摘要:

Micro and Surface Analysis in Art and Archaeology† Plenary Lecture F. ADAMS*, A. ADRIAENS, A. AERTS, I. DE RAEDT, K. JANSSENS AND O. SCHALM University of Antwerp, Department of Chemistry, Universiteitsplein 1, B-2610 Antwerpen, Belgium A variety of instrumental analytical techniques can be applied be readily solved. Archaeological applications require an important arsenal of instrumental tools for the often very to the physical and chemical examination of works of art and archaeology.In this paper, a few examples are discussed of the complex materials to be studied. Archaeology and analysis of art objects were added to this list of application areas of application of micro-analytical chemistry in this interdisciplinary field. The following subjects from the MiTAC fairly recently. Indeed, the laboratory had acquired some interest in this particular applied research area at the experience of our laboratory, in collaboration with several specialized institutes, were selected: Early Bronze Age ceramic beginning of the 1980s but it was only in the last few years that a more systematic development of this particular research crucibles, residues and powders from Go�ltepe, South Central Turkey, have been analysed using surface analytical techniques endeavour was beginning to be explored.This paper describes a number of different recent appli- to investigate potential evidence of tin smelting. The study indicates that the crucibles were used for processing of tin and cations of trace and microscopical analysis in art and archaeology.gives clear evidence of a local tin industry. Roman glass from a collection of objects discovered in Qumra�n near the Dead Sea was used to study the corrosion of glass objects in a EARLY BRONZE AGE ARTEFACTS FROM particularly stable environment over a period of nearly 2000 TURKEY: AN APPLICATION OF SURFACE years. The corrosion of a series of glass-in-lead windows from ANALYSIS IN ANCIENT METALLURGY St.Michael and St. Goedele’s Cathedral, Brussels, was studied This section demonstrates the use of micro and surface analysis using electron probe microanalysis and micro X-ray methods for the study and analysis of ancient metallurgical fluorescence. New views can be formulated on the corrosion artefacts. The project that will be considered runs in collab- mechanism, which appears to be a complex multiphase process oration with The Oriental Institute in Chicago. It concerns under the influence of atmospheric pollution.A few preliminary the excavation of a mining complex containing tin ore and results are discussed for the analysis of glass paintings, in an associated production/habitation site in South Central particular carnation red glass paints. Turkey. Both sites have been dated to the Early Bronze Age Keywords: Microanalytical techniques ; surface analysis; art (3000–2200 BC). objects; archaeology The analysed samples originate from the production site in Go�ltepe.Excavations have disclosed pit house structures, which were filled with debris of a similar nature: grindstone Nowadays, a variety of analytical techniques can be applied to tools, ceramic crucibles, powdered materials, ore fragments the physical and chemical examination of works of art and and charcoal. It is presumed that these materials are related archaeological objects. In this paper, a few examples of the to roasting and smelting activities of cassiterite from the mine application of a number of trace and microscopical techniques in Kestel.1 The mine is the first tin mine to be located in (X-ray fluorescence analysis, electron probe microanalysis and Turkey.Until the discovery, it was solely believed that the tin secondary ion microscopy) will be discussed in order to illusused for producing bronze in Anatolia during the Early Bronze trate the possibilities of microanalytical chemistry in this inter- Age was imported from distant places, such as Afghanistan.disciplinary field. All thetechniques evolvedfrom the laboratory The tin in the mine, however, is not obvious as the cassiterite experience acquired in the Micro and Trace Analysis Centre is only low grade ore, presumably a remainder of richer (MiTAC) of the University of Antwerp. The main goal of this deposits. This raises the question of whether the mine was research group involves development of methods for trace and once a productive metallurgical site.microchemical analysis and application of these methods in In order to expand the archaeological evidence for a tin environmental chemistry and materials science. industry, the present study focuses on the analysis of a set of In basic analytical chemistry, MiTAC has access to a number residues, crucible fragments and powders from Go�ltepe. Using of different analytical methodologies in trace analysis and several analyticaltechniques such as electron microprobe X-ray microscopical analysis.These are summarized in Table 1. In analytical (EPXMA), secondary ion mass spectrometry (SIMS) addition, the laboratory has access to a number of other and X-ray fluorescence (XRF), our intention was to examine methodologies, in other laboratories, for its applied research. whether any remains of tin smelting activities could be found. We refer to other publications for the details on all these This work is related to other investigations, in which the issue techniques.is approached from a different angle.2,3 In general, hardly any applied analysis depends on the application of a single analytical tool. Very often, it is by the application of a wide range of techniques, and the synergy of Residues answers that they provide, that complex problems can often One of the crucible fragments excavated contains on its inner surface a distinct visible layer of shiny accretion (a few square † Presented at the Eighth Biennial National Atomic Spectroscopy Symposium (BNASS), Norwich, UK, July 17–19, 1996.centimetres). The matrix of this residue, which basically Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12 (257–265) 257Table 1 Instrumental techniques used at MiTAC Instrumental trace analysis— Spark source (Jeol JM 1-BM2, 1974) and glow discharge (VG 9000, 1989) mass spectrometry Energy-dispersive (Tracor 5000, 1985) and glancing angle wavelength dispersive X-ray fluorescence analysis (Philips prototype, 1995) Instrumental neutron activation analysis ICP (Perkin-Elmer 6500 XR, 1982) and microwave-induced plasma optical emission spectrometry Atomic absorption spectrometry (Perkin-Elmer 5100, 1988) Microscopical analysis— Scanning electron microscopy and electron probe microanalysis (Jeol 733, 1983, and Jeol JSM 6300, 1993) Laser microprobe mass spectrometry (FTMS) (Bruker Spectrospin, 1992) Dynamic and (soon) static secondary ion microprobe analysis (Cameca 4F, 1985, and TOF-SIMS IV) Microscopical X-ray fluorescence (laboratory source based on Siemens rotating anode tube) Micro X-ray fluorescence with synchrotron radiation sources (at different sources) Scanning transmission electron microscopy (Jeol 1200, 1988) and electron energy loss spectrometry (Zeiss ESI 902) FT infrared microscopy, scanning Auger microscopy, X-ray photoelectron spectrometry, particle induced X-ray emission (through collaboration) consists of an iron–calcium–tin–alumina silicate (FeO–CaO– ably due to the prolonged burial of the material.This layer does not cover the entire inner surface but is observed at SnO2–Al2O3–SiO2), contains two types of crystals (Fig. 1). The equiaxed crystals are iron–tin oxides, with an average size localized areas on the surface with varying thickness. Between the CaCO3 layer and the ceramic material a layer of silicate of roughly 10 mm2. Some of these crystals contain small inclusions (a few micrometres in diameter), which appear to material exists, of which its matrix contains 2–3% tin oxides.This layer can clearlybe observed in the back-scattered electron be a different type of iron–tin oxide phase, with a higher concentration of tin. The longitudinal crystals are composed image of a crucible cross-section (Fig. 2), where it is represented bht band several micrometres thick. This material of SnO2 . These crystals are 0.5–2 mm wide and can be up to 50 mm long.The two types of crystals are primarily slag contains small inclusions, which appear to consist of a different silicate phase with up to 65% of tin oxides. minerals originating from the ore and gangue, which have been trapped in the matrix. The results indicate that the residue In addition, line scans with SIMS were performed across the cross-section of the crucible fragment. In these experiments, corresponds in composition to a metallic tin slag.4 These conclusions were verified by comparing the material in detail the bombarding ion beam is moved across the sample in distinct steps of 10 mm, covering roughly 800 mm in total.At with a typical tin slag from Cornwall.5 each ion beam position on the crucible, compositional data of the sample are acquired. Fig. 3 shows the results, in which the Ceramics signals of Ca+, Sn+ and SiO+ are plotted as a function of distance. The left-hand side of Fig. 3 portrays the edge of the Most of the samples on the site, however, do not show any sample with the epoxy resin, the right-hand side the bulk of visible indication of tin residues.Many crucible fragments have the crucible material. Between, the Ca signal can be observed been found which possess a smooth grey interior surface, of originating from the CaCO3 layer. The SiO signal is present which it is nevertheless believed that they were used for in the ceramic material, but is also observed in the CaCO3 roasting or smelting cassiterite. layer.Again, a tin peak is clearly present at the interface of In order to analyse these samples, the crucible sherds were cross-sectioned and embedded in epoxy resin. Electron microprobe analyses have shown that the body of these ceramic materials mainly consists of alumina silicates with fragments of quartz and iron oxides. The inner surface of some crucible fragments contains an accretion layer of CaCO3 of roughly 200–500 mm in thickness, the presence of which is most prob- Fig. 2 Back-scattered electron image of a crucible cross-section. The tin-bearing layer is represented by the bright band a few micrometres thick across the image. The dark grey material beneath the layer containing tin is the ceramic material (alumina silicates). The bright spots within the ceramic are iron oxides. The darker areas in the ceramic material are quartz fragments. The medium grey amorphous Fig. 1 Back-scattered electron image of accretion material.Tin oxide structure above the tin-bearing layer is the CaCO3 layer. The black area on the upper side of the image is epoxy resin. Scale bar, 500 mm. crystals (longitudinal) and iron–tin crystals (equiaxial) are visible in an alumina silicate matrix. Magnification, ×390. Scale bar, 100 mm. Magnification, ×65. 258 Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12containing materials were selectively transported to Go�ltepe for ore dressing and smelting.3 It is further believed that there is a correlation between the concentration of tin in the powders and their site context.This still needs to be investigated in detail, but there are already a few significant indications. Sample MRN 2298 for instance was recovered from a garbage dump, a place where one would expect samples to be the least rich in tin. From Table 2, it can be observed that the tin concentration of this sample is significantly lower than that of the other samples.Samples MRN 3834, MRN 3858 and MRN 3830 are three samples which have been located in the same pit house structure, but in different ceramic pots next to each other, presumably meant for storage. In addition, the tin Fig. 3 SIMS line scan of crucible cross-section. The measurement content varies between the three samples, indicating that that starts on the epoxy resin (left) and is performed towards the body of samples may represent different stages from an ore concen- the ceramic, crossing the CaCO3 and the tin-bearing layer.The tration process. Sample MRN 4573 contains 2.93% Sn, which absolute signals of Ca+, SiO+ and Sn+ are plotted as a function of is, to date, the highest tin concentration encountered in pow- distance. The left-hand side of the figure portrays the edge of the dered material from the sites. Sample MRN 3697 has been sample, the right hand side the bulk of the crucible material. located next to a hearth in a pit house.The material is strongly magnetic, indicating the presence of magnetite. Experiments the ceramic material and the CaCO3 layer and, therefore, at have shown that whenever hematite (a-Fe2O3) is heated above the inner surface of the crucible fragment. a temperature of 550 °C, it is partially converted to magnetite (Fe3O4) through a maghemite phase (c-Fe2O3).3 It is, therefore, clear that the magnetically attracted material found in Go�ltepe Powders is a by-product of a heating process. This is confirmed by electron microscopy analysis of MRN The pit house structures in Go�ltepe have also yielded dense 2836, which is also highly magnetic.The sample is dark brown concentrations of variously coloured ground ore powders, to black, again indicating the presence of magnetite. The main often in excess of 10 kg. The colours range from purple/ purpose of the analyses using electron microscopy is to charac- burgundy, pink to beige and black and are readily distinguished terize the particles containing tin.As the cassiterite is located from the surrounding soil. The powders were analysed mineralin a hematite-rich environment, iron is likely to be reduced to ogically and found to be the same as material from the Kestel some extent, which will give rise to a range of iron/tin mine.6 The site of Go�ltepe contains no mineral bearing veins compounds, called hardhead. Particles composed of Fe and and analysis of the host rock shows no resemblance to the Sn therefore form an additional pointer to heating processes.archaeological material. The latter thus establishes the idea In addition to SnO2 particles, MRN 2836 shows the presence that the materials were taken to the site from their source a Fe–Sn particles (Fig. 4), indicating that the sample is very few kilometres away.6 likely to be a product of roasting or smelting processes. The powders have been located in ceramic cups, apparently meant for storage, as deposits on the floor in the neighbourhood of hearths and pots, and in garbage dumps.It is assumed Conclusions that the powdered materials are in turn related to the tin processing that was occurring at the production site. The The present analyses have revealed the presence of tin at the hypothesis is that some of these powders may be unprocessed interior surface of the crucible fragments, indicating that these ground ore material, while others may be the residue from an ceramics were probably used for the processing of tin-bearing ore concentration process (the enriched portion having been materials.The fact that the composition of the crucible residue extracted) and a third group may be waste materials such as corresponds to a metallic tin slag gives additional evidence for metallic slags. potential tin smelting activities. In addition, some of the The intention, therefore, was to perform a further characterization of the powders using XRF and electron microscopy. Bulk analyses were performed using XRF to obtain a general idea of the composition of the samples.Table 2 shows the XRF results of the main elements in nine different powders and show a significant difference in composition. The amount of tin ranges between 0.3 and 2.9%, which is much higher than for similar material found at the Kestel mine (0.07% in hematite samples).3 The latter suggests that only high tin- Table 2 XRF results of Go�ltepe powders (concentration in % m/m) Sample Ca Si Fe As Sn MRN 2298 9.34 20.45 24.31 0.43 0.28 MRN 2836 2.10 2.55 54.40 0.09 0.85 MRN 3032 9.70 22.77 16.60 0.05 0.34 MRN 3697 7.30 15.04 33.10 0.22 0.64 MRN 3738 4.90 17.84 28.90 0.15 0.70 MRN 3830 7.30 11.02 41.00 0.33 1.18 MRN 3834 15.80 22.06 6.90 0.66 0.43 MRN 3858 7.60 18.15 34.30 0.10 0.85 Fig. 4 Back-scattered electro a powder particle composed MRN 4573 4.20 9.10 21.80 0.08 2.93 of iron and tin. Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12 259powders show sufficient evidence of being heated and confirm the hypothesis that they are, therefore, metallurgical waste materials. In conclusion, all data provide positive evidence for tin smelting activities at Go�ltepe. The results of the present study are directly related to investigations involving the search for ancient tin production in Turkey and metal trade during the Early Bronze Age in the Near East. APPLICATIONS OF MICROANALYTICAL TECHNIQUES TO THE ANALYSIS OF HISTORICAL GLASS AND JADE Glass, a material of interest to archaeologists, is rarely found in its original state.It has usually been altered by different external factors over a long period of time. Its state of preservation also depends on its own nature. Roman and Fig. 5 Hierarchical cluster analysis dendrogram obtained on the basis medieval glass, for example, differ profoundly with respect to of both major and trace element composition.their chemical composition. The chemical durability of medieval glass, for instance, decreases due to the use of potash as a raw material instead of soda, which was used in ancient times.7 Table 3 Average major element composition of two groups of Qumra�n glass samples as determined by SEM (all values in % m/m)* Determination of the composition of various glass samples using EPXMA and micro-synchrotron radiation induced X-ray Group I Group II fluorescence (m-SRXRF) can reveal information on the prov- Na2O 16.5±0.5 17.0±0.5 enance and history of objects or windows and/or the site in MgO 0.2±0.1 0.1±0.1 which they were found.In addition, the correlation with the Al2O3 2.5±0.1 2.4±0.3 present day appearance can be studied. In this section, two SiO2 69.6±0.7 71.6±0.8 applications of trace analysis and microanalysis to the charac- P2O5 0.1±0.04 <0.1 terization of ancient glass will be described: a Roman glass SO3 0.1±0.1 0.2±0.1 collection from Khirbet Qumra�n and medieval glass windows Cl 0.8±0.1 1.1±0.1 K2O 0.8±0.1 0.6±0.1 from St.Michael and St. Goedele’s Cathedral in Brussels. CaO 8.4±0.5 5.9±0.8 TiO2 0.1±0.02 <0.1 MnO 0.4±0.1 0.8±0.3 Investigation of Glass Objects from Qumra�n Fe2O3 0.5±0.1 0.4±0.1 In addition to an extensive series of terra-cotta oil lamps and a collection of stone objects, a group of 90 fragments of various * Significant differences between groups I and II are underlined.glass objects was also recovered during the excavations of Khirbet Qumra�n. This archaeological site consists of a collec- Table 4 Average trace element composition of two groups of Qumra�n tion of ruins of various buildings situated at the eastern edge glass samples as determined by m-SRXRF (all values in ppm by mass)* of the Judean desert, on the north-western shore of the Dead Sea in Israel. It used to be a Roman settlement in ancient Group I Group II times. The glass objects are assumed to date back to the period Cr2O3 22±5 30±20 4–68 AD.They show corrosion phenomena on their surface NiO 8±1 9±4 due to contact with water in the soil for about 1900 years.8 CuO 143±36 13±9 ZnO 32±21 19±7 Rb2O 12±2 12±2 Analysis of the bulk glass SrO 595±99 540±38 Y2O3 9±2 7±2 EPXMA and microscopical XRF were applied to the analysis ZrO2 86±13 71±9 of the bulk glass. All quantitative results of the Qumra�n bulk Mo2O3 3±2 2±2 glass were kept together and hierarchical cluster analysis was SnO2 117±33 52±29 Sb2O5 281±127 17±26 employed to emphasize the structure in the data.The resulting BaO 165±56 129±56 dendrogram, obtained on the basis of both major and trace Ta2O5 18±3 3±3 element composition (only 36 samples), is shown in Fig. 5. A PbO 128±26 16±14 clear distinction between a large and a small group can be observed. There seems to be no straightforward correlation * Significant differences between groups I and II are underlined.between the structure in the dendrogram and the typology of the objects except for the fact that the small group is exclusively composed of ointment vessels and goblets. the concentrations of CuO, PbO, SnO2 and Sb2O3. In addition, the trace element signature of each sample within a group is The difference between these two groups becomes much clearer when the average composition within each group is almost identical. This strongly indicates that all objects in one group originate from the same batch of bulk glass and even calculated. In Table 3 the major composition is listed together with the corresponding standard deviation found in each the two groups seem to be closely related. Regarding the question of provenance of the objects, the group.The glass is clearly natron-based. Natron is a deposit from the Wadi Natru�n (Egypt) which was extensively used in composition information discussed above suggests that almost all objects found at Qumra�n were either made locally from the ancient times for glass manufacture.7 In the major element composition only the CaO abundance exhibits a significant same batch of glass or imported ready-made in large numbers from elsewhere.In either case there seems to have been a difference between the two groups. When the trace element composition is considered (Table 4), relatively large demand for glass vessels of various types, which appears to be consistent with the hypothesis that at this place a much clearer distinction can be made, more specifically in 260 Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12some industrial activity in the field of perfume manufacture Investigation of Glass Fragments from Thirteenth Century Oculi from Brussels’ Cathedral was located, as suggested in ref. 8. During the restoration of the choir of St. Michael and St. Goedele’s Cathedral in Brussels some round gaps were found Corrosion behaviour in the backwall of the triforium. Many of these gaps contain Secondary electron images of polished cross-sections of the their original glass windows.These glass windows, initially Qumra�n samples show three well defined regions in the glass: used for the lighting of the choir, probably date back to 1273, a central part of unaffected glass, a leached layer which has the year when the gallery of the triforium was built. However, formed on the inside of the original glass pane and a precipi- they lost their original function during the sixteenth century, tation crust on top of the surface.EPXMA maps taken across when the roof of the gallery was heightened by the building of such cross-sections show marked changes in composition two chapels.10–12 All glass panels are severely weathered on between these different areas. both sides. The lightly pitted glass surface is covered with an Overall, the leached layer is depleted in elements such as opaque crust. A thorough study of the chemical composition Na and Ca whereas Al and K are more concentrated in that of the glass and corrosion products and of the durability is part, the Si-level is almost the same throughout the whole necessary to consider further treatment of the glass windows.glass and Mg rises gradually towards the surface. The increase in K in the leached layer is unusual but may be related to the high concentration of this element in the soil the glass had Analysis of the bulk glass been buried in.The crust is very rich in Ca, which might point to the presence of CaCO3 or Ca(HCO3)2. Two glass windows, IVB and VIA, each consisting of multiple panels, were analysed. Five panels of each glass window were Inside the leached layer, the precipitation of an almost black substance rich in Mn, assumed to be MnO2, is sometimes sampled. EPXMA measurements on polished cross-sections of the 10 glass samples reveal that all glass samples are potash observed.9 In order to obtain a better intion of various elements between the glass and the surrounding glasses with compositions that are broadly similar.In Table 5 the average chemical composition of five samples taken from soil, trace element distributions of such an area were also observed. A number of m-PIXE (micro-particle induced X-ray window IVB and of five samples taken from window VIA is summarized. However, several factors point to the use of emission) maps are shown in Fig. 6. Broad alternate bands enriched in different elements can be observed. Since the order recycled glass: the use of large pieces of uncut glass covered with lead frames, the fact that many panels have a different and relative thickness of these sub-layers vary considerably among the different samples, it is still very hard to understand shade of green and the presence of panels with unfinished grisaille paintings. Principal component analysis (PCA) was all the processes taking place during leaching.Fig. 6 m-PIXE maps from a 1200×1200 mm region of the corrosion layer with a clear MnO2 deposit on one of the Qumra�n samples. Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12 261Table 5 Average major and trace element composition of the bulk glass belonging to two glass windows, IVB and VIA, from Brussels’ Cathedral as determined with SEM and m-SRXRF Glass window 1 Glass window 2 IVB1 IVB2 IVB3 IVB4 IVB5 VIA1 VIA2 VIA3 VIA4 VIA5 Major element composition (% m/m): MgO 5.46 4.98 4.57 4.40 3.94 3.90 3.85 4.28 3.55 3.49 Al2O3 4.05 3.42 3.82 3.63 3.93 3.98 4.01 3.04 1.20 1.55 SiO2 47.25 46.20 46.47 46.29 47.42 47.70 48.33 48.12 52.19 47.96 P2O5 2.05 2.23 1.94 1.87 1.17 1.73 1.18 2.52 4.94 1.57 K2O 11.57 12.21 11.48 11.69 12.41 11.95 12.23 14.62 13.78 19.21 CaO 27.49 28.89 29.31 29.73 28.39 28.16 27.77 25.67 23.44 23.70 MnO 1.07 1.14 1.18 1.21 1.48 1.24 1.43 0.90 0.32 1.92 Fe2O3 0.54 0.44 0.58 0.54 0.67 0.60 0.64 0.54 0.54 0.46 BaO 0.52 0.50 0.62 0.65 0.58 0.74 0.57 0.31 0.04 0.14 T race element composition (ppm): TiO2 1503 1336 1670 1837 1670 1837 1503 1336 1002 1169 NiO 42 35 36 36 39 33 43 29 8 26 CuO 98 218 96 97 100 101 126 211 87 163 ZnO 123 158 163 162 163 127 167 197 270 388 Ga2O3 7 5 5 5 7 5 8 4 — 4 Rb2O 291 318 286 291 368 271 501 438 132 427 SrO 2478 1888 2714 2832 2360 2360 3422 1888 1062 1652 Y2O3 15 6.35 — 12 7 10 13 10 7 14 ZrO2 265 128 234 257 199 240 362 163 100 160 PbO 32 100 41 24 80 36 19 100 23 383 performed to detect glasses of different origin.Fig. 7 shows a observed with SEM (scanning electron microscopy). The crust is rather powdery and can be easily scraped off. Fig. 8 Ca–Mn plot since Ca and Mn are highly correlated with PC1 shows a secondary electron image of a crater on the glass and PC2. All other major elements show a similar correlation. surface. The crystalline products with a plate-like structure in From these plots, samples VIA5, VIA4 and VIA3 are disand around the crater can be characterized as gypsum tinguished as outliers.The difference in major element composi- (CaSO4 2H2O).13 EPXMA studies of the weathering crust tion confirms the statement of the use of recycled glass.The reveal that the products mainly consist of calcium and sulfur. seven other panels have the same composition and are probably Far-IR measurements of the white powder present on the made from the same glass batch.The placement of the windows surface and in the craters confirm this statement. at a significant height, the cost of glass in the thirteenth century The actual condition of medieval glass depends on the and the function of the oculi as sealers, allowing light to get chemical composition of the bulk glass. Glasses with high through, explain why recycled glass was used. potash and low silica content will have a severely crusted surface while a higher silica content will lead to corrosion by Description of the weathered surface pitting.Nevertheless, other parameters, such as the environ- The seriously deteriorated glass samples are all covered with mental conditions and the manufacturing process of the glass, can play an important role.14 a weathering crust and numerous small and large craters are Fig. 7 Ca–Mn plot showing three outliers among the ten measured glass panels from Brussels’ Cathedral. 262 Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12Fig. 8 Secondary electron image of gypsum formed in and around a crater on the pitted and crusted surface. Surface Patina on Buried Nephrite Objects PIGMENT ANALYSIS OF GLASS PAINTINGS17 One of the methods of colouring glass is to cover a colourless Jade (nephrite) artefacts from burial sites of the Han dynasty in China are commonly covered with a white powdery layer glass pane with a thin layer of paint; two types exist: grisailles and enamels. The grisaille paint consists of a flux (e.g., a crystal on parts of the objects.Several explanations have been proposed regarding the origin of this particular surface layer, glass of 4PbO 5SiO2) and a pigment (e.g., Fe2O3). The mixture is ground, painted onto the glass and fired at approximately either explaining it as (i) an organic deposit from decomposing bodies or other objects in the grave or (ii) as a calcification of 630°C in order to fuse it to the surface (Fig. 9). In contrast, an enamel is a coloured glass with a low melting-point (e.g., the jade objects or finally (iii) as due to leaching of the mineral in the medium of decomposing bodies.As shown in Table 6, Fe2O3–PbO–SiO2). It consists of one type of grain and gives after firing a more homogeneous paint layer than a grisaille electron microanalysis of the altered surface layers on some of the objects shows readily that the surface layer has the same (Fig. 10). The glass paint Rouge Jean Cousin is a collective name for composition as the nephrite from which it originates.This clearly points to an alteration process of the jade objects under all the paints with a carnation red colour used for colouring heads, hands and bodies on stained glass windows. This type the influence of the alkaline medium of decomposing human bodies. This explanation has been advanced earlier by Gaines of paint was used from the fifteenth century until the beginning of the twentieth century.In the nineteenth century the name and Handy in 1975.15 Hence, the process occurs during the first few months after burial and the alteration layer cannot Rouge Jean Cousin was given to the carnation red paints after their presumed discoverer Jean Cousin the elder (1490–1560) be used to prove that an object has been buried for a long period of time. The regions of alteration are correlated with or Jean Cousin the younger (1522–1594). A study of these carnation red glass paints was performed at MiTAC by analys- the surface state of the objects as it occurs most often on unpolished or mineralogically weaker parts of the objects.16 ing historical stained glass windows and Rouge Jean Cousin Table 6 Comparison of composition of bulk nephrite and three deposit layers.* All values in % m/m Oxides Nephrite: Literature Nephrite: Electron microscopy Deposit: Sample A† Deposit: Sample B‡ Deposit: Sample C§ Na2O 1.3 — — — — MgO 24.2 21.5±0.2 24.4±2.3 23.7±2.7 25.2±2.4 Al2O3 1.3 0.7±0.1 1.1±0.5 1.1±0.7 0.9±0.6 SiO2 58.0 57.5±0.3 60.2±0.6 60.3±1.0 60.3±0.7 P2O5 — 0.1±0.1 0.1±0.1 0.2±0.2 0.1±0.1 SO3 — 0.9±0.1 0.3±0.1 0.1±0.1 0.1±0.1 Cl — 0.2±0.1 0.4±0.2 0.1±0.1 0.1±0.1 K2O — 0.9±0.1 0.3±0.2 0.1±0.1 0.2±0.1 CaO 13.2 13.7±0.1 12.2±1.7 13.3±2.1 12.5±2.0 FeO 2.1 4.5±0.1 1.1±0.4 1.0±0.3 0.7±0.3 * Chemical analysis of polished nephrite and of the deposit of three samples from the Dongxi Collection, Brussels, compared with the average composition of nephrite.† Sample A: Huang with whorl pattern, Late Spring and Autumn period, 6th/5th century BC. ‡ Sample B: Green and brown conge� flat walls, Zhou period, c. 1050–256 BC. § Sample C: Small round disc in the form of an animal, Late Zhou–early Western Han dynasty, 3rd/2nd century BC. Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12 263Table 7 Average composition of Rouge Jean Cousin particles of Lacroix (% m/m) as determined with SEM Composition determined with Composition calculated Oxide EPXMA from the recipe B2O3 Present 6.25 Na2O 3.00±0.4 2.79 MgO 0.91±0.05 — Al2O3 2.9±0.10 — SiO2 17.11±0.60 16.95 K2O 0.41±0.05 — CaO 2.30±0.3 — Fe2O3 34.3±0.5 19.70 CuO 0.38±0.07 — ZnO 0.29±0.04 — PbO 38.4±0.7 54.31 SiO2, PbO and Fe2O3 at the same time, which indicates that pigment and glass were melted together and ground afterwards.Table 7 shows the average composition of the Rouge Jean Cousin powders of Lacroix and the composition calculated from the original recipe of Lacroix.The two are not the same, but there is a resemblance. Fig. 9 Schematic diagram of a grisaille paint layer. In conclusion, a change in production of carnation red glass paints from grisaille to enamel paints seems to have taken place at the end of the nineteenth century. CONCLUSIONS All these examples from our laboratory’s projects were selected with the aim of showing the potential of instrumental analytical techniques in archaeology and art.The kind of applications discussed in this paper are multidisciplinary par excellence. In the first place they rely heavily on the experience of the laboratory in the field of environmental analysis as the corrosion of buried objects often critically depends on the local environment from which they are recovered. Second, the characterization of archaeological and art objects poses similar problems of sample handling and analytical procedures as that of modern technological materials.Finally, this type of work is extremely specialized and multidisciplinary with a necessity of collaboration between analytical chemistry and various other disciplines (history, art history, restoration, archaeology, etc.). In such circumstances, collaboration with specialized institutes is mandatory. In the examples given in this paper we worked in collaboration with the Oriental Institute of the University of Chicago for the early tin metallurgy, with the Royal Academy of Fine Arts (Antwerp), the National Institute for the Cultural Heritage (Brussels) and the Fraunhofer Institut fu�r Silicatenforschung (Wu� rzburg) for the glass corrosion, and with several art collectors and art Fig. 10 Schematic diagram of an enamel paint layer. historians for the other problems discussed. powders and the results were compared with recipes found in The authors thank K. A. Yener (Oriental Institute, Chicago), the literature from the seventeenth to the nineteenth century.H. O�zbal (Bogazici University, Istanbul), B. Earl (Cornwall), This carnation red paint exists as a grisaille in the period N. Van Hout, A. Balis (Rubenianum, Antwerpen), H. Wouters between the fifteenth and the beginning of the nineteenth (Royal Institute for Cultural Heritage, Brussels) and J. Caen century. A red grisaille was obtained when Fe2O3 grains with (Royal Academy of Fine Arts) for the samples and interesting a particle size smaller then 0.5 mm were used and when the discussions.We are also indebted to L.d’Alessandro, L. Vincze paint layer had a thickness of about 1 mm. These small pigment and P. Veny who assisted in the analysis of the trace elements grains were collected by mixing pigment, flux and arabic gum and to C. Yang, R. Utui and K. Malmqvist for assistance with in a glass of water. The largest particles sank to the bottom the m-PIXE measurements. after 3–5 d. The upper part, containing the smallest particles, was decanted and evaporated to obtain the red paint. REFERENCES At the end of the nineteenth century, glass paints were produced industrially by Lacroix & Cie in Paris (France) and 1 Yener, K. A., O� zbal, H., Kaptan, E., Pehlivan, A. N., and Kielblock in Arnstadt (Germany) for example. The analysis of Goodway, M., Science, 1989, 244, 200. Rouge Jean Cousin powders of Lacroix and Kielblock reveals 2 Yener, K. A., and Vandiver, P., Am. J. Archeol., 1993, 97, 207. 3 Earl, B.,and O� zbal, H., Archaeometry, 1996, 38, 289. that these paints are enamels, because a single grain contained 264 Journal of Analytical Atomic Spectrometry, March 1997, Vol. 124 Bachmann, H., T he Identification of Slags from Archaeological 14 Schreiner, M., Glastech. Ber., 1988, 61, 197. Sites, Institute of Archaeology, London, 1982. 15 Gaines, A. M., and Handy, J. L., Nature (L ondon), 1975, 253, 433. 5 Adriaens, A., Mikrochim. Acta, 1996, 124, 89. 16 Aerts, A., Janssens, K., and Adams, F., Orientations, 1995, 79. 6 Yener, K. A., The Go� ltepe/Kestel Project: 1994–1995 Annual 17 Schalm, O., Janssens, K., Albert, J., Peeters, K., Caen, J., and Report, Oriental Institute, Chicago, IL, 1996. Adams, F., in Dossier de la Commision Royale de Monuments, 7 Turner, W. E. S., J. Soc. Glass T echnol., 1956, XL, 194 and 277. Sites et Fouilles, Proceedings of Forum pour la Conservation et la 8 Donceel-Vou�te, P., Archeologia, 1994, 98, 24. restauration des V itraux, L ie` ge, June 19–22, 1996, ed. Barlet, J., 9 Geilmann, W., Glastech. Ber., 1956, 29, 145. Commission Royale des Monuments, Sites et Fouilles, Lie`ge, 10 Caen, J., Art Restorers Assoc., 1993, 9. Belgium, 1996, vol. 3, pp. 155–162. 11 Berckmans, W., Research Report, Royal Academy for Fine Arts, Antwerp, 1990. Paper 6/06091I 12 Genicot, F.-F., and Coomans, T., Revue des Arche�ologues et Received September 9, 1996 Historiens d’art de L ouvain, 1992, 11. 13 Gillies, K. J. S., and Cox, A., Glastech. Ber., 1988, 61, 75. Accepted January 6, 1997 Journal of Analytical Atomic Spectrometry, March 1

