Fig. 2. The measured diameters of the craters ranged from compiled, in-house written, Microsoft BASIC program which 60 to 80 mm. translated the data into a Lotus formatted spreadsheet file. All subsequent data manipulations were done o-line using commercial spreadsheet software on personal computers. Precision and accuracy Following background correction of the data, average signals obtained during laser sampling were converted into concen- Since LAM-ICP-MS is a spatially sensitive technique, a homogeneous material is required to demonstrate the precision and trations by calibrations against NIST synthetic glass reference materials. Drift, matrix eects, changes in laser sampling yield accuracy of the technique.In this study, NIST 610 glass was analysed for its trace element concentrations over a period of and transport eciency during LAM-ICP-MS analysis cannot 654 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12Table 2 Comparison of LAM-ICP-MS and solution nebulization (SN) ICP-MS analysis of NIST 610 silicate glass (concentrations in mg g-1) LAM-ICP-MS SN Element (n=6) RSD (%) LOD (n=3) Dierence (%) Sr 497 3.1 0.022 440 13 Y 439 2.8 0.036 430 2 Zr 389 2.0 0.080 433 -10 La 418 6.7 0.035 410 2 Ce 419 3.7 0.146 423 -1 Pr 414 5.1 0.141 420 -1 Nd 402 4.5 0.316 417 -3 Sm 417 3.5 0.033 423 -1 Eu 427 5.5 0.158 420 2 Gd 392 4.6 0.081 427 -8 Tb 448 5.0 0.019 420 7 Dy 409 3.7 0.061 413 -1 Ho 447 4.6 0.017 420 7 Er 449 6.1 0.050 420 7 Tm 443 4.5 0.013 427 4 Yb 432 3.9 0.031 427 1 Lu 438 4.8 0.010 410 7 Th 430 5.5 0.051 423 2 U 434 6.3 0.012 430 1 6 d.The NIST 612 glass was used as an external calibration standard and Ca as an internal standard. Results from this (a) (b) study obtained using LAM-ICP-MS are presented in Table 2. Fig. 3 Scanning electron micrograph of laser sampling craters of (a) The RSDs are<7% for all trace elements. Accuracy, calculated apatite (upper) and (b) monazite (lower).as the relative dierence between LAM-ICP-MS and solution ICP-MS values, is also given in Table 2. The relative dierence is 13% for all elements and <5% for most trace elements. Table 3 Comparison of LAM-ICP-MS and solution nebulization (SN) ICP-MS analysis of apatite (concentrations in mg g-1) Although one can reasonably argue that accurate calibration cannot be made in every case using synthetic glass samples, LAM-ICP-MS SN we confirmed the LA-ICP-MS results with EPMA results Element (n=14) RSD (%) (n=3) wherever possible.Y 332 30 460 La 2838 19 3200 Ce 4310 7 4200 L imits of detection Pr 310 7 340 Limits of detection are defined as three times the SD (s) of the Nd 822 19 990 Sm 101 28 130 dierence between the count rate of n replicates of a sample Eu 17 39 20 and n replicate determinations of the count rate of the back- Gd 84 37 120 ground.The limit of detection is calculated as 3sÓ(1/n+1), Tb 12 47 20 where s is expressed in concentration units. Limits of detection Dy 58 39 76 using LAM-ICP-MS are given in Table 2. Ho 14 58 15 Er 37 62 36 Tm 5.6 66 5.3 RESULTS AND DISCUSSION Yb 27 67 20 Lu 7.6 93 4.3 Trace Element Analyses of Geological Materials Apatite and monazite Apatite and monazite are two dissolution-resistant minerals and are therefore dicult to analyse by solution nebulization ICP-MS.A scanning electron micrograph of a laser sampling crater in apatite is shown in Fig. 3(a). As has been reported by Jackson et al.,12 most grains ablated uncontrollably, and rapid partial or complete destruction of the grain resulted because of very low laser light absorption. Calcium was used as an internal standard whose concentration was determined by EPMA. Analytical results for the apatite sample are presented in Table 3. Chondrite-normalized REE patterns indicate good agreement between LAM-ICP-MS and solution ICP-MS analyses (Fig. 4). Chondrite-normalized plots are constructed by dividing the REE concentrations obtained for the sample by those for a standard meteorite. A chondrite-normalized plot showing a smooth curve (except for Eu) indicates a geologically consistent result and serves as an independent verification of the validity of the chemical analysis. The high values for the RSDs indicate that the grains are very Fig. 4 Comparison of mean chondrite-normalized data for the geological material apatite using LAM-ICP-MS and solution ICP-MS. heterogeneous, which was also observed in a previous study.12 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 655Table 4 Elemental composition of monazite determined by LAM-ICP-MS and EPMA (concentrations in mg g-1; n=3) Spot 1 Spot 2 Spot 3 Spot 4 Element Mean RSD (%) Mean RSD (%) Mean RSD (%) Mean RSD (%) Average RSD (%) EPMA Sr 560 5 533 8 523 14 273 20 472 25 Y 21232 6 24895 33 15956 13 38535 10 25154 33 12620 Zr 2.1 12 1.5 5 13.5 44 2.2 15 4.8 105 Nb 765 9 592 7 2915 35 293 17 1141 91 Ba 507 8 582 5 442 12 333 65 466 20 La 98552 2 97821 1 100732 1 104638 3 100436 3 85860 Pr 23827 0 23846 4 25206 2 26427 3 24826 4 25660 Nd 100044 2 99951 3 99063 2 104715 2 100943 2 102870 Sm 24699 2 24915 3 23210 2 25719 3 24636 4 16520 Eu 102 5 105 2 118 2 117 5 110 6 510 Gd 12265 2 12296 4 10890 2 13074 3 12131 6 10790 Tb 1303 5 1322 9 1111 5 1543 4 1320 12 Dy 4055 4 4281 19 3242 7 5456 6 4258 19 2580 Ho 489 8 551 29 351 16 729 9 530 26 Er 1003 10 1219 36 659 20 1712 12 1148 33 Tm 162 6 199 35 97 26 270 13 182 34 Yb 1160 15 1388 37 672 25 1819 12 1260 33 Lu 127 6 153 47 84 31 225 11 147 35 Th 132580 5 124660 3 105604 7 75866 8 109677 20 73610 U 1680 15 1894 4 1372 21 2801 11 1937 27 Elemental concentrations in monazite range from several mg g-1 for Zr to 10% for La.It is very dicult to determine these elements in a single solution because the signal intensities for some elements rise above the linear dynamic range of the detector.In order to accommodate highly concentrated elements within a single analytical run, OmniRange settings were used for certain elements. This option reduces the sensitivity of the mass spectrometric system by momentarily changing conditions such that the ion throughput is diminished for certain elements. In this study, OmniRange settings of 11, 13, 4, 5 and 17 were used for La, Ce, Pr, Nd and Th, respectively.Cerium was used as an internal standard whose concentration was determined by EPMA. Analytical results of monazite are presented in Table 4. Four spots were analysed, and three replicates were completed for each spot. A scanning electron micrograph of a laser sampling crater in monazite is shown in Fig. 3(b). Chondrite-normalized REE patterns of Fig. 5 Comparison of chondrite-normalized data for the geological mean data for each spot in addition to the average of four material apatite using (A–D) LAM-ICP-MS and (E) EPMA.spots and EPMA data are shown in Fig. 5. Heavy rare earth elements are heterogeneously distributed and fractionated in dierent sample spots. Table 5 Trace element composition of chromites determined by LAM-ICP-MS (concentrations in mg g-1) Sample Ti V V* Mn Co Ni Ni* Cu Zn Ga Ge Zr Nb Hf chro20 49 1692 1904 2358 321 423 550 90 649 25 1.76 2.32 2.68 0.15 chro21 275 1533 1224 1834 417 279 157 32 2683 19 2.31 7.11 6.03 0.35 chro23 65 1157 1224 1572 314 944 1414 32 741 64 2.44 4.96 4.54 0.24 chro24 63 268 476 4061 142 397 864 7 521 17 0.51 0.86 0.73 0.07 chro25 98 797 1020 1310 153 1538 1807 27 364 64 1.41 1.71 2.31 0.14 chro26 337 829 816 2489 209 1474 1022 499 1558 43 16.62 17.06 37.44 2.46 chro27 122 745 1088 1834 154 708 1414 61 498 35 2.18 3.02 2.75 0.28 chro28 26 1716 1768 1899 366 642 707 137 837 41 1.68 1.75 2.15 0.35 chro29 190 394 340 1637 154 1384 2043 48 249 35 1.67 1.68 1.43 0.14 chro30 306 799 1156 1375 170 1435 2043 79 361 53 3.73 3.13 5.25 0.48 chro43 133 342 272 1441 158 1298 1493 46 168 25 3.34 1.99 4.87 0.46 chro44 116 903 1088 1441 359 1216 1414 38 651 67 2.76 2.63 5.69 0.50 chro45 43 1274 1564 2292 408 505 393 40 816 37 1.85 2.44 3.96 0.29 chro46 166 1098 884 2620 290 688 786 45 609 38 1.38 1.69 1.90 0.20 chro48 97 1363 1564 1965 297 756 1022 41 736 42 2.17 2.68 5.63 0.45 chro50 266 910 816 1441 178 1378 1572 64 373 44 2.96 4.05 15.55 0.47 chro52 175 225 1088 1441 40 270 1179 9 78 12 0.31 1.55 7.22 0.10 * EPMA data. 656 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12Table 6 Trace element composition of olivines determined by LAM-ICP-MS (concentrations in mg g-1) Element Olivine 1 Olivine 2 Olivine 3 Olivine 4 Ti 108 229 94 150 V 10 16 14 15 Cr 238 199 251 195 Mn 807 1120 806 1123 