ISSN:0267-9477    

DOI:10.1039/a606091i

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

(Boro)Hydride Techniques in Trace Element Speciation† Invited Lecture ALAN G. HOWARD Department of Chemistry, University of Southampton, Southampton, UK, SO17 1BJ A number of applications of the sodium tetrahydroborate(iii ) present as the phosphate ion; lead, to be measured colorimetrically by the dithizone method, must be present as the Pb2+ (sodium borohydride) reagent in the determination of arsenic, selenium, sulfur and tin species are reviewed. The reaction of ion. In order to overcome this species selectivity, classical analytical methods, employed to determine the total amount an analyte species with aqueous tetrahydroborate(iii ) is most frequently employed to yield a volatile hydride product which of an element, have had to incorporate a sample preparation stage (such as combustion or wet ashing) designed to break can be readily removed from the bulk matrix.This results in the isolation of the analyte from interferents and gives a down the differing species into a single common, determinable form.Measurement methods having intrinsic species-selectivity species which can be readily concentrated and separated from other species. Some simple speciation analyses, such as the have therefore been carried out for many years, but it is only comparatively recently that the objective of measuring individ- differentiation of arsenite and arsenate, can be performed by careful control of reaction conditions. With more complex ual species has become a specific goal.This paper sets out to provide a short review of the molecules some of the original chemical structure of the target species (such as a C–As bond) can be conserved in the reaction applications of sodium tetrahydroborate(III) (sodium borohydride) in speciation analysis, drawing on personal examples of products, permitting the deduction of further speciation information. Arsenic oxy-anions [RnAsO(OH)3-n, where its use in techniques employed in the study of chemical processes taking place in the aquatic environment.Readers R=alkyl or aryl] and alkyltin species, for example, yield different arsines (RnAsH3-n) and stannanes, respectively, which requiring more specialized reviews of the instrumentation employed in both speciation and total element analyses are can be cryogenically trapped and subsequently separated by distillation or gas–liquid chromatography. More recently, referred to the reviews by Fang et al.1 on flow injection analysis, Harrison and Rapsomanikis2 on the interfacing of analyte reactions with tetrahydroborate (iii) have been employed to provide a powerful link between the components chromatography with atomic spectrometry, Pyrzynska3 and Olivas et al.4 on selenium determination, and Nakahara5 and employed in hyphenated instrumentation.One of its most significant roles is in the development of HPLC Dedina and Tsalev6 on hydride generation AAS. instrumentation where it can be employed to generate gas phase analyte species which are compatible with sensitive and HYDRIDE METHODS highly selective gas/vapour detection systems such as AAS, Sodium Tetrahydroborate(III) Reagent AES and ICP-MS.Examples are given in which the tetrahydroborate(iii ) reaction link is employed in the coupling Sodium tetrahydroborate(III ) is a particularly versatile reagent of HPLC with AAS. The addition of photochemical oxidation which has become widely used for its reducing and hydrideor microwave digestion steps prior to the tetrahydroborate(iii) transfer properties.Whilst it has many applications, it has reaction stage further extends the range of detectable found particular value in the conversion of species in aqueous compounds. Whilst the tetrahydroborate (iii) reagent is largely solution into volatile hydrides (hydride generation, HG). In associated with the generation of volatile analyte hydrides, this this role, the tetrahydroborate(III ) reagent can be thought of is not always the case.Cold-trap methods, for example, can be as acting as both a reductant and as a hydride source. In its employed to measure some trimethylarsenic compounds by the reaction with the arsenic oxy-anions, which contain arsenic in production of trimethylarsine and alkyltin hydrides can be the +5 oxidation state, the first step is believed to be the preconcentrated by solvent extraction. HPLC instrumentation reduction of the arsenic to the +3 state.is also described which employs the tetrahydroborate(iii ) gas RnAs(O)(OH)3-n+H++BH4-�RnAs(OH)3-n+H2O+BH3 phase/liquid phase link to generate volatile sulfur species which are compatible with flame photometric detection. In most current speciation studies, R is a methyl group and n ranges from 0 to 3. Keywords: T etrahydroborate(III) (borohydride); speciation ; Subsequent reaction with the tetrahydroborate(III) takes the arsenic; selenium; sulfur ; tin; hydride generation; arsenic compound through to the corresponding arsine: interferences; dimethylsulfoniopropionate; high-performance liquid chromatography; flame photometric detection ; RnAs(OH)3-n+(3-n)BH4-+(3-n)H+�RnAsH3-n hyphenated techniques ; ultraviolet photolysis; microwave; +(3-n)BH3+(3-n)H2O review The borane generated by these reactions hydrolyses giving boric acid and gaseous hydrogen.Analytical methods which are selective towards certain chemical species are not new; examples abound in the literature of BH3+3H2O�H3BO3+3H2 classical methods which are dependent on the analyte being present as a particular species.The colorimetric determination pH Selectivity and Speciation of phosphorus, for example, depends on the element being The reaction between NaBH4 and an ion in solution is sensitive to pH and it appears that, for the reaction to proceed rapidly, † Presented at the Eighth Biennial National Atomic Spectroscopy Symposium (BNASS), Norwich, UK, July 17–19, 1996.the target species must not be present in solution as a negatively Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12 (267–272) 267charged species. In arsenic speciation analyses this means that be fully protonated at readily achievable pH values (selenate pK2=2.0 implying a pK1 value of approximately-3). Selenite, arsenate must be fully protonated if it is to be converted to arsine. which is fully protonated below pH 2 (pK1=2.5), can, however, be determined in the presence of selenate and then ‘total As pK1 for arsenic acid is 2.3 (Table 1), the reaction must therefore be carried out at very low pH; 1–2 M hydrochloric inorganic’ selenium can be assessed after the conversion of selenate to selenite (normally by heating with hydrochloric acid is commonly employed.Arsenite, on the other hand, is protonated under most conditions (pK1=9.2) and will react acid). This pre-reduction step can be carried out on-line by microwave heating of the acidified sample prior to the HG with the tetrahydroborate(III ) under conditions which are only mildly acidic.The yields of arsines under various conditions step.12,13 of acidity are shown in Fig. 1. In the absence of other arsenic species, differentiation of TRAP SPECIATION METHODS arsenate and arsenite can therefore be achieved simply by exploiting the pH dependency of the tetrahydroborate(III ) The discovery of dissolved methylated arsenic and tin species reaction. Carrying out the reaction first at pH 5 gives a measure in the marine environment, and most of the subsequent work of arsenite and then a repeat reaction in 2 M hydrochloric acid which has been carried out to understand the presence and can be used to assess both arsenite and arsenate.It should be significance of related selenium, antimony and tin species, has noted, however, that the presence of methylated arsenic species, resulted from the development of cryogenic trap hydride common in environmental samples, is liable to interfere in methods.Cryogenic trapping (CT) of the volatile hydrides such an analysis unless steps are taken to separate the methyl- resulting from the HG reaction, at liquid nitrogen temperaated arsines which are formed from the arsenite/arsenate- tures, results in a focusing of the products as a narrow plug of derived AsH3. material which can then be separatther by a distillation- Whilst the arsenic oxy-anion species have received great based mechanism using a trap consisting of a short column of attention, a large proportion of the arsenic which is found in glass beads, or by gas–liquid chromatography (GLC) using a biological tissues is present in chemical forms such as arsenob- trap containing a conventional coated support material.The etaine and arseno-sugars, which do not yield volatile hydrides. first system, presented by Braman and Foreback in 1973,14 Breakdown of such compounds, to yield products which are employed a distillation trap concentrating the arsines in a susceptible to HG, can be carried out by in-line persulfate U-tube containing glass beads, prior to detection by AES.oxidation enhanced by either microwave heating10 or UV There have been a number of subsequent developments of the radiation.11 Without prior separation of such species, how- technique including the introduction of continuous-flow ever, such an approach destroys the available speciation hydride generation and a move to AAS detection.9,15,16 information.In its most recent forms the apparatus employed for cryo- Whilst pH-based species differentiation is particularly useful genic trap arsenic speciation uses a continuous-flow reagent for the study of arsenic species, it is not widely applicable to system mixing sample, acid/buffer and NaBH4 (Fig. 2). The other elements. It cannot, for example, be fully extended to the resulting arsines are then collected at -196 °C on a trap filled measurement of selenite and selenate as the latter ion cannot either with etched glass beads or a chromatographic packing material such as the silicone OV-3 coated on Chromosorb W-HP.17,18 Table 1 Acid dissociation constants of some acids7,8 The collected arsines differ significantly in their boilingpoints (Table 2) and can be sequentially evolved from the trap Abbreviation pK by heating.With the glass bead traps this is simply achieved As(OH)3 AsIII pK1=9.2 by allowing the trap to warm to room temperature but, if a As(O)(OH)3 AsV pK1=2.3 GLC packing is employed, it must be heated to elute the pK2=6.8 arsines.The arsines are then atomized in a quartz furnace tube pK3=11.6 (CH3)As(O)(OH)2 MMAA pK1=4.0 (heated to approximately 900 °C) aligned in the light path of pK2=8.6 an atomic absorption spectrometer. When calibrated using (CH3)2As(O)OH DMAA pK=6.3 1 ml samples containing less than 1 ng of arsenic, the system Se(O)(OH)2 SeIV pK1=2.5 pK2=7.3 Se(O)2(OH)2 SeVI pK1=-3 (estimated) pK2=2.0 Fig. 2 Current cryogenic trap apparatus based on instrumentation reported in ref. 16. Table 2 Boiling-points of arsines Compound bp/°C Fig. 1 pH dependency of the arsine yield obtained from the reaction AsH3 -55 of a number of arsenic oxy-anions with NaBH4 (adapted from ref. 9). CH3AsH2 2 Small vertical arrows indicate the pH values at which pH=pK1 for (CH3 )2AsH 35.6 each acid. 268 Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12offers very low detection limits (<50 ng l-1). Larger volumes formation of arseno-cysteine complexes. The resulting cysteine complexes are uncharged under weakly acidic conditions and of sample can be employed, reducing these detection limits further. significantly less acid therefore needs to be used for them to react with the tetrahydroborate(III). This results in a reduction The results of trap experiments are often misinterpreted and it is often implied that such techniques can identify species in the acid-derived blank contributions to the inorganic arsenic response and detection limits (3s) for the four arsenic oxy- such as selenite and dimethylarsinate (DMAA).This is not the case as the procedure can neither distinguish selenite from anions in the range 30–99 pg with relative standard deviations of 3.1–7.9%.33 selenide nor distinguish DMAA from other ‘dimethylarsenic’ (DMA) species as the anionic nature of the species is not A second, yet equally important, effect of the cysteine pretreatment procedure is the improved performance of the conserved in the hydride product. Interferences, largely from transition metals,19–27 can influ- hydride trap procedure when measurements are made in the presence of metals such as iron(II ), nickel(II), cobalt(II), manga- ence the results obtained from the technique but a number of sample pre-treatments and masking agents have been devel- nese(II) and copper(II) as these interfere in the conventional cryogenictrap HGAAS speciation procedures.Detailed investi- oped to overcome such problems.9,28,29 Such problems normally only occur in samples containing high levels of interfering gations of the use of cysteine in the measurement of total Sb, As, Bi and Sn have recently been reported by Tsalev et al.35 ions. Some of these masking agents, most notably L-cysteine and L-cystine, have, however, been shown to alter the character- Gas chromatographic separation of trapped hydrides has been extensively employed for the measurement of alkylated istics of the HG reaction significantly.Not only does their presence in the reaction mixture lead to a reduction in inter- selenium, lead, germanium and tin compounds. Whilst interference problems have been widely investigated for total ferences, but also altered reaction conditions are required when they are used.When such changes are made in the determi- element determinations, comparatively little is known about such problems in speciation analyses. A CT-HGAAS system nation of total arsenic the responses to different species are equalized.30–32 Le et al.32 pointed the way to the utilization of presented by Donard and co-workers,36,37 employing a chromatographic separation, has been widely adopted for the cysteine pre-derivatization in a cold trap technique, which we have subsequently developed into a technique, which gives measurement of alkyltin species. This remains one of the few speciation systems for which studies have been carried out into more uniform sensitivities to arsenite, arsenate, monomethylarsonate (MMAA) and DMAA and greater resistance to inter- the effects of metal ion interferences.38 Interference effects were found to be significant,but they could be successfully controlled ferences from metal ions (Fig. 3).33 This pre-treatment with cysteine can be carried out at room by the use of the masking agents EDTA and L-cysteine.temperature but occurs more rapidly by microwave heating. A number of suggestions have been made regarding the mechanism of this effect. Brindle and Le34 proposed the HPLC SPECIATION METHODS—HYDRIDE GENERATION AS A HYPHENATION LINK formation of an H3B–SR- intermediate which was a more efficient reductant than tetrahydroborate(III ). This did not, Whilst HG can be successfully employed as an initial sample however, explain the fact that arsine yields depend on the pre- pre-treatment stage and to yield simple speciation information, reaction between the arsenic species and cysteine being allowed its use inevitably results in the loss of structural information.to reach completion prior to the addition of the tetrahydrobo- Chromatographic separation of the unaltered analyte species rate(III ). Whilst at present the mechanism behind the effective- is therefore always preferable, as this permits the use of ness of the cysteine treatment is unproven, the most consistent retention time information in species identification.The prob- explanation is that the reaction proceeds by the initial lem is then to provide a means of detecting the eluting species reduction of arsenic(V) species to arsenic(III ) (Fig. 4) and the and it is here that the HG reaction can provide a valuable link between the liquid effluent from an HPLC column and the enhanced performance of gas phase element-specific detection systems.Simple Tetrahydroborate(III) Post-column Reduction Systems Hydride generation has been employed in the linking of HPLC to a number of element-specific detectors employed in the measurement of the simple arsenic and selenium oxy-anions. Four common arsenic species, arsenite, arsenate, MMAA and DMAA, yield volatile hydrides on reaction with tetrahydroborate( III). For their determination, simple hyphenated HPLC– HGAAS instrumentation can be constructed by connecting Fig. 3 Comparison of trap HGAAS measurements of arsenic the HPLC column directly to the inlet of a continuous-flow species.33 hydride generation module such as that described by Arbab- Zavar and Howard.9 The main considerations involved are the careful selection of the HPLC eluent, measures to limit band-broadening and compensation for the buffer capacity of the eluent. Whilst all of the examples of the use of tetrahydroborate(III ) mentioned so far have involved the generation of volatile hydrides, this is not always the case.The HPLC instrumentation shown in Fig. 5, for example, can be used for the measurement of trimethylarsine oxide with its detection being due to the generation of trimethylarsine. In this case the tetrahydroborate(III ) has performed just a simple reduction. Fig. 4 Proposed reaction mechanism for the cysteine-based HG from various methylated arsenic oxy-anions (R=Me, CysSH=cysteine).The system can, however, be taken further and we have recently Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12 269Fig. 5 Schematic illustration of HPLC–AAS instrumentation employing a HG link.39 Fig. 7 HPLC trace showing the detection of a number of dimethyl sulfide (DMS) precursors [dimethylsulfoniopropionate (DMSP), 2-(methyl)dimethylsulfoniopropionate (DMSP-2-MP), dimethylsulfonioacetate (DMSAcet), 4-(dimethylsulfonio)butanoate (DMSBut), 5-(dimethylsulfonio)pentanoate (DMSPent), S-methylmethionine (SMM)] (Russell and Howard, unpublished results).the arsenic metabolites arsenobetaine, arsenocholine and arseno-sugars; organoselenium compounds such as selenocysteine and selenomethionine and the butyltin antifoulants. In some such cases the compounds can be detected in the HPLC eluate if they are first photochemically broken down to compounds which, on reaction with NaBH4 , yield volatile species. Difficulties with the detection of p-aminophenyl arsonate by Fig. 6 HPLC-FPD instrumentation employing in-line post-column coupled ion chromatography–HGAAS led Ricci et al.43 to reaction with NaBH4 (Russell and Howard, unpublished results). incorporate an acid-persulfate oxidation step in their system. Later, Atallah and Kalman44 demonstrated the use of in-line developed dedicated instrumentation for the measurement of photo-oxidation for the determination of organoarsenic com- sulfur species by HPLC. pounds by AAS. The tributyltin ion is also difficult to detect Sulfate metabolism by marine algae results in the formation in HPLC studies as it produces a hydride product which has of a number of organosulfur compounds, the most commonly restricted volatility at room temperature.It can, however, be reported being dimethylsulfoniopropionate (DMSP), a com- broken down to inorganic SnIV by exposure to UV light, which pound which is believed to be an osmoregulator.40 Its chemical, then gives stannane when reacted with tetrahydroborate(III ).45 bacterial and/or enzymic breakdown releases volatile dimethyl The apparatus employed for the measurement of organoarsenic sulfide (DMS) into the atmosphere where it is rapidly oxidized compounds is a simple adaptation of the HPLC–HGAAS to methanesulfonic acid, sulfur dioxide and sulfate.This can system shown in Fig. 511 incorporating a post-column treat- cause the nucleation of cloud droplets and increased aerosol ment of the eluate with alkaline persulfate and exposure to and precipitation acidity.41 UV radiation prior to the HG step (Fig. 8).11 Incorporation of a tetrahydroborate(III ) reaction stage into An additional degree of confidence in compound identifi- HPLC instrumentation originally developed for the highly cation can be achieved with this system by running the specific measurement of DMSP42 has permitted us to extend apparatus with and without the UV irradiation (Fig. 9). the range of compounds which can be detected to other An example of the application of the HPLC–UV–HGAAS potential DMS-precursors. The instrument incorporates iso- instrumentation is shown in Fig. 10 in which an arsenic cratic ion-exchange HPLC of the DMS-precursors followed compound, with retention behaviour similar to that of known by post-column reaction with tetrahydroborate(III ), which releases volatile sulfur-containing products from the compounds. After separation from the liquid stream, these can be detected using a custom-designed flame photometric detector (Fig. 6). The separation of a number of potential DMS-precursors is shown in Fig. 7. Tetrahydroborate(III) Link with In-line Redox Pre-treatment Whilst tetrahydroborate(III ) alone is capable of producing volatile products from a number of important species, it is not universally applicable and many compounds, whilst separable by HPLC, cannot be readily detected either due to their inertness towards the tetrahydroborate(III ) reagent or due to the formation of hydride products which have limited volatility Fig. 8 HPLC with AAS detection employing in-line photolysis and HG links.11 at room temperature. Examples of such compounds include 270 Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12as arsenate and DMAA. With the development of specialized HPLC systems it is now possible to carry out compound identification using the chromatographic step and then to perform a post-column photolysis treatment of tetrahydroborate( III)-resistant compounds to convert them into forms which are amenable to the tetrahydroborate(III ) reagent.Such approaches will inevitably result in the application of the reagent to a much broader range of chemical species. REFERENCES 1 Fang, Z., Xu, S., and Tao, G., J. Anal. At. Spectrom., 1996, 11, 1. 2 Harrison, R. M., and Rapsomanikis, S., Environmental Analysis Fig. 9 Compound differentiation by HPLC–UV–HGAAS showing using Chromatography Interfaced with Atomic Spectroscopy, Ellis the effect of photolysis on the range of detected arsenic species (PhAs= Horwood, Chichester, 1989.phenylarsonic acid, AsC=arsenocholine, AsB=arsenobetaine, 3 Pyrzynska, K., Analyst, 1996, 121, 77R. TMAS=tetramethylarsonium, p-As=p-arsanilic acid, o-As=o-arsan- 4 Olivas, M. R., Donard, O. F. X., Camara, C., and Quevauviller, P., ilic acid).11 Anal. Chim. Acta, 1994, 286, 357. 5 Nakahara, T., Prog. Anal. At.Spectrosc., 1983, 6, 163. 6 Dedina, J., and Tsalev, D. L., Hydride Generation Atomic Absorption Spectrometry, Wiley, Chichester, 1995. 7 Ricci, J. E., J. Am. Chem. Soc., 1948, 70, 109. 8 Douglas, B., McDaniel, D. H., and Alexander, J. J., Concepts and Models in Inorganic Chemistry, Wiley, New York, 2nd edn., 1983. 9 Arbab-Zavar, M. H., and Howard, A. G., Analyst, 1980, 105, 744. 10 Le, X.-C., Cullen, W. R., and Reimer, K. J., Appl. Organomet. Chem., 1992, 6, 161. 11 Howard, A.G., and Hunt, L. E., Anal. Chem., 1993, 65, 2995. 12 Cobo-Fernandez, M. G., Palacios, M. A., Chakraborti, D., Quevauviller, P., and Camara, C., Fresenius’ J. Anal. Chem., 1995, Fig. 10 Possible identification of arseno-sugars in sediment from 351, 438. the tidal mudflats of Langstone Harbour (Hampshire, UK). 13 Pitts, L., Worsfold, P. J., and Hill, S. J., Analyst, 1994, 119, 2785. Chromatograms obtained using the system shown in Fig. 9 (Perrett 14 Braman, R. S., and Foreback, C.C., Science, 1973, 182, 1247. and Howard, unpublished results). 15 Andreae, M. O., Anal. Chem., 1977, 49, 820. 16 Edmonds, J. S., and Francesconi, K. A., Anal. Chem., 1976, algal arseno-sugars, was found to be present in estuarine 48, 2019. 17 Howard, A. G., and Comber, S. D. W., Mikrochim. Acta, 1992, sediments. 109, 27. The enhancement of LC–HGAAS systems by the incorpor- 18 Howard, A. G., and Arbab-Zavar, M. H., Analyst, 1981, 106, 213. ation of a photolysis step is mirrored in ICP-OES systems.A 19 Welz, B., and Melcher, M., Analyst, 1984, 109, 573. related LC–UV–HG-ICP-OES system has been developed46,47 20 Smith, A. E., Analyst, 1975, 100, 300. for the measurement of arsenic species and employed for the 21 Van Cleuvenbergen, R. J. A., Van Mol, W. E., and Adams, F. C., analysis of the fish reference materials CRM 278 (Mussel), J. Anal. At. Spectrom., 1988, 3, 169. 22 Verlinden, M., and Deelstra, H., Fresenius’ Z. Anal. Chem., 1979, CRM 422 (Cod) and DORM-1 (Dogfish).48,49 296, 253.In recent years, post-column microwave-induced breakdown 23 Pierce, F. D., and Brown, H. R., Anal. Chem., 1976, 48, 693. and reduction has become a useful tool to overcome the 24 Pierce, F. D., and Brown, H. R., Anal. Chem., 1977, 49, 1417. resistance of a number of selenium species to HG. It has been 25 Welz, B., and Schubert-Jacobs, M., J. Anal. At. Spectrom., 1986, employed in the on-line speciation of selenite, selenate and 1, 23.trimethylselenium detected by AAS12 and in the determination 26 Hershey, J. W., and Keliher, P. N., Spectrochim. Acta, Part B, 1986, 41, 713. of selenite and selenate.50 Camara and co-workers have devel- 27 Welz, B., and Melcher, M., Analyst, 1984, 109, 569. oped both thermo-oxidation51 and on-line microwave-assisted 28 Kirkbright, G. F., and Taddia, M., Anal. Chim. Acta, 1978, oxidation52,53 to break down difficult arsenic compounds prior 100, 145. to HGAAS detection of arsenic species. 29 Belcher, R., Bogdanski, S. L., Henden, E., and Townshend, A., Analyst, 1975, 100, 522. 30 Boampong, C., Brindle, I. D., Le, X.-C., Pidwerbesky, L., and CONCLUSIONS Ceccarelli Ponzoni, C. M., Anal. Chem., 1988, 60, 1185. 31 Chen, H., Brindle, I. D., and Le, X.-C., Anal. Chem., 1992, 64, 667. The NaBH4 reagent has proven to be an exceptionally reliable 32 Le, X.-C., Cullen, W. R., and Reimer, K. J., Anal. Chim. Acta, reagent for the conversion of a number of elements to volatile 1994, 285, 277.forms. The most fundamental result is that an element can be 33 Howard, A. G., and Salou, C., Anal. Chim. Acta, 1996, 333, 89. transferred from the liquid phase to the gas phase. This has 34 Brindle, I. D., and Le, X.-C., Anal. Chim. Acta, 1990, 229, 239. 35 Tsalev, D. L., D’Ulivo, A., Lampugnani, L., Di Marco, M., and resulted in the transfer of the element to the detector more Zamboni, R., J. Anal. At. Spectrom., 1996, 11, 989. efficiently than would have been possible using, for example, 36 Donard, O.F. X., Rapsomanikis, S., and Weber, J. H., Anal. a spray nebulization system. Being now in the gas phase the Chim., 1986, 58, 772. analyte element has become compatible with a wide range of 37 Randall, L., Donard, O. F. X., and Weber, J. H., Anal. Chim. element-specific detection systems which would not have been Acta, 1986, 184, 197. suitable for use with the original aqueous solutions. Once the 38 Martin, F. M., and Donard, O.F. X., J. Anal. At. Spectrom., 1994, 9, 1143. volatile analyte species has left the solution, it has also left 39 Comber, S. D. W., PhD Thesis, University of Southampton, 1990. behind many of the species which might act as interferents in 40 Dickson, D. M., Wyn Jones, R. G., and Davenport, J., Planta, the analysis. 1980, 150, 158. The tetrahydroborate(III ) reagent is fairly selective in the 41 Andreae, M. O., Mar. Chem., 1990, 30, 1. compounds with which it will react to yield a volatile product. 42 Howard, A. G., and Russell, D. W., Anal. Chem., 1995, 67, 1293. In the analysis of arsenic species the reagent has, until recently, 43 Ricci, G. R., Shepard, L. S., Colovos, G., and Hester, N. E., Anal. Chem., 1981, 53, 610. been restricted to the analysis of the simple oxy-anions such Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12 27144 Atallah, R. H., and Kalman, D. A., T alanta, 1991, 38, 167. 51 Lopez, M. A., Gomez, M. M., Palacios, M. A., and Camara, C., 45 Ebdon, L., Hill, S. J., and Jones, P., T alanta, 1991, 38, 607. Fresenius’ J. Anal. Chem., 1993, 346, 643. 46 Rubio, R., Padro, A., Alberti, J., and Rauret, G., Anal. Chim. 52 Lo�pez Gonza�lvez, M. A., Go�mez, M. N., Camara, C., and Acta, 1993, 283, 160. Palacios, M. A., J. Anal. At. Spectrom., 1994, 9, 291. 47 Rubio, R., Alberti, J., and Rauret, G., Int. J. Environ. Anal. Chem., 53 Martin, I., Lopez Gonzalvez, M. A., Gomez, M., Camara, C., and 1993, 52, 203. Palacios, M. A., J. Chromatogr. B, 1995, 666, 101. 48 Alberti, J., Rubio, R., and Rauret, G., Fresenius’ J. Anal. Chem., 1995, 351, 415. Paper 6/05050F 49 Rubio, R., Alberti, J., Padro, A., and Rauret, G., T rends Anal. Received July 22, 1996 Chem., 1995, 14, 6. Accepted December 19, 1996 50 Pitts, L., Fisher, A., Worsfold, P., and Hill, S. J., J. Anal. At. Spectrom., 1995, 10, 519. 272 Journal of Analytical Atomic Spectrometry, March 1997, V

ISSN:0267-9477    

DOI:10.1039/a605050f

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Characterization of an Ultrasonic Nebulizer–Membrane Separation Interface with Inductively Coupled Plasma Mass Spectrometry for the Determination of Trace Elements by Solvent Extraction† Invited Lecture I. B. BRENNER*a , A. ZANDERa, M. PLANTZb AND J. ZHUc aGinzton Research Center, Varian Associates, 3075 Hansen Way, Palo Alto, CA 94305–1025, USA bVarian Optical Spectroscopy Instruments, 201 Hansen Court, Wood Dale, IL 60191, USA cCETAC T echnologies, Inc., 5600 South 42nd Street, Omaha, NE 68107, USA The problems of using an ultrasonic nebulizer–membrane solvent extraction procedures for analyte preconcentration and desolvation separation interface (USN–MEMSEP) for the matrix elimination with direct ICP-MS determination has been determination of trace elements by solvent extraction and constrained owing to the deleterious effects of organic solvent ICP-MS were studied.The interference effects of chloroform loading on the plasma discharge characteristics.The introducand the behavior of trace metal chelates as a function of tion of volatile solvents into an ICP results in plasma instability MEMSEP temperature and sweep gas were studied. In due to energy withdrawal, formation of C molecular species comparison with conventional nebulization, the use of the such as C2, CN and COin the plasma, formation of polyatomic interface resulted in an approximately 10-fold increase in ions, and carbon deposition on the torch and sample cone.3–8 analyte signals.Although the interface removed much of the Hence, a system which would allow the direct and routine chloroform vapor from the aerosol stream by selective nebulization of volatile solvents into the ICP would greatly permeation and argon counter gas purging, residual solvent increase the prospects for solvent extraction as a method of resulted in polyatomic ion interferences that affected the limits preconcentration, with the benefits of removal of salt matrices of detection. The addition of a small flow of oxygen to the and elimination of spectroscopic and non-spectroscopic auxiliary gas minimized these interferences and prevented interferences.carbon deposition on the torch tubes and sampler cones. The The effects of organic solvent load have been reduced by 40Ar12C+ signal was attenuated, but those of 40Ar16O+ and addition of oxygen to the plasma, which converts C and the CeO+/Ce+ increased slightly. An increase in MEMSEP various C-molecular species into CO and CO2.9,10 The solvent desolvation temperature resulted in a decrease in 35Cl16O+ and load was minimized using a cooled spray chamber10–12 and by 40Ar12C+ signals due to enhanced rejection of chloroform. transporting the heated aerosol produced with an ultrasonic Thermal desolvation of the metal organic compound vapors nebulizer (USN) through adesolvating condenser at-10 °C.13 and aerosols resulted in a decrease in the ion counts of the It was demonstrated that detection limits improved with chelated analytes with increasing temperature, probably due to decreasing spray chamber and increasing heating stage temtheir volatilization and rejection from the membrane. An peratures.Overloading was also minimized by desolvating the internal standard could be used to compensate for the aerosol using crycondensers which were cooled by dry-ice to responses to changes in temperature. Signal responses of the –77 °C.14 Although this desolvation system removed most of metal dithiocarbamates to changes in MEMSEP desolvation the solvent, a significant amount was passed on to the plasma.temperature were significantly different to those in chloroform In a complex arrangement consisting of an ultrasonic nebulizer solutions of oil-based standards, and as a consequence the and multiple heating and cooling cryogenic loops, a larger latter were unsuitable for calibration. The advantages of the proportion of volatile organic solvents was removed from the technique include matrix elimination, marked reduction in aerosol stream by repetitive heating at 100 °C and cooling polyatomic carbide ions, enhanced LODs and reduced plasma with dry-ice and with ethanol at -80 °C.15,16 Brotherton interferences.et al.12 described a low volume interface for flow injection of volatile solvents into the ICP, consisting of a water cooled Keywords: Inductively coupled plasma mass spectrometry; mini-spray chamber, a thermal desolvation interface and a ultrasonic nebulization ; membrane separation; solvent partial-suction injector to remove solvent excess by utilizing extraction; trace elements differences in momentum between aerosol droplets and the solvent vapor. The complications of direct analysis of volatile Chemical preconcentration of trace elements by solvent extrac- organics can be inelegantly overcome by back-extracting the tion is a convenient approach for improving limits of determi- analytes into an acidic aqueous solution before ICP analy- nation and eliminating interference effects due to bulk chemical sis.17–19 This is a time-consuming approach with potential constituents.Although methods of solvent extraction have sample contamination. been widely employed with flame spectroscopy,1,2 the scope of Gustavsson and co-workers20,21 developed a membrane interface for ICP-AES to desolvate both aqueous and volatile organic solvents.The membrane separated the nebulizer gas † Presented at the Eighth Biennial National Atomic Spectroscopy Symposium (BNASS), Norwich, UK, July 17–19, 1996. and the aerosol from the the solvent vapors, which were purged Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12 (273–279) 273through the membrane and removed using an argon counter EXPERIMENTAL gas. Tao and Miyazaki22 described a polyimide membrane Reagents, Purification and Preparation separator operating at 80 °C to decrease the water loading in All plastic and glassware were cleaned by soaking in 20% an ICP-MS system.As a result, 40Ar16O+ and 35Cl16O+ nitric acid for 24 h, followed by three rinses with ultrapure decreased by one and two orders of magnitude, respectively, water. Stock standard solutions (Spex Industries, Edison, NJ, and the CeO+/Ce+ ratio at the maximum 108Ce+ signal USA) were diluted to produce the tuning solutions and the intensity amounted to 0.11% compared with 11% with their spiked trace element water samples. conventional system.Organic solvents were not studied. Botto and Zhu23 evaluated a commercial USN–membrane desolvation separation interface (MEMSEP) for the analysis of Extraction Procedure volatile organic solvents using ICP-AES. Approximately 90% Chloroform (Fluka, Buchs, Switzerland) was purified by extrac- of the water mass was removed but only about 30% for tion with 50% nitric acid (Select Plus Mallinckrodt, St.Louis, various organic solvent vapors. In the case of highly volatile MO, USA); 1 l of chloroform was poured into an acid washed solvents, the efficiency of removal must be high in order to PTFE separating funnel, 100 ml of the acid were added and overcome the problems described above. the mixture shaken for 10 min. The funnel was allowed to The extraction chemistry of the combined sodium diethyldi- stand for 5 min for separation of the acid layer, which was thiocarbamate–ammonium tetramethylenedithiocarbamate then poured off.This was repeated three times. The chloroform (pyrrolidine-1-dithiocarbamate) (NaDDC–APDC) system was was then washed three times with ultrapure water (18.3MV) studied by McLeod et al.,18 Wyttenbach and Bajo24 and (MilliQ Plus Millipore, Bedford, MA, USA.) Kinrade and Van Loon.25 They demonstrated that these chelat- In accordance with the results obtained by Kinrade and ing agents, used in a ratio of 151, permit the selective extraction Van Loon,25 1 g of NaDDC (Fluka) and 1g of APDC (Fluka), of numerous transition metals over the pH range 2–6.Citrate dissolved in 200 ml of ultrapure water, were used. These or acetate buffers of pH 4–5 were employed, the former being solutions were purified by mock extraction with 10 ml of preferred. This extraction system in chloroform and IBMK purified chloroform for 10 min. Citrate buffer was prepared by has been widely applied in flame atomic absorption1,2,25–28 dissolving 120 g of citric acid and 44 g of sodium citrate in and in, ETV-AAS,29 but very rarely using direct injection into ultrapure water and diluting to 500 ml.The pH was approxi- the ICP with AES and MS owing to the interferences described mately 4. The buffer was purified by extraction with 10 ml of above. Thus, Zhuang et al.19 used an elaborate on-line flow NaDDC–APDC in an acid-washed PTFE separating funnel. injection manifold using Co-APDC coprecipitation for the A 10 ml volume of purified chloroform was added, mixed for preconcentration of trace amounts of heavy metals in rain 10 min and allowed to stand for 5 min.The organic layer was water. Although IBMK dissolved the precipitate, the solvent separated and discarded. This was repeated three times. could not be introduced directly into the plasma. Consequently, Aqueous solutions containing the trace elements of interest the precipitate, collected on a PTFE membrane, was dissolved were processed in order to determine the recoveries.The in nitric acid and hydrogen peroxide and the solution aspirated extraction was performed by adding 4 ml of citrate buffer to into the ICP. 200 ml of sample containing 0.1% (v/v) nitric acid, followed In a previous investigation,30 we examined the feasibility of by 5 ml of NaDDC–APDC reagent, and agitating the mixture separating a volatile solvent containing extracted metal che- for 10 min and then allowing it to stand for 5 min.A 10 ml lates using a Cetac (Omaha, NE, USA) ultrasonic nebulizer volume of purified chloroform was added and the agitation USN–MEMSEP. Several trace metals were extracted using process repeated. The organic layer was then drained and the the Aliquat 336–IBMK extraction system31,32 and the recover- extraction repeated. The final extraction volume was 20 ml. ies determined using ICP-AES. IBMK was aspirated directly into the USN to increase the aerosol production efficiency, Calibration decrease population of the C-species in the plasma by desolvation and then deliver the organic solvent vapors and aerosols Calibration standards were prepared by mixing a multi-element to a thermal membrane desolvator where residual volatile oil-based standard (Spex 21, Spex Industries) with purified organic components were separated from the metal chelate chloroform to provide a series of standards containing a blank aerosols.The MEMSEP was operated at 160 °C.The ability (chloroform only) and 5, 20, 50 and 100 mg l-1 of the trace of the MEMSEP to separate the volatile organic flows from elements of interest (V, Cr, Mn, Ni, Cu, Fe, Mo, Ag, Zn, Sn, metal aerosols was demonstrated by determining the recoveries Tl, Cd, Bi and Pb). In order to determine whether these oil- of several transition metals in an oil-based IBMK standard based standards could be used to analyze chloroform extrac- and in the Aliquat 336–IBMK extract relative to an aqueous tions, ultrapure deionized water and Canadian National solution.It was observed that recoveries of several metal Research Council NAAS-2 sea-water were spiked with 5, 10, chelates were low, evidently owing to volatilization of the 20 and 50 mg l-1 of trace metals, processed and analyzed. organic metal ketonates in either the heating stage of the USN (140 °C) or/and the membrane. The multi-element capability Instrumentation and Operating Conditions and limits of determination were limited owing to sequential atomic emission detection. The instrumentation and ICP operating conditions used are In this investigation, we conducted several experiments to listed in Table 1.The chloroform extracts were delivered peri- evaluate the interference effects when several metal dithiocarba- staltically to the USN and removed to the drain using Viton mates and a multi-element oil-based standard in chloroform chloroform resistant tubing kits.The USN–MEMSEPinterface was introduced into a USN–MEMSEP. Chloroform (bp about was a Cetac U-6000AT system consisting of an ultrasonic 60 °C) was selected because of its low solubility in water and nebulizer to increase aerosol production efficiency and a 1.5 m because most of the solvent would be separated in the interface, microporous tubular PTFE membrane desolvator to remove resulting in an attenuated solvent load in the plasma. The the organic solvents by a counter flow of argon sweep gas, NaDDC–APDC extraction system was selected owing to its while the aerosol particles passed through the center of the multi-element extraction capability.The effects of membrane membrane owing to low permeation and were conveyed to the desolvation temperature, oxygen and argon sweep gas flow plasma. The schematic configuration of the MEMSEP and details of its construction were provided by Botto and Zhu,23 rates were determined. 274 Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12Table 1 Instrumentation and ICP operating conditions Spectrometer Varian Ultramass Rf generator Varian 40.68 MHz Torch Demountable, positioned in coil with placement tool Injector Alumina, 1.8 mm id Plasma gas 15.5 l min-1 Auxiliary gas 1.1 l min-1 Oxygen bleed gas Varian AGM-1 oxygen supply manifold, 30 ml min-1 Rf power 1.2 kW Depth of sampling 10 mm for aqueous, 6 mm for organic solvents Sample delivery 1 ml min-1 Delivery and drain tubes Viton; orange for sample, violet for drain Aerosol carrier flow rate 0.4 l min-1 Stabilization time 5 s Wash-out time 200 s USN Cetac U 6000 AT Desolvation 140 °C Cooling Peltier, –10 °C MEMSEP CETAC MDX-100 Heating tube 60–160 °C Sweep gas 0.5–0.7 l min-1 aqueous; 2 l min-1 chloroform Brenner et al.30 and the Cetac instruction manual.33 MEMSEP operating conditions are listed in Table 1.Data were collected using peak hopping with three points per peak. The measurement time was 1 s, the dwell time 10 ms and the number of replicates three.The plasma was ignited in the usual way without difficulty. Oxygen was then added slowly to the auxiliary argon gas flow via a T-junction using a Varian (Palo Alto, CA, USA) AGM-1 oxygen delivery manifold (Table 2). Water was then rapidly replaced by propan-2-ol and then by chloroform. The plasma was slowly doped with oxygen until the violet–green C2 emission surrounding the base and periphery of the plasma and the sampler cone became invisible to the naked eye.The optimum flow rate was 30 ml min-1. During this procedure, operator observation of the sampler cone and the torch is essential; an inadequate oxygen flow resulted in the deposition of carbon on the sampler cone, while an excess amount caused its rapid corrosion. Fig. 1 Comparison of relative ion counts with and without the MEMSEP. USN attached, aqueous solution. RESULTS Optimization of the ICP-MS System 2–3% to approximately 0.1% and the Ce2+/Ce+ ratio was Aqueous solutions 0.6%. The reduced oxide ratios are due to low water loading and as a result the desolvated plasma contained more energy The Varian Ultramass was initially optimized with a concentric pneumatic nebulizer which was then replaced with the USN. which previously was needed to dissociate water molecules.The effect of the sweep gas flow rate on the ion counts of In comparison with conventional pneumatic nebulization, the ion intensities obtained with the USN increased by approxi- 24Mg+, 59Co+, 232Th+, 40Ar16O+ and 108Ce+ is illustrated in Figs. 2 and 3. The data in Fig. 2 indicate that the flow rate mately one order of magnitude, while the effect on the background was insignificant. This enhancement is due to higher required to obtain maximum intensity appeared to be mass dependent, the flow rate for 232Th+ exceeding those for 24Mg+ nebulization efficiencies. The USN–MEMSEP interface was then optimized with an aqueous solution containing 10 mg l-1 and 59Co+, implying that the zone of optimum ion intensity for 232Th+ was located lower in the plasma.These sweep gas Mg, Co, In, Ce, Th and Pb in 2% v/v nitric acid. With the MEMSEP attached, the relative ion counts of 24Mg+, 59Co+, flow rates are higher than those observed in a previous study when the MEMSEP was coupled with ICP-AES.30 With 115In+, 208Pb+ and 232Th+ increased by factors of 5–25 (Fig. 1). With a sampling depth of 10 mm, an rf power of 1.2 kW and increasing sweep gas flow rate, 40Ar16O+ decreased signifi- cantly, whereas CeO+/Ce+ passed through a maximum at aerosol and sweep gas flow rates of 0.5 and 0.7 l min-1, respectively, the CeO+/Ce+ ratio dramatically decreased from about 0.75 l min-1 (Fig. 3). Table 2 Plasma ignition routine with USN and MEMSEP attached Gas flow rates/l min-1 Rf power/ Plasma phase Plasma Auxiliary Nebulizer kW Time/s Purge 20 2 1 0 30 Delay 20 2 0 0 0 Ignition 18 2 0 0 10 Aspiration 15 1.5 0.8 1–1.2 10 Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12 275Fig. 2 Effect of sweep gas on 24Mg+, 59Co+ and 232Th+ intensities. Aqueous solution with USN–MEMSEP attached. Fig. 4 Effect of oxygen on 40Ar12C+, 108Ce+ and CeO+. USN– MEMSEP; oxygen flow rate 30 ml min-1; chloroform. Fig. 3 Variation of CeO+/Ce+ and 40Ar16O+ as a function of the argon sweep gas flow rate. Aqueous solution. Fig. 5 Effect of sweep gas flow rate on 24Mg+, 59Co+ and 232Th+ Chloroform intensities.Chloroform; oxygen flow rate 30 ml min-1; aerosol flow rate 0.4 l min-1. The MEMSEP was then optimized with a chloroform–acetone solution containing the analytes mentioned previously using Effect of MEMSEP desolvation temperature an aerosol flow rate of 0.4 l min-1. Despite desolvation, organic gases and vapors were observed in the plasma, and oxygen While numerous potential matrix induced interferences were was introduced into the auxiliary gas flow in order to prevent eliminated by solvent extraction (Table 3), numerous C and carbon formation in the torch and on the sampler cone.Cl interferences (Table 4) were observed when chloroform was injected into the MEMSEP at 65 and 160 °C. In particular, high background counts were observed at 51V+, 53Cr+, 56Fe+, Effect of oxygen 63Cu+ and 65Cu+ due to 35Cl16O+, 37Cl16O+, 40Ar16O+, The effect of oxygen while chloroform was aspirated was 35Cl12C16O+ and 37Cl12C16O+, respectively.The desolvation determined using an oxygen flow rate of 30 ml min-1 and a temperature was observed to be a controlling factor in the sweep gas flow rate of 2 l min-1. A comparison of the relative intensity of C and Cl molecular ions that interfere with these ion and oxide counts with and without oxygen is illustrated in Fig. 4. While the 140Ce+ count was not affected by adding Table 3 Potential interferences curtailed by applying solvent extrac- oxygen, the 40Ar12C+ and 40Ar12CH+ signals decreasedsharply tion techniques and membrane separation by a factor of 10, an important benefit for the determination of Cr using 52Cr+ and 53Cr+.The marked decrease in 40Ar12C+ Mass Potential molecular interference is due to the enhanced pyrolysis of chloroform vapor and 52Cr 40Ca12C aerosol. Notwithstanding the addition of oxygen to the plasma, 54Cr 40Ca14N the CeO+/Ce+ ratio increased only by a factor of 8, probably 55Mn 39K16O, 23Na32S 56Fe 40Ca16O, 39K16OH, 40Ar16O, 44Ca12C owing to oxygen consumption in chloroform pyrolysis. 58Ni 40Ca18O, 23Na35Cl, 26Mg32S, 44Ca14N 60Ni 44Ca16O, 25Mg35Cl, 23Na37Cl 61Ni 44Ca16OH, 24Mg37Cl, 23Na38Ar Effect of sweep gas 62Ni 24Mg38Ar, 25Mg37Cl The effect of the sweep gas flow rate on the relative ion counts 63Ni 26Mg37Cl, 24Mg38Ar, 23Na40Ar 64Ni 23Na218O of 24Mg+, 59Co+, 115In+, 208Pb+ and 232Th+ is illustrated in 59Co 24Mg35Cl, 40Ca18OH, 43Ca16O Fig. 5. As in the case of the aqueous solution, the flow rates 63Cu 26Mg37Cl, 23Na40Ar required to obtain maximum intensity for high mass 208Pb+ 64Zn 24Mg40Ar and 232Th+ were higher than those for 24Mg+, 59Co+ and 66Zn 26Mg40Ar 115In+ when chloroform was aspirated. 276 Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12Table 4 Typical interferences due to the presence of chloroform and organic vapors in the aerosol delivered to the plasma Mass Molecular interference 51V 35Cl16O, 32S18OH 50Cr 32S18O 52Cr 40Ar12C, 35Cl16OH, 53Cr 37Cl16O, 40Ar12CH 54Cr 35Cl18OH 60Ni 28Si32S 63Cu 28Si35Cl, 12C16O35Cl 65Cu 28Si37Cl, 12C16OH37Cl 66Zn 12C16OH37Cl trace metal analytes.For example, Fig. 6 shows the influence of MEMSEP temperature on 40Ar12C+, 108Ce+ and CeO+ counts when pure chloroform was aspirated. Ce in the inorganic form was employed in order to avoid the possibility of volatilization from the membrane. With increasing temperature from 90 to 160 °C, 40Ar12C+ decreased by an order of magnitude owing to the removal of chloroform at high temperature, Fig. 7 Variation of 35Cl16O+ and 40Ar12C+ as a function of MEMSEP whereas 108Ce+ and CeO+ remained essentially constant, the desolvation temperature. Chloroform, chloroform–oil and latter owing to the absence of water. chloroform–DTCs. Sweep gas flow rate 2 l min-1; aerosol flow rate In an attempt to devise a universal method of calibration, 0.4 l min-1. based on a chloroform solution of oil-soluble metal organic compounds, it was necessary to examine the effect of the desolvation temperature on the signals of several polyatomic ions and analyte masses, and to determine whether the responses are similar for the oil, chloroform and dithiocarbamate matrices.Fig. 7 shows that the intensities of the 35Cl16O+ and 40Ar12C+ (and 40Ar12CH+) ions produced by nebulizing dithiocarbamate–chloroform and oil–chloroform solutions decreased by 1–3 orders of magnitude with increasing temperature, the decrease being more prominent between 80 and 90 °C.It should be noted, however, that the signal reductions were not identical for all the matrices. Fig. 8 shows the abundance corrected net ion counts for an oil–chloroform solution using MEMSEP temperatures of 65 and 160 °C, a 2 l min-1 sweep gas flow rate and the conditions listed in Table 1. In general, the ion signals for 160 °C were lower than those obtained at 65 °C, evidently owing to the volatility of the metal organic compounds used for the prep- Fig. 8 Background corrected and abundance normalized intensities aration of these standards. In detail, Fig. 8 indicates that the for a chloroform–oil solution. MEMSEP temperatures 65 and 160 °C; 65-to-160 °C signal ratios can be divided into four groups: (a) sweep gas flow rate 2 l min-1; aerosol flow rate 0.4 l min-1. Note 66Zn+ and 111Cd+ minima. light elements, 51V+, 52Cr+, 53Cr+, 55Mn+,58Ni+, 60Ni+, 63Cu+, 65Cu+, 56Fe+ and 57Fe+, with count ratios varying from 15 to about 25; (b) 95Mo+, 96Mo+, 107Ag+ and 109Ag+, ratios; and (d) heavy elements, 209Bi+ and 208Pb+, with ratios with ratios of about 5; (c) 66Zn+ and 111Cd+, which have low of about 1.In order to distinguish between ionization and possible volatilization processes, these ratios were plotted as a function of the first ionization potentials (IP) (Fig. 9). Although there is some spread in the relationship between the ratios and IP, Fig. 9 shows that under conditions of moderate plasma overloading, the ion signal of elements that have higher IP (111Cd+ and 66Zn+) are more depressed than those that have lower IP.The position of 95Mo+, 107Ag+, 209Bi+ and 208Pb+ may reflect their volatilities. The effect of desolvation temperature on the signals of the metal chelates was evaluated by nebulizing a DDTC–NaDDC extract in chloroform (Fig. 10). Noteworthy was the moderate decrease in the 208Pb+ signal and the large decrease in 60Ni+ counts at 60–100 °C.The similar behavior of the other analytes indicates that any one could be employed as an internal standard to compensate for these variations in the USN– MEMSEP interface. It is concluded that metal dithiocarbamates in chloroform are volatilized in the heating stages of the interface. Analyte losses in thermal desolvation devices, especially for those that have high vapor pressures, have been Fig. 6 Effect of MEMSEP desolvation temperature on 40Ar12C+, observed frequently in previous studies, e.g., on Hg, B, Os and 108Ce+ and CeO+ intensities.Chloroform; sweep gas flow rate 2 l min-1; aerosol flow rate 0.4 l min-1. Re in aqueous solutions and metal dithiocarbamate and keton- Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12 277ate complexes.34 Fig. 11 compares the ion count ratios obtained spiked water samples and the Canadian sea-water standard NAAS-2. at 165 and 65 °C for a 20 mg kg-1 chloroform–oil solution and a dithiocarbamate–chloroform extract.It is evident that the responses for the analytes in these media differ, thus Limits of Detection explaining the unsatisfactory recoveries of these elements from The 3s LODs are listed in Table 5. The LODs of 51V+ and 52Cr+ were constrained by background interferences induced by 35Cl16O+ and 40Ar12C+. The degraded LOD of 56Fe+ was due to 40Ar16O+ and those of 66Zn+ and 63Cu+ were due to contamination and interferences from 12C16OH37Cl and 12C16O35Cl.In the case of 63Cu+ and 65Cu+, the background increase might also be due to 28Si35Cl+ formed by thermal ablation of the quartz torch. 118Sn+ showed interference by 86Kr16O+2. Short-term Variations The short-term variation was determined by aspirating an oil– chloroform solution into the MEMSEP and recording the intensities over a period of about 120 min (Fig. 12). In general, the accuracy for the 10 mg g-1 standard varied from -30 to 50%, for 40 mg kg-1 it varied from -28 to 40% and for 100 mg kg-1 it varied from -22 to 28%.More than one internal standard would be required to compensate for these Fig. 9 Relationship between the first ionization potential and the ion variations. count ratios obtained at MEMSEP temperatures of 65 and 160 °C. Chloroform–oil solution; sweep gas flow rate 2 l min-1; aerosol flow rate 0.4 l min-1; oxygen flow rate 30 ml min-1. Table 5 LODs (3s) in chloroform. Data in mg kg-1 Element LOD/mg kg-1 51V 0.75 52Cr 0.17 53Cr 2 55Mn 0.011 56Fe 0.12 57Fe 0.20 58Ni 0.042 60Ni 0.015 63Cu 0.20 65Cu 0.22 66Zn 0.30 95Mo 0.012 96Mo 0.007 107Ag 0.011 109Ag 0.018 111Cd 0.015 118Sn 0.16 205Tl 0.012 208Pb 0.002 Fig. 10 Effect of MEMSEP desolvation temperature on the ion 209Bi 0.003 counts for trace metal dithiocarbamates in chloroform. Sweep gas flow rate 2 l min-1; aerosol flow rate 0.4 l min-1; oxygen flow rate 30 ml min-1 . Fig. 11 Response of ion counts obtained at 165 and 65 °C for a case, the MEMSEP would be located between the HPLC column and the ICP.The MEMSEP described in this paper fulfils the requirements of such an interface, namely small memory effects, minimum effects of the mobile phase on the plasma and phase flow rates of 1–2 ml min-1. The technique would also have potential for determining metal organic species in the environment where the integrity of the complex can only be assured by using solvent extraction procedures. Hence we can expect a resurgence of solvent extraction techniques with the use of this USN–MEMSEP interface and ICP-MS. REFERENCES 1 Cresser, M.S., Solvent Extraction in Flame Spectroscopic Analysis, Butterworth, London, 1978. 2 Skougstad, M. W., Fishman, M. J., Friedman, L. C., Erdmann, D. E., and Duncan, S. S., in T echniques of Water-Resources Fig. 13 Memory effect in the USN–MEMSEP interface. Chloroform Investigations of the United States Geological Survey, US dithiocarbamate extract (DTC) and oil–chloroform (oil) solution con- Geological Survey.Washington, DC, 1978, ch. A1. taining 100 mg kg-1 of trace elements. Desolvation temperature 65 °C; 3 Boorn, A., and Browner, R. F., Anal. Chem., 1982, 54, 1402. seep gas flow rate 2 l min-1; aerosol flow rate 0.4 l min-1; oxygen 4 Tang, Y. Q., Du, Y. P., Shao, J. C., Tao, W., and Zhu, M. H., flow rate 30 ml min-1. Spectrochim. Acta, Part B, 1992, 47, 1353. 5 Maessen, F. J. M. J., Kreunig, G., and Balke, J., Spectrochim. Acta, Part B, 1984, 41, 3.Memory Effects 6 Blades, M. W., and Caughlin, B. L., Spectrochim. Acta, Part B, 1985, 40, 579. The memory effect due to possible deposition of analytes in 7 Weir, D. G., and Blades, M. W., J. Anal. At. Spectrom., 1994, the interface and MS cone was determined using a tuned and 4, 1323. calibrated instrument. A 100 mg l-1 mixed trace element oil 8 Boorn, A. W., Cresser, M. S., and Browner, R. F., Spectrochim. Acta, Part B, 1980, 35. 823 standard and a dithiocarbamate extract, both in chloroform, 9 Ohls, K.D., Flock, J., and Loepp, H., ICP Inf. Newsl., 1988, 14, 83. were aspirated for 120 s. A blank solution was then introduced 10 Magyar, B., Lienemann, P., and Vonmont, H., Spectrochim. Acta, after 60 s and analyzed after taking into account the time Part B, 1986, 41, 27. needed for sample uptake. Fig. 13 indicates that the ion count 11 Hausler, D. W., and Taylor, L. T., Anal. Chem., 1981, 53, 1223. decreased by almost three orders of magnitude after 60 s.This 12 Brotherton, T. J., Pfannerstill, P. E., Creed, J. T., Heitkemper, pattern is similar to our previous observations using ICP- D. T., Caruso, J. A., and Pratsinis, S. E., J. Anal. At. Spectrom., 1989, 4, 341. AES.30 13 Botto, R. I., J. Anal. At. Spectrom., 1993, 8, 51. 14 Wiederin, D. R., Houk, R. S., Winge, R.K., and D’Silva, A. P., Anal. Chem., 1990, 62, 1150 CONCLUSIONS 15 Alves, L. C., Wiederin, D. R., and Houk, R. S., Anal. Chem., 1992, The aim of this work was to evaluate interference effects 64, 1164. 16 Alves, L. C., Minnich, M. G., Wiederin, D. R., and Houk, R. S., observed when metal chelates and volatile organic solvents are J. Anal. At. Spectrom., 1994, 9, 399. introduced into a USN–MEMSEP ICP-MS interface. On the 17 Bradford, G. R., and Bakhtar, D., Environ. Sci. T echnol., 1991, one hand, the use of high desolvation temperatures results in 25, 1704. the efficient removal of interfering organic solvent vapors. On 18 McLeod, C.W., Otsuki, A., Okamoto, K., Haraguchi, H., and the other hand, operation at high temperature results in the Fuwa, K, Analyst, 1981, 106, 419. partition of volatile chelatesinto the USN–MEMSEP interface, 19 Zhuang, Z., Wang, X., Yang, P., Yang, C., and Huang, B., J. Anal. At. Spectrom., 1994, 9, 779. depending on the lability of the complexes. In this respect, oil- 20 Gustavsson, A., Spectrochim. Acta, Part B, 1987, 42, 111. based standards appear to be unsuitable for universal cali- 21 Backstrom, K., Gustavsson, A., and Hietala, P., Spectrochim.bration for the determination of the metal organic species. Acta, Part B, 198, 44, 104. Because of the volatility of the metal complexes, even at 22 Tao, H., and Miyazaki, A., J. Anal. At. Spectrom., 1995, 10, 1. moderate temperatures, the organic solvent used for extraction 23 Botto, R. I., and Zhu, J. J., J. Anal. At. Spectrom., 1994, 9, 905. should have a very low boiling-point. Despite these difficulties, 24 Wyttenbach, A., and Bajo, S., Anal.Chem., 1975, 47, 1813. 25 Kinrade, J. D., and Van Loon, J. C., Anal. Chem., 1974, 46, 1894. internal standards similar in thermal behavior can be used to 26 Brooks, R. R., Presley, B. J., and Kaplan, I. R., T alanta, 1967, compensate for these losses. We are currently evaluating 14, 809. alternative calibration strategies using solvents such as hexane 27 Kojima, I., Inagaki, K., and Kondo, S., J. Anal. At. Spectrom., to avoid interferences from 35Cl16O+ and calibration standards 1994, 9, 1161. produced by solvent extraction, and the addition method. 28 Koirtyohann, S. R., and Wen, J. W., Anal. Chem., 1986, 45, 1973. Even though a large amount of the solvent can be removed 29 Sturgeon, R. E., Berman, S. S., Desaulnier, A., and Russell, D. S., T alanta, 1980, 27, 85. in the USN–MEMSEP, residual organic vapors and aerosols 30 Brenner, I. B., Zander, A., and Zhu, J., Fresenius’. J. Anal. Chem., are introduced into the plasma. Hence a controlled flow of 1996, 355, 774. oxygen is necessary to enhance pyrolysis, minimize polyatomic 31 Motooka, J. M., Appl. Spectrosc., 1998, 42, 1293. interferences and prevent the deposition of carbon on the torch 32 Viets, J. G., Anal. Chem., 1978, 50, 1097. tubes and sampler cone. By removing matrix elements such as 33 Cetac MDX 100 Membrane Desolvator Instruction Manual, 1995, alkali and alkaline earth metals, salt effects are eliminated, and Cetac, Omaha, NE. 34 Castillo, J. R., Delfa, J., Mir, J. M., Bendicho, C., De la Guardia, analytes can be preconcentrated provided that their behavior M., Mauri, A. R., Mongay, C., and Martinez, E., J. Anal. At. on the membrane interface is well understood. Optimization Spectrom., 1990 5, 325. of the measurement and extraction systems could result in at least 100-fold concentration factors, depending on the amount Paper 6/05533H of sample used and blank contamination. Received August 8, 1996 The technique will be highly suited to HPLC methods which Accepted December 5, 1996 require volatile organic solvents as the mobile phase. In this Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12 279