Mn* 829 1161 929 1239 Co 107 131 111 129 Ni 2404 2870 2269 2399 Ni* 2986 2829 2907 2672 Cu 138 708 131 405 Zn 265 1237 249 390 Y 0.6 1.8 1.1 0.7 Zr 3.9 8.6 2.7 4.5 Nb 1.5 12.1 3.2 1.9 * Data from EPMA.Fig. 6 Time variation of the signal of selected elements from two dierent chromite samples: (a) a homogeneous sample and (b) a sample showing element zoning. Fig. 8 Spatial variations of concentrations of (a) Ca, (b) Ni and Co and (c) Cu, Pb and Zn in a concretion sample. Chromite Trace element (Ti, V, Cr, Ga, Ni, Co, Cu, Zn and Zr) concentrations in chrome spinel are important indicators for diamond exploration.Chromite is a dissolution-resistant mineral and determinations of trace elements in chromite are Fig. 7 Scanning electron micrograph of laser sampling craters of a concretion sample showing sampling locations. usually carried out using proton microprobes.18 In this study, Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 657Table 7 Trace element composition across the length of a single sample exhibiting constant signals during the laser sampling concretion sample determined by LAM-ICP-MS (concentrations period, whereas Fig. 6(b) shows significant zoning for Ti, Zr, in mg g-1) Nb, Hf and Ta in this particular chromite sample. Sample spot Ca Co Ni Cu Zn Pb 1 935 71 372 24 290 42 Olivine 2 611 57 384 21 285 39 3 836 59 402 24 265 32 Trace element (Ti, V, Cr, Ga, Ni, Co, Cu, Zn and Zr) 4 499 69 598 35 308 35 concentrations in olivine are important for petrogenetic studies. 5 840 74 651 43 347 52 Silicon was used as an internal standard whose concentration 6 428 78 770 70 261 56 7 852 97 1161 70 321 64 was determined by EPMA. Analytical results for four dierent 8 1035 91 984 39 237 30 olivine samples are given in Table 6. Manganese and Ni 9 689 79 805 36 185 22 concentrations determined by LAM-ICP-MS are in good 10 868 94 1010 41 225 31 agreement with those obtained by EPMA. 11 800 84 846 53 275 45 12 410 84 781 51 264 50 13 403 68 554 42 221 46 14 543 77 501 31 216 34 Concretion sample 15 347 74 321 15 166 17 16 983 60 395 26 299 61 A concretion till sample was analysed for Ca, Co, Ni, Cu, Zn 17 1792 64 363 27 320 53 and Pb.The sample displayed a layered structure, as shown in Fig. 7, and there were no significant variations in major element concentrations. EPMA showed that MnO concentrations are constant at 0.5% throughout the sample, and so chromite samples of dierent origin were analysed by LAMICP- MS. Magnesium was used as an internal standard whose Mn was used an internal standard.In order to provide spatial variations of element concentrations, 17 spots at intervals of concentration was determined by EPMA. Analytical results for dierent chromite samples are given in Table 5. Vanadium 1 mm (Fig. 7) spanning the width of the sample were analysed by LAM-ICP-MS. The analytical results are given in Table 7. and Ni concentrations determined by LAM-ICP-MS are in good agreement with those obtained by EPMA. Time vari- This example demonstrates that LA-ICP-MS can be used to provide zone analysis at a scale not handled easily by electron ations of the signal (counts per second) of selected elements from two dierent chromite samples probably indicate sample microprobe techniques, which are limited to the analysis of spots of only a few micrometres in diameter.homogeneity (Fig. 6). Fig. 6(a) indicates a homogeneous Table 8 Rare earth element composition of fused standard reference material powders determined by LAM-ICP-MS (concentrations in mg g-1, n=8) Sy-2 AGV-1 Element Mean RSD (%) SN Lit.19 Mean RSD (%) Lit.19 Y 127 6 128 129 13 26 20 La 72 3 75 67 30 20 38 Ce 156 8 175 159 53 14 67 Pr 19 5 19 19.5 5.7 23 7.6 Nd 74 6 73 73.7 25 22 33 Sm 17 18 16 15.7 3.4 47 5.9 Eu 2.2 16 2.4 2.5 1.2 33 1.6 Gd 14 11 17 16.8 2.9 40 5.0 Tb 3.1 10 2.5 3.1 0.5 47 0.7 Dy 21 5 18 21.3 2.5 32 3.6 Ho 4.9 10 3.8 5.1 0.7 36 0.7 Er 17 11 12 15.2 1.7 35 1.7 Tm 2.6 22 2.1 2.4 0.3 36 0.3 Yb 19 5 17 18.1 1.1 38 1.7 Lu 3.0 16 2.7 2.9 0.3 78 0.3 Unkn-1* Unkn-2* Mean RSD (%) Lit.19 Mean RSD (%) Lit.19 Y 59 12 106 25 7 46 La 26 11 37 122 7 167 Ce 93 15 97 271 7 329 Pr 11 13 14 29 5 36 Nd 43 17 59 106 8 138 Sm 11 27 15 14 17 19 Eu 1.6 27 2.0 2.3 14 3.0 Gd 8.8 18 16 8.4 12 10.8 Tb 1.9 18 2.9 1.3 11 1.4 Dy 10.7 15 17 5.3 16 6.8 Ho 2.3 16 3.7 1.1 14 1.4 Er 6.6 14 10.1 2.8 26 3.5 Tm 1.1 24 1.5 0.4 28 0.5 Yb 6.3 7 9.7 2.3 13 3.1 Lu 0.9 27 1.4 0.3 20 0.4 * Unkn-1=1+1 mixture of ACE and BHVO-1; Unkn-2=1+1 mixture of GSP-1 and STM-1. 658 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12Silicate rock powder collaboration with Perkin-Elmer. J. Stirling and S. Ballantyne are thanked for providing some of the samples used in this Four rock powder standard reference materials were prepared study. The authors acknowledge the assistance of R. Meeds as silicate glass buttons by fusing a mixture of rock powder and N. Bertrand for completing ICP-OES and MS analyses and lithium metaborate (LiBO2).Standard reference material on solutions and A. Tsai for SEM photography. This paper is SY-2 is a syenite rock and AGV-1 is an andesite. Samples GSC Publication No. 1996240. designated Unkn-1 and Unkn-2 were used as ‘unknown’ samples to monitor the quality of LAM-ICP-MS analyses as were 1+1 mixtures of AC-E and BHVO-1 and of AC-E and REFERENCES STM-1. Reference material AC-E is a granite, BHVO-1 a basalt and STM-1 a syenite rock.Since Ca is a major element for all 1 Ready, J. F., Phys. Rev. A, 1965, 137, 620. of the matrices studied, this element was used as an internal 2 Darke, S. A., and Tyson, J. F., J. Anal. At. Spectrom., 1993, 8, 145. standard with values for its concentration taken from the 3 Moenke-Blankenburg, L., L aser Microanalysis, Wiley, New York, 1989, pp. 1–288. literature.18 Results for the determination of 15 REEs are given 4 Dittrich, K., and Wennrich, R., Prog.Anal. At. Spectrosc., 1984, in Table 8. The agreement between these data and literature 7, 139. values is reasonable, with the RSDs for most elements being 5 Ishizuka, T., and Uwamino, Y., Anal. Chem., 1980, 52, 125. <20%. Poorer RSDs were obtained for those elements whose 6 Mitchell, P. G., Sneddon, J., and Radziemski, L. J., Appl. concentrations are close to the limit of detection. Chondrite- Spectrosc., 1987, 41, 141 normalized plots for these samples were constructed (not 7 Darke, S.A., Long, S. E., Pickford, C. J., and Tyson, J. F., J. Anal. At. Spectrom., 1989, 4 715. shown) and gave petrologically valid patterns over the range 8 Thompson, M., Chenery, S., and Brett, L., J. Anal. At. Spectrom., of REE elements. 1989, 4, 11. 9 Gray, A. L., Analyst, 1985, 110, 551. CONCLUSION 10 Arrowsmith, P., Anal. Chem., 1987, 59, 1437. 11 Cromwell, E. F., and Arrowsmith, P., Anal. Chem., 1995, 67, 131. This study has shown that combining EPMA with laser 12 Jackson, S. E., Longerich, H. P., Dunning, G. R., and Fryer, B. J., sampling ICP-MS is a powerful analytical tool which can yield Can. Mineral., 1992, 30, 1049. very rapid and accurate trace analyses of a wide variety of 13 Marshall, J., Franks, J., Abell. I., and Tye, C., J. Anal. At. Spectrom., 1991, 6, 145. geological materials. Important results of this study are that 14 Perkins, W. T., Fuge, R., and Pearce, N. J. G., J. Anal. At. little or no sample preparation is required and that samples in Spectrom., 1991, 6, 445. dierent physical formats can be successfully handled if 15 Jarvis, K. E., and Williams, J. G., Chem. Geol., 1993, 106, 251. major compositional elements (concentrations pre-determined 16 Chenery, S., and Cook, J. M., J. Anal. At. Spectrom., 1993, 8, 299. by EPMA) are used as internal standards. The use of both 17 Fryer, B. J., Jackson, S. E., and Longerich, H. P., Can. Mineral., analytical technologies allows the complete determination of 1995, 33, 303. both major and trace elements in a single sample. The obvious 18 Ryan, C. G., Cousens, D. R., Sie, S. H., Grin, W. L., and Suter, G. F., Nucl. Instrum. Methods, Sect. B, 1990, 47, 55. disadvantage, of course, is the requirement for access to two 19 Govindaraju, K., Geostand. Newsl., 1994, 18, 1. relatively expensive analytical instruments, although these need not be located within the same laboratory. Paper 6/06599F Received September 25, 1996 This research was financially supported by the Industrial Partners Program of National Resources Canada (NRCan) in Accepted November 21, 1996 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 659