ISSN:0267-9477    

DOI:10.1039/a605533h

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Comparison of Axially and Radially Viewed Inductively Coupled Plasma Atomic Emission Spectrometry in Terms of Signal-to-Background Ratio and Matrix Effects Plenary Lecture† CENDRINE DUBUISSON, EMMANUELLE POUSSEL AND JEANMICHEL MERMET* L aboratoire des Sciences Analytiques (UMR 5619), Universite� Claude Bernard, L yon I, 69622 V illeurbanne Cedex, France A comparison was made of axial and radial viewing modes in of ICP-AES systems, improvements were recently observed for terms of the signal-to-background ratio for several most figures of merit,1 particularly the LOD.commercially available ICP-AES systems. The average It is convenient to use the SBR-RSDB approach derived by improvement factor was found to be in the range of 2–3 for Boumans2 to discuss any possible improvement in the LODs. ionic lines when the axial viewing mode was used. To obtain The approach is based on the use of ratios such as the signalhigher improvement factors, other modifications of the ICP to-background ratio, SBR, and the relative standard deviation systems must be associated with axial viewing, such as better of the fluctuations of the background, RSDB .The LOD is: practical resolution of the dispersive system and better efficiency of the sample introduction system. Matrix effects LOD =k.c. RSDB SBR (1) with axial viewing, such as those resulting from the presence of Na, can be minimized to the same extent as for radial viewing provided that a high rf power and a low carrier gas flow rate where k is a statistical constant, usually equal to 3, and c is are used to optimize the exchange of energy between the the concentration of the analyte. From this relationship, it is surrounding plasma and the central channel. An increase in the simple to deduce that it is possible to improve the LOD by id of the torch injector is also beneficial for this purpose.increasing the value of the SBR without increasing the value Keywords: Inductively coupled plasma; atomic emission of the RSDB.Although the SBR is directly related to the spectrometry ; axial viewing; radial viewing; signal-to- sensitivity rather than to the LOD, its value is widely used background ratio; matrix effects because the SBR is dimensionless and is independent of the detection system (e.g. the voltage or the amplifier gain). The SBR value is, therefore, used as an easy means of comparison Improving LOD has always been a challenge for the analyst. between various ICP-AES systems.1 SBR values can be nor- The naturally occurring concentration of many elements malized to a concentration of 1 mg l-1 and are equal to the requires LOD in solids or solutions of below the ppm reciprocal value of the background equivalent concentration or ng ml-1 level, respectively. Lowering the LOD also results (BEC), which is also widely used in ICP-AES.in lower limits of quantitation, i.e. the lowest concentration An increase in the SBR can be achieved through use of a obtained for a given precision. more efficient sample introduction system (better selection of Among the various analytical methods, ICP-AES has gained the pneumatic nebulizer and the spray chamber, ultrasonic wide acceptance because of its ease of use, its multi-element nebulizer associated with a desolvation system, and thermo- capability, its high sample throughput, its tolerance to high spray system), through better characteristics of the hf generator salt concentrations and the relatively few matrix effects.LOD (frequency and waveform), by improving the practical reso- are generally at the ng ml-1 level for most elements. However, lution of the dispersive system, and by changing the conven- ICP-AES systems that were commercially available a few years tional mode of observation, named radial, lateral or side-on. ago exhibited LOD which were unsuitable for some appli- The use of axial viewing (or end-on) as an alternative to the cations.For example, for lead the LOD was in the range radial viewing of the plasma was first described in 1976.3,4 30–50 ng ml-1, whereas a value of 3 ng ml-1 is currently One important feature of the ICP is the confinement of the required by the Contract Laboratory Program of the US sample in a central channel surrounded by the argon plasma. Environmental Protection Agency, EPA method 200.7. Conventional radial observation results in an intense and The use of the ICP as an ionization source for mass unwanted Ar background around the central channel and the spectrometry could solve the problem of insufficiently low need to optimize the observation height.The aim of axial LOD. However, the investment cost is higher than that of viewing is to improve the efficiency of observation of the ICP-AES, and it is generally considered that more expertise is central channel, while avoiding the surrounding intense Ar required for ICP-MS.There was, therefore, a need for a system plasma. It could be expected that the signal would increase that exhibits LODs between the conventional ICP-AES and and the background would decrease. The first publications the ICP-MS systems. Because of a strong and active compeclaimed that a significant improvement was observed in terms tition between the manufacturers involved in the production of LOD.3–6 Since then, several papers have been published on the use of axial viewing either for AES7–16 or for AAS17–19 † Presented at the Eighth Biennial National Atomic Spectroscopy Symposium (BNASS), Norwich, UK, July 17–19, 1996.where such a use seems logical. Axial viewing has also been Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12 (281–286) 281Table 1 Pneumatic nebulizers, spray chambers, and uptake rates used for the observation of the ball-shaped plasma formed by (ml min-1) used with the Perkin-Elmer Optima 3000 DV ICP system a low-flow torch.20,21 The use of axial viewing implies some technological modifi- Uptake rate/ Observation cations.An interface must be designed to eliminate the hot Pneumatic nebulizer Spray chamber ml min-1 height/mm* air–argon unstable recombination zone formed at the tip of Perkin-Elmer cross-flow Double-pass 1 4 the plasma so as to improve light transmission, and to prevent Perkin-Elmer conespray Double-pass 1 5 damage of the optical components (lens or mirror).Moreover, Perkin-Elmer V-71 Cyclonic 1.5 7 the collimating system must also be modified to cope with the Cetac MCN† Double-pass 1 10 Cetac MCN† Double-pass 0.04 10 central channel depth and circular shape, i.e. a long depth of focus and a change in the optics because of the tall, narrow * Observation height above the load coil was selected so as to entrance slit. Several designs are currently used to remove the obtain the maximum value of the Mg II 280 nm5Mg I 285 nm line unstable zone: (i ) removal of the hot air by means of a powerful intensity ratio.fan system; (ii) use of a shear gas;4,6 and (iii ) use of counter † Concentric micronebulizer designed for low uptake rates. gas flow. A consequence of the use of axial viewing is that no observation height adjustment is necessary. However, an accurate manual or motor-driven centring of the torch with respect evaluated with similar Glass Expansion concentric pneumatic nebulizers and Sturman-Masters spray chambers.to the axis of observation is usually required. Most manufacturers involved in the production of ICP-AES The values of the power and the carrier gas flow rate were adjusted according to the experiments and will be mentioned systems have introduced at least one system based on the use of axial viewing. It is claimed that the overall improvement in below. Besides these three systems for which detailed experiments terms of LOD can reach one order of magnitude. However, the use of axial viewing is generally combined with some other have been performed, other commercially available ICP systems have been used for the evaluation of the SBR values. changes in the system such as a better resolution or a more efficient sample introduction system.It is therefore important These include the Baird 2070, Jobin-Yvon JY 38 Plus, JY 38S, JY 138 and JY Panorama, Fisons-ARL 3520, Perkin-Elmer to evaate the actual contribution of axial viewing to the LOD improvement through the SBR value.This is the purpose P2000, Unicam PU 7000, Spectro Modula, and Thermo Jarrell Ash 25, for radial viewing and the Fisons Maxim, Thermo of this work. The improvement in LOD is interesting from an analytical Jarrell Ash 61E Trace, and Spectro EOP, for axial viewing. This selection can be considered as being representative of the point of view if this can be achieved without degradation of the analytical performance, in particular the occurrence of ICP market over the past and present years.Usually, the operating conditions were optimized to obtain the best SBR adverse matrix effects. Axial viewing has a rather bad reputation since the matrix effects that may occur in the atomization values. Results concerning the SBR values were normalized to a zone within the coil are incorporated by the observation system. A significant amount of time was found to be necessary concentration of 1 mg l-1 and are summarized for three elements, Ni II 231 nm, Cd II 226 nm and Pb II 220 nm. Ni to optimize the power and the carrier gas flow rate to obtain robust conditions with radial viewing, i.e.conditions where a was selected because the Ni experiment was a simplified version of the diagnostic experiment based on the use of an LOD change in the matrix concentration does not result in a significant variation in the analyte signal. It is, therefore, scale.1 Actually, the sum of ionization and excitation energies for the Ni II 231 nm line, 14.01 eV, is located within the energy interesting to verify the actual level of effects due to the presence of elements at high concentrations, in particular the range (11–15 eV) used for the LOD scale.1,22 Both Pb and Cd were selected because of their importance as heavy elements.easily ionizable elements, or reagents such as acids, when axial viewing is used. Moreover, the LOD of Pb is one of the worse for ICP-AES. Although similar conclusions would be found for most atomic lines (e.g.Mg I 280 nm or P I 213 nm), the optimization of the operating parameters, power and carrier gas flow rate, can be EXPERIMENTAL significantly different from the other lines for lines such as Al I 396 nm. For instance, using the Optima 3000 DV system, the Three ICP systems were used for the comparison of the SBR values and the matrix effects, a Perkin-Elmer Optima 3000 maximum value of the SBR of the Al I 396 nm line was obtained for a carrier flow rate of 0.95 l min-1, in both viewing DV, A Varian Liberty 220 and a Varian Liberty 150 AX Turbo.The Perkin-Elmer ICP system is equipped with a 40 MHz modes, instead of 0.6–0.55 l min-1 for the other lines, as seen below. It is therefore difficult to make a fair comparison of the generator and an echelle grating-based simultaneous dispersive system associated with an SCD (segmented-array charge- influence of the viewing mode for some atomic lines if compromise operating conditions are used.Therefore, the study of coupled device) detector. The Perkin-Elmer system makes it possible to use either the axial viewing or the radial viewing atomic lines such as Al I 396 nm was not considered in the present work. mode by means of a periscope. A fair comparison of the possibilities of axial viewing was thus possible. The system is The magnitude of the matrix effects due to the presence of Na at a concentration of 10 g l-1 and 1 g l-1 was evaluated equipped with a shear gas to eliminate the unstable zone.A 2 mm id injector was used. Several pneumatic nebulizers and by using simple experiments based on the use of five ionic lines: Ba II 455 nm (energy sum of 7.93 eV), Ba II 233 nm spray chambers were evaluated (Table 1). The SBR values were measured using the area mode, i.e., the sum of the signals (11.22 eV), Mg II 280 nm (12.07 eV), Ni II 231 nm (14.03 eV) and Zn II 206 nm (15.40 eV). Their sum of their ionization produced by three adjacent pixels of the SCD detector. The use of the peak mode, i.e., a scanning of the line profile by and excitation energies covers almost the entire range of energy of the ionic lines, and their relative behaviour, particularly moving the entrance slit could slightly enhance the SBR value.Both Varian ICP systems are equipped with a sequential that of the Ba II 233 nm line was helpful in explaining the possible origin of the effect.23 For instance, it was found that dispersive system and a 40 MHz generator.The axial-viewing based Liberty 150 AX Turbo system is equipped with a counter an effect leading to a flat response of the variation in the line intensities could be assigned mainly to the aerosol formation gas flow system through a cone. The id of the injector is 2.3 mm in contrast to that of the radial-viewing based Liberty and transport system.23 In other words, the matrix effects due to the plasma were minimized. 220 system which is only 1.4 mm. The two systems were 282 Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12Table 2 SBR values obtained for a concentration of 1 mg l-1 and using the Ni II 231 nm, Cd II 226 nm and Pb II 220 nm lines, for axial and radial viewing, and the axial/radial SBR ratio* Ni SBR Cd SBR Pb SBR Pneumatic nebulizer Axial Radial Ni ratio Axial Radial Cd ratio Axial Radial Pb ratio Perkin-Elmer cross-flow 6.7 2.6 2.6 16.4 9.2 1.8 1.8 0.6 2.9 Perkin-Elmer conespray 3.0 1.7 1.8 9.1 6.4 1.4 0.8 0.4 2.0 Perkin-Elmer V-71 7.5 5.3 1.4 22.8 16.5 1.4 2.2 1.3 1.7 Cetac MCN, 1 ml min-1 8.5 4.8 1.8 21.4 14.7 1.5 2.3 1.2 1.9 Cetac MCN, 40 ml min-1 4.6 2.3 2.0 12.7 7.2 1.8 1.3 0.6 2.1 * A Perkin-Elmer Optima 3000 DV ICP system was used.The carrier gas flow rate was 0.5 and 0.55 l min-1 for radial and axial viewing, respectively. RESULTS AND DISCUSSION viewing. The Spectro EOP system developed for axial viewing uses an id of 2.5 mm instead of 1.7 mm for the conventional Comparison of the SBR Values for the Perkin-Elmer Spectro ICP systems.It is interesting to note that the SBR ICP System measurements were reproduced with seven Varian Liberty 220 systems belonging to customers. Although the systems were As mentioned above, the Optima allowed the determination of the SBR values using either axial viewing or radial viewing equipped with different nebulizers, the SBR values remained within the range 6–11.5, which is good evidence of the on the same system.The results, therefore, led to a fair evaluation of the improvement resulting from the use of the homogeneity of the various systems. Clearly, the use of axial viewing cannot result in an improve- axial viewing. A power of 1100W was used for both observation modes. The observation height was optimized (Table 1) ment in the SBR values of an order of magnitude. A factor of 3–5 can be considered as excellent. Some preliminary investi- to obtain the maximum value of the Mg II 280 nm5Mg I 285 nm line intensity ratio, while the carrier gas flow rate was gations using an ICP system equipped with a high resolution dispersive system have shown that the improvement due to optimized by optimizing the SBR values.Values of 0.6 and 0.55 l min-1 were found to be the most suitable for radial and axial viewing can be of a factor of 2 at the very best. This can be explained by the fact that the contribution of the back- axial viewing, respectively.Results are summarized in Table 2 for the set of nebulizers and spray chambers given in Table 1. ground has been already minimized when a high resolution is used.24,25 In most cases several nebulizers of the same type were available, the results presented are those obtained for the best nebulizer of each type. From Table 2, it can be seen that the average Comparison of the SBR Values for a Large Set of improvement factor is around 1.9, 1.6 and 2.1 for Ni, Cd and Commercially Available ICP Systems Pb, respectively.This result is similar to those previously published for the same ICP system.15 It is interesting to note Based on the use of the commercially available systems menthat a factor of 3 can be observed between various nebulizers, tioned previously, SBR values were obtained for the three this factor being larger than the improvement factor obtained elements, Ni, Cd, and Pb. Results are summarized in Figs. 1, by using axial viewing. In any case, the improvement of the 2 and 3, for Ni, Cd, and Pb, respectively. The range of the SBR values obtained with the Optima ICP system results from actual SBR values for both radial and axial viewings is given. a more important increase in the signal than in the background From these figures, it can be seen that a large range of SBR as both signal and background increased. When LOD are values could be observed for each mode. Obviously, a compariconcerned, it is obvious that a careful selection of the nebulizer son of the best SBR value obtained with axial viewing with and the spray chamber must be carried out.Note also that the worse SBR value obtained with radial viewing, results in even when a low uptake rate of 40 ml min-1 was used, the a large improvement factor, i.e., 38, 17 and 37 for Ni, Cd and MCN nebulizer gave a suitable SBR value. The only constraint Pb, respectively. However, it is important to note that the use of the MCN nebulizer used in this experiment was the need of axial viewing does not necessarily imply an improvement in to use a high carrier gas back pressure, e.g.an absolute pressure the SBR value. Some ICP systems still using radial viewing of 4.3 bar for a flow rate of 0.7 l min-1. Comparison of the SBR Values for the Two Varian ICP Systems Results are summarized in Table 3. In this case, the improvement factor was about 5. But it is important to note that a major modification had to be made to the ICP system equipped with axial viewing: the id of the injector was dramatically increased from 1.4 to 2.3 mm.This is actually a trend for axial Table 3 Comparison of the SBR values obtained using the Varian Liberty 220 ICP system (radial viewing) and the Varian Liberty 150 AX Turbo ICP system (axial viewing) for Ni II 231 nm, Cd II 226 nm and Pb II 220 nm Element* SBR radial SBR axial Ratio Ni 9 40 4.4 Fig. 1 SBR values observed for a large range of commercially avail- Cd 25 120 4.8 Pb 3.5 20 5.7 able ICP-AES systems using axial viewing or radial viewing. The SBR values were obtained using the Ni II 231 nm line and were normalized to a concentration of 1 mg l-1.* 1 mg l-1 of each. Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12 283leading to a plasma state close to robustness, the variation in the analyte signal becomes similar for each ionic line, and a slightly depressive, flat response is usually observed.This remaining effect may be assigned to the aerosol formation and transport. The use of non-robust conditions, i.e. a low power and a high carrier gas flow rate, permits the ICP user to enhance the effect due to the plasma processes and to facilitate its study. Therefore, the use of robust and non-robust conditions was compared for the radial and axial viewings for the Perkin-Elmer ICP system and the Varian ICP systems. Operating conditions are summarized in Table 4.The concentration of the elements was 1 mg l-1. Since the trends were similar for each type of nebulizer, the results are given in Fig. 4 for the Perkin-Elmer ICP system equipped with the MCN nebulizer (with an uptake rate of 1 ml min-1) under non-robust conditions and robust conditions using radial viewing. The use of robust conditions Fig. 2 SBR values observed for a large range of commercially available ICP-AES systems using axial viewing or radial viewing. The SBR minimized the effect of Na.A further dilution of Na to 1 g l-1 values were obtained using the Cd II 226 nm line and were normalized made the effect almost negligible. Similar results were obtained to a concentration of 1 mg l-1. when axial viewing was used (Fig. 5). The only difference was that, under non-robust conditions, every line was depressed, with the exception of Ba II 233 nm. By enlarging the variation scale (Fig. 6) it can be seen that the magnitude of the Na effect is similar irrespective of the viewing mode and is close to unity when a Na concentration of 1 g l-1 is used.It should be added that the %RSD of the repeatability was in the range 5–10%, which means that the two curves in Fig. 6 are not significantly different. These results are in good agreement with similar experiments.15 The same conclusions were reached when the two Varian systems were used. Results are summarized in Figs. 7 and 8 where the operating conditions given in Table 4 were used with an Na concentration of 10 g l-1.Under non-robust Table 4 Operating conditions leading to robust and non-robust conditions for the study of the Na effect on ionic lines Fig. 3 SBR values observed for a large range of commercially avail- ICP system Robust conditions Non-robust conditions able ICP-AES systems using axial viewing or radial viewing. The SBR Perkin Elmer Optima 1500 W, 0.6 l min-1 900 W, 1.0 l min-1 values were obtained using the Pb II 220 nm line and were normalized 3000 DV to a concentration of 1 mg l-1.Varian Liberty 220 1500 W, 0.4 l min-1 1000 W, 0.8 l min-1 Varian Liberty 150 1400 W, 0.6 l min-1 1000 W, 1.0 l min-1 AX Turbo provide SBR values which are better than some systems based on the use of axial viewing. This can be explained by improvements brought to the resolution, the generator design, and the sample introduction system. Actually, a comparison of the best SBR values obtained by using axial viewing with those obtained with radial viewing leads to an improvement factor in the range of 2–3, which confirms the values given above.Note that the improvement factor was 4.2 and 2.6 for Cd and Ni, respectively, in the work of Nakamura et al.14 In contrast, the improvement factor in the SBR was rather poor or negligible in the work of Demers (1 and 1.5 using the Cd II 226 nm and Pb II 220 nm lines, respectively). The improvement in the LOD was 7.5 and 4 for Cd and Pb, respectively, which can only be assigned to an improvement in the RSDB value.Matrix Effects Due to the Presence of Na It has been shown that the effect of Na on ionic lines is not necessarily the same over a large range of energy sum.23 Actually, the effect depends on the power and the carrier gas Fig. 4 Effect of Na on the intensity of ionic lines plotted as a function of the energy sum (eV) of the ionic lines. The following lines were flow rate. The concept of robustness is used to describe the selected: Ba II 455 nm (energy sum of 7.93 eV), Ba II 233 nm (11.22 eV), capability of the plasma to accept achange in the concentration Mg II 280 nm (12.07 eV), Ni II 231 nm (14.03 eV) and Zn II 206 nm of the matrix (or a reagent) without producing a significant (15.40 eV).The Perkin-Elmer Optima 3000 DV ICP system was used variation in the analyte signal.23 Robustnesscan be approached with the radial viewing mode. A MCN nebulizer with an uptake rate when high power and low carrier gas flow rate are used.23 The of 1 ml min-1 was used.($) 10gl-1 of Na; 900 W; carrier gas, value that is required for the carrier gas flow rate depends 1.0 l min-1; (&) 10 g l-1 of Na; 1500 W; carrier gas, 0.6 l min-1; and (2) 1 g l-1 of Na; 1500 W; carrier gas, 0.6 l min-1. strongly on the id of the injector.23 Under operating conditions 284 Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12Fig. 8 Effect of Na on the intensity of ionic lines plotted as a function Fig. 5 Same study as Fig. 4, except that the Perkin-Elmer Optima 3000 DV ICP system was used with the axial viewing mode. of the energy sum (eV) of the ionic lines. The following lines were selected: Ba II 455 nm (energy sum of 7.93 eV), Ba II 233 nm (11.22 eV), Mg II 280 nm (12.07 eV), Ni II 231 nm (14.03 eV) and Zn II 206 nm (15.40 eV). The Varian Liberty 220 ICP system was used with the axial viewing mode. ($) 10 g l-1 of Na; 1000 W; carrier gas, 1.0 l min-1; (&) 10 g l-1 of Na; 1400 W; carrier gas, 0.6 l min-1; and (2) 1 g l-1 of Na; 1400 W; carrier gas, 0.6 l min-1.id of the injector. It has been observed that an increase in the id of the injector corresponds to a longer residence time and, probably, to a larger zone of energy exchange between the central channel and the surrounding plasma.23,26 However, there was no significant difference due to the dilution of the Na matrix (Fig. 8). These experiments indicate that the magnitude of the Na effects can be minimized to the same extent with both viewing modes provided that robust conditions, i.e.high power and low carrier gas flow rate, are used. The current trend to increase the id of the injector seems to result in a reduction in Fig. 6 Comparison of the effect of Na (1 g l-1) under robust con- matrix effects. ditions (1500 W, 0.6 l min-1) with axial ($) and radial viewing (&) Similar conclusions were obtained when studying the effect using the Perkin-Elmer Optima 3000 DV ICP system.of an acid such as HNO3. Under robust conditions and for an acid concentration of 20% (v/v), both axial and radial viewings led to a similar, usual depressive effect (Fig. 9). Note that %RSD was about 3%. There was, therefore, no drastic difference in the effect, regardless of the viewing mode. CONCLUSIONS Although it is commonly believed that the use of axial viewing can result in a dramatic increase in SBR values, and Fig. 7 Effect of Na on the intensity of ionic lines plotted as a function of the energy sum (eV) of the ionic lines. The following lines were selected: Ba II455 nm (energy sum of 7.93 eV), Ba II233 nm (11.22 eV), Mg II 280 nm (12.07 eV), Ni II 231 nm (14.03 eV) and Zn II 206 nm (15.40 eV). The Varian Liberty 220 ICP system was used with the radial viewing mode. ($) 10gl-1 of Na; 1000 W; carrier gas, 0.8 l min-1; and (&) 10 g l-1 of Na; 1500 W; carrier gas, 0.4 l min-1.conditions the depressive effect observed with axial viewing was more pronounced than that observed with the Perkin- Elmer system. It is interesting to note that, under robust Fig. 9 Comparison of the effect of HNO3 (20% v/v) on the intensity conditions the Na effect was weaker with axial viewing (average of ionic lines under robust conditions (1500 W, 0.6 l min-1) with axial 0.91) than with radial viewing (average 0.8) with the Varian ($) and radial viewing (&) using the Perkin-Elmer Optima 3000 DV ICP system.A cross-flow nebulizer (Table 1) was used. systems. This is probably due to the significant increase in the Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12 2857 Apel, C. T., Duchane, D. V., Palmer, B. A., Bieniewski, T. M., consequently in LOD, the actual improvement factor seems to Pena, H. V., Cox, L. E., Gallimore, D. L., Vincent, K., Lopez, M., be in the range of 2–3. A more significant overall improvement Kline, J.V., and Steinhaus, D. W., in Developments in Atomic may be observed when axial viewing is associated with other Plasma Spectrochemical Analysis, ed. Barnes, R. M., John Wiley, technological modifications such as better practical resolution, Chichester, 1983, p. 383. a more efficient sample introduction system, a longer residence 8 Faires, L. M., Bieniewski, T. M., Apel, C. T., and Niemczyk, time and change in the hf generator characteristics. Compared T. M., Appl. Spectrosc., 1985, 39, 5. 9 Pan Fuxing, You Suling, He Qinghua, Wang Xiaoping, Ma with the results obtained ten years ago, this combination of Yeying, Huang Yanmin, Xu Yuxin, Xu Wi, and Wu Tingfang, technological changes have led to an overall improvement of Spectrochim.Acta, Part B, 1986, 41, 1211. a factor of about 20. For instance, Pb, with values of the RSDB 10 Pan Fuxing, You Suling, He Qinghua, Kang Daming, Xu Yuxin, less than 0.5%,1 can exhibit an LOD as low as 1 ng ml-1, in and Li Wenmei, Spectrochim.Acta, Part B, 1987, 42, 853. comparison with values of 20–40 ng ml-1 obtained several 11 Kato, T., Uehiro, T., Yasuhara, A., and Morita, M., J. Anal. At. years ago. Some light elements that are difficult to determine Spectrom., 1992, 7, 15. 12 Xiao-Jin Yang, and Jing-Su Guan, Anal. Chim. Acta, 1993, with ICP-MS due to isobaric interferences, e.g. Fe, Ca, Be, can 279, 261. exhibit very low LOD, particularly when an ultrasonic 13 Pan Fuxing, Tong Dezhi, Ren Ming, and Ma Heying, T alanta, nebulizer is used.The LOD of Fe can be at the 10 pg ml-1 1993, 40, 1107. level, while that of Be can be at the pg ml-1 level. 14 Nakamura, Y., Takahashi, K., Kujirai, O., Okochi, H., and Although axial viewing had a poor reputation concerning McLeod, C. W., J. Anal. At. Spectrom., 1994, 9, 751. the matrix or reagent effects, these effects seem to be of the 15 Ivaldi, J. C., and Tyson, J. F., Spectrochim. Acta, Part B, 1995, same magnitude than those observed for radial viewing. 50, 1207. 16 O’Hanlon, K., Ebdon, L., and Foulkes, M., J. Anal. At. Spectrom., Nevertheless, it is important to use so-called robust conditions. 1996, 11, 427. Moreover, an increase in the id of the injector seems to be 17 Rayson, G. D., and Shen, D. Y., Anal. Chem., 1990, 62, 1239. highly beneficial to minimize the matrix effects. More change 18 Rayson, G. D., and Shen, D. Y., Appl. Spectrosc., 1991, 45, 706. in the torch design and a careful study of the volume probed 19 Rayson, G. D., and Shen, D. Y., Spectrochim. Acta, Part B, 1995, by the focusing system are probably the next steps to take full 46, 1237. benefit of axial viewing. 20 De Loos-Vollebregt, M. T. C., Tiggelman, J. J., and De Galan, L., Spectrochim. Acta, Part B, 1988, 43, 773. The authors wish to thank Perkin-Elmer Germany for the 21 De Loos-Vollebregt, M. T. C., Tiggelman, J. J., Bank, P. C., and Degraeuwe, C., J. Anal. At. Spectrom., 1989, 4, 213. loan of the Optima 3000 DV ICP system and Varian France 22 Marichy, M., Mermet, M., Murillo, M., Poussel, E., and Mermet, for the loan of the Liberty 220 ICP system and the access to J. M., J. Anal. At. Spectrom., 1989, 4, 209. the Liberty 150 AX Turbo. 23 Romero, X., Poussel, E., and Mermet, J. M., Spectrochim. Acta, Part B, in the press. 24 Boumans, P. W. J. M., Spectrochim. Acta, Part B, 1991, 46, 431. REFERENCES 25 Mermet, J. M., Carre�, M., Fernandez, A., and Murillo, M., 1 Mermet, J. M., and Poussel, E., Appl. Spectrosc., 1995, 49, 12A. Spectrochim. Acta, Part B, 1991, 46, 941. 2 Boumans, P. W. J. M., Anal. Chem., 1994, 66, 459A. 26 Mermet, J. M., Spectrochim. Acta, Part B, 1989, 44, 1109. 3 Abdallah, M. H., Diemiaszonek, R., Jarosz, J., Mermet, J. M., Robin, J., and Trassy, C., Anal. Chim. Acta, 1976, 84, 271. Paper 6/06445K 4 Lichte, F. E., and Koirtyohann, S. R., presented at FACSS 1976. Received September 18, 1996 5 Danielsson, A., ICP Inf. Newsl., 1978, 4, 147. 6 Demers, D. R., Appl. Spectrosc., 1979, 33, 584. Accepted December 6, 1996 286 Journal of Analytical Atomic Spectrometry, March 1997, Vol.

ISSN:0267-9477    

DOI:10.1039/a606445k

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Toward the Next Generation of Atomic Mass Spectrometers† Plenary Lecture GARY M. HIEFTJE*, DAVID P. MYERS, GANGQIANG LI, PATRICK P. MAHONEY, THOMAS W. BURGOYNE, STEVEN J. RAY AND JOHN P. GUZOWSKI Department of Chemistry, Indiana University, Bloomington, IN 47405, USA Atomic mass spectrometry, embodied principally as ICP mass Given these capabilities, it is perhaps not surprising that spectrometry (ICP-MS) and glow discharge mass many researchers have turned to ICP-MS or glow-discharge spectrometry (GDMS), has enjoyed rapid growth during the mass spectrometry for routine analysis.However, neither last decade, yet both methods exhibit shortcomings that would method is without its shortcomings. ICP-MS, for instance, still be desirable to reduce or eliminate. Prominent among these suffers from a number of matrix and spectral interferences. shortcomings are drift and limited precision, several During the past several years, a great deal of effort has been troublesome spectral and matrix interferences, and moderate expended in understanding the nature of many matrix interatom- detection efficiency.This last limitation is particularly ferences. As a result of those studies, modifications have been troublesome when ICP-MS, for example, must be interfaced made in the design of the interface between the ICP and mass to analytical systems that deliver extremely small sample spectrometer in order to reduce the incidence of space charge volumes or low flow rates or when extremely limited sample (coulombic repulsion) and the effect it has on analyte signal sizes must be examined.Such situations are projected to be levels. However, enough concern about interferences remains increasingly common in the next decade because of the that most commercial laboratories rely upon standard importance of biotechnology and nanostructured materials. additions (spiking) or matrix matching in order to ensure an Overcoming these limitations will require substantial acceptable level of accuracy.Similarly, spectral overlaps with modifications in both sources and mass-spectrometer designs. polyatomic species continue to be a problem, despite a number Sources will be required that are more efficient at sample of clever schemes to overcome them. In addition, GDMS utilization, aerosol volatilization and atomization and that suffers from a difficulty in using internal standards, in blank provide multidimensional information.Similarly, mass subtraction, and in a susceptibility to contamination. spectrometers of the future must be more atom-efficient, Difficulties are also encountered when there is a need to should measure all elements and isotopes simultaneously, and measure transient or time-dependent samples or signals. In must operate on a time scale that is compatible with ICP-MS such signals are encountered when one employs microsampling and transient-sampling technology. Possible electrothermal atomization, laser ablation, flow injection or alternative systems that meet these criteria will be outlined chromatographic separation.In GDMS, time-dependent and their likely performance assessed. Greatest emphasis is signals must be measured if depth profiling is desired. placed on time-of-flight mass spectrometry coupled with an Unfortunately, virtually all commercial ICP-MS and GDMS ICP source. instruments are scanned devices; that is, they can measure only Keywords: Inductively coupled plasma mass spectrometry; a single mass at a time.As a result, there is a necessary trade time-of-flight mass spectrometry; glow discharge mass off between elemental coverage and signal-to-noise ratio. On spectrometry ; instrumentation ; plasma-source mass the one hand, one can limit the elemental coverage and thereby spectrometry accumulate enough signal to yield low detection limits or high precision. On the other hand, one might prefer more complete elemental coverage but will have to accept a loss in sensitivity Over the past decade, plasma-source mass spectrometry or reproducibility as a result.(PSMS) has emerged as an attractive technique for elemental The limitation of measuring only a single isotopic peak at analysis. Despite its relatively high cost, the method enjoys a a time also constrains the precision that can be realized in a high degree of popularity because of a number of important ratioing mode.It is generally recognized that the dominant attributes. For example, the most widely used of the PSMS source of fluctuations in a PSMS are multiplicative, that is, techniques, ICP-MS, offers sub-part per trillion detection they seem to affect all signals in much the same way. This is limits, a linear range that extends over approximately seven especially true of ICP-MS. Variations in nebulizer or spray- orders of magnitude, isotope analysis and isotope-ratioing chamber performance, flutter in the plasma tail flame and the capability, modest precision (1–5%), excellent performance in sampling of inhomogeneous plasma zones all lead to signal a semi-quantitative (standardless) mode, limited spectral (iso- fluctuations that more or less track each other.That is, when baric) and matrix interferences, virtually complete elemental one elemental signal rises, others tend to follow it. As a result, coverage, high speed on a per-sample and per-isotope basis, precision can best be improved by a ratioing method such as and the convenience of solution-based sampling.This last internal standardization or isotope dilution. Unfortunately, the feature, of course, requires sample dissolution in many cases. signal fluctuations are often rapid enough that ratioing-based However, it also simplifies standardization, blank subtraction, compensation is effective only if the target isotopes or elements and the use of standard additions or internal-standardization are measured within microseconds of each other.Although approaches. this sort of speed is possible with quadrupole-based mass spectrometers, only a few elements or isotopic peaks can then † Presented at the Eighth Biennial National Atomic Spectroscopy Symposium (BNASS), Norwich, UK, July 17–19, 1996. be measured at a time. The problem is even more difficult in Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12 (287–292) 287sector-field instruments, which often cannot be scanned (or limited storage capacity, just as does the IT.In addition, to realize the extremely high resolution of which the FT–ICR is peak-hopped) as rapidly. In recognition of this limitation, most sector-field instruments designed for high-precision isotope capable requires an extended trapping time, rendering the system unsuitable for use with transient-sampling devices. ratioing employ multiple collectors and an extremely stable (thermal ionization) source.Lastly, the FT–ICR instruments available today are extremely expensive, made so by the need for cryogenically cooled From the foregoing comments, the requirements of the nextgeneration PSMS instrument seem evident. It should offer magnets, extremely low operating pressures, and sophisticated signal-processing systems. Overall, it seems unlikely that they virtually complete freedom from matrix and spectral interferences, simultaneous detection of the entire atomic mass will be widely used in the future.The third candidate for a future system is a sector-field mass range of interest, high precision in a ratioing mode and high spectral-acquisition rates in order to be amenable to coupling spectrometer equipped with a focal-plane detector array. Such a system has been in development in our laboratory9,10 and in with transient-sampling devices. Furthermore, such an instrument must not sacrifice any of the important figures of merit others11 for several years and offers substantial promise.This device can be considered to be the mass-spectrometric analog that have become associated with PSMS in general. Lastly, the next-generation device should possess the attributes that of modern emission spectrometers based on charge-transfer device technology. However, unlike an atomic emission spec- all in the field have come to know and love: low cost, high reliability, user-friendliness and a high degree of automation.trum, an atomic mass spectrum consists of a relatively modest number of peaks whose locations are well established and whose ratios, for a given element, do not vary greatly. In THE OPTIONS particular, unit-mass resolution is all that is required for routine atomic mass spectrometry measurements. Also, in most Some of the features to be present in the next-generation of PSMS instruments will no doubt be the result of modification applications it is necessary to detect only the mass range from lithium (7 amu) to uranium (238 amu). Thus, in the best of or re-design of sample-processing or sample-introduction equipment, or from improved plasma sources.For example, circumstances, fewer than 250 detector elements (pixels) would be required to provide full AMS coverage. Thus, it would unless extremely high mass-spectral resolution is employed, the most likely way to overcome residual spectral and matrix seem that PSMS is almost ideally suited for detector-array technology.interferences is through more efficient solvent removal, improved sample-introduction arrangements and modification Unfortunately, the mass-spectral display produced by most magnetic-sector mass spectrometers is quadratic. As a result, of the plasma environment. In large part, these improvements will be possible with only slightly modified glow-discharge or a detector array that serves well on one end of the mass spectrum does not provide adequate resolution on the other.ICP sources. However, it seems likely also that more substantial departures from the conventional systems might be neces- In addition, problems can arise from extremely intense massspectral peaks, such as that produced by Ar+ in the ICP, sary. One such approach, the tandem source, will be described in more detail later. However, to constrain the length of this causing detector saturation, blooming or non-linearity. In recognition of this situation, the system we designed9,10 treatment to a manageable level, the rest of our discussion will focus mainly on mass-spectrometer options for PSMS.is configured to record the full atomic mass-spectral range in two segments that straddle, but avoid, 40Ar+. The low-mass In identifying options for the next-generation mass spectrometer to be used in elemental analysis, one must recognize segment, extending from 7 to 38 amu, is recorded first and the accelerating voltage in the instrument then switched abruptly that the desirable features outlined above do not require that the mass spectrum be recorded in a truly simultaneous fashion.to cover the high-mass segment from 42 to 238 amu. Unfortunately, our current system suffers from problems Rather, it is necessary only that ions be sampled from the plasma source simultaneously. In an ion trap (IT), for example, that have been traced to the detector array. Regrettably, it is difficult to detect atomic ions directly on a linear detector ions can all be accumulated at the same time from a continuously operating ion source, so the stored population represents array of the kind intended for optical sensing.In a sector-field instrument, ions achieve an extremely high kinetic energy, the composition of the plasma at a particular point or period in time. Determining the mass spectrum of the ions stored in sufficient to sputter the surface of the detector. Consequently, it is necessary to turn to several inter-domain conversions in an IT then can be achieved in a much slower, sequential fashion, by expelling ions from the trap one mass at a time, so order to preserve detector integrity.The most common such approach is to employ a microchannel plate to convert each they can be detected by a suitable ion multiplier. Indeed, a substantial amount of success has already been achieved with incoming ion to a shower of electrons. In turn, the electrons are converted to photons by impinging on a phosphor screen.both ICP and GD sources with an IT mass spectrometer.1–4 Unfortunately, ion traps can accommodate populations of only Lastly, the photons are detected with a conventional linear photodetector array. Although cumbersome, this approach has about 106 ions before space charge repulsion becomes a problem. Furthermore, ion traps can contain only approxi- been found successful by a number of earlier workers.12–14 Unfortunately, such an arrangement is extremely susceptible mately 104 ions and remain analytically useful.Thus, at least 100 experiments would have to be performed if the dynamic to fringing fields from the magnetic sector. Unless these fields are reduced or the detector array shielded from them, the range expected of moderate PSMS instruments (106) were to be matched. Furthermore, this multiple-storage requirement consequence is a peak shaped such as that seen in Fig. 1. Apparently, the electrons produced by the microchannel plate will make the system rather unattractive for coupling with transient-atom sources.are affected sufficiently by fringing magnetic fields that they add a broad and almost triangular base to the mass-spectral Another candidate instrument for next-generation systems is the Fourier transform ion-cyclotron resonance (FT–ICR) peaks. As a result, the mass-spectral resolution (defined in terms of peak half-width) is adequate for elemental analysis, mass spectrometer.5–8 Like the ion trap, an FT–ICR instrument accumulates ions from an external source but then reads them although abundance sensitivity (the ‘leakage’ of one mass into the next) is unacceptable.Work continues in an effort to in a truly simultaneous fashion. Furthermore, the tremendous resolution of which the method is capable should completely overcome this problem. A final candidate for the next generation of PSMS instru- eliminate any concerns about isobaric overlaps.Indeed, recent experiments have shown that resolving powers as high as ments is the time-of-flight mass spectrometer (TOFMS). Like the ion trap, the TOFMS accepts all ions at the same time, 1800 000 can be realized.8 Unfortunately, the FT–ICR has 288 Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12OVERCOMING THE LIMITATIONS OF TOFMS The resolution ‘problem’ in TOFMS can be solved by applying a number of technologies developed in the past.For example, an input ion beam of extended width can be collapsed upon itself, and thereby will produce a short detector pulse, by means of a technique termed ‘space focusing’, developed first by Wiley and McLaren.15 The technique involves using a twostage ion-input system, consisting of an ‘extraction zone’ and a subsequent acceleration region. Ions are initially fed into the extraction zone, from which they are pushed by means of a ‘repeller’ voltage into the acceleration region.Ions at the rear of the extraction zone begin farther from the detector and therefore would ordinarily arrive later in time than those ions Fig. 1 Image of a mechanical slit placed in front of a microchannel in the front of the extraction zone. However, ions in the rear plate–CCD detector system situated at the focal plane of a Mattauch– of that zone experience a greater ‘push’ from the repeller Herzog mass spectrograph.10 The measured half-width of the peak is 510 mm, whereas the expected width is 495 mm.The broadened, almost electrode behind them (since they are closer to it) and therefore triangular base of the peak is probably the result of fringing fields catch up with ions ahead of them. Placing a detector at the produced by the magnetic sector. location where the ions meet then results in improved resolution. Similarly, any kinetic-energy variation among analyte ions in the direction of the flight tube can be overcome by means but in an extremely short burst that lasts for only a few of a ‘reflectron’.In essence, the reflectron configuration involves nanoseconds. However, the mass spectrum then requires more the use of an ion mirror which consists of suitably applied time to record, ordinarily of the order of 50 ms or so. Still, this fields that repel ions sent into it. A relatively low-energy ion spectral-acquisition time is considerably shorter than that of penetrates the field for only a short distance before being which any of the other devices discussed above is capable.reflected to a distant detector. In contrast, a higher-energy ion Stated differently, it allows as many as 20000 elemental mass penetrates the reflecting field farther and therefore must travel spectra to be accumulated each second, making it applicable a greater distance before it reaches the detector. Because the to all but the most rapid transient-sampling systems. path traveled by the higher kinetic energy (higher velocity) ion Also, the TOFMS is among the simplest of all mass is greater than that of the low-energy ion, the kinetic-energy spectrometers, consisting basically of an input section, an open variation can be almost completely overcome.16 evacuated tube and a detector.To perform a mass-spectral Raising the duty factor of a TOFMS is also possible with analysis, a packet of ions is accelerated through a fixed voltage well established technology. In the 1960s, workers in the drop, typically on the order of 2000 V or so, imparting to all Bendix Corporation17 realized that an orthogonal acceleration ions the same kinetic energy (KE).The ions therefore achieve (OA) configuration would achieve the desired end. When ions a velocity (v) that is inversely proportional to the square root are sampled from an atmospheric-pressure source, they first of their respective mass-to-charge ratios (m) (i.e., KE=0.5 mv2). form a supersonic expansion, in which they all achieve roughly The arrival time of each ion at the end of a field-free region the same velocities.Obviously, that velocity should correspond (termed the ‘flight tube’) can then be determined and related to their original thermal temperatures which, in the case of to mass-to-charge ratio. The result is an instrument that is the ICP, are of the order of 0.5 eV. In fact, the energies are a rather maintenance-free and relatively inexpensive to produce little higher because of the existence of a plasma offset voltage.and to operate. Furthermore, because of its open structure, the The ions therefore begin their flight into the vacuum system TOFMS offers extremely high efficiency. In some applications, at a relatively low energy (velocity) of from 1 to 5 eV. In such as secondary-ion mass spectrometry, transmission contrast, when the ions are accelerated into the TOFMS flight efficiencies approaching unity have been achieved.This factor tube, they reach velocities that correspond to 2000 eV, if the alone should make the TOFMS as much as 100 times more accelerating voltage used in propelling them into the flight sensitive than competitive quadrupole-filters or sector-field tube is 2000 V. This factor of roughly 1000 difference in devices. velocities enables one to overcome all but about 10% of the Not surprisingly, the TOFMS is also not without its weak- duty factor of the TOFMS. The initial, slowly moving ion nesses.Because ions of a given mass to charge ratio must beam first fills the extraction zone, from which ions are sent arrive at the detector at virtually the same time, the TOFMS in a perpendicular direction down the flight tube. As the mass is often viewed as a low-resolution instrument. Any factor that spectrum of that ion packet is being recorded, the extraction can contribute to a broadened ion packet will degrade reso- zone is slowly refilled by the continuously flowing input beam.lution. Such factors include a spatially broad input ion pulse, The resulting limited duty factor, roughly 10%, is more than a temporally long input ion pulse, a spread in kinetic energies compensated by the high transmission efficiency of the of ions that enter the drift tube, and disparate ion paths TOFMS. In addition, the duty factor can be further improved through the drift tube. Although these problems can be largely either by electrostatically decelerating the input ion beam or overcome through use of established techniques, they offer by shortening the TOFMS flight tube (i.e., so that a higher challenges to instrument design.repetition rate can be employed). The latter approach produces Perhaps even more of a concern, however, is the relatively a loss in resolving power, of course, but not to the point where low duty factor of a TOFMS. In order to achieve high massthe device becomes unsuitable for atomic mass spectrometry.spectral resolution, only an extremely short input pulse of ions Interestingly, the open structure that imparts a high trans- can be used. Ordinarily, that pulse is on the order of 5 ns. It mission efficiency to TOFMS also raises background noise. then requires as long as 50 ms for a full atomic mass spectrum Of course, in an orthogonal acceleration TOFMS arrangement to be recorded; only after this time will the next 5-ns pulse of there should be little difficulty caused by either neutral species ions be introduced.Consequently, only a tiny fraction of the or photons, since the detector can be effectively shielded from input ion beam can be used for the analysis; in this case the them. However, because of the spread of the input-ion beam, fraction is (5×10-9)/(50×10-6)=10-4. In essence, the device will waste 99.99% of a continuously operating input ion beam. ion–atom and ion–ion collisions, and fringing fields in the Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12 289extraction zone, the input ion beam leaks almost continuously from the extraction zone into the acceleration region of the TOFMS.These leakage ions are then accelerated into the flight tube and ultimately reach the detector, where they produce a continuous background-noise level. Fortunately, it is possible to overcome this ion-induced background noise by means of energy discrimination.18 It will be recalled that an input ion pulse is created by ‘pushing’ a segment of the input beam from the extraction zone into the acceleration region.This ‘push’ imparts to the ion packet a kinetic-energy component that is added to the energy the ions receive in the acceleration region. For example, it is common to employ an accelerating voltage of 2000 V and an extraction pulse of 500 V. The analyte ions therefore receive a kinetic Fig. 2 The precision of isotope-ratio measurements improves as the energy between roughly 2150 and 2350 eV, depending upon square root of the number of signal counts that are accumulated.their initial positions in the extraction zone. In contrast, Closed circles represent precision values obtained by ion counting for background ions leak into the acceleration zone without the ICP-TOFMS measurement of the Ag 107:109 isotope ratio. The dashed line indicates the precision expected from counting statistics. experiencing the same ‘push’. As a result, their kinetic energies cannot exceed 2000 eV.A kinetic-energy barrier placed immediately in front of the ion detector then serves to discriminate efficiently against background ions but to pass the analyte Because of its extremely high-speed spectral acquisition, the ICP-TOFMS is also well suited to the measurement of transi- packets with high efficiency. ent signals. In fact, as many as 100 mass spectra can be accumulated during a single laser-ablation event lasting no PERFORMANCE OF AN ICP-TOFMS longer than 10 ms.23 This capability enables the spatially resolvedcomposition of aheterogeneous solid to be determined Because of the high throughput (transmission efficiency) of the by repetitive laser-ablation sampling.In these experiments, as TOFMS and use of the energy-discrimination scheme described in the continuous-nebulization experiments outlined above, above,18 detection limits achieved with a laboratoryprecision can be improved by isotope ratioing or internal constructed ICP-TOFMS instrument lie mostly below 1 ppt, standardization.competitive with commercial instrumentation (see Table 1).18,19 Perhaps the time-resolution capability of an ICP-TOFMS Moreover, these detection limits are all achievable in the same is best exploited when it is used to add dimensionality to an ten-second interval. In contrast, scanning-based ICPMS instruanalysis. 20 This added dimensionality is illustrated, for example, ments require a separate 10 s interval to measure the detection in gas chromatography–mass spectrometry, wherein gas- limit of each isotope of interest.Moreover, our current system chromatographic separation provides information along one is not equipped with an advanced high-efficiency interface such axis (dimension) and mass analysis provides orthogonal (inde- as can be found on most commercial instruments. It seems pendent) information along a second dimension. In ICP- likely, therefore, that detection capabilities should be able to TOFMS, the same capability could be achieved in combination improve appreciably.Indeed, in a recent analysis20 it was with chromatography. Importantly, unlike with scanning mass shown that the simplicity of a TOFMS instrument enables ion spectrometers, no mass-spectral skew will then exist, since all losses to be tracked and possible areas of improvement to be atomic masses of interest will be taken from the same portion identified.That evaluation estimated that between 100 and of the chromatographic peak. Consequently, empirical formu- 1000 atoms of a target element present in the sample should lae should be more reliable to determine and the overlap of ultimately be able to be detected. chromatographic peaks simpler to detect. As was suggested earlier, the virtually simultaneous extrac- An example of this added dimensionality taken from our tion of all ions from a continuous source enables precision to own work24 arises when the ICP-TOFMS is coupled with an be improved by isotope ratioing or by internal standardization.electrothermal atomizer (ETA). In this experiment, the ETA As is shown in Fig. 2, this precision improves as the square receives a 10-ml sample which is subsequently dried and ashed, root of the number of ion counts that are accumulated,21,22 much as is done in atomic absorption spectrometry. The enabling it to be improved to virtually any desired level just furnace temperature is then ramped upward; along that ramp, by employing a sufficiently long integration time.Furthermore, elements are volatilized in accordance with their appearance this high-precision capability is available for all elements or temperatures. That is, the more volatile elements appear earlier isotopes at once, an impossibility with scanned mass than those of lower volatility. This temperature ramp then spectrometers. provides a time axis that enables elements to be separated in part on the basis of their volatility differences.Because a TOF Table 1 Limits of detection obtained with an ICP-TOF mass spectrometer18 mass spectrum can be obtained at virtually every point during the temperature ramp, what results is a two-dimensional plot Detection limit/ with volatilization time (temperature) on the vertical axis and Element ng l-1 a flight time (the mass spectrum) on the horizontal axis. An Li 1.5 example of such a plot is shown in Fig. 3. Mg 4.2 From Fig. 3, it can be seen that all isotopes of a given Mn 1.5 element volatilize at the same time, which would clearly be Co 1.1 expected. However, the locations in the two-dimensional plot, Sr 0.6 where different elements can be found, are separated by a Rh 0.5 Ag 0.9 greater distance than they would be in a simple one- Cs 0.5 dimensional mass-spectral display. This added dimensionality Ho 0.4 allows species to be distinguished that would otherwise overlap Bi 0.6 isobarically.For example, the ArO+ peak (not shown in the U 0.6 particular display of Fig. 3) appears very early in the volatiliz- 290 Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12therefore follow a different trajectory in the perpendicular flight tube, and some miss the detector. The duty factor of the current system is also only about 10% at present. It could be improved, however, by shortening the flight tube, with an acceptable sacrifice in resolving power. The same change would increase the spectral-acquisition rate from the current 20 kHz to roughly twice as fast.Interestingly, several of these problems can be overcome by abandoning the OA geometry in favour of a more traditional but significantly modified on-axis configuration. In this alternative arrangement, the goal is to retain the advantages of the OA geometry (mainly its improved duty factor), but to avoid its limitations (mass bias and low transmission efficiency). The way in which this goal is being achieved is to ‘tag’ (by means of its higher energy) each introduced ion packet before it is accelerated and to use second-order (higher efficiency) space focusing27 to collapse that tagged ion packet upon itself.The Fig. 3 Two-dimensional plot illustrating the ability of ICP-TOFMS packet so tagged then enters an acceleration zone, but travels to resolve atomic isobars (spectral overlaps) on the basis of volatility in the same direction in which it had been moving.Thus, the differences. An electrothermal atomizer is used to provide a tempera- difficulties of the OA geometry are avoided. However, the high ture ramp (vertical axis), along which elements volatilize sequentially, in the order of their appearance temperatures. The high speed of the duty factor is retained because ions are tagged and collapsed TOFMS enables this time of appearance to be followed and the upon themselves while they are moving slowly rather than otherwise overlapping isotopes individually measured (horizontal axis).after they are accelerated to the drift-tube velocity. In this new Taken from reference 24. instrument, the ‘tagging’ is accomplished by means of a voltage pulse applied to an in-line repeller grid, much as has been used ation cycle, whereas the 56Fe+ peak, with which it overlaps, in the past in our OA geometry. As a result, only the tagged appears much later. Similarly, Cd, In and Sn have isotopes ions have the extra velocity (energy) required to penetrate an that mutually overlap.Separating them by mass spectrometry energy barrier placed just before the ion detector and only alone would require resolving powers between 90000–300 000. they contribute to the mass spectrum. The rest of the ion In contrast, they can be resolved with the tandem ETA–ICP- beam, allowed to travel unimpeded from the ion source into TOFMS combination quite readily (see Fig. 3). the TOFMS flight tube, is ignored. Of course, the TOFMS can be used also with alternative Although still in its relative infancy, we feel this new on-axis ion sources. We have already reported its coupling with a glow PS-TOFMS might truly be the next generation of atomic mass discharge (GD) source25 and are pursuing its use with an spectrometers. We will look forward to reporting in greater electrospray ionization (ESI) system.26 With both the GD and detail on its performance in the future.ESI source, the advantages outlined above for ICP use are Supported in part by the National Institutes of Health through once again realized: high precision in a ratioing mode, excellent grant GM 53560. detection limits and applicability to transient-sampling analysis. However, with the ESI source an additional advantage appears. The TOFMS has often been touted as the best REFERENCES mass spectrometer for the separation of high-mass species. 1 McLuckey, S. A., Glish, G.L., Duckworth, D. C., and Marcus, Theoretically, there is no upper mass limit in a TOFMS. R. K., Anal. Chem., 1992, 64, 1606. Consequently, the TOFMS is particularly well suited to speci- 2 Duckworth, D. C., Barshick, C. M., Smith, D. H., and McLuckey, ation studies, in which complex, clustered ions must be S. A., Anal. Chem., 1994, 66, 92. observed. This flexibility is not often available with quadrupole 3 Koppenaal, D. W., Barinaga, C. J., and Smith, M. R., J. Anal.At. Spectrom., 1994, 9, 1053. or sector devices intended for use in atomic spectrometry. 4 Eiden, G. C., Barinaga, C. J., and Koppenaal, D. W., J. Anal. At. Spectrom., 1996, 11, 317. 5 Barshick, C. M., and Eyler, J. R., J. Am. Soc. Mass Spectrom., CONCLUSIONS 1992, 3, 122. From the foregoing account, a TOFMS would seem to be a 6 Marcus, R. K., Cable, P. R., Duckworth, D. C., Buchanan, M. V., Pochkowski, J. M., and Weller, R. R., Appl. Spectrosc., 1992, highly attractive alternative to currently available commercial 46, 1327.systems for atomic mass spectrometry. It offers the same low 7 Watson, C. H., Wronka, J., Laukien, F. H., Barshick, C. M., and detection limits on a relative (concentration) basis, low absolute Eyler, J. R., Spectrochim. Acta, Part B, 1993, 11, 1445. detection limits (on an atom or mass basis) because of its rapid 8 Watson, C. H., Wronka, J., Laukien, F. H., Barshick, C. M., and response, nearly simultaneous mass measurement, precision Eyler, J.R., Anal. Chem., 1993, 65, 2801. that is limited only by counting statistics if ratioing is employed, 9 Burgoyne, T. W., Hieftje, G. M., and Hites, R. A., ‘Design of a Mattauch–Herzog Array Detector Mass Spectrometer for Atomic extremely rapid spectral-acquisition rates, good resolution Mass Spectrometry’, 41st ASMS Conference on Mass Spectrometry (between 1600 and 2300), the ability to measure small samples and Allied T opics, San Francisco, CA, ASMS, Santa Fe, NM, such as those accepted by an electrothermal atomizer or USA, 1993, paper #WP200.produced by laser ablation, and a virtually unlimited upper 10 Burgoyne, T. W., Hieftje, G. M., and Hites, R. A., J. Am. Soc. mass range, useful for speciation studies. Mass Spectrom., 1996, in the press. Yet, the system is not without its drawbacks. The trans- 11 Cromwell, E. F., and Arrowsmith, P., J. Am. Soc. Mass Spectrom., 1996, 7, 458. mission efficiency of our current system is not as high as would 12 Hill, J.A., Biller, J. E., Martin, S. A., Biemann, K., Yoshidome, K., be desired; it appears to be limited by divergence of the initial and Sato, K. Int. J. Mass Spectrom. Ion Process, 1989, 92, 211. slowly moving ion beam as it travels in the perpendicular 13 Murphy, D. M., and Mauersberger, K., Rev. Sci. Instrum., 1985, direction down the flight tube. The same orthogonal geometry 56, 220. produces a mass bias, since ions in the initial ion beam all 14 Murphy, D. M., and Mauersberger, K., Int. J. Mass Spectrom. Ion have roughly the same velocity and therefore a kinetic energy Process, 1987, 76, 85. 15 Wiley, W. C., and McLaren, I. H., Rev. Sci. Instrum., 1955, 26, 1150. that is proportional to their mass. Ions of different mass Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12 29116 Myers, D. P., Li, G., Mahoney, P. P., and Hieftje, G. M., J. Am. 23 Mahoney, P. P., Li, G., and Hieftje, G. M., J. Anal. At. Spectrom., 1996, 11, 401. Soc. Mass. Spectrom., 1995, 6, 400. 24 Mahoney, P. P., Ray, S. J., Li, G., and Hieftje, G. M., Anal. 17 O’Halloran, G. J., and Walker, L. W., T echnical Document Report Chem., 1996, submitted for publication. No. ASD TDR 62–644, Parts I and II, Nov., 1964, Bendix 25 Myers, D. P., Heintz, M. J., Mahoney, P. P., Li, G., and Hieftje, Corporation, Research Laboratories Division, Southfield, MI. G. M., Appl. Spectrosc., 1994, 48, 1337. 18 Mahoney, P. P., Li, G., Ray, S. J., and Hieftje, G. M., J. Am. Soc. 26 Mahoney, P. P., Guzowski, J. P., Ray, S. J., and Hieftje, G. M., Mass Spectrom, 1996, in the press. Appl. Spectrosc., 1996, submitted for publication. 19 Myers, D. P., Li, G., Mahoney, P. P., and Hieftje, G. M., J. Am. 27 Li, G., and Hieftje, G. M., U.S. Patent Application no. 08/434,931, Soc. Mass Spectrom., 1995, 6, 411. filed May 5, 1995. 20 Hieftje, G. M., J. Anal. At. Spectrom., 1996, 11, 613. 21 Myers, D. P., Mahoney, P. P., Li, G., and Hieftje, G. M., J. Am. Paper 6/05067K Soc. Mass. Spectrom., 1995, 6, 920. Received July 22, 1996 22 Mahoney, P. P., Doctoral Dissertation, Indiana University, 1996. Accepted November 25, 1996 292 Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12