ISSN:0267-9477    

DOI:10.1039/a606599f

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Study of a Partial Least-squares Regression Model for Rare Earth Element Determination by Inductively Coupled Plasma Mass Spectrometry W. ZHUa , E. W. B. DE LEERb, M. KENNEDYc , P. KELDERMANc AND G. J. F. R. ALAERTSc aTNO Institute of Environmental Sciences, Energy Research and Process Innovation, Delft, T he Netherlands bNMi Van Swinden L aboratorium, Delft, T he Netherlands cInternational Institute for Inf rastructure, Hydraulic and Environmental Engineering, Delft, T he Netherlands A partial least-squares regression (PLSR) model was cessing. Spectral fitting procedures to reduce the eects of spectral interference in ICP-MS have been described in detail developed for rare earth element (REE) determination by inductively coupled plasma mass spectrometry (ICP-MS), in by De Boer et al.21 and Van Veen et al.28 Vaughan and Horlick29 proposed PCA correction procedures for REE deter- order to correct interferences from REE oxides, hydroxides and isobaric spectral overlap.The total variance was explained mination by ICP-MS and reported satisfactory results for specific REEs. by a 14 factor PLSR model. The square error of prediction was less than 0.005 and the oxide/hydroxide/isobaric In this study, a PLSR model was developed and aspects of experimental design, model optimization and model appli- interferences were almost completely removed. It was found that Nd, Sm, Gd, Dy and Yb played a more significant role in cation were addresssed.Using this model, the concentrations of elements in unknown samples can be predicted accurately the model than the other REEs because they possess multiple isotopes which require repeated calibration. REE and interferences from all REE oxides and hydroxides, isobaric overlap and Ba can be removed. concentrations could be accurately predicted despite barium interference, even in samples with high ratios of Ba to REE, when the weights of certain isotopes in the model were set THEORY below 0.1.The PLSR model was compared with the normal PLSR performs simultaneous and interdependent PCA in both calibration method (NCM) and the Gauss elimination method [X] and [Y] matrices, in such a way that the information in (GEM). The results indicated that the PLSR model was more the [Y] matrix is used directly as a guide for optimum accurate than the NCM and exhibited greater flexibility than decomposition of the [X] matrix, and subsequently performs the GEM.regression of [Y].1,9,10,30 Keywords: Partial least-squares regression; rare earth PCA is a method for the extraction of systematic variations elements; inductively coupled plasma mass spectrometry in a single data set (also called data decomposition). The general purpose of PCA is to perform data description and interpretation. Original data are decomposed into a model Principal component analysis (PCA), a technique for studying comprising both signal and error (noise) parts.To illustrate matrices of data, is one of the most powerful mathematical how PCA can be applied to REE mass spectra in ICP-MS, a methods in chemistry and environmental chemistry,1–4 which hypothetical data matrix is considered, i.e., [X], comprising can be applied whenever a measurement can be expressed as the signals of a series of objects/samples of the same measured a linear sum of product terms. PCA has been applied in areas mass spectral components. For example, suppose the [X] of chemistry such as spectrophotometry and mass spec- matrix contains i rows (objects or mixtures) and k columns trometry, and numerous applications of the technique have (variables or masses), the matrix can be written as been described in detail.5–8 Both principal component regression (PCR) and partial least squares regression (PLSR) are projection methods resulting in principal component models.PLSR algorithms9–13 are becoming more and more x11 x12 … x1k x21 x22 … x2k [X]= e e e e x(i-1)1 x(i-1)2 … x(i-1)k xi1 xi2 … xik (1) popular in analytical chemistry.ICP-MS is a multi-element analytical method for the determination of REEs. In the past decade, many publications dealing with the determination of REEs in geological samples have appeared.14–18 Likewise, many publications on the deter- Assuming m factors influence the above data set system, s mination of REEs in aqueous samples, e.g., surface, ground (score) is a measure of how much a factor is present in an and sea water, have been published.19–22 However, ICP-MS object and l (loading) expresses how much each original suers from polyatomic ion interference from REE oxides and variable contributes to each factor.The data x can be written hydroxides, isobaric spectral overlap and interference from as follows: doubly charged ions, which reduce the sensitivity and accuracy xik=sk1l1i+sk2l2i+…+sk(m-1)l(m-1)i+skmlmi (2) of measurements.23–25 Several correction approaches for the determination of REEs The [X] matrix can also be expressed in terms of [S] and [L] by ICP-MS have been developed.Longerich et al.26 and matrices: Dulski27 provided a detailed discussion of REE oxide forma- [X]=[S]×[L]+R(m) (3) tion and potential interferences during analysis. Lichte et al.14 minimized oxide formation by instrumental optimization and where R(m) is defined as the residual of the [X] matrix for m factors, and can be regarded as the model error. [S] and [L] employed an algebraic correction scheme during data pro- Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 (661–665) 661are the score and loading matrices of the [X] matrix, model data library. Software packages used for experimental design, data acquisition, model establishment and data predic- respectively. In the same way, the concentration matrix, [Y], can be tion included Unscrambler, version 4.0 (CAM, Norway) and Statgraphic Plus, version 6.0 (Manugistics, USA).expressed as [Y]=[S]¾×[L]¾+R¾(m) (4) Experimental Design where [S]¾ and [L]¾ are the score and loading matrices of [Y], respectively. R¾(m) is the model error in the [Y] matrix The multivariate calibration model which relates [Y] and [X] for m factors, and can be calculated by subtracting [S]¾×[L]¾ variables has to be determined from a set of calibration objects/ from [Y]. samples. In order to ensure reliable prediction, the program The PLSR model is based on the non-linear iterative partial Statgraphic Plus was employed and factorial design was chosen least-squares (NIPALS) algorithm.9 The algorithm extracts for the design of the mixed REE solutions.Two-level designs one factor at a time and each factor is obtained iteratively by were considered in this study. Fifteen factors were selected, repeated regression of [X] on scores [S] to obtain an improved corresponding to the number of REEs (Pm was omitted) and loading [L].barium in the model system. Forty mixed samples were rec- In summary, PLSR theory can be used to establish a cali- ommended as the optimum number for the system employing bration model based on a representative set of corresponding afolded Plackett–Burman31 design method. In order to develop measurements of both X-data and Y -data: the best model and test its applicability, four dierent concentration ranges were studied: (a) 5–30, (b) 50–500, (c) 0–100 X-data+Y -data�calibration model and (d) 0.5–5 mg l-1.Fifteen pure element standards were Once the calibration model is available, the model can be use prepared for both the lowest and highest concentration of each to predict new Y -data from the measurement of new X-data: range (i.e., 30 standards per concentration range). All mass spectral data (uncorrected counts measured new X-data+calibration model�new Y -data per second) from the ICP-MS system were corrected by 115In (internal standard) and further transformed into data matrices.The data matrix files were converted into an acceptable format using Unscrambler. A data matrix set is given in Fig. 1. The EXPERIMENTAL data set consists of a number of objects/samples (rows/lines), Reagents and Standards each object/sample corresponding to a set of variables/REE ipes (columns). The matrix dimensions are 70 lines/samples REE solutions were prepared from 1000 mg ml-1 stock standard solutions (Perkin-Elmer, U� berlingen, Germany).Single ×52 columns/isotopes. The matrix consists of 3640 data points. element standards were prepared for La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu and Ba, and in all cases RESULTS AND DISCUSSION indium (115In) was added as an internal standard. De-ionized water (18 MV purity) and super-pure HNO3 (Merck, Eect of the Number of Factors on the Model Darmstadt, Germany) were used for all dilution and digestion PLSR, the model calibration method of interest in this study, experiments.Reference materials employed in this study com- is characterized by the number of factors required for accurate prised marine sediment (MAG-1) and soil (GBW07313) mate- data prediction. When the correct number of factors is rials, from the US Geological Survey (Washington, DC, USA) employed, the resulting data matrix should equal the original and from the Chinese National Research Centre for Certified data matrix, within experimental error.If too few factors are Reference Materials (Beijing, China), respectively. employed in the model, the data will not be reproduced with sucient accuracy, and if too many factors are used, the extra factors will introduce additional model error. Instrumentation Variance is often expressed as the mean square error of the Measurements were performed on a VG Plasmaquad Plus mean value of residual squares. To assess the eect of the ICP-MS, system (VG Elemental, Winsford, Cheshire, UK).number of factors on the model, explained variance is Indium-115 was used for optimization of the operating param- employed. Explained variance is expressed as a percentage, eters and ion lens voltages. Typical ICP-MS operating param- indicating the accuracy of the model. Each additional new eters optimized for the analysis of aqueous REEs are given in factor explains more variance in the original data set. In Table 1. A Vectra 50 MHz 486-DX microcomputer (Hewlett- general, before model calibration, 0% of the original variance Packard Avondale, PA, USA) with 8 Mb RAM, mathematical is explained, whereas 100% of the original variance is explained coprocessor and 210 Mb hard disk was employed for the by the model when the optimum number of factors is chosen.Table 1 Typical ICP-MS operating parameters for aqueous REE determination Rf power— Forward/kW 1.439 Reflected power/W 0 Gas flow rates/l min-1— Inner 1.1 Intermediate 1.18 Outer 13.5 Scanning mode Elemental Measurement mode Multi-channel Measurements per peak 3 Measurement time/s 1 Repeats per integration 6 Mass range (m/z) 100–180 Fig. 1 A data matrix set for PLSR modelling. 