ISSN:0267-9477    

DOI:10.1039/a605067k

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Use of Image Processing to Aid Furnace Set-up in Electrothermal Atomic Absorption Spectrometry† PHILIPPE R. BOULOa , JOHN J. SORAGHANa, DARAN A. SADLERb , DAVID LITTLEJOHN*b AND ANDREW CREEKEc aSignal Processing Division, Department of Electronics and Electrical Engineering, University of Strathclyde, Glasgow, G1 1XL UK bDepartment of Pure and Applied Chemistry, University of Strathclyde, Glasgow, G1 1XL UK cUnicam Atomic Absorption, York Street, Cambridge, CB2 2PX UK A charge-coupled device (CCD) camera has been used to (GFTV) fitted to a SOLAAR 939 atomic absorption spectrometer( Unicam Atomic Absorption, Cambridge, UK) .The obtain high-definition video images of the cross-section of the tube in a graphite furnace atomizer.Image processing imaging system, which is shown schematically in Fig. 1, consists of a computer-controlled movable mirror, a fixed plane mirror, algorithms have been developed to extract automatically from the images certain features which are of use during the initial a focusing lens and a CCD camera.The system is situated between the furnace head and the monochromator and is not set-up and alignment of the atomizer in an atomic absorption spectrometer. These features include the alignment of the tube, visible to the user. To view the cross-section of the tube, the movable mirror is rotated to divert the undispersed light from the angle of a platform with respect to the vertical axis and the height of an autosampler capillary tip above the tube wall the hollow cathode lamp (HCL) onto the CCD via the plane mirror and the lens.The HCL provides illumination for the or the platform. In addition, algorithms have been developed to analyse images recorded after sample injection, to detect system and any object placed inside the tube, such as a platform, the capillary tip or the sample volume, will cast a whether all of the liquid has been deposited correctly into the tube. Detection of the poor injection of a blood serum sample shadow onto the CCD.Hence, the presence of an object in the tube is associated with an absence of signal on the image is given to illustrate the usefulness of the algorithm. recorded by the CCD camera. Keywords: Electrothermal atomic absorption spectrometry; The camera is a Pulniz type TM 580 (Pulniz Europe, charge-coupled device camera; image processing ; atomizer Alzenau, Germany), which consists of an interline transfer alignment CCD with 500×582 pixels.Each pixel is 12.7×8.3 mm, giving an optically sensitive area for the CCD of 6.35×4.8 mm. The It is well known that the accuracy and precision of an analysis output from the camera is a standard composite video signal. by ETAAS are significantly affected by the initial alignment A Video Blaster SE frame grabber (Creative Labs, Milpitas, and set-up of the instrument. For example, incorrect alignment CA, USA), situated in a personal computer, receives the video of the furnace head increases signal noise because a greater and displays the images on the computer monitor.The software intensity of emission from the tube wall passes into the to display the GFTV images, which runs under Microsoft spectrometer. Also, incorrect positioning of the autosampler Windows 3.1 (Microsoft, Seattle, WA, USA) is incorporated, capillary tip in the tube causes poor repeatability of injection, and may be run concurrently with the SOLAAR software.which adversely affects the precision of the AA signals. To Examples of images obtained from the GFTV system are perform tasks such as adjustment of the capillary tip height shown in Fig. 2. In Fig. 2(a), a platform and the autosampler and monitoring of liquid injection, a dental mirror is often capillary tip can be seen, whilst Fig. 2(b) shows a poorly used to view the interior of the tube. This is not an ideal aligned platform. procedure, as the image of the tube cross-section is small and Images obtained with the GFTV were saved to floppy positioning of the mirror is often inconvenient. Observation of diskette in an 8-bits-per-pixel (256 grey levels) tagged image the tube is simplified if a charge-coupled device (CCD) camera file format (TIFF). These images were then loaded into the is used to obtain images of the tube cross-section.In this VISILOG (Noesis Orsay, France) image processing package, paper, a number of image processing algorithms are described in which all of the algorithms were developed and tested.which can automatically determine from the tube images a Additional code was written in the interpreted C language number of features which are useful for instrument alignment which accompanies the VISILOG package. The time required during the set-up of the atomizer. Furthermore, by comparing images obtained prior to injection and after injection it is possible to determine the profile of the sample droplet.Hence, it is possible to detect the presence of residual liquid on the autosampler capillary tip, the movement of liquid around the tube wall under capillary action, or any contact between the sample droplet and the tube when using a platform for atomization. If uncorrected, all of these problems can cause poor accuracy and/or precision in analysis by ETAAS. INSTRUMENTATION The images shown in this paper were all obtained with the CCD camera system known as graphite furnace television † Presented at the Eighth Biennial National Atomic Spectroscopy Fig. 1 Schematic diagram of the GFTV set-up. Symposium (BNASS), Norwich, UK, July 17–19, 1996. Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12 (293–300) 293Fig. 2 Typical CCD images showing (a) a capillary tip and platform and (b) a misaligned platform. to process an image using the VISILOG package on a personal (1) The image is converted from an 8-bit grey-level image computer with a Pentium processor is approximately 2 s.into a binary image, containing only zeros and ones. All the pixel intensity values below a threshold value are set to zero and all other pixels are set equal to one. The value of the IMAGE ANALYSIS PROCEDURE threshold will vary depending on the HCL and current used The complete image analysis system consists of two distinct for illumination and hence must be determined at the beginning sections. The first section analyses images during the instru- of each set-up procedure.The equation used to determine the mental set-up, the second section provides information on the threshold is given by: deposition of a liquid into the tube and the presence of any residual sample left on the capillary tip. Each section consists threshold= p�+pmax+pmin 3 (1) of an appropriate pre-processing stage, followed by an analysis stage. The resulting information is then presented to the user. where p� is the mean value of the pixel intensities, calculated Instrumental Set-up over the whole image; pmax is the maximum pixel intensity and pmin is the minimum pixel intensity.Fig. 4(a) and (b) shows an Overview example of a GFTV image before and after thresholding, The instrumental set-up analysis flow diagram is shown in respectively. Fig. 3. It was designed to identify the following key features: (2) Isolated points around the edge of an object, which may (1) Alignment of the furnace and camera.(2) Type of tube cause problems in detecting the boundary edge, are removed. (straight-edged or part-ridged) and presence of any platform. This is achieved using a morphological filter, such as an (3) Alignment of the tube (if part-ridged) or platform to the openingoperator.1 This is a smoothing operator which discards horizontal axis. (4) Presence of the capillary tip. (5) small objects in the image, but keeps the largest ones with a Characteristics of the capillary tip, i.e., depth and position. shape similar to the original object.Fig. 4(c) shows the effect The first function is to determine the presence of a platform. of the opening operator on the GFTV image shown in This is required as different algorithms are used to assess the Fig. 4(b). The individual, detached pixels around the edge of alignment of the furnace and camera, depending on the pres- the image of the tube wall, visible in Fig. 4(b), are removed by ence of atform in the tube. Once the alignment of the the opening operator and are not visible in Fig. 4(c). Further furnace has been checked, the angle to the horizontal of any analysis of the CCD images is then carried out on the platform can be determined. If no platform is present then the pre-processed image. type of tube (straight-edged or part-ridged) is identified. This information is used by the instrument to ensure that the optimum heating rate, which is tube-type dependent, is applied. If the tube is part-ridged it is possible to use the ridges to determine the angle of the tube to the horizontal axis, which Detection of platform is a measure of the correct alignment of the tube in the To find the boundary of an object in a given direction, the atomizer.If the tube is not accurately aligned in the furnace image data are scanned in this direction and the boundary head, the user can be warned and the tube re-aligned. Once edges of an object are given by a transition from zero to one the type and alignment of the tube have been determined, the or from one to zero.The detection of the presence of a platform location of the autosampler capillary tip is found. The first is based on scanning the image vertically. The first transition function is to determine whether or not the tip has actually from one to zero, encountered when scanning from the top of entered the tube. Once the capillary tip has entered the tube, the image, at three equidistant positions across the tube is the depth of the tip and the angle, to the vertical axis, is found denoted by A, B and C.This is shown schematically in Fig. 5. and displayed numerically to the user. Hence, it is possible to A platform is present only if these three points lie on a straight know the exact depth of the capillary for any particular line. The choice of the horizontal separation of the three points, analysis and to reproduce accurately the same depth at a Dx in Fig. 5, is obviously important. If Dx is too small, the later date. points A, B and C may appear to lie on a straight line, even if no platform is present. However, if Dx is too large, points Pre-processing outside the platform may be detected and hence A, B and C may not lie on a straight line, even if a platform is present. A Prior to image analysis, all images undergo a pre-processing satisfactory value for Dx has been found to be 0.4 mm for a step, the primary aim of which is to obtain an image which is easier to analyse.It consists of two separate steps: tube diameter of 5 mm. 294 Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12Fig. 3 Flow chart of the overall instrument set-up system. Fig. 4 (a) Original CCD image of a tube; (b) binary image obtained after thresholding the image in (a); and (c) image (b) filtered by an opening operator. Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12 295Fig. 5 Procedure for detection of the presence of a platform.(a) A, B and C are not aligned: there is no platform; and (b) A, B and C are aligned: there is a platform. Alignment of the f urnace head When the furnace head and the camera are not aligned properly, the image of the tube cross-section is non-circular. Thus, the extraction of quantitative information from the CCD images may not be accurate due to the inability to relate the physical dimensions of the tube to the image dimensions. The Fig. 7 Schematic diagram showing the two methods used to detect user can be made aware of this situation and decide whether the alignment of the atomizer with the CCD camera. (a) Method 1: case of cuvette with platform; and (b) Method 2: case of straight or or not to re-align the furnace head. Fig. 6 gives real GFTV ridged cuvette. Left: the camera and furnace head are aligned; and images showing images from aligned and misaligned furnace right: the camera and furnace head are not aligned.heads, respectively. It is clear from Fig. 6(a) that, in the aligned case, the CCD image of the tube cross-section is a circle whereas in Fig. 6(b), the misaligned case, it is an ellipse. aligned and misaligned atomizers the cross-section of the tube is assumed to be circular if 86°<a<94°. Detecting the alignment of the furnace head is then equivalent to detecting whether the image is circular or elliptical. Method 2. This method has been designed to detect the Depending on the presence of a platform, a different method alignment of the furnace and the camera when the tube is was devised.The two different methods are illustrated in Fig. 7. straight or part-ridged, and no platform is present. It is illustrated in Fig. 7(b). Eight points along the boundary of the Method 1. This method, shown in Fig. 7(a), is used when a platform has been detected. The method uses Apollonius’ top of the object are found by scanning the image horizontally at four different levels.The circle, C1, fitting these eight points theorem,2 which states that if parallel chords of an ellipse are drawn along a direction other than along the axes of the is computed. This operation is repeated on the bottom of the object to obtain the circle C2. For a normal, straight-edged ellipse, and their centres joined by a line d, the line d is perpendicular to the direction of the chords only if the ellipse tube, the image of the cross-section is circular if all 16 points lie on the same circle.Hence, if the length l between the centres is circular. The angle a between the line d and the chords is therefore computed to check the alignment of the tube. If of C1 and C2 is computed, then l#0 implies a correctly aligned furnace head. This method also works with part-ridged tubes a#90°, the furnace head is therefore aligned. This method is satisfactory only when the chords span the area of the ellipse because the centre of the circle defining the inner diameter of the tube and the circle defining the edge of the ridges are as much as possible; therefore, it is necessary to use the information already obtained about the platform position to concentric, even though the diameters are different.Even for a perfectly aligned atomizer the value of l may not determine the lowest position on the tube edge which may be used to construct a chord. To compute the angle a, five parallel be exactly equal to zero due to noise and imperfections in the imaging process. A lower limit is therefore required which, if chords are constructed at equidistant points along the direction given by d, and linear regression of the chord centres is used exceeded, indicates that the tube cross-section is significantly non-circular.From the examination of a number of images, a to determine a. From the analysis of a number of images of Fig. 6 (a) Example of correctly aligned head and (b) incorrectly aligned furnace head. 296 Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12lower level for l of 10 pixels (0.3 mm) was considered appro- platform to the horizontal is then the tangent of the slope of this line. Using this procedure an analyst can assess conditions priate. In Fig. 6(a), for example, the value of l is 8 pixels, which is below the threshold and therefore indicates a well aligned to give an adequate sample injection, which may typically be a platform angle of, for example, no more than 2° to the atomizer.However, for the misaligned case, shown in Fig. 6(b), the value of l is 39 pixels, indicating that the tube cross-section horizontal axis. For Fig. 2(a), the angle of the platform to the horizontal axis was calculated to be -0.6°, which is satis- is elliptical. factory, whereas the platform in Fig. 2(b), which is clearly misaligned, was calculated to be at an angle of -6.5° to the Determination of tube type horizontal.The method for determining the type of tube is illustrated in Fig. 8. The two circles C1 and C2, computed during the furnace head alignment process, are used in this step. The radii of the Detection and characteristics of capillary tip two circles are compared, and if the two circles have the same To detect the presence of the autosampler capillary tip, the radius, the tube is a normal, straight-edged tube, whereas if CCD image is scanned horizontally at a number of different the radii are different, the tube is part-ridged.vertical positions, and the number of boundary points for each When the tube is part-ridged, the alignment of the tube is scan, n, is counted. If n is equal to 2, only the edges of the computed by finding the intersections of the ridge with the tube have been detected and there is no tip at this level. inner diameter of the tube. The image is scanned along a circle However, if n>2, one of the boundary points corresponds to whose centre is the centre of C2 and whose radius is the a boundary of the tip, implying the presence of the tip at this average of the radii of C1 and C2.The intersection points, i.e., position. The end of the tip can be found by using a binary points A and B in Fig. 8, are detected as boundary points recursive bisection method,3 which is a computationally fast along this circle. The angle of the tube to the horizontal is the search procedure which eliminates the need to scan the angle of the line AB joining the end points of the ridges.whole image. When the position of the capillary tip end is known, the Determination of the angle of the platform depth and the angle of the tip to the vertical axis may be computed as shown in Fig. 10. The depth of the tip is easily The angle of the platform can be approximated by calculating calculated by scanning the image vertically from the end of the angle of the line passing through the points A, B and C, the tip to the bottom of the tube, or the platform when it shown in Fig. 5. To obtain a more accurate measure of the exists, and by calculating the distance between these two angle, more points along the top edge of the platform are points. By knowing the internal diameter of the tube, the depth found by scanning across the platform. This is done by of the capillary can be accurately calculated. With a tube producing an orthogonal projection from the centre of the diameter of 5 mm the position of the tip end can be calculated tube cross-section to the platform, using the approximate value to approximately 0.1 mm.For Fig. 2(a), the capillary tip was of the angle of the platform as a guide. The image of the calculated to be at a height of 0.8 mm above the platform. platform is then scanned either side of this projection to The angle of the tip is obtained after having found two determine the boundary points. This is illustrated in Fig. 9, points on each side of the tip.Scanning the image horizontally where five equidistant points are shown scanned across the just above the end of the tip gives two points on each side of centre of the platform. Linear regression of these points is then the tip, labelled A and A¾ on Fig. 10, and the two others, used to give the best line fitting these points. The angle of the labelled B and B¾, are found by analysing the image along a circle whose centre is the centre of the tube and whose radius is smaller than the radius of the tube.The angle of the tip to the vertical axis is then the mean of the angle of each line, i.e., AB and A¾B¾, joining the detected points on each side of the tip. The tip is considered central when the position of the end of the tip is on the central vertical axis of the tube, i.e., d=0 in Fig. 10. Fig. 8 Detection of the tube type and the tube angle. Fig. 10 Characteristics of the autosampler capillary tip. Fig. 9 Computation of the angle of the platform.Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12 297Sample Injection Analysis System General algorithm The flow diagram for the algorithm is shown in Fig. 11. Two The second component of the CCD image processing software analyses images taken after injection to determine whether the images are required, the first is a reference image which contains an image of the capillary tip prior to injection. The liquid has been deposited correctly into the tube, and if there is any residue left on the capillary tip.The GFTV system can second image is of the tip at the same position, but after injection of the sample. Each image is then pre-processed and record images during the sample injection and drying phase of the furnace programme and then the movable mirror is the two images are subtracted to produce a difference image. This image is then further processed, resulting in an image of automatically rotated out of the optical axis.Hence, images of each injection in an analysis sequence can be obtained. the sample left in the tube. Pre-processing the CCD images The two CCD images are first pre-processed using the same method previously described, i.e., using a threshold operator and an opening operator. In order to remove objects in the image of the capillary tip, due to transmission of light, further processing is required. In this discussion an ‘object’ is taken to mean a collection of adjacent pixels whose value is equal to one.Firstly, an erosion operator1 is used to remove any small objects. The erosion operator works by moving a disc, of diameter d, around the inside of the boundary of an object. If the disc is entirely included within the object, the boundary point is replaced by the centre of the disc. The effect of such an operator is to remove small objects (of a size <d), to shrink other objects, to discard peaks on the object boundary and to disconnect some parts of the object.The parameter d is chosen to be equal to half of the width of the capillary tip. Small regions of transmitted light within the image of the capillary tip will then be removed in the eroded image, however, the size and shape of the image of the capillary tip, the sample droplet, the tube wall and any platform will be distorted. This image is then used in a reconstruction operation1 as a ‘marker Fig. 11 Flow chart of general analysis system for sample injection.image’. A reconstruction operator requires two images, the Fig. 12 Processing of a capillary tip image to remove the effects of transmitted light. (a) Raw unprocessed CCD image; (b) tip image after processing by thresholding and opening operator; (c) tip image after erosion; and (d) reconstruction of image (b) using image (c) as a marker. 298 Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12original image and an image containing the markers.The to induce capillary action and cause the sample to flow up the outside of the tip. These raw images are then pre-processed reconstructed image is formed from the original image by including only those objects whose boundary contains the and subjected to the erosion and reconstruction operators, the results of which are shown in Fig. 13(c) and (d). The logical position of a marker in the second image. The marker may be a single pixel or a collection of pixels. This operation keeps subtraction of the two images, Fig. 13(e), results in an image which shows the presence of the serum sample around the end objects in the original image which are larger than d, so that the size and shape of the image of the capillary tip, the sample of the capillary tip and also around the injection hole.The use of the opening operator cleans up the subtracted image by droplet, tube wall and any platform remain unaffected. The effect of the erosion and reconstruction operators on a reference removing the individual pixels around the inner diameter of the tube, and it is this image, Fig. 13(f ), which is shown to the tip image is shown in Fig. 12, where it can be seen that the transmitted light in the original reference tip image has been user. In Fig. 14, an example of a good injection is shown. The raw images of both the reference image and of the capillary removed, without degrading the overall shape of the image of the capillary tip. Both the reference image and the image after tip after injection, have been omitted.The capillary tip was adjusted to the correct depth and the delivery of the serum sample injection are subjected to this pre-processing. sample into the tube was acceptable. The subtracted image, Fig. 14(c), is completely blank except for some pixels around Sample visualization process the inner diameter of the tube. Once the opening operator has been used the resulting image, Fig. 14(d), is completely blank.To determine the position of the liquid in the tube, the logical The sample cannot be seen as a part-ridged tube was used and subtraction on a pixel-by-pixel basis of the pre-processed image the sample is lying below the height of the ridge. (after injection of the sample) and the pre-processed reference image is computed. The resulting difference image is processed, with an opening operator, to remove pixels due to any small CONCLUSION difference of position between the two images.The final image is an image of the position of the liquid in the tube and any A number of image processing algorithms have been developed for use with a CCD camera system that provides cross- residual liquid left on the tip. This process is illustrated in Fig. 13. A reference image, of the tube with the capillary tip sectional images of a graphite furnace atomizer tube. The algorithms are designed to provide the user with information prior to injection, and an image after injection of a blood serum sample with the capillary tip at the same position as in which is useful when initially setting up the atomizer and spectrometer prior to an analysis.The alignment of the furnace the reference image, are shown in Fig. 13(a) and (b), respectively. The capillary tip was deliberately set too low in the tube head to the CCD camera is checked, which is important to Fig. 13 Processing of images for sample visualization, poor injection of blood serum.(a) Raw capillary tip reference image; (b) raw capillary tip image after injection; (c) pre-processed version of image (a); (d) pre-processed version of image (b); (e) logical subtraction of images (c) and (d); and (f ) image (e) after application of an opening operator. Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12 299Fig. 14 Processing of images for sample visualization, good sample injection. (a) Pre-processed capillary tip reference image; (b) pre-processed capillary tip image after injection; (c) logical subtraction of images (a) and (b); and (d) image (c) after application of an opening operator.minimize the passage of continuum emission from the tube used. Both of these effects will lead to problems during an analysis. The algorithms developed here can be used to warn wall into the monochromator. Furthermore, a measure of the the user when a poor injection has occurred, or to instruct the angle to the horizontal axis of a platform is provided, which instrument to disregard the signal and repeat the measurement. can ensure that the platform is correctly aligned. For optimum control of the furnace heating, the type of tube used is automatically determined. A quantitative measure of the depth REFERENCES of the capillary tip in the tube is also obtained. This is a useful 1 Serra, J., Image Analysis and Mathematical Morphology, Academic measure as the analyst can set the depth to an accuracy of Press, New York, 1983. #0.1 mm, which means that once the optimum depth for a 2 Bronshtein, I. N., and Semendyayev, K. A., A Guide-book to particular sample type has been found, the capillary tip can Mathematics, Verlag Harri Deutsch, Frankfurt, 1971. be reset to exactly this depth each time the sample type 3 Press, W. H., Teukolsky, S. A., Vetterling, W. T., and Flannery, is analysed. B. P., Numerical Recipes, Cambridge University Press, In addition, algorithms have been developed to analyse Cambridge, 1992. images recorded during injection with an autosampler. These may be used to monitor the liquid delivery and to provide Paper 6/06255E warnings if, for example, there is residual liquid on the tip, or Received September 10, 1996 Accepted December 10, 1996 sample in contact with the tube wall if a platform is being 300 Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12