662 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12Here, the first factor explained about 65% of the total variance to explain the eect of weights on the model: and the first six factors together described about 90% of the RD=[(Cprediction-Ccertified)/Ccertified]×100% (5) total variance; further factors (from 7 to 14) explained only where Cprediction is the concentration calculated by the model marginally more of the total variance (approximately 1% more and Ccertified is the certified (known) concentration. total variance per factor).Thus, 14 factors together explained Fig. 3 shows the RD obtained using dierent weights. The nearly 100% of the total variance. When fewer than 14 factors PLSR model was very sensitive to added weights. When the were used for data prediction, the results always showed lower weights for all X and Y variables were set at 1.0, a large error concentrations than the reference samples.On the other hand, resulted (maximum RD of 60%), owing to large dierences when more than 14 factors were used, the prediction error between variable values. For example, 142Ce16O1H interferes increased. Furthermore, part of the results was considered with the determination of 159Tb, and if the weights of all [X] abnormal, because higher order factors increased the unexvariables were set at 1.0, the RD in the prediction of 159Tb plained variance in the model.The eect of the number of would be about 59%. Setting the weights equal to the inverse factors on dierent REE concentration ranges, i.e., 0.5–5, 0–100 of the standard deviation (1/s) of the absolute noise for all [X] and 0–30 mg l-1 (see Experimental Design) was also investiand [Y] variables yielded an RD of approximately 8% for the gated.The results are shown in Fig. 2. The eect of the number low REE concentration range (below 1 mg l-1) and less than of factors on all the concentration ranges was very similar and 5% for concentration ranges between 1 and 100 mg l-1. nearly 0% total residual variance was obtained with 14 factors From the results of residual variance, it can be seen that the for all concentration ranges investigated. model employing weights equal to 1/s performed better than the model with weights equal to 1.The noise (residual variance) Eect of Weights on the Model from ICP-MS depends on the interference level. If 1/s weight is used for the prediction, (i) the eect of normal noise In order to reduce the eects of dierent variables and noise (interference signal) can be easily reduced and (ii) the influence in a data set, a priori weights can be placed on the [X] and from accidental cases (unstable instrumental cases or sample [Y] variables. The term relative deviation (RD) was employed contamination) can also be minimized.It was found that 136Ba,137Ba and 138Ba seriously influence the REE prediction, especially if high concentration ratios (CBa/CREE) are present. As a result, it is very dicult to predict accurately the concentrations of Ce, La and some other isotopes because of strong interference from Ba. However, it was found that barium interference could be overcome by changing the model weights. For example, if the weights set for 136Ce, 138Ce and 138La in the [X] matrix were decreased, the RD of prediction was reduced for these isotopes.Moreover, when weights below 0.1 were used for these isotopes, a minimum RD was obtained (<5%). Loading and Score Loading expresses the relationship between variables and factors in a model (Table 2). From the loading, the importance Fig. 2 Comparison of the model variance explained for dierent REE of each isotope/variable in each factor can be established.concentration ranges: A, 5–30; B, 50–500; C, 0–100; and D, 0.5–5 mg Factor 1 primarily described all the REEs and was the most l-1. significant factor in the model. Factors 2–6 were important for the description of REEs which possess multiple isotopes, i.e, Nd, Sm, Er, Dy, Yb and Gd. Factors 7–14 were important for the description of REEs which possess individual isotopes or isotopes in abundance, i.e, Pr, Tb, Ho, Tm and 139La, 140Ce and 175Lu (see Table 2 for summary).The term ‘score’ can be used to study variations between objects in the data set and to interpret which properties or characteristics of the objects are important in the model. Score results indicated that 66% of the calibration variance was described by factor 1 and 6, 5, 5, 4, 4 and 2% were described by factors 2–7, respectively. Factors 8–14 described 1% of the variance, implying that higher order factors influence the model Fig. 3 Eect of weights on the PLSR model (RD is the relative by less than 1%.Factor score plots showed that factor 1 is deviation, 1/s is the inverse of the standard deviation of the absolute noise). 1, Weight=1/s; &, weight=1.0. the most important factor in describing all the samples/objects Table 2 Relationship between variables (REE isotopes) and factors in the PLSR model Factor no. Description of Factor No. Description of 1 All REEs 8 140Ce, 151Eu and 153Eu 2 Nd, Sm, Gd, Er, Dy and Yb isotopes 9 139La, 140Ce, 159Tb, 165Ho, 169Tm and 175Lu 3 Nd and Sm isotopes 10 139La and 175Lu 4 Gd and Yb isotopes 11 139La, 140Ce, Tb159, 165Ho, 169Tm and 175Lu 5 Gd, Dy and Yb isotopes 12 141Pr 6 Er, Dy and Yb isotopes 13 165 Ho and 169Tm 7 140Ce, 159Tb, 165Ho, 169Tm and 175Lu 14 159Tb,65Ho and 169Tm Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 663in the model. If higher concentrations of Nd, Sm, Gd, Dy, Er and Yb were present in the sample/object, the importance of the sample/object in this factor is increased.From factors 2 to 14, most score contribution came from samples/objects containing higher concentrations of the elements Nd, Sm, Gd and Dy. For these important elements each factor in the score plot is closely related to the result of the loading plot. Elements such as Nd, Sm, Gd, Dy and Yb played a more important role in the model because they posses multiple isotopes which require repeated calibration. Accuracy of the Model Fig. 4 compares the concentrations of 140Ce, 152Sm, 164Dy and 174Yb predicted with the PLSR model with their certified concentrations.The correlation coecients were 0.9967, 0.9995, 0.9998 and 0.9999, respectively. 152Sm, 164Dy and 174Yb showed Fig. 5 Predicted concentrations of 140Ce at (a) 500 and (b) 50 mg l-1 good correlation between the certified and predicted values. and 152Sm at (c) 500 and (d) 50 mg l-1. The average correlation coecient for all 52 REE isotopes was 0.9982. It can be concluded that all variables were well additions of all REEs and increasing the model weights of modelled and a good correlation existed between predicted these samples.and known (certified) values. The concentration of REEs in a set of standard solutions was predicted with the PLSR model. The solutions contained 10 mg l-1 of each REE and were measured by ICP-MS. The Error of the Model results indicated (Fig. 6) that dierent isotopes could be meas- In order to define the expected error of future predicted values, ured with the same accuracy even when their abundance the uncertainty limits of the model (prediction error) were diered by a factor of 1000.The recovery of the REEs ranged estimated. Fig. 5 shows the predicted values for Ce and Sm at between 10.3 and 9.8 mg l-1 and standard deviations were less 50 and 500 mg l-1; the prediction error is represented as the than 0.41 mg l-1. Both the relative error and RSD were less vertical bars on the graph.The prediction error was higher at than 5%. the lower (50 mg l-1) concentration level, being about 7–10% for all elements. On the other hand, the prediction error at the Comparison of the PLSR Model and the Normal Calibration higher concentration level (500 mg l-1) was less than 5% for Method all elements. A comparison of REE concentrations predicted with the PLSR model and the normal calibration method (NCM) for the soil Prediction of REE Concentrations Using the PLSR Model reference material indicated that the PLSR model was more The prediction of REE concentrations in samples was carried accurate (Fig. 6). The average RSD of prediction was 4.7 and out as follows: the sample mass spectral signals were obtained 15.2% for the PLSR model and the NCM, respectively. The from ICP-MS, then normalized by daily calibration (using relative deviation of prediction was less than 5% in most cases standard solutions or reference materials) to ensure the same instrumental sensitivity as the model data set.If a large variation in instrumental sensitivity between the model input data and sample input data was found or if the sample matrix diered greatly from the model matrix, the model must be re-calibrated using extra input data from samples with standard Fig. 6 Comparison of the PLSR model and the NCM for various Fig. 4 Accuracy of the PLSR model: comparison of predicted and REE isotopes. Prediction was made using GBW07313 soil reference material, diluted 1000-fold with HNO3.certified concentrations of (a) 140Ce, (b) 152Sm, (c) 164Dy and (d) 174Yb. 664 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12for the PLSR model. However, for Lu, Tm and Ho the relative REFERENCES deviations were 11, 10.2 and 8.9%, respectively. This may be 1 Malinowski, E. R., Factor Analysis in Chemistry, Wiley, New attributed to the fact that their concentrations in the soil York, 1980, p. 251. reference samples were close to the detection limit. 2 Brereton, R. G., Chemometrics Applications of Mathematics and Statistics to L aboratory Systems, Ellis Horwood, Chichester, 1990, p. 307. 3 Meloun, M., Militky, J., and Forina, M., Chemometrics for Analytical Chemistry, Ellis Horwood, Chichester, 1992, p. 330. 4 Sjogren, M., Li, H., Rannug, U., and Westerholm, R., Environ. Comparison of the PLSR Model with the PCA and Gauss Sci. T echnol., 1996, 30, 38. Elimination Methods 5 Devaux, M.F., Bertrand, D., Robert, P., and Qannari, M., Appl. Spectrosc., 1988, 6, 1020. PCA only decomposes the original X data into a model 6 Machado, A. A. S. C., and da Silva, J. C. G., Chemom. Intell. L ab. comprising signal and error (noise) parts, whereas the PLSR Syst., 1993, 19, 155. model simultaneously and interdependently decomposes both 7 Sahota, R. S., and Khaledi, M. G., Anal. Chem., 1994, 66, 2374. the [X] and [Y] matrices so that the information in the [Y] 8 Smith, R.M., and Burford, M. D, Chemom. Intell. L ab. Syst., 1993, 18, 285. matrix is used directly as a guide for the optimum decompo- 9 Geladi, P., and Kowalski, B. R., Anal. Chim. Acta, 1986, 185, 1. sition of the [X] matrix, and then performs regression of Y. 10 Helland, I. S., Scand. J. Stat., 1990, 17, 97. In this way, the concentration of new unknown samples can 11 De Jong, S., Chemom. Intell. L ab. Syst., 1993, 18, 251. be predicted accurately using the PLSR model and interference 12 Ferre, J., and Rius, F.X., Anal. Chem., 1996, 68, 1565. from all REE oxides and hydroxides, isobaric overlap and Ba 13 Wang, H. Y., Wang, D. X., Wang, Y. H., Chen, S. G., and Zhang, F. J., Analyst, 1995, 5, 1603. interference can be removed. 14 Lichte, F. E., Meier, A. L., and Crock, G., Anal. Chem., 1987, The Gauss elimination method (GEM) is a method of 59, 1150. algebraically solving n equations in n unknowns; a detailed 15 Jarvis, K. E., J. Anal. At. Spectrom., 1989, 4, 563.discussion was presented by Vaughan and Horlick.29 This 16 Doherty, W., Spectrochim. Acta, Part B, 1989, 44, 263. 17 Croudace, I. W., and Marshall, S., Geostand. Newsl., 1991, 15, 139. method provides a correction scheme for spectral overlap of 18 Balaram, V., Curr. Sci., 1995, 8, 640. REE oxides. The accuracy is comparable to that with the 19 Moller, P., Dulski, P., and Luck, J., Spectrochim. Acta, Part B, PLSR model. However, it is a complicated and time-consuming 1992, 47, 1379.method for the correction of REE spectral overlap and oxide/ 20 Stetzenbach, K. J., Amano, M., Kreamer, D. K., and Hodge, hydroxide interferences. In contrast, the PLSR method is fast, V. F., Ground Water, 1994, 6, 976. 21 De Boer, J. L. M., Verweij, W., van der Velde-Koerts, T., and flexible and simple to use. Mennes, W., Water Res., 1996, 30 190. 22 Kreamer, D. K., Hodge, V. F., Rabinowitz, I., Johannesson, K. H., and Stetzenbach, K. J., Ground Water, 1996, 34, 95. 23 Tan, S. H., and Horlick, G., Appl. Spectrosc., 1986, 4, 445. 24 Vaughan, M. A., and Horlick, G., Appl. Spectrosc., 1986, 40, 434. CONCLUSIONS 25 Evans, E. H., J. Anal. At. Spectrom., 1993, 8, 1. 26 Longerich, H. P., Fryer, B. J., Strong, D. F., and Kantipuly, C. J., Employing optimum modelling parameters (e.g., 14 factors, Spectrochim. Acta, Part B, 1987, 42, 75. model centre at the origin and model weights equal to 1/s), 27 Dulski, P, Fresenius’ J. Anal. Chem., 1994, 350, 194. almost 0% total residual variance was achieved with the PLSR 28 Van Veen, E. H., Bosch, S., and De Loos-Vollebregt, M. T. C., model. The square error of prediction was less than 0.005 and Petrochim. Acta, 1994, 49B, 1347. 29 Vaughan, M. A., and Horlick, G., Appl. Spectrosc., 1990, 44, 587. the oxide/hydroxide/isobaric interferences were almost com- 30 Wold, S., Esbensen, K., and Geladi, P., Chemom. Intell. L ab. Syst., pletely (100%) removed. Nd, Sm, Gd, Dy and Yb played a 1987, 2, 37. more significant role in the model because they possess multiple 31 Deming, S. N., and Morgan, S. L., Experimental Design: a isotopes which require repeated calibration. Barium inter- Chemometric Approach, Elsevier, Amsterdam, 1993, p. 437. ference could be overcome by setting the weights of certain isotopes in the model below 0.1. The results indicated that the Paper 6/07621A PLSR model was more accurate than the NCM and was Received November 8, 1996 Accepted February 12, 1997 simpler and more flexible than the GEM. Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 665

ISSN:0267-9477    

DOI:10.1039/a607621a

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Microconcentric Nebulizer for the Coupling of Micro Liquid Chromatography and Capillary Zone Electrophoresis With Inductively Coupled Plasma Mass Spectrometry ANDERS TANGEN*, ROGER TRONES, TYGE GREIBROKK AND WALTER LUND Department of Chemistry, University of Oslo, P.O. Box 1033, N-0315 Oslo, Norway A concentric nebulizer designed to deliver less than 10 The analytical figures of merit of the DIN are well documented, 19,24 but there is little documentation for the use of the ml min-1 directly into an ICP-MS system is described.The DIN at flow rates lower than 10 ml min-1 . Wiederin et al.17 influence of the liquid flow rate, the nebulizer gas flow rate and concluded that the optimum flow rate for the DIN-ICP-MS the distance from the nebulizer tip to the plasma was system is 120 ml min-1. This was nearly identical with the flow evaluated with regard to sensitivity and precision. The rate found to produce the best signal-to-noise ratio for a DIN- nebulizer was tested at liquid flow rates of 3–7 ml min-1. ICP-AES system.25 A number of papers18,20,22,23 have reported Samples of tetramethyllead and tetraethyllead dissolved in the use of liquid flow rates in the range 10–200 ml min-1.A dimethylformamide were used for the evaluation of the recent paper28 evaluates a commercially available microcon- nebulizer. The detection limit for Pb (m=208) with a direct centric nebulizer (MCN-100, Cetac Technologies, Omaha, NE, injection system was found to be 0.3 pg (S/N=3) or 5 mg l-1, USA) coupled to an ICP-MS system through a standard spray based on an injection volume of 60 nl, a liquid flow rate of 5 chamber.Figures of merit were given for liquid flow rates ml min-1 and a dwell time of 10 ms. The precision was found between 6 and 80 ml min-1. RSD values 1% were reported to be 1.9% (n=10) based on peak area measurements. The for 10 consecutive measurements with a total measurement responses of tetramethyllead and tetraethyllead were not time of 30 min.However, the use of such long MSmeasurement significantly dierent (p=0.05 ). The time needed for one times (3 min) may not be representative for the narrow peaks determination was less than 15 s, owing to the quick rinse out possibly obtained by CZE or packed capillary HPLC. of the nebulizer. The nebulizer can be used as part of a direct Considering the use of the MCN-100 for the coupling of CZE injection instrumental set-up. An application of the nebulizer to and ICP-MS or micro flow LC and ICP-MS, we expect the coupling of packed capillary liquid chromatography with problems with long wash-out times and large dead volumes ICP-MS detection is demonstrated.due to the use of a spray chamber. Keywords: Inductively coupled plasma mass spectrometry; The aim of this investigation was to develop a nebulizer for direct injection nebulization ; concentric nebulizer; speciation; the coupling of packed capillary high-temperature LC29 and micro liquid chromatography; capillary zone electrophoresis CZE with ICP-MS.Both techniques are based on mobile phase flow rates less than 10 ml min-1. The advantages of After the commercial introduction of ICP-MS in 1983, a micro liquid chromatography (mLC) include low consumption number of sample introduction modes have been tested for the of both mobile and stationary phases, which facilitates the use introduction of liquid samples, including standard pneumatic of exotic or expensive phases, increase in mass sensitivity, nebulizers,1–3 ultrasonic nebulization,4 thermospray nebuliz- achievement of high resolution with long columns and applicaation, 5 high-pressure nebulization,6–8 oscillating capillary neb- bility of temperature programming.30 Packed capillary columns ulization,9 monodisperse microparticulate injection,10,11 high- provide more ecient heat transport because of their smaller eciency nebulization (HEN)12 and electrothermal vaporiz- dimensions, and thereby faster response to temperature proation. 13 Each mode of sample introduction represents dierent gramming than conventional sized columns (4.6 mm id). Another important factor is that small diameters prevent radial properties with regard to parameters such as transport temperature gradients which will decrease the eciency. eciency, precision, detection limits, sample size demand, Detection by ICP-MS will make it possible to identify liquid flow demand, spectral and non-spectral interferences organometallic compounds.and washout time. Concerning the coupling of chromato- The nebulizer was examined with both an open tubular mLC graphic systems to ICP-MS,14,15 the direct injection nebulsystem and a packed capillary HPLC system. The open tubular izer16–27 (DIN) deserves special attention in this paper. ‘column’ was used to facilitate short analytical runs with a low In 1984, Lawrence et al.16 described the first DIN for ICPargon consumption when the nebulizer was tested.The current AES. Wiederin et al.17 described an improved model for instrumental set-up can be used as both a direct injection ICP-MS in 1991. The advantages of the DIN include improved system and a mLC–ICP-MS system. precision,17 reduced rinse-out time,17 low absolute detection limits,17 reduced peak broadeningdue to the low dead volume16 EXPERIMENTAL and good compatibility with organic solvents.17 The DIN is particularly useful for interfacing an LC system with an ICP Apparatus system; LaFreniere et al.18 described an HPLC–ICP-AES Open tubular mL C–ICP-MS system used for elemental speciation studies based on a DIN interface.Later the DIN nebulizer was also used for a number A Waters (Milford, MA, USA) Model 590 pump with modified of separation techniques combined with ICP-MS, includ- inlet and outlet check-valves and a Valco (Houston, TX, USA) ing columns of microbore dimensions18–21 and recently also Model CI4W manually operated injection valve with a 60 nl internal loop volume were connected to a Perkin-Elmer capillary zone electrophoresis (CZE).23 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 (667–670) 667(Norwalk, CT, USA) SCIEX ELAN 5000 ICP-MS system via hours for gradual release of the pressure through the steel screen. The column was cut to the desired length and supplied the concentric nebulizer. The ICP-MS system was controlled by an IBM PS/2 77 486 DX2 computer equipped with ELAN with a similar steel screen at the inlet end before it was placed in the chromatographic system. 5000 (Xenics) software. The ICP-MS operating conditions and data acquisition parameters are given in Table 1. A 115 cm×15 mm id fused silica capillary (Polymicro Nebulizer Design Technologies, Phoenix, AZ, USA) was used for the connection The nebulizer is shown in Fig. 1. It is positioned inside the of the injector to the nebulizer.A mixture of oxygen standard torch and injector liner provided by the instrument (40 ml min-1) and argon (800 ml min-1) was used as the nebul- manufacturer. The distance from the tip of the nebulizer to the izer gas. The two gases were mixed in a 1/4 in Swagelok tee. bottom of the rf coil (this distance is marked L in Fig. 1) was The flow rate of oxygen was controlled with a Bronkhorst varied by using dierent lengths of nebulizer steel tubing. The HI-TEC (Ruurlo, The Netherlands) mass flow meter.nebulizer gas is forced into the nebulizer steel tubing through the four holes (0.5 mm) shown in Fig. 1. The end of the fused Instrumental set-up for X–Y adjustment silica capillary was located 0.2 mm behind the end of the steel The system was the same as above, except that a Rheodyne tubing. The thread on the nebulizer is dimensioned to fit into (Cotati, CA, USA) Model 7010 injector with a 20 ml external the thread of the connection nut that is placed on the original loop was placed between the pump and the Valco injectionvalve. injector liner of the ELAN 5000 ICP-MS instrument.Packed capillary mL C–ICP-MS Reagents For the mLC experiments, the open tubular capillary was Tetramethyllead (TML) and tetraethyllead (TEL) were replaced by a 70.5 cm×320 mm id×450 mm od column packed obtained from Associated Octel (Milton Keynes, UK). TML with ODS particles (5 mm). The column was connected to the and TEL were dissolved in dimethylformamide (DMF).injector by a 25 cm×50 mm id fused silica capillary and to the Dimethylformamide (glass distilled), chloroform, methanol and nebulizer by a 50 cm×15 mm id fused silica capillary. acetonitrile (AN) (all of HPLC quality) were obtained from Rathburn (Walkerburn, UK). Oxygen (99.7%) and argon Column Preparation (99.999%) were purchased from AGA (Oslo, Norway). All fused silica capillaries were obtained from Polymicro The length of the empty column was approximately 80 cm.Technologies. The capillary with a 0.5 mm steel screen (Model 24000; Keystone Scientific, Bellefonte, PA, USA) at one end was Procedure placed in a 10.4 l sonicating bath (NEY 300; Bloomfield, CT, USA) containing water. The temperature of the water was Dierent lengths of the nebulizer steel tubing were tested by kept at 60°C during the packing procedure. The open end of injecting 60 nl of TML and TEL in a carrier stream/mobile the capillary was connected to a 70×2 mm id stainless steel phase of 10% DMF in AN.For the packed capillary chromato- tube which served as the reservoir for the packing material graphic experiments, an eluent of water–chloroform–methanol (5 mm ODS particles, Hypersil; Shandon, UK). (5+5+90) was used. Peak heights and areas were calculated An Isco (Lincoln, NE, USA) Model 100 DM pump was with Microcal (Northampton, MA, USA) Origin version 3.5 used to pack the column.Liquid carbon dioxidewas introduced software. into the packing reservoir at a starting pressure of 100 bar.The pressure was gradually increased by 10 bar min-1 to RESULTS AND DISCUSSION 550 bar, which was held for 30 min. This procedure, in conjunc- Useful Experimental Experiences tion with the ultrasonic vibration produced by the sonicator, led to the formation of a uniformly packed bed inside the A 115 cm×15 mm id fused silica capillary was necessary to capillary. The valves between the pump and the packing obtain a suitable working pressure for the check valves on the reservoir were then closed and the column was left for several pump.Occasionally, this capillary clogged at the injector side owing to the small id of 15 mm. The inlet and outlet check valves of the Waters Model 590 pump had to be modified to Table 1 ICP-MS operating parameters make it work properly at the low liquid flow rates used in this study. The total time of an analytical run was approximately Plasma gas flow rate 15 l min-1 15 s.The injected sample required 3–7 s (depending on the Auxiliary 1.0 l min-1 liquid flow rates used) from the point of injection to the Nebulizer (variable) 0.8 l min-1 Ar and 40 ml min-1 O2 Forward power 1400 W Sampler Pt; aperture diameter 1.15 mm Skimmer Pt; aperture diameter 0.89 mm Injector liner Alumina (2.0 mm id) Resolution Normal Transfer frequency Measurement Baseline time 0 ms Polarity + Data acquisition parameters — Nebulizer test mLC–ICP-MS Replicate time/ms 10 100 Dwell time/ms 10 100 Scanning mode Peak hop Peak hop Sweeps per reading 1 1 Fig. 1 Design of the nebulizer. 1, sampler core; 2, coil; 3, torch; 4, Readings per replicate 1 1 nebulizer tip; 5, screw; 6, D union; 7, threads for connection to the Number of replicates 1200 25000 injector liner; 8, four holes in the 1/16 in stainless steel tubing; 9, fused Points per spectral peak 1 1 silica capillary, 1.15 m×15 mm id; 10, nebulizer gas; 11, Valco injector; Isotope 208 (Pb) 208 (Pb) 12, pump; 13, length (L ) between the nebulizer tip and the bottom of the rf coil; and 14, injector liner. 