ISSN:0267-9477    

DOI:10.1039/a606255e

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Slurry Sampling Electrothermal Atomic Absorption Spectrometric Determination of Lead, Cadmium and Manganese in Human Hair Samples Using Rapid Atomizer Programs† PILAR BERMEJO-BARRERA*, ANTONIO MOREDA-PIN� EIRO, JORGE MOREDA-PIN� EIRO AND ADELA BERMEJO-BARRERA Department of Analytical Chemistry, Nutrition and Bromatology, Faculty of Chemistry, University of Santiago de Compostela, Avenida de las Ciencias, s/n E-15706 Santiago de Compostela, Spain Rapid methods for the determination of lead, cadmium and such as the atmosphere, dirt, dust, sweat, cosmetic and pharmaceutical preparations occur.In addition, the concentration of manganese in human scalp hair by ETAAS using rapid atomizer programs and deuterium arc background correction a metal varies along the length of a hair from the root to tip, and for lead, a slight increase towards the tip has been have been developed. Samples were powdered by means of a zirconia vibrating ball mill over a period of 20 min (mean reported.6 Another drawback that usually occurs when a metal determination is performed, is the fact that the concentration particle size less than 1 mm).Then 0.1 g of the powder was suspended in a few ml of water and diluted to 25 ml. An of a certain metal in hair is dependent on the sampling site.7 To overcome problems related to contamination, an adequate optimum drying temperature of 250 °C was found for lead, cadmium and manganese. Optimum atomization temperatures hair washing stage prior to measurement is needed.In this context, the concentration of some metals, such as lead, has of 2200 and 1900 °C were found for cadmium and manganese, and for lead, respectively. The program cycles were 38 s for been demonstrated to be dependent on the washing procedure carried out,4 the most appropriate being the washing process cadmium and lead, and 37 s for manganese. Glycerol, at an optimum concentration of 0.4% m/v, was used as a stabilizing recommended by the IAEA.8 To overcome the problems associated with variation along the length of a hair and the agent.Simple aqueous calibration was used for all analytes. The limits of detection were 0.03, 0.05 and 0.04 mg kg-1 for sampling site, it has been demonstrated that the segments of hair follicles that are still biologically active are the most cadmium, lead and manganese, respectively. Accuracy was studied by analysing CRM 397 human hair, and the cadmium representative and should be collected from the occipital region of the head, as close as possible to the scalp.7 and lead levels were found to be in accordance with the certified accuracy values.The levels of metals obtained were in However, as hair is a solid material, an initial sample decomposition step is necessary, and various procedures such agreement with those previously reported for healthy people in Europe, i.e., less than 3.0 and 0.3 mg kg-1 for lead and as acid4,7,9–11 or alkali12 digestions have been described.In order to avoid problems introduced by this sample pre- cadmium, respectively. treatment, such as contamination or loss of volatile elements, Keywords: Electrothermal atomic absorption spectrometry; the slurry sampling technique appears to be a suitable alterna- slurry sampling; cadmium; lead; manganese; human hair tive. This approach has been applied successfully with ETAAS for the determination of several metallic elements in a wide range of biological and environmental samples.13 Lead, cadmium and manganese are classified as prevalent toxic metals which tend to be concentrated in environmental systems In the present paper, methods are proposed for the determination of lead, cadmium and manganese in slurries of scalp and humans.Absorption from the air in the local environment and intake from the diet are the major sources of human hair using atomizer programs without a charring stage. The omission of the charring stage offers a rapid ETAAS determi- exposure.Although toxic metal accumulation in humans is dependent upon the nature of the toxic metal compound and nation, as the program cycles have been shortened.14 This fact, in conjunction with the rapid slurry preparation procedure, the environmental metal load, lead has been demonstrated to be accumulated in bone and in some soft tissues, such as the makes the proposed methods faster than those reported previously for determination of metals in hair.4,7,9–12 liver, kidneys and brain, while cadmium is known to damage organs such as the kidneys, liver and lungs.1 Moreover, although excess of manganese is a relatively rare phenomenon, some manganese poisoning has been reported and it has been EXPERIMENTAL established that the manganese is assimilated by the liver, Apparatus heart and skeleton.2 Hair has been shown to be a major vehicle for the excretion All absorbance measurements were performed on a Perkinof toxic metals,3 and thus hair analysis has been used as an Elmer (U� berlingen, Germany) 1100B atomic absorption specappropriate technique in forensic science4 and as an index of trometer equipped with a deuterium lamp for background environmental exposure.5 Because the hair is a vehicle for correction, a Perkin-Elmer HGA-400 graphite furnace and a excretion of metals, the concentration of toxic metals in hair Perkin-Elmer AS-40 autosampler.Spectrometric operating is up to ten-fold higher than the levels found in blood or urine conditions are given in Table 1.A Laser Coulter Series LS100, samples.3,4 Nevertheless, external contamination from sources Fraunhofer Optical Model particle sizer (Coulter Electronics, Hialeah, FL, USA) was used to obtain the particle size distributions. A vibrating ball mill (Retsch, Haan, Germany), † Presented at the Eighth Biennial National Atomic Spectroscopy Symposium (BNASS), Norwich, UK, July 17–19, 1996.equipped with zirconia cups and zirconia balls (7 mm Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12 (301–306) 301Table 1 ETAAS temperature programs and spectrometer operating of the hair powder was transferred into a polyethylene vial conditions (Zinsser Analytic, Frankfurt, Germany) and 1 ml of ultrapure water was added, with gentle shaking to suspend the hair Lead Cadmium Manganese particles. Finally, the hair slurry was adjusted to a final volume Operating conditions— of 25 ml by addition of ultrapure water. The hair slurries were Lamp current/mA 10 4 30 kept in polyethylene vials at 4 °C.Wavelength/nm 283.3 228.8 279.5 Spectral bandwidth/nm 0.7 0.7 0.2 D2 arc background correction On On On Measurement Injection volume/ml 20 20 20 Measurement mode Peak area Peak area Peak area Portions of 0.50 or 0.75 ml of the slurry, for lead and manga- Integration time/s 3 3 2 nese, or for cadmium, respectively, with appropriate volumes of palladium, magnesium nitrate and glycerol solutions, were ETAAS temperature programs— transferred into autosampler cups and then made up to 1 ml and stirred magnetically before measurement.The concen- Temperature Ramp Hold Ar flow /°C /s /s /ml min-1 tration of glycerol was 0.2% v/v in all cases. A concentration of 5 mg l-1 of palladium was used for the determination of Drying 250 20 10 300 manganese, while concentrations of 15 and 10 mg l-1 of Atomization 1900* 0† 3‡ 0 (read) Cleaning 2500 1 2 300 palladium and magnesium nitrate, respectively, were used for the lead and cadmium determinations.In all instances, volumes * 2200 °C for cadmium and manganese. of 20 ml were introduced into the atomizer, the sequential dry– † 1 s for cadmium. atomize–clean programs of the graphite furnace (Table 1) were ‡ 2 s for cadmium and manganese. run and the integrated absorbance recorded. Platform atomization was used for lead and manganese but wall atomization diameter), was used to pulverize samples and to reduce the was chosen for cadmium, because better performances were particle size of the powdered samples.An Agimatic magnetic obtained.16,17 agitator from Selecta (Barcelona, Spain) was used to suspend the slurry particles just prior to measurement. RESULTS AND DISCUSSION Reagents The slurry preparation procedure was the same for lead, m and manganese. The particle size distribution meas- Acetone. 99.7%, Carlo Erba Analytical, Milan, Italy.ured by laser diffraction showed that the mean particle size Cadmium nitrate stock standard solution. 1.000 g l-1, Merck, was less than 1 mm for slurries prepared from the powdered Darmstadt, Germany. hair. The time chosen was 20 min, as for lower times portions Glycerol. Sigma, St. Louis, MO, USA. of unpowdered hair were obtained. L ead nitrate stock standard solution. 1.000 g l-1, BDH, An initial acid pre-digestion of the slurries, recommended Poole, UK.by some workers to improve the repeatability,18–21 was carried Magnesium nitrate stock standard solution. 2.000 g l-1, soluout, adding nitric acid at varying concentrations of between 1 tion prepared from magnesium nitrate, BDH. and 15% v/v, to different slurries. For nitric acid concentrations Manganese nitrate stock standard solution. 1.000 g l-1, greater than 2% v/v, lead, cadmium and manganese were Roimil, Cambridge, UK. mobilized into the liquid phase, and similar analyte Nitric acid. 35.0% v/v, prepared from AnalaR nitric acid, absorbances were obtained when the slurry and the liquid 69.0–70.5% v/v, BDH. phase were injected. The within-run precision corresponding Palladium stock standard solution. 3.000 g l-1, solution preto 11 replicate injections of slurries, prepared with and without pared according to Welz et al.15 from palladium, 99.999%, nitric acid, was calculated and similar RSD values were Aldrich Chemicals, Milwaukee, WI, USA.obtained for both, 3.2 and 3.5% for slurries with and without Reference material, CRM 397 human hair. From the nitric acid, respectively, for lead measurements. Similar results Commission of the European Communities Community were achieved for cadmium and manganese. Thus the addition Bureau of Reference (BCR), with certified lead and cadmium of nitric acid does not improve precision, and therefore, use of contents of 33.0±1.2 mg kg-1 and 0.521±0.024 mg kg-1, aqueous slurries was preferred.respectively. T riton X-100. Merck. V iscalex HV 30. Allied Colloids, Bradford, UK. Optimization of the ETAAS Temperature Programs Drying temperatures of between 150 and 400 °C were tested Procedures for experiments on hair slurries. These studies were made Hair washing without chemical modification, and with palladium, magnesium nitrate and palladium–magnesium nitrate as the chemi- Approximately 0.5 g of hair sample was cut with stainless-steel cal modifiers for lead, cadmium and manganese, respectively.scissors from the scalp region, and the hair length varied In Fig. 1 (a–c), the background signals obtained for different between 1 and 3 cm. Washing of hair samples was required to drying temperatures in the presence and absence of chemical provide an accurate assessment of the endogenous metal modification are shown. (NH4)2HPO4, which has been content. The washing procedure carried out was that proposed reported to be an appropriate chemical modifier for cadmium, by the IAEA,8 and thus hair samples were first washed with does not enable efficient volatilization of the matrix to take ultrapure water, then washed three times with acetone, and place, and thus, high background signals were obtained even finally, they were again washed with ultrapure water (three at high drying temperatures (0.720 s for 400 °C), deuterium times).The samples were then oven-dried at 100 °C. arc background correction being unsuitable [Fig. 1(b)]. Similar results were obtained for hydrogen peroxide,20,21 which has Pulverization and slurry preparation been proposed recently for oxidizing the matrix sample at low temperatures. Therefore, palladium or palladium–magnesium The hair samples were pulverized in a vibrating zirconia ball mill over a period of 20 min using 75% power. A 0.1 g portion nitrate appear to be the best chemical modifiers. However, the 302 Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12Fig. 2 Effect of various synthetic air ($) and oxygen (&) flows on the cadmium absorbance (solid lines) and background (dotted lines) signals for a hair slurry spiked with 0.5 mg l-1 of Cd2+. Table 2 Within-run precision for different drying temperatures for a hair slurry in the determination of lead, cadmium and manganese in the absence of chemical modifiers and also in the presence of Pd–Mg (NO3 )2 RSD in the presence RSD in the absence of Pd–Mg(NO3)2 of chemical (%) modifiers (%) Drying temperature/ °C Pb Cd Mn Pb Cd Mn 150 5.0 7.6 2.0 7.1 12.4 3.5 200 3.0 6.6 1.8 6.3 8.6 2.6 250 3.2 4.7 1.7 6.3 7.7 2.9 300 3.5 5.2 1.8 5.8 7.8 2.9 400 3.1 4.8 1.1 5.1 7.8 3.1 with different ramp and hold times for the drying step with lead and manganese.However, higher cadmium absorbance and lower background signals were achieved (Fig. 3) when the ramp and hold times were increased. Therefore, for cadmium 10 and 20 s were chosen as optimum for the hold and ramp times. These values were also selected for lead and manganese.Other parameters for the ETAAS temperature programs are shown in Table 1. Optimum concentrations of 10 and 15 mg l-1 of magnesium nitrate and palladium, respectively, were obtained for lead and cadmium. For manganese, an important decrease in the Fig. 1 Effect of the drying temperature on the background signal in absorbance was observed when magnesium nitrate was present, the determination of lead (a), cadmium (b) and manganese (c) in the absence of chemical modification ($), and in the presence of Pd (|), Mg(NO3)2 (*), Pd–Mg(NO3)2 (&), (NH4)2HPO4 (x) and hydrogen peroxide (2).background signals obtained for cadmium were high, about 0.250 s. In order to decrease the background signals obtained for cadmium measurements, synthetic air and oxygen as alternative purge gases were tried, and the drying step thus converted into an oxidative process.22 The results obtained for flows in the range 30–150 ml min-1 are shown in Fig. 2. An important decrease in the background signal is achieved even for low flows, however, the cadmium absorbances are also decreased, so this approach is unsuitable. From studies on within-run precision (Table 2) and background signal (Fig. 1), 250 °C was chosen as a suitable drying temperature. The ramp rate and the hold time for the drying step were Fig. 3 Dependence of the combined effects of different ramp and hold studied through experiments involving different combinations times for the drying step on the cadmium absorbance (solid lines) and of ramp rates and hold times.Results showed that there is no background (dotted lines) signals for a hair slurry spiked with 0.5 mg l-1 of Cd2+. significant variation in the absorbance and background signals Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12 303calibration and standard additions slopes are very similar, any statistical difference when the t-test23 was applied for a confi- dence level of 95.0% not being found.Thus, it can be assumed that there is no matrix effect and the calibration graphs can be used when taking measurements in the determination of lead and manganese. However, there is a significant difference between the two slopes for cadmium, and thus, the matrix effect becomes important, so the standard additions method is necessary when determining cadmium. Detectability The limits of detection and quantification (LOD and LOQ) are defined as 3s/m and 10s/m, respectively, where s is the standard deviation and m is the slope of the calibration graph (lead and manganese) or standard additions graph (cadmium); for the determination of lead these were 0.05 and 0.18 mg kg-1, Fig. 4 Dependence of the combined effects of various amounts of Pd respectively. For cadmium, the LOD and LOQ were 0.03 and and Mg(NO3)2 on the manganese absorbance signal for a hair slurry. 0.09 mg kg-1, and 0.04 and 0.13 mg kg-1 for manganese.The characteristic masses (n=4) obtained were 18.5±0.6, 0.9±0.2 as shown in Fig. 4 where results for manganese absorbance as and 2.1±0.3 pg for lead, cadmiumand manganese, respectively. a function of various combinations of palladium and mag- To obtain the lowest LOD and LOQ using the slurry nesium nitrate are plotted. Therefore, the optimum amount of preparationmethod described, an increase in the slurry concenpalladium for determining manganese was found to be 20 mg tration was necessary.However, an increase in the slurry l-1. concentration can worsen the analytical performance. Therefore, the effect of sample concentration on the precision was studied, slurries of 0.05, 0.075, 0.10, 0.20 and 0.40 g of a Effect of Wetting Agents on the Stability of the Slurries hair sample were prepared, with slurry concentrations of 0.02, 0.03, 0.04, 0.08 and 0.16% m/v, respectively. To study the The effect of different wetting agents, such as glycerol, Triton precision, 11 replicate injections of each slurry were made, X-100 and Viscalex HV40, on the stability of the slurries was different volumes of Pb2+, Cd2+ and Mn2+ aqueous standard carried out by determining the within-run precision (n=11).solutions being added to slurries prepared from smaller RSD values related to each wetting agent and for a slurry amounts of hair sample in order to achieve similar absorbances. without wetting agent are shown in Table 3.Glycerol and Results obtained for lead were 2.2, 1.7, 1.1, 2.0 and 1.2% RSD Triton X-100 give the smallest RSDs in comparison with for 0.02, 0.03, 0.04, 0.08 and 0.16% m/v, respectively. For Viscalex HV40 and a slurry without a wetting agent. Moreover, manganese, RSDs of 1.5, 1.7, 1.5, 1.6 and 1.8% were obtained the use of these agents does not increase the background for slurry concentrations of between 0.02 and 0.16% m/v. For signal, as these are similar to those obtained for slurries cadmium, RSDs of less than 5% were achieved for slurry without any agent present.Glycerol was chosen as the most concentrations lower than 0.08% m/v, 3.8, 3.9, 4.0 and 4.7% suitable agent, 0.4% m/v being the optimum concentration to for 0.02, 0.03, 0.04 and 0.08% m/v, the background signal for achieve a good stabilizing effect on the slurry. a slurry concentration of 0.16% m/v being too high to be suitable for the deuterium arc background correction system. Calibration and Standard Additions Graphs Using a slurry concentration of 0.16% m/v and taking 20 ml of this solution the LODs will therefore be 13.4 and 9.9 mg kg-1 Equations for calibration and standard additions graphs are for lead and manganese, respectively.For cadmium, the LOD given in Table 4. For lead, cadmium and manganese, the is reduced to 14.2 mg kg-1 (0.08% m/v). Table 3 Within-run precision achieved for different wetting agents added at a concentration of 0.1% m/v Precision and Accuracy RSD (%) The within-batch precision and the analytical recovery of the methods, obtained for 11 replicates of samples of hair slurries Wetting agent Lead Cadmium Manganese spiked with 0, 10, 20 and 30 mg l-1 of Pb2+, 0, 0.5, 1.0, 1.5 Triton X-100 3.3 5.6 2.4 and 2.0 mg l-1 of Cd2+ and 0, 1, 2, 4 and 5 mg l-1 of Mn2+, Viscalex HV40 7.3 10.2 5.7 are shown in Table 5.In addition, the accuracy of the methods Glycerol 3.4 4.8 1.5 was evaluated by analysis of a reference material, CRM 397 Without wetting agent 7.5 13.3 8.1 human hair, from the Commission of the European Table 4 Calibration and standard additions equations Analyte Calibration Standard additions Lead QA*=0.002+4.74 10-3 [Pb] QA=0.058+4.55 10-3 [Pb] r†=0.999 r=0.999 Cadmium QA=0.001+0.110 [Cd] QA=0.053+0.096 [Cd] r=0.999 r=0.998 Manganese QA=0.002+0.037 [Mn] QA=0.069+0.034 [Mn] r=0.998 r=0.999 * QA=integrated absorbance. † r=regression coefficient. 304 Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12Table 5 Within-batch precision and analytical recovery of the methods [Pb] Analytical [Cd] Analytical [Mn] Analytical /mg l-1 RSD (%) recovery (%) /mg l-1 RSD (%) recovery (%) /mg l-1 RSD (%) recovery (%) 0 5.2 — 0 5.8 — 0 2.5 — 10 2.7 97.0±2.1 0.5 2.3 98.0±0.1 1 1.2 107.0±1.7 20 2.6 97.0±2.0 1.0 1.8 102.0±0.2 2.5 1.2 97.0±1.2 30 2.2 101.7±1.8 2.0 1.9 100.0±0.3 5.0 1.3 100.7±1.1 Table 6 Analysis of reference materials [Pb]/mg kg-1 [Cd]/mg kg-1 [Mn]/mg kg-1 Reference material Certified Found* Certified Found* Certified Found* CRM 397 33.0±1.2 32.0±0.4 0.521±0.024 0.530±0.009 11.2±0.3† 11.9±0.4 DOLT-1 1.36±0.29 1.26±0.10 4.18±0.28 4.17±0.06 8.72±0.53 8.73±0.10 DORM-1 0.40±0.12 0.38±0.06 0.086±0.012 0.082±0.005 1.32±0.26 1.30±0.09 * Value expressed as x± ts ÓN where N=number of determinations, s=standard deviation, x=mean value, t=Student’s t; t=2.23 (N-1=10).† Informative value only.Table 7 Analytical recovery (n=11) for different slurry concentrations Analytical recovery (%) Slurry Lead Cadmium Manganese concentration (% m/v) +10 mg l-1 +20 mg l-1 +0.5 mg l-1 +1.0 mg l-1 +1.0 mg l-1 +2.5 mg l-1 0.02 98.6±2.0 100.3±0.8 98.0±1.1 100.1±0.7 102.4±2.3 99.4±1.2 0.03 103.7±2.0 98.0±2.4 102.0±1.7 99.3±0.9 98.8±2.4 100.4±1.1 0.04 98.2±2.4 100.6±1.9 97.2±0.8 101.1±0.6 99.2±1.9 102.3±0.9 0.08 100.0±2.8 100.2±2.5 105.5±2.1 103.3±1.5 105.6±1.3 98.6±1.5 0.16 98.6±4.8 102.9±2.8 121.0±5.4 113.3±3.8 106.8±3.7 98.2±3.2 Table 8 Effect of different elements on the lead, cadmium and manga- Communities Community Bureau of Reference (BCR).The nese absorbance signals reference materials DOLT-1 and DORM-1 (dogfish liver and dogfish muscle) from the National Research Council of Canada Concentration Maximum Variation in were also investigated with each method. It is shown in Table 6 added (referred concentration absorbance (%) that the lead and cadmium values found are in agreement with to sample) reported Species /mg kg-1 /mg kg-1* Pb Cd Mn certified lead and cadmium contents.For manganese, the value found in the CRM 397 material is different from that reported, Al3+ 500 10 -3.2 -4.9 +3.5 however, it must be noted that this value is not certified, only Ba2+ 500 10 -3.2 -4.3 +2.2 Ca2+ 1000 1000 -3.1 -1.2 +0.6 informative. Cd2+ 500 1 +1.1 — +2.9 Moreover, a study on the effect of the slurry concentration Cu2+ 500 500 +5.3 -1.8 +3.4 on analytical recovery was also performed.Analytical recover- Cr3+ 500 1 +4.2 -4.9 +0.6 ies were thus obtained relating to hair slurries prepared from Fe3+ 500 500 -2.1 -1.8 +2.9 0.05, 0.075, 0.10, 0.20 and 0.40 g of sample, to achieve the same K+ 25000 1000 -1.1 -3.1 +3.4 slurry concentrations as indicated previously. The analytical Mn2+ 5000 10 -1.1 +4.9 — Na+ 25000 1000 -1.1 +1.2 +1.7 recoveries shown in Table 7 were close to 100% for all sample Ni2+ 500 10 -4.3 +1.2 +1.7 concentrations and for the two concentration levels investi- Pb2+ 500 100 — -3.7 +1.7 gated for lead and manganese. However, for cadmium, a good Zn2+ 500 500 -1.1 -5.2 +2.9 analytical recovery was not achieved for 0.4 g of hair, owing Cl- 25000 1000 -4.3 -3.1 +3.4 to the high background signal obtained.Therefore, the amount PO42- 50000 1000† +5.2 +3.7 -1.2 of hair sample can be increased up to 0.4 g without loss of SiO32- 5000 >0.001† +2.1 -4.9 +0.6 SO42- 50000 1000† +1.1 +0.6 +4.1 analytical recovery for lead and manganese, and up to 0.2 g for cadmium.* Data from ref. 7. † Expressed as total P, Si and S, respectively. Interferences Various amounts of several cationic and anionic species were lower than the within-run precision compared with unspiked added to a hair slurry at a higher concentration than would hair slurries, 5.2, 5.8 and 2.5% for lead, cadmium and mangabe found in this type of sample. The results obtained for the nese, respectively.concentration of each interferent added and the concentration equivalent in the hair sample, together with the variation in Applications the absorbances for lead, cadmium and manganese, are shown in Table 8. In addition, the concentration intervals reported in The proposed methods were applied to the rapid determination of lead, cadmium and manganese in 14 human scalp hair the literature7 for some of the species studied are also shown. As can be seen, the variation in absorbances obtained was samples from healthy people.Each hair sample was prepared Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12 305Table 9 Lead, cadmium and manganese levels in several hair samples, the deuterium arc system. Chemical modifiers such as pal- expressed as mean±s, corresponding to two analyses ladium or magnesium nitrate produce beneficial effects at the drying temperatures used, thus better precision and lower Sample [Pb] /mg kg-1 [Cd] /mg kg-1 [Mn] /mg kg-1 background signals can be obtained.It appears that 1 16.79±0.04 160.2±0.25 1.09±0.01 (NH4)2HPO4 and hydrogen peroxide offer no improvement in 2 11.72±0.02 30.9±0.04* 0.45±0.01 terms of matrix volatilization. 3 11.12±0.02 42.1±0.06* 0.18±0.01 The proposed method, rapid ETAAS determination com- 4 10.72±0.02 64.3±0.09* 0.51±0.01 bined with slurry sampling, is simpler and faster than those 5 20.36±0.06 40.5±0.06* 6.16±0.02 6 8.34±0.01 78.6±0.13* 0.72±0.01 previously reported for trace elements in hair samples, and 7 19.14±0.06 62.8±0.10* 1.17±0.01 thus is attractive for the routine determination of low trace 8 22.90±0.06 26.5±0.04* 0.40±0.01 element levels in hair as an index of environmental exposure. 9 26.56±0.08 114.3±0.20 0.92±0.01 10 12.93±0.02 30.5±0.04* 0.35±0.01 11 5.50±0.01 213.0±0.39 0.72±0.01 REFERENCES 12 20.45±0.05 10.3±0.01* 0.15±0.01 13 19.54±0.05 48.8±0.07* 0.22±0.01 1 Metals and their Compounds in the Environment, ed.Merian, E., 14 21.40±0.06 136.2±0.27 0.14±0.01 VCH, New York, 1991. 2 Rennert, O. M., and Chan, W. Y., Metabolism of T race Metals in * Slurry concentration of 0.08% m/v. Man, CRC Press, Boca Rato�n, FL, 1984. 3 Capel, I. D., Pinnok, M. H., Donell, H. M., Williams, D. C., and Grant, E. C. G., Clin. Chem., 1981, 27, 879. Table 10 Results of the application of the t-paired test for the 4 Burguera, J. L., Burguera, M., Rondo�n, C. E., Rivas, C., comparison of the conventional and the rapid ETAAS method; Burguera, J.A., and Alarco�n, O. M., J. T race Elem. Electrolytes t(n=9)=2.26 Health Dis., 1987, 1, 21. 5 De Antonio, S. M., Katz, S. A., Scheiner, D. M., and Wood, J. D., Analyte xd*/mg kg-1 s/mg kg-1 tcal Clin. Chem., 1982, 28, 2411. Lead 0.065 0.402 0.51 6 Alder, J. F., and Batoreu, M. C. C., Anal. Chim. Acta, 1983, Cadmium -0.032 0.455 -0.22 155, 199. 7 Bencze, K., Fresenius’ J. Anal. Chem., 1990, 338, 58. * xd is the difference between the concentrations obtained applying 8 Report on the Second Research Co-ordination Meeting of IAEA, the conventional and the rapid methods.Neuherberg, 1985. 9 Voellkopf, U., and Grobenski, Z., At. Spectrosc., 1984, 5, 115. 10 Castillo, J. R., and Ferna�ndez, A., Microchem. J., 1989, 39, 224. 11 Srikumar, T. S., Ka�llgard, A., Lindeberg, S., O� ckerman, P. A., in the form of a slurry, and each one was subjected to ETAAS and Akesson, B., J. T race Elem. Electrolytes Health Dis., 1994, analysis twice.The slurry concentration chosen was 0.04% 8, 21. m/v for all lead and manganese determinations. As the cad- 12 Uchida, T., Isoyama, H., Yamada, K., Oguchi, K., Nakagawa, G., mium concentrations in hair samples were low, a slurry Sugie, H., and Iida, C., Anal. Chim. Acta, 1992, 256, 277. concentration of 0.08% m/v was used for some cadmium 13 Bendicho, C., de Loos-Vollebregt, M. T. C., J. Anal. At. Spectrom., determinations. The results are shown in Table 9. 1991, 6, 353. 14 Halls, D. J., J. Anal. At. Spectrom., 1995, 10, 169. 15 Welz, B., Schlemmer, G., and Mudakavi, J. R., J. Anal. At. Comparison With Conventional ETAAS Determinations Spectrom., 1988, 3, 695. 16 Bermejo-Barrera, P., Moreda-Pin�eiro, A., Romero-Barbeito, T., Results for the application of the rapid method to the determi- Moreda-Pin�eiro, J., and Bermejo-Barrera, A., Clin. Chem., 1996, nation of lead and cadmium in hair slurries were compared 42, 1287. with those obtained from conventional ETAAS methods.16,24 17 Bermejo-Barrera, P., Barciela-Alonso, M. C., and Bermejo- Barrera, A., Mikrochim. Acta, 1996, 124, 251. The t-paired test was applied to compare the results of each 18 Miller-Ihli, N. J., Spectrochim. Acta Part B, 1989, 44, 1221. method for lead and cadmium. In Table 10 the results obtained 19 Miller-Ihli, N. J., Fresenius’ J. Anal. Chem., 1990, 337, 271. for nine hair samples and for a confidence interval of 95.0% 20 Vin�as, P., Campillo, N., Lo�pez-Garcý�a, I., and Herna�ndez- (P=0.05) are shown. The results from the rapid method are Co�rdoba, M., T alanta, 1995, 42, 527. comparable to those obtained with conventional ETAAS 21 Vin�as, P., Campillo, N., Lo�pez-Garcý�a, I., and Herna�ndez- methods. Co�rdoba, M., At. Spectrosc., 1995, 16, 86. 22 Stephen, S. C., Ottaway, J. M., and Littlejohn, D., Fresenius’ Z. Anal. Chem., 1987, 328, 346. CONCLUSIONS 23 Miller, J. C., and Miller, J. N., Statistics for Analytical Chemistry, Wiley, New York, 1984. Rapid temperature programs have been shown to be suitable 24 Bermejo-Barrera, P., Moreda-Pin�eiro, A., Romero-Barbeito, T., for the determination of trace (lead and manganese) and Moreda-Pin�eiro, J., and Bermejo-Barrera, A., T alanta, 1996, ultratrace (cadmium) metals in complex samples without prior 43, 1099. decomposition procedures. The methods are limited by the background absorbance signals when ultratrace metals such Paper 6/05073E as cadmium are determined. This can be easily overcome using Received July 22, 1996 the Zeeman-effect background correction system rather than Accepted September 27, 1996 306 Journal of Analytical