668 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12detector. The signal was back at the background noise level after 10–15 s. Experiments were performed without the pump in order to check the self-uptake rate of the nebulizer. No self uptake was detected, probably owing to the small inner diameter of the capillary (15 mm) and the length (30–100 cm). Oxygen was added at a flow rate of 40 ml min-1, which was sucient to prevent carbon deposits at the sampler and skimmer cones.An rf power of 1400 W provided the most stable plasma. At lower power the plasma had a tendency to flicker. The X–Y adjustment of the nebulizer tip was found to be more critical than with normal sample introduction. Working with transient signals, the X–Y adjustment of the nebulizer tip is time consuming, so a coarse adjustment was made prior to plasma ignition by extending the sample introduction fused silica capillary to the hole of the sampler cone and then turning the adjustment knobs on the sampler interface to centre the nebulizer.After plasma ignition, only minor adjustments were performed prior to the experimental run. One may also inject a larger sample volume (20 ml) in the direct injection system to obtain a signal that is stable for a time long enough to optimize the X–Y direction and nebulizer gas flow rate. Both procedures maintained a satisfactory reproducibility.Obviously, the nebulizer steel tubing must be straight. Commercially available steel tubing of 0.5 mm id is made of soft steel. This problem was reduced by the use of a harder steel tube outside the nebulizer steel tubing, welded directly on the connection screw. Influence of Inner Diameter and Length of the Nebulizer Steel Tubing Preliminary investigations showed that nebulizer steel tubing of 0.5 mm id provided a more reproducible signal with less Fig. 2 Signals obtained after injection of 60 nl of TML (100.5 mg l-1 noisy spikes than tubing of 0.7 mm id. We believe that the as Pb) when the distance (L ) from the tip of the nebulizer steel tubing inner fused silica capillary (0.375 mm od) has a tendency to to the bottom of the rf coil was altered. L =(a)11.5, (b) 7.5 and (c) 3.5 cm. vibrate inside the nebulizer steel tubing of 0.7 mm id, giving rise to noisy signals. Therefore, 0.5 mm id nebulizer steel tubing rate tested, six injections of a TML sample (100.5 mg l-1 as was used for further studies.The end of the fused silica Pb) were made. The peak height increased linearly by 23% capillary was located 0.2 mm behind the end of the steel tubing when the liquid flow rate was increased from 4 to 7 ml min-1. to make sure that the nebulizer gas was forced to mix with It was not possible to use liquid flow rates higher than the mobile phase. This distance was chosen based on previous 7 ml min-1 without changing the experimental set-up, owing experiments with a laser light scattering detector.31 to the maximum pressure limit of the pump.An increase in Six dierent lengths of the nebulizer steel tubing were tested. flow rate was accompanied by a decrease in the peak width, The distances (L ) from the bottom of the rf coil to the tip of which gave rise to a higher signal. The precision was 4.5% or the nebulizer steel tubing (see Fig. 1) were 11.5, 9.5, 7.5, 5.5, better for all liquid flow rates tested.The influence of the 3.5 and 2.2 cm, respectively. Six injections of TML (100.5 mg l-1 nebulizer gas flow rate is shown in Fig. 3. As for ordinary as Pb) were performed for each length of steel tubing. The nebulizers, a well defined optimum nebulizer gas flow rate was signals obtained for L=11.5, 7.5 and 3.5 cm are shown in found. A nebulizer gas flow rate of 800 ml min-1 Ar with an Fig. 2. L=7.5 cm gave the highest sensitivity. However, noise additional 40 ml min-1 of O2 provided the best sensitivity for spikes were abundant using this condition (Fig. 2). For L= all liquid flow rates tested. 11.5 cm the peak area was half of that obtained for L=7.5 cm whereas for L=3.5 cm, the signal was very low (Fig. 2). The Response of TML and TEL shorter nebulizer steel tubes provided a more reproducible signal which was less prone to noise spikes. A possible expla- In order to test whether the response of TML and TEL diered nation for this observation is that the aerosol formed from the significantly, five standard solutions in the concentration range shortest nebulizer steel tubing contains the smallest droplets when it reaches the plasma: the distance from the tip of the nebulizer steel tubing to the plasma increases as the length of the nebulizer steel tubing decreases, which favours the evaporation process.The fact that the longest nebulizer steel tubing gave a very unstable plasma supports this explanation.In contrast to the DIN,17 which is positioned as close to the plasma as possible, the tip of this nebulizer is positioned at a longer distance from the plasma. Liquid Flow Rate and Nebulizer Gas Flow Rate Liquid flow rates of 3, 4, 5, 6 and 7 ml min-1 were evaluated Fig. 3 Eect of nebulizer gas flow rate on peak height for 100.5 with regard to sensitivity and precision. At each liquid flow mg l-1 (as Pb) TML. Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 669In the future, applications of high-temperature LC–ICP-MS and CZE–ICP-MS using the new nebulizer will be investigated. The authors thank Yngve Kristiansen and Bjørn Dalbakk in the engineering workshop in the Department of Chemistry for valuable technical assistance. REFERENCES 1 Browner, R. F., and Boorn, A. W., Anal. Chem., 1984, 56, 786A. 2 Browner, R. F., and Boorn, A. W., Anal. Chem., 1984, 56, 875A. Fig. 4 Separation of 10.5 mgml-1 (as Pb) TML and TEL by mLC– 3 Sharp, B.L., J. Anal. At. Spectrom., 1988, 3, 613. ICP-MS. Column, 70.5 cm×320 mm id×450 mm od capillary packed 4 Fassel, V. A., and Bear, B. R., Spectrochim. Acta, Part B, 1986, with ODS (5 mm); mobile phase, water–chloroform–methanol 41, 1089. (5+5+90); liquid flow rate, 3 ml min-1 ; injection volume, 60 nl. 5 Vanhoe, H., Moens, L., and Dams, R., J. Anal. At. Spectrom., 1994, 9, 815. 6 Jakubowski, N., Feldmann, I., Stuewer, D., and Berndt, H., Spectrochim. Acta, Part B, 1992, 47, 119. 40.2–402 mg l-1 (as Pb) were made for both TML and TEL 7 Luo, S. K., andBerndt, H., Spectrochim. Acta, Part B, 1994, 49, 485. dissolved in DMF. This nebulizer was developed for the 8 Berndt, H., and Ya� n� ez, J., J. Anal. At. Spectrom., 1996, 11, 703. 9 Wang, L., May, S. W., Browner, R. F., and Pollock, S. H., J. Anal. purpose of coupling high-temperature LC and CZE with At. Spectrom., 1996, 11, 1137. ICP-MS, so 10% DMF in acetonitrile was chosen as mobile 10 French, J.B., Etkin, B., and Jong, R., Anal. Chem., 1994, 66, 685. phase because it is useful for packed capillary high-temperature 11 Olesik, J. W., and Hobbs, S. E., Anal. Chem., 1994, 66, 3371. LC. Three injections were made of each standard solution and 12 Nam, S., Lim, J., and Montaser, A., J. Anal. At. Spectrom., 1994, calibration curves were constructed (method of least squares) 9, 1357. for both TML and TEL, based on estimates of peak area and 13 Park, C.J., and Hall, G. E. M., J. Anal. At. Spectrom., 1987, 2, 473. 14 Vela, N. P., Olson, L. K., and Caruso, J. A., Anal. Chem., 1993, height. The slopes were as follows (the standard deviation of 65, 585A. the slopes and the number of measurements are given in 15 Vela, N. P., and Caruso, J. A., J. Anal. At. Spectrom., 1993, 8, 787. parentheses): for peak area measurements, the slope for TML 16 Lawrence, K. E., Rice, G. W., and Fassel, V. A., Anal. Chem., was 8.117 (s=0.082, n=15) and for TEL the slope was 8.156 1984, 56, 289.(s=0.044, n=15), and for peak height measurements, the slope 17 Wiederin, D. R., Smith, F. G., and Houk, R. S., Anal. Chem., for TML was 6.168 (s=0.095, n=15) and for TEL the slope 1991, 63, 219. 18 LaFreniere, K. E., Fassel, V. A., and Eckels, D. E., Anal. Chem., was 6.189 (s=0.12, n=15). The responses of TML and TEL 1987, 59, 879. are not significantly dierent (p=0.05). 19 Zoorob, G., Tomlinson, M., Wang, J., and Caruso, J., J.Anal. At. TML and TEL were used to test the present system because Spectrom., 1995, 10, 853. these compounds have dierent responses in a normal nebulizer 20 Shum, S. C. K., Pang, H., and Houk, R. S., Anal. Chem., 1992, system used for atomic spectrometry.32–34 As shown above, 64, 2444. the observed response (sensitivity) was the same for TML and 21 Powell, M. J., Boomer, D. W., and Wiederin, D. R., Anal. Chem., 1995, 67, 2474. TEL with the nebulizer used in this work. This is an important 22 Shum, S.C. K., and Houk, R. S., Anal. Chem., 1993, 65, 2972. advantage when the element is present in dierent chemical 23 Liu, Y., Lopez-Avila, V., Zhu, J. J., Wiederin, D. R., and Beckert, forms in the sample. A species-specific response complicates W. F., Anal. Chem., 1995, 67, 2020. the quantitative determination of individual species. For chemi- 24 Shum, S. C. K., Johnson, S. K., Pang, H., and Houk, R. S., Appl. cal speciation studies by chromatographic methods, quantifi- Spectrosc., 1993, 47, 575.cation is much simplified when a single chemical species of 25 LaFreniere, K. E., Rice, G. W., and Fassel, V. A., Spectrochim. Acta, Part B, 1985, 40, 1495. each element can be used for calibration. Also, the quantitative 26 Avery, T. W., Chakrabarty, C., and Thompson, J. J., Appl. determination of an unknown species is impossible if the Spectrosc., 1990, 44, 1690. instrumental response is species-specific. 27 Wiederin, D. R., Smyczek, R.E., and Houk, R. S., Anal. Chem., The detection limit for Pb (m=208) with a direct injection 1991, 63, 1626. system was found to be 0.3 pg (S/N=3) or 5 mg l-1 based on 28 Vanhaecke, F., van Holderbeke, M., Moens, L., and Dams, R., an injection volume of 60 nl, a liquid flow rate of 5 ml min-1 J. Anal. At. Spectrom., 1996, 11, 543. 29 Trones, R., Iveland, A., and Greibrokk, T, J. Microcol. Sep., 1995, and a dwell time of 10 ms. No blank signal was detected with 7, 505.the present set-up and the organic solvents used in this work. 30 Ishii, D., Introduction to Microscale High-Performance L iquid Chromatography, VCH, New York, 1988, pp. 1 and 28. Coupling of mLC and ICP-MS 31 Homann, S., Norli, H. R., and Greibrokk, T., J. High Resolut. Chromatogr., 1989, 12, 260. An application of the new nebulizer demonstrating 32 Kashiki, M., Yamazoe, S., and Oshima, S., Anal. Chim. Acta, mLC–ICP-MS for speciation studies is shown in Fig. 4. TML 1971, 53, 95. and TEL were separated on a packed capillary column 33 Bagdi, G., Lakatos, J., and Lakatos, I., J. Anal. At. Spectrom., 1992, 7, 769. (70.5 cm×0.32 mm id). The column was packed with 5 mm 34 Al-Rashdan, A., Heitkemper, D., and Caruso, J. A., J. Chromatogr. porous ODS particles and the mobile phase was water– Sci., 1991, 29, 98. chloroform–methanol (5+5+90). This mobile phase has been 35 Hajlbrahim, S. K., J. L iq. Chromatogr., 1981, 4, 749. used to separate petroporphyrins.35 Fused silica capillaries of 36 Lee, M. L., and Markides, K. E., Analytical Supercritical Fluid 50 mm id were used as the connection tubing from the injector Chromatography and Extraction, Chromatography Conferences, to the column and from the column to the tip of the nebulizer, Provo, UT, 1990, p. 26. 37 Al-Rashdan, A., Vela, N. P., Caruso, J. A., and Heitkemper, D. T., in accordance with the guidelines given in the literature.30 The J. Anal. At. Spectrom., 1992, 7, 551. liquid flow rate was 3 ml min-1 , which provides the optimum linear velocity36 (0.1 cm s-1) on a 70 cm×0.32 mm id packed Paper 6/07623H capillary column (5 mm ODS particles). As can be seen from Received November 8, 1996 Fig. 4, the peaks of TML and TEL are well separated. The Accepted March 5, 1997 unknown peaks in the chromatogram are probably due to contamination or degradation of the tetraalkyllead products.37 670 Journal of Analytical Atomic Spectrometry, June 1997, Vol

ISSN:0267-9477    

DOI:10.1039/a607623h

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

On-line Removal of Anions for Plant Analysis by Inductively Coupled Plasma Mass Spectrometry AMAURI A. MENEGA� RIO AND MARIA FERNANDA GINE� * Centro de Energia Nuclear na Agricultura, Universidade de Sa�o Paulo, Caixa Postal 96, 13400–970, Piracicaba, S.P., Brazil The multi-element determination of B, Al, Cu, Zn, Mn, Fe, system. Subsequently, the retained species were eluted and sulfate was determined at m/z 48. Cr, Cd, Pb and S in plant digests by ICP-MS with on-line removal of anions is presented.A column with the anionic resin AG1-X8 was placed in a flow system to remove mainly EXPERIMENTAL sulfate and chloride while the solution was flowing to the ICP-MS system for the determination of the other elements. A VG-Plasma Quad PQ-2 ICP-MS system with a concentric Subsequently, the sulfur retained by the column was eluted and nebulizer installed in a water-cooled (10°C) spray chamber determined at m/z 48. The eect of perchlorate on sulfate was used. The experimental conditions are given in Table 1. retention was studied.Interferences on isotopes 65Cu, 64Zn, The ICP-MS system was optimized by continuous nebulization 67Zn and 53Cr were minimized using this approach. The of a multi-element solution (10 mg l-1 B, Mg, Co, In and Pb) analytical range was from 0.2 to 1.0% S for the vegetable covering the analytical range of interest. samples analyzed. The detection limit was 0.02% S (dry basis). The main components of the flow system are a Minipuls 3 Results from reference materials (NIST SRM 1573a Tomato peristaltic pump (Gilson, Worthington, OH, USA), a Model Leaves and ORSTOM palmier and pommier golden) were in 352 time-controlled injector (Micronal, Sa�o Paulo, S.P., Brazil) agreement with certified values at the 95% confidence level.described earlier6 and a resin column connecting the injector and the nebulizer. All the system connections were made of Keywords: Inductively coupled plasma mass spectrometry; polyethylene tubing (0.8 mm id).The tube connecting the resin anion exchange; flow system; matrix removal; plant analysis column to the nebulizer was 5 cm long. The chloride form of the anionic resin AG-1-X8 from Bio- Sulfur is a macronutrient of plants occurring at concentrations Rad Labs. (Richmond, CA, USA) (200–400 mesh) was used. from 0.06 to 1.5% of the dry matter.1 The content of sulfate The resin column was built in a Perspex block by drilling a in samples is a potential interferent in the determination of hole 5 mm wide and 20 mm long.The wetted resin was loaded some elements by ICP-MS, mainly due to the spectral overlap inside the column by pushing with a syringe and both ends from sulfur polyatomic species such as SO2+/ S2+ on 64Zn, were plugged by disks of porous polyethylene fixed with OSO 2H+ on 65Cu and SO+ on 50Cr. Chlorine is also present in ring seals into Perspex disks. Small holes on the disks were plant material at levels up to 6% of the dry matter,1 producing connected to the polyethylene tubing as described in a previous species such as ClOH+, ClO+ and ClO2+, which interfere paper.7 The resin was converted into the nitrate form by with the determination of 52Cr, 53Cr and 67Zn, respectively.pumping 3 ml min-1 of 1.0 M nitric acid through the column Interferences due to Cl and S have been alleviated in a for 1 h. number of ways. For example, Cl interference has been partially Multi-element standard solutions were prepared from SPEX reduced by using a high flow rate of the nebulization gas, but (Metuchen, NJ, USA) stock standard solutions (Multi-element to remove ClO+ and ClO2+interference eectively the addition Plasma Standard ICP-MS 2 and ICP-MS 4).Water purified of 8%nitrogen to the aerosol sample carrier gas was necessary.2 with a Milli-Q system (Millipore, Bedford, MA, USA) was A flow system for on-line removal of interferences for the used throughout.Working standard solutions containing 2.5, determination of V, Mn, Cu, Zn, Cd and Pb in biological 5.0, 25.0, 50.0 and 100.0 mg l-1 of B, Al, Cr, Mn, Fe, Cu, Zn samples by ICP-MS was described by Ebdon et al.3 An and Pb in 0.05 M HClO4 and standard solutions containing iminodiacetate-based resin was used to chelate the cations to 0.00, 20.0, 40.0 and 60.0 mg l-1 of sulfate were prepared. Nitric be separated from the matrix. Later, the same flow system was acid purified by sub-boiling distillation, perchloric and sulfuric proposed using a column with activated alumina (acidic form) acid (Suprapur) were obtained from Merck (Darmstadt, for the separation of anionic forms of As, Cr, Se and V from the biological digests.The determination was achieved after matrix elimination by ICP-MS.4 However, these methods were Table 1 ICP-MS operating conditions restricted to certain elements, thus reducing the multi-element capabilities of ICP-MS. An approach based on the retention Forward power 1350 W of anionic species by a strongly basic anion-exchange resin in Argon gas flow rates— Coolant 13 l min-1 the NO3- form while the analytes were eluted with diluted Auxiliary 0.55 l min-1 nitric acid was described earlier.5 This method allowed the Nebulizer 0.85 l min-1 simultaneous multi-element determination of V, Cr, Cu, Zn, Scan measurements— As and Se in biological and environmental samples.The main Detection mode PC advantage was that analytes in the anionic and cationic forms Range m/z 6–210 were separated from Cl and S.However the separation was Channels per u 19 Dwell time 320 ms performed in columns placed o-line. Sweep 0.5 s In this paper, a flow system with a column of anion-exchange Single ion measurements— resin (AG1-X8) placed on-line with the ICP-MS system is Detection mode PC proposed. The S and Cl species in the plant digests were Dwell time per channel 10.24 ms retained by the resin while the analytes followed to the ICP-MS Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 (671–674) 671Germany) and 30% hydrogen peroxide (puriss) from Fluka when increasing concentrations of perchloric acid were added to a 100 mg l-1 sulfate solution. The results (Fig. 2) indicate (Buchs, Switzerland). no losses of sulfate on adding perchloric acid up to 0.05 M. Ions at m/z 34 and 48 were monitored. The higher signal at Sample Preparation m/z 34 was probably due to interference from O2+.All signals increased after addition of 0.1 M perchloric acid, indicating Plant samples of dierent origins and reference material from that sulfate passed through the resin. After collecting each National Institute of Standard and Technology (Gaithersburg, euent, the anions were eluted with 1.0 M nitric acid, producing MD, USA) (NIST SRM 1573a Tomato Leaves) and two others transient signals at m/z 48. The smallest transient peak corre- from the Oce de la Recherche Scientifique et Technique sponded to the elution of sulfate after addition of 0.1 M Outre-Mer (ORSTOM) (palmier and pommier golden)8 were perchloric acid, confirming the loss of more than 90% of digested using nitric acid, hydrogen peroxide and perchloric sulfate. The retention of perchlorate from plant digests in acid, following a procedure similar to that described earlier.9 AG-1-X8 was reported earlier.10 The dilution factor was 1000 to leave the solutions in 0.025 M The capability of the resin to remove anions from plant perchloric acid media.Blanks were prepared together with digests was proved by adding increasing concentrations of the samples. sulfate to the NIST-1573a reference sample solution. The results at m/z 64 and 65 for the determination of Zn and Cu, Flow System respectively, are shown in Fig. 3 At m/z 64 the calculated concentration of Zn in the sample was in agreement with the The flow system proposed to perform the removal of anions certified value for Zn of 30.9±0.7 mg g-1, indicating no inter- is presented in Fig. 1 with the injector at the elution (top) and ference from the 0.96% m/m of S in the sample (approximately sampling (bottom) positions. The top design represents the 30 mg l-1 sulfate in solution). Interference occurred with sulfate sample S pumped at 1.4 ml min-1, filling the tubes and going concentrations higher than 100 mg l-1 and increased lto the discarded waste D, while the eluent solution passed up to 1.0 g l-1.The results at m/z 65 showed similar behavior through the column towards the ICP-MS system. On