ISSN:0267-9477    

DOI:10.1039/a605073e

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Modulated Glow Discharge Source With Supplementary Microwave Excitation† EDWARD B. M. STEERS*a AND FRANZ LEISb aSECEAP, University of North L ondon, 166–220 Holloway Road, L ondon,UK N7 8DB bInstitut fu� r Spektrochemie und Angewandte Spektroskopie (ISAS), Bunsen-Kirchhoffstrasse 11, D-44139 Dortmund, Germany Spectral interferences occurring in a microwave-boosted glow charge, for a constant dc current in the primary discharge. Typical values of F for atomic lines of the sample material discharge source due to emission lines of the carrier gas and to band systems of molecular species were investigated.A method range between 10 and 100; the value depends not only on the energy of the upper level of the transition, but also on the is described which greatly reduces such interferences. It is shown that analytical lines are strongly dependent on the atomic state involved. Ionic lines from the sample may be enhanced by a smaller amount, but often decrease in intensity current in the dc discharge whereas Ar I lines are scarcely influenced and OH bands are influenced to only a small extent.if excited by charge transfer. For the carrier gas, F is typically 5 for atomic lines and <1 for ionic lines. Band systems from By modulating the dc current with the microwave discharge running continuously and using a lock-in amplifier, selective impurity molecules present in the source (e.g., OH, NO, NH and N2 systems) are often strongly enhanced.Limits of detec- amplification of analyte atomic lines is achieved. In this way, spectral interferences by Ar I lines on Fe I lines and tion depend not only on the line intensity but also on the background level, and the improvement with the boosted interferences by lines in OH bands on the Bi I 306.771 nm and Al I 308.216 and 309.271 nm analytical lines, which occur even discharge is about one order of magnitude. Although the ‘enhancement factor’ is a convenient quantity when using a high resolution monochromator, are greatly reduced.to describe the behaviour of a source at a particular current, it fails to distinguish between two effects. Lines from species Keywords: Boosted glow discharge sources; modulated glow produced in the primary discharge (e.g., sputtered atoms, discharge sources; spectral interferences; microwave excitation carrier gas ions) become more intense with increasing current; F remains constant or may change with the current to a greater or lesser extent.On the other hand, the carrier gas Glow discharge sources (GDS) of various types have been used for many years in atomic spectroscopy. In 1967, Grimm1,2 atoms and molecular species can be strongly excited in the supplementary discharge (and are still excited if the primary introduced a special type of source, in which the sample to be analysed, usually a plane metallic sample, forms the cathode discharge is switched off ), and the total intensity of the lines produced does not depend greatly on the current in the and is situated very close to the anode tube.In operation, a carrier gas, usually an inert gas, most commonly argon, flows primary discharge, particularly if it is low; F will therefore be high for low dc currents in the primary discharge and fall through the source at a pressure of a few hPa. Ions and high velocity atoms of the carrier gas (and metallic ions from the rapidly with increasing current.Spectral interferences from the carrier gas and from molecu- cathode itself ) bombard a restricted area of the cathode; sample atoms are sputtered into the discharge and excited and lar species can degrade analytical results. This is particularly serious if low resolution spectrometers are used,4 but such ionized there. This type of source was later manufactured by various companies, and has been extensively used for the problems can often be overcome by using a higher resolution instrument which will also give a lower background. However, analysis of metals, both for bulk samples and also for depth profile analysis of coated metals.However, the ever-increasing difficulties remain, e.g., Wagatsuma and Hirokawa5 have disdemand for better analytical performance, above all for lower cussed the effect of the Ar II 437.97 nm line on the determilimits of detection, has led to many attempts to develop nation of vanadium using the V I 437.92 nm line. Interferences improved glow discharge sources.from molecular bands can be more troublesome than those One approach has been the incorporation of a supplemen- from the carrier gas, in view of the greater complexity of band tary discharge to provide more efficient excitation; such sources spectra (e.g., OH bands in the 300–310 nm region, NH bands are often described as ‘boosted’ glow discharge sources. The at 336 nm and N2 bands in various parts of the spectrum). main glow discharge producing the sputtered atoms is usually Further, the impurities can be released from the sample, and a dc glow discharge operated in constant voltage or constant the intensity of molecular bands may vary considerably with current mode and the two discharges can be described as the time.dc and the supplementary discharge. In this work the current In general, the intensity of carrier gas lines increases roughly was modulated and the main glow discharge which produces linearly with current, particularly when the anode to cathode sputtering will therefore be referred to as the primary discharge potential difference does not change greatly, whilst the lines (either dc or modulated).from sputtered atoms show a greater dependence on current. The characteristics of various forms of boosted sources have Various attempts have been made to reduce interference effects been reviewed by Leis and Steers.3 The spectral changes in a Grimm GDS by modulating the current (or voltage) and produced by any supplementary discharge are complex, using a lock-in amplifier system.6,7 A small modulation around although there are marked similarities between various forms the normal operating voltage can increase the intensity of of boosted discharges; one way to quantify these changes is to sample lines relative to the carrier gas line, but the absolute calculate the ‘enhancement factor’ F for individual spectral signals are lower.Wagatsuma and Hirokawa5,8 also successlines.This may be defined as the ratio of the intensity with the fully used a source with an intermediate hollow cathode to supplementary discharge to the intensity without that dis- allow selective modulation of sample lines. Dogan and U� lgen4 used an auxiliary electrode to balance the intensity of argon lines in the two time periods of a modulated source. In a † Presented at the Eighth Biennial National Atomic Spectroscopy Symposium (BNASS), Norwich, UK, July 17–19, 1996.microwave boosted GDS, the intensities of the Ar I lines are, Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12 (307–312) 307to a first approximation at least, independent of the current in Electronics, Bracknell, UK) was used; the dc amplifier incorthe primary discharge, and the intensities of molecular bands porated in the monochromator was initially used for measuredo not vary greatly with the current whereas the intensity of ments with a constant current in the primary discharge, but sample lines is dependent on the current.It was therefore subsequently a light chopper was used when the source was decided to investigate the use of a modulated current in the operated in an unmodulated mode (A, C or E), so that the primary discharge and it was found that interference due to Brookdeal amplifier could be used for all measurements to argon lines and molecular bands can be very significantly ensure direct comparability.All values for intensity given in decreased by this technique. this paper have been converted to a common arbitrary scale which includes the changes in the photomultiplier voltage and amplifier settings, but does not take into account variations EXPERIMENTAL with wavelength of the photomultiplier and amplifierresponses. Results from the microwave boosted GDS using a slab line The amplifier output was fed to a pen rder and to an cavity9 first showed the possibilities of this technique, but the analogue to digital converter for subsequent processing of the results reported here were obtained at the Institut fu�r data. The Gremlin package developed for handling high reso- Spektrochemie und Angewandte Spektroskopie (ISAS), lution Fourier transform spectra11 proved a convenient tool Dortmund, Germany, using the source developed by Leis for rapid comparison of spectra but was not sufficiently flexible et al.,10 which incorporates a Beenakker coaxial microwave to produce the spectral diagrams for this paper and other cavity.The source, the power supply, the computer controlled packages were also used. scanning monochromator and photomultiplier amplifier incor- The pressure of the carrier gas was measured with an MKS porated in it had all been constructed at ISAS. The power Baratron capacitative pressure gauge (M.K.S. Instruments, supply was used in a current controlled mode, either giving a Burlington, MA, USA) connected directly to the main body of constant dc current or a modulated current, the frequency and the source, and a Pirani gauge was used to check that the waveform of which were controlled by a waveform generator. normal ultimate pressure (2–4 Pa) had been obtained.The microwave power was supplied by a Microtron 3 microwave generator (ElectroMedical Supplies, Wantage, UK, 2.45 GHz, maximum power 200W). The source could be operated in five distinct modes: A, RESULTS AND DISCUSSION primary (sputtering) discharge only, constant current; B, primary discharge only, with modulated current; C, boosted General discharge, with constant primary discharge current; D, boosted Throughout, an approximately 151 duty cycle was used for discharge, with modulated current for the primary discharge; the modulation.The maximum frequency was limited by the and E, microwave discharge only (obtained by switching off power supply to about 150 Hz for sine wave modulation and the primary current after using a boosted discharge); in this to about 40 Hz or lower for square wave modulation.No case, only Ar I lines and molecular bands are observed. significant changes were observed with frequency or waveform, With the slit widths used, the monochromator, focal length and 40 Hz was normally used. The effects of the depth of 0.8 m and theoretical resolving power 360 000, had in practice modulation and limitations imposed in practice are discussed a bandwidth of 10 pm.It could be programmed to scan below. The experimental parameters used were those giving successively through a number of preselected spectral regions, the best limits of detection for the microwave boosted GDS, adjusting the photomultiplier voltage to control the sensitivity viz., a primary discharge current of 20 mA and a pressure of for each group of spectral lines. For the pulsed measurements, a Brookdeal Type 401 lock-in amplifier (EG & G, Brookdeal about 2–3 hPa.Table 1 Experimental conditions for results presented in the figures Current/mA Anode–cathode potential/V Microwave power/W Figure No. Sample Spectrum Minimum Maximum Average* Minimum Maximum Average* Forward Reflected 1(a) Mild steel — — Various — — Various 0/40 0/28 1(b) — — Various — — Various 0 0 1(c) — — Various — — Various 40 28 2 Mild steel a — — 20 — — 1260 0 0 b 13 27 20 780 1600 1150 0 0 c — —- 20 — — 800 40 28 d 10 30 20 400 1300 800 40 28 3 Pressed copper; baked, Upper — — 20 — — 750 40 28 100 mg g-1 Al and Bi Lower 9 31 20 500 1100 700 40 28 4 Pressed copper, — — Various – – Various 40 30 1000 mg g-1 Al and Bi 5 Pressed copper, 1 5 30 14 400 1000 640 40 28 100 mg g-1 Al and Bi 2 — — 0 — — 0 40 30 3 — — 18 — — 720 40 30 6 Pressed copper, 1 5 30 14 400 1000 640 40 28 100 mg g-1 Al and Bi 2 — — 0 — — 0 40 30 3 — — 18 — — 720 40 30 7 Pressed copper, 1 4 32 16 400 1200 950 40 28 1000 mg g-1 Al and Bi 2 — — 0 — — 0 40 30 3 — — 20 — — 820 40 30 8 Pressed copper, Upper 0 30 16 300 1000 640 40 30 10 mg g-1 Al, 500 mg g-1 Bi Pressed copper blank Lower 0 30 16 300 1000 640 40 30 * Average or steady-state value. 308 Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12Reduction of Interference from Neutral Lines of the Carrier Gas (Ar I Lines) To study this, a plane mild steel cathode was used and a small section of the spectrum between 394.500 and 394.940 nm was recorded; this contained Ar I, Ar II and Fe I lines.The intensities of these lines were measured as a function of the dc current in the primary discharge for boosted and unboosted conditions and for various modulation depths in modes B and D described above. The Fe I 394.877 nm and Ar I 394.897 nm lines are very close and, although this particular iron line is weak and would not be used for analytical work, the lines provided a clear test of the potential of the system. Fig. 1(a) shows the voltage–current characteristics of the primary dis- Fig. 2 Fe I 394.877 nm line (left) compared with Ar I 394.897 nm line charge under unboosted and boosted conditions, and Fig. 1(b) (right) for various discharges. Peak intensities on common arbitrary scale: a, dc primary discharge (mode A) 410; b, modulated primary discharge (mode B) 330; c, boosted discharge (mode C) 2620; and d, modulated boosted discharge (mode D) 2600. Steady or average current in the primary discharge 20 mA. Detailed experimental conditions are given in Table 1.and (c) show the corresponding intensity variations of typical lines without and with the supplementary discharge. The general experimental conditions for the results presented are summarized in Table 1. The pressure used for the various experiments varied slightly, in the range 2.7–2.9 hPa. For the modulated primary discharge, the maximum and minimum values for the current and potential were read from an oscilloscope, the current ±1 mA and the voltage ±50 V.For sine wave modulation, the average current as recorded on the moving coil meter on the power supply should be the mean of the minimum and maximum values. Owing to the non-linear relationship between current and anode–cathode potential difference (PD), the waveform of the PD is distorted and the average value is not the mean of the minimum and maximum values. For a given current, the anode–cathode potential difference is higher without the supplementary microwave discharge, and the maximum current that could be used at a given pressure was therefore lower; further, at a given current, the lower PD with a boosted discharge means that the spreading of the profile in the wings of some lines is reduced.To emphasize the differing dependence of intensity on current for different lines shown in Fig. 1(b) and (c), the intensities for each line have been normalised to its maximum value. Fig. 2a–d show the close lying Fe I 394.877 nm and Ar I 394.897 nm lines, recorded in modes A–D.From Fig. 1(b), it can be seen that in an unboosted discharge the intensity– current relationships for Ar I and Fe I lines differ significantly, so that by modulating the current between appropriate limits, the Fe I line may be favoured in comparison with the Ar I line. However, in practice there are considerable problems in this case; ideally, the modulation should be about the optimum current (20 mA); a small modulation, although favouring Fe I lines, gives small absolute intensities but with a large modulation of about 20 mA, the voltage is excessively high at the peak current.If the modulation is based on a lower current, absolute intensities are again lower. In practice, with this sample it was not possible to choose conditions which gave a useful increase in the intensity of the Fe I line relative to the Ar I line (see Fig. 2a and b). A much more significant effect could be obtained with a slightly recessed cathode (a ‘hollow Fig. 1 (a) Voltage rsus current relationship for (&) unboosted and cathode’, 2 mm deep and 8 mm in diameter) for which there is (%) boosted discharges. Experimental conditions as in Table 1. (b) Normalized intensity versus current relationships in the unboosted a current range where the voltage rises much less steeply, and discharge. Experimental conditions as in Table 1. Maximum intensities it is probable that this technique is also more successful for on common arbitrary scale: &, Ar II 394.608 nm, 760; V, Ar I plane samples at higher pressures, where the voltage again 394.897 nm, 410; %, Fe I 394.809 nm, 60; and #, Fe I 394.877 nm, 20.rises less steeply with current. (c) Normalized intensity versus current relationships in the boosted The use of the supplementary microwave discharge enhances discharge. Experimental conditions as in Table 1. Maximum intensities the Fe I 394.877 nm line far more than the Ar I 394.897 nm on common arbitrary scale: &, Ar II 394.608 nm, 760; V, Ar I 394.897 nm, 1660; %, Fe I 394.809 nm, 930; #, Fe I 394.877 nm, 2620.line (see Fig. 2c). Further, as discussed above, in the boosted Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12 309Fig. 4 Normalized intensity versus current relationships in the Fig. 3 Sections of spectra showing analytical lines of Bi I and Al I. boosted discharge. Experimental conditions as in Table 1.Maximum Spectra displaced vertically for clarity. Upper spectrum, boosted intensities on common arbitrary scale: $, Ar II 308.308 nm, 190; #, discharge (mode C); lower spectrum, modulated boosted discharge Al I 308.216 nm, 800; %, OH 306.722 nm, 230; &, OH 306.765 nm, + (mode D). Spectral regions (nm) and full-scale intensity (common Bi I 306.771 nm, 260; 1, OH 306.792 nm, 160; and 2, OH arbitrary scale) are as follows: section 1, 306.630–306.903 nm, 300; 308.328 nm, 640.section 2, 308.150–308.423 nm, 600; section 3, 309.090–309.454 nm, 600; and section 4, 309.350–309.480 nm, 20 000. samples, severe interferences from OH lines were observed in discharge, the argon intensity is almost independent of the this region. current in the primary discharge [Fig. 1(c)], so the Fe I line Fig. 3 shows sections of spectra from the boosted dc discan be selectively enhanced if the primary current is modulated charge using a compressed copper disc containing 100 mg g-1 (Fig. 2d). In the unboosted GDS, the intensity of the iron line of Bi and Al. The figure is divided into four sections, recorded could not be accurately measured owing to the wings of the at different sensitivities. Sections 1, 2 and 3 include the Bi I adjacent Ar I line; with a modulated, boosted source, the 306.771 nm, the Al I 308.216 nm and the Al I 309.271 nm lines, relative signal from this line is reduced by a factor of at least respectively, whilst a part of section 3 is repeated with greatly 20.It should be pointed out that as Ar II lines are not strongly reduced sensitivity in section 4 to record the Cu I 309.399 nm excited in the microwave discharge on its own, the Ar II lines line; the intensities of the two spectra, from the modulated are dependent on the primary discharge current and so are boosted discharge and from the dc boosted discharge, have modulated in a similar way to the sample lines. This technique been adjusted to give the same signal for this line.The therefore does not eliminate interference from Ar II lines, in dependence of the intensities of various spectral lines in this contrast to other methods.4,5,8 However, in the microwave region on the dc current in a boosted discharge is shown in boosted GDS, the Ar II 394.608 nm line is already reduced by Fig. 4, using results obtained from a sample containing a factor of 130 relative to the Fe I 394.877 nm line compared 1000 mg g-1 of Bi and Al.The intensities of the OH lines vary with the unboosted case, for a current of 20 mA in the primary more with current than does the intensity of Ar I lines discussed discharge, and interference from Ar II lines is a much less in the previous example. Moreover, they do not all behave in serious problem than interference from Ar I lines. the same way; the intensity distribution among the lines in the OH band is not the same in a dc primary discharge and in a microwave only discharge, i.e., the rotational temperature is Lines Subject to Interference from OH Lines in the 306–310 nm different, and this means that in a boosted discharge the change Region in the intensity of OH lines with the current in the primary discharge varies with the line: three examples are shown in The most sensitive Bi line (306.771 nm) and the second most sensitive Al I doublet (308.216 and 309.271 nm) lie in this Fig. 4. Fig. 4 also shows the intensity variation for the signal at 306.765 nm.In this case, the Bi line coincides with an OH region. The modulated boosted system was used both with compressed copper samples and with solid metallic samples. line, so although the intensity is not zero at zero current, it changes more with current than do the intensities of other OH The copper discs contained various amounts of Al2O3 and Bi2O3 , compressed with high purity copper powder at a lines. Under the conditions used, the enhancement factor for the Al I lines was about 10, of the same order as that for the pressure of 6 t cm-2.They are slightly porous, so that the ultimate pressure in the source was limited to about 10–20 Pa. OH lines, although F for the Cu I 327.4 nm line recorded at the same time had its typical value of about 60. Further, the Initially, severe contamination with nitrogen bands was observed and it was more difficult to obtain a satisfactory intensities of some of the OH lines were much greater than some of the analytical lines; it is also likely that the waveform boosted discharge (the appropriate conditions are clearly indicated when the bluish discharge changes to a bright green of the OH intensity differs from that of the analytical lines.These factors meant that the selection of an appropriate colour). A water-cooled cap of the type described by Dogan et al.12 was therefore placed behind the sample, so that argon magnitude of current modulation was more difficult than in the Fe I–Ar I case discussed above.In some cases there was at atmospheric pressure could be maintained at the back of the sample. The ultimate pressure was not improved, but the either a small negative signal or increased noise in the modulated boosted spectrum at wavelengths corresponding to the gas entering through the sample was argon and did not affect the operation of the discharge so that the full effects of the OH lines. Further work is needed to reduce these effects.The interferences to which aluminium and bismuth lines are auxiliary microwave discharge could be obtained. Strong OH bands were observed, the intensity of which decreased rapidly subject are shown in more detail in Figs. 5–7. These compare the dc boosted spectrum, the modulated boosted spectrum and over the first 10 min, and then more slowly with time. The band intensities could be reduced if the samples were heated a pure microwave spectrum (mode E); in the last case, only OH and Ar I lines are recorded.In Figs. 5 and 6, recorded at about 90 °C for 3 h before use. Even with solid metallic 310 Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12Fig. 8 Bi 306.771 nm line and adjacent lines, modulated boosted discharge (mode D). Spectra displaced vertically for clarity. Upper spectrum, 500 mg g-1 Bi; lower spectrum, blank, for same experimental Fig. 5 Section of spectrum from 308.150 to 308.423 nm; 100 mg g-1 parameters.The same intensity scale was used, without adjustment. Al. Spectra displaced vertically for clarity. 1, Modulated boosted discharge (mode D); 2, microwave discharge only (mode E); and 3, dc boosted discharge (mode C). Using a polychromator set on this line, the measured Al content could be in error by a factor of 2 at the 200 mg g-1 level; with modulation, the OH signal and hence the concentration at which such errors occur can be reduced by a factor of at least 25. Fig. 7, obtained with a sample containing 1000 mg g-1 of Bi, shows how the Bi I line (a complicated partially resolved line with hyperfine structure) coincides with two of a group of three partially resolved OH lines.Even at this concentration, the analysis would be in error by a factor of 2 with the unmodulated discharge; again, the interference can be reduced by a factor of at least 25 by modulating the primary discharge. In Fig. 8, for the modulated boosted discharge, the spectrum from a sample containing 500 mg g-1 of Bi is compared with that from a blank, indicating the limits of detection that may Fig. 6 Section of spectrum from 309.090 to 309.454 nm; 100 mg g-1 be obtained with this technique. It should be pointed out that Al. Spectra displaced vertically for clarity. 1, Modulated boosted with a dc boosted discharge (mode C), it may be possible to discharge (mode D); 2, microwave discharge only (mode E); and 3, dc eliminate interference by inert gas lines by comparing the boosted discharge (mode C).sample spectrum with the spectrum from a blank, but this is not a possibility with OH interference, as the OH line intensity may differ from sample to sample, and certainly changes with time. A sample was used for measurements with a boosted discharge; it was kept in the source, in argon at atmospheric pressure, and then used for the determination of dc intensity versus current dependence. Running the sample at various currents in the 0–25 mA range, it was found that over a period of 20 min the PD at 20 mA decreased from 1070 to 1000 V, presumably owing to sample erosion; the intensity of Ar II and Al I lines changed by less than 5%, but the intensity of OH lines decreased by almost 50% of their initial value. Preliminary measurements on bands in the nitrogen second positive system show that these bands behave in a very similar way to Ar I lines, and so interferences due to these bands should be easily eliminated.This work shows that the use of a modulated boosted GDS can overcome severe problems caused by interfering lines of Ar I and molecular bands. Further work is planned to quantify the improvements in limits of detection that may be obtained. Fig. 7 Section of spectrum from 306.630 to 306.903 nm; 1000 mg g-1 Bi. Spectra displaced vertically for clarity. 1, Modulated boosted discharge (mode D); 2, microwave discharge only (mode E); and 3, dc E.B.M.S. thanks Dr. Leis and Professor Klockow, the Director boosted discharge (mode C).of the Institut fu�r Spektrochemie und Angewandte Spektroskopie, for the opportunity to carry out experimental work at ISAS. The work on glow discharge sources there is using a copper disc containing 100 mg g-1 of Al and Bi, it is clear that both the aluminium lines suffer from interferences; supported by the Ministerium fu�r Wissenschaft und Forschung des Landes Nordrhein-Westfalen and the Bundesministerium the line at 308.216 nm (Fig. 5) is affected with interference from the weak OH 308.207 nm line, which would be serious fu�r Bildung, Wissenschaft, Forschung und Technologie (the Ministry for Science and Research of Nordrhein-Westfalen and at lower aluminium concentrations. In Fig. 6, it can be seen that even at the 100 mg g-1 level, the 309.271 nm Al I line the Federal Ministry for Training, Science, Research and Technology). cannot be resolved from the stronger 309.278 nm OH line. Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12 3119 Outred, M., Ru�mmeli, M. H., and Steers, E. B. M., J. Anal. At. REFERENCES Spectrom., 1994, 9, 381. 1 Grimm, W., Naturwissenschaften, 1967, 54, 588. 10 Leis, F., Broekaert, J. A. C., and Laqua, K., Spectrochim. Acta, 2 Grimm, W., Spectrochim. Acta, Part B, 1968, 23, 443. Part B, 1987, 42, 1169. 3 Leis, F., and Steers, E. B. M., Spectrochim. Acta, Part B, 1994, 11 Learner, R. C. M., Thorne, A. P., Wynne-Jones, I., Brault, J. W., 49, 289. and Abrams, M. C., J. Opt. Soc. Am., A: Opt. Image Sci., 1995, 4 Dogan, M., and U� lgen, A., Fresenius’ J. Anal. Chem., 1996, A12, 2165, and references cited therein. 355, 651. 12 Dogan, M., Laqua, K., and Massmann, H., Spectrochim. Acta, 5 Wagatsuma, K., and Hirokawa, K., Anal. Sci., 1991, 7, 289. Part B, 1972, 27, 65. 6 U�lgen, A., Dogan, M., Go�kmen, A., and Yalcin, S., Spectrochim. Acta, Part B, 1993, 48, 65. Paper 6/06773E 7 Wagatsuma, K., and Hirokawa, K., Anal. Chem., 1984, 56, 2732. Received October 3, 1996 8 Wagatsuma, K., and Hirokawa, K., Spectrochim. Acta, Part B, 1988, 43, 831. Accepted November 29, 1996 312 Journal of Analytical Atomic Spectrometry, March

ISSN:0267-9477    

DOI:10.1039/a606773e

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Development of a Multi-element Hydride Generation–Inductively Coupled Plasma Mass Spectrometry Procedure for the Simultaneous Determination of Arsenic, Antimony and Selenium in Waters† JUSTINE BOWMAN, BEN FAIRMAN* AND TIM CATTERICK L aboratory of the Government Chemist, Queens Road, T eddington,Middlesex, UK TW11 OLY A multi-element hydride generation–inductively coupled steps of any hydride generation procedure), need optimising, in turn, to establish the ideal conditions for the simultaneous plasma mass spectrometry (HG–ICP-MS) method for the simultaneous determination of arsenic, antimony and selenium determination of these analytes.At the present time the analysis of As, Sb and Se in our in water matrices has been developed. The method involves an off-line pre-reduction procedure for the reduction of SeVI to laboratory is carried out using a single-element, HG–AFS system. With the analysis of over 3000 samples per annum this SeIV by HCl, combined with an on-line reduction of AsV and SbV to the trivalent state with thiourea and generation of the restriction to single-element analysis results in approximately 9000 determinations.In addition to As, Sb and Se, the instru- hydrides. Analytical characteristics include detection limits of 0.08 ng g-1 As, 0.06 ng g-1 Sb and 0.10 ng g-1 Se, linearity of mentation is also used for the determination of Hg. Problems are encountered when changing from one elemental system to four orders of magnitude and short and long term reproducibility of between 8 and 12%.Results from four another due to differences in the chemical strategies used. Carry-over of the reducing agents often leads to a significant reference water samples for As, Sb and Se showed data which were all within 10% of the target values. Interferences were decrease in sensitivity as the system is switched between elements. The time needed to flush out the system and achieve minimal for As and Sb, whereas Cu2+, and to a lesser extent Ni2+ and Cd2+, caused signal suppression effects on Se. stability of response can often take a considerable time (e.g., 2–3 h).Advantages over an alternative, single element, HG–AFS technique include speed of analysis (by a factor of two) and To overcome these problems a multi-element procedure using HG–ICP-MS for the determination of As, Sb and Se in elimination of the conflicting chemistry requirements, traditionally found with sequential single element hydride water has been developed for use in a commercial laboratory environment.The method involves a pre-reduction stage of generation methods. heating the sample with acid. This reduces the SeVI to SeIV. Keywords: Simultaneous multi-element analysis; hydride Thiourea is added on-line to reduce pentavalent As and Sb to generation; inductively coupled plasma mass spectrometry; the trivalent species, and then the gaseous hydrides are formed water; arsenic; antimony; selenium during the reaction with NaBH4.Final determination is performed by use of ICP-MS. The unified sample pre-treatment leads to a streamlined system, which has dramatically reduced Hydride generation coupled with atomic spectrometry has become a widely used analytical technique. Its popularity is the analysis time and eliminated the problems when alternating between elements using the single element AFS system. due to the fact that this coupled technique is both highly selective and sensitive, produces low detection limits and, due to its ability to remove the analyte from the sample matrix, is EXPERIMENTAL relatively free from spectral interferences.Instrumentation The efficiency of the hydride generation step is dependent upon the experimental conditions and particularly the oxi- All determinations were carried out using a Perkin-Elmer dation state of the analytes.1 When As and Sb are present as ELAN 5000A ICP-MS instrument (Perkin-Elmer, the pentavalent species they have a reduced propensity towards Beaconsfield, UK).The on-line reduction was fully automated hydride generation (including reduction schemes involving with the inclusion of a Perkin-Elmer FIAS 400 flow injection sodium tetrahydroborate). Using this system Se, in the hexaval- system. The instrumentation was computer controlled using ent state, is unable to form a hydride.2 The lower oxidation Perkin-Elmer integrated software. The operating conditions states of AsIII, SbIII and SeIV are the most favourable for the for the ICP-MS instrument and the FIAS system are given in formation of the hydrides, hence a conversion of the analytes Tables 1 and 2, respectively.A schematic diagram of the entire to these valencies prior to reaction with NaBH4 is required. HG–ICP-MS system is shown in Fig. 1. Iodide and L-cysteine can be used to reduce AsV and SbV to A Tecam water-bath fitted with a Techne TE-8A thermoreg- AsIII and SbIII,3 but they also reduce SeIV to elemental Se which ulator [Techne (Cambridge) Ltd., Duxford, Cambridgeshire, is unable to form a hydride.4 Because of these oxidation state UK] was used for the pre-reduction stage.Sample preparation difficulties, few methods have been published that can be called and pre-reduction was performed using 25 ml sterilin tubes true simultaneous hydride generation methods for these three (Bibby Sterilin, Walton, Staffordshire, UK). elements.5 As well as the oxidation state question many other experimental parameters, such as acid concentration, reaction Reagents and Standard Solutions time, reducing agent concentration, and temperature (critical All solutions were prepared with high purity deionized water (18 MV, Elga, High Wycombe, Buckinghamshire, UK).† Presented at the Eighth Biennial National Atomic Spectroscopy Symposium (BNASS), Norwich, UK, July 17–19, 1996. Stock solutions of the elements (1000 mg g-1) were prepared Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12 (313–316) 313Table 1 Operating conditions for the ELAN 5000A ICP-MS (Fisons, Loughborough, Leicestershire, UK). A 0.3 M solution of thiourea (Aldrich) in high purity water was prepared daily. Forward power 1050 W Carrier gas 15 l min-1 Nebulizer gas 0.850 l min-1 Analytical Procedure Auxillary gas 0.80 l min-1 All standards and samples solutions are acidified by the Parameter file— addition of concentrated HCl in the ratio 12.5 ml of sample to 10 ml of concentrated HCl to give a final acid concentration Points across peak 1 Resolution Normal of 5