moving to 64Zn. The initial value for Cu was in agreement with the the central part of the injector unit, the configuration shown certified value of 4.70±0.14 mg g-1. On adding sulfate to the at the bottom design is obtained, where the sample flowed sample up to 1.0 g l-1, a linear signal increase up to 8 mg g-1 through the column to the spectrometer for 2 min.of Cu was observed, probably due to 32SO2H+ interference. Simultaneously with the injector movement, the acquisition In both instances the interference was totally removed by using process with a 40 s uptake and two replicates of 25 s each by the proposed procedure. These results showed the eciency of scanning the analyte masses in the ICP-MS system was the resin in removing sulfate at concentrations more than 20 implemented. After the sample, water (W) was passed through times the maximum level found in plant materials.At m/z 67 the column at 2.5 ml min-1 to remove sample residues from the overlap due to the 35ClO2+ and 34SO2H+ species produced the resin. After finishing the multi-element data acquisition process, the injector was moved to the elution position and 1.0 M nitric acid at E was passed through the column at a flow rate of 1.4 ml min-1 to elute the anions. The sulfur species SO+ was determined at m/z 48 using the single ion mode with a 22 s uptake and 8 s of data measurement.At this position the sample could be changed. RESULTS AND DISCUSSION The eciency of sulfate removal in the presence of other anions was determined by analyzing the euents leaving the column Fig. 2 Losses of sulfate from the euents on adding perchloric acid. The average of three values monitored at m/z 34 and 48 produced the upper and lower curves, respectively. The values at m/z 34 were 10 times higher.The sulfate concentration was 100 mg l-1. Fig. 1 Flow diagram of the proposed system. The top and bottom diagrams represent the elution and sampling steps, respectively. The Fig. 3 Eect of sulfate addition to the NIST SRM 1573a sample to assembled three-rectangular design represents the injector port with moving central part. The sample (S) or water (W) flows into the determine Zn at m/z 64 (upper curves) and Cu at m/z 65 (lower curves). [C] indicates the analyte concentration.The dotted and solid lines injector. The peristaltic pump action is represented by the black arrows. E corresponds to the eluent solution, C indicates the resin were obtained with and without the use of the anion-exchange resin, respectively. Means and deviations were obtained with n=3. column and D the discarded waste. 672 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12Table 2 Results of routine analysis and certified or reported values (mg g-1) NIST SRM 1573a Palmier Pommier golden Tomato Leaves (ORSTOM) (ORSTOM) Element Determined* Certified Determined* Reported Determined* Reported 10B — — 12.04±0.13 14.51±2.5 31.83±0.49 28.35±3.58 27Al — — 234±19 150 195±2 200 50Cr 10.50±0.49 1.99† — — — — 52Cr 1.86±0.07 1.99† — — — — 53Cr 3.59±0.55 1.99† — — — — 55Mn 245±10 246±8 650±35 630 44±1 47 56Fe 318±30 368±7 249±19 200 325±0.4 311 63Cu 4.30±0.23 4.70±0.14 8.31±0.20 8.70 6.96±0.06 8.40 65Cu 4.44±0.34 4.70±0.14 8.60±0.28 8.70 7.05±0.33 8.40 64Zn 29.3±1.0 30.9±0.7 22.8±0.1 23.0 29.9±3.1 30.0 66Zn 28.9±0.7 30.9±0.7 23.1±0.7 23.0 30.6±1.1 30.0 67Zn 32.2±1.66 30.9±0.7 — — — — 68Zn 29.9±0.7 30.9±0.7 21.9±1.1 23.0 31.5±2.1 30.0 114Cd 1.28±0.16 1.52±0.04 — — — — 208Pb — — 2.5±0.3 3.0 11.6±0.8 16.0 48S (%) 0.95±0.03 0.96† 0.23±0.02 0.21±0.03 0.23±0.01 0.34±0.05 Cl (%) — 0.66† — 0.611±0.015 — 0.100±0.007 * Mean±standard deviation (n=3).† Reported values. an initial value of 36.7 mg g-1 of Zn for the reference sample, which was also corrected using the anion-exchange resin.This was not observed at m/z 50 where the interference eect is produced by SO+, ArN+ and ArC+ ions and isobaric species of Ti and V. Even when the anion-exchange resin produced a considerable decrease in the m/z 50 signals, the value obtained for Cr was not in agreement with the reported value of 1.99 mg g-1. At m/z 53 the interferences of ClO+ and ArNH+ were observed in the results presented in Table 2. The initial Cr concentration of 14.9 mg g-1 was reduced to 3.5 mg g-1 when using the resin, indicating good eciency of chloride removal.The reported content of chlorine in this sample was 0.66% m/m. Sulfate from standard solutions was separated and eluted Fig. 5 Calibration curves for sulfate obtained with standard solutions from the anion-exchange resin producing the transient peaks (0.0, 20.0, 40.0 and 60.0 mg l-1 ) . The curves correspond to %, direct at m/z 48 shown in Fig. 4.Calibration curves obtained by nebulization and using the anion-exchange resin at dierent uptake direct nebulization and using the proposed approach showed and measurement times: &, 0 and 60 s; $, 15 and 30 s; and 2, 22 good linearity, as can be observed in Fig. 5. Sulfate determi- and 8 s. Points indicate the means of three values. nation was performed by measuring the dierent areas under the transient peak depicted in Fig. 4. The signal acquisition was programmed using uptake and measurement times of 0 results observed with direct nebulization showed poor sensi- and 60, 15 and 30, and 22 and 8 s, respectively.Sensitivity was tivity. Using the anion-exchange resin, the slope of the curves enhanced for measurements made at the peak maximum. The increased by a factor of 10 when the acquisition procedure was well adjusted to the transient signal. A sample volume of 2.8 ml was consumed per experiment. A limit of detection of 0.6 mg l-1 of sulfate was calculated using the 3s criterion.Routine analysis using the proposed procedure allowed the determination of the analytes in plant samples, as shown in Table 2. The paired t-test at the 95% confidence level was applied to compare the two sets of results (n=32) corresponding to the determined and reported values. The calculated t was 0.85, which was lower than the tabulated value (t=2.042), demonstrating good assessment of accuracy. CONCLUSIONS The flow scheme presented allowed the sequential determination of several analytes free from anion interferences and the total sulfur content in digested plant solutions. Interferences from 64Zn, 67Zn and 65Cu were completely removed, but not those from 50Cr and 53Cr.The proposed approach was very ecient for removing Fig. 4 Transient peaks corresponding to sulfate eluted from the resin. anions commonly found in plant digests. The strongly basic From the baseline to the signal peaks the sulfate concentrations were 0.0, 20, 40 and 60 mg l-1.anion-exchange resin used was robust in removing up to 30% Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12 6735 Goossens J., and Dams R., J. Anal. At. Spectrom., 1992, 7, 1167. m/m S, indicating the possibility of eliminating S in the analysis 6 Bergamin Fo., H., Reis, B. F., Jacintho, A. O., and Zagatto, of materials with high sulfur contents by ICP-MS. E. A. G., Anal. Chim. Acta, 1980, 117, 81. 7 Reis, B. F., Gine�, M. F., Santos Filha, M. M., and Baccan, N., Grateful acknowledgements are made to FAPESP (Fundac�a� o J. Braz. Chem. Soc., 1992, 3, 80. de Amparo a` Pesquisa do Estado de Sa�o Paulo) for financial 8 Pinta, M., Van Schouwenberg, A., Bonvalet, M., Lachica, M., and support and to Dr. Victor Vitorello for language improvement. Herman, P., in Proceedings of the 4th International Colloquium on the Cf Plant Nutrition, Rijksuniversiteit, Ghent, 1976, vol. II, p. 41. REFERENCES 9 Jarvis, I., in Handbook of Inductively Coupled Plasma Mass 1 Martin Prevel, P., Plant Analysis as a Guide to the Nutrient Spectrometry, ed. Jarvis, K. M., Gray, A. L., and Houk, R. S., Requirements of T emperate and T ropical Crops, Lavoisier, New Blackie, Glasgow, 1992. York, 1987. 10 Gomes Neto, J. A., Bergamin Fo., H., Sartini, R. P., and Zagatto, 2 Laborda, F., Baxter, M. J., Crews, H. M., and Dennis, J., J. Anal. E. A. G., Anal. Chim. Acta, 1995, 306, 343. At. Spectrom., 1994, 9, 727. 3 Ebdon, L., Fisher A. S., Worsfold, P. J., Crews, H., and Baxter, Paper 7/00444C M., J. Anal. At. Spectrom., 1993, 8, 691. Received January 20, 1997 4 Ebdon, L., Fisher A. S., andWorsfold, P. J., J. Anal. At. Spectrom., 1994, 9, 611. AcceptedMarch 18, 1997 674 Journal of Analytical Atomic Spectrometry, June 1997, Vol. 12

ISSN:0267-9477    

DOI:10.1039/a700444c

出版商:RSC    

年代:1997

数据来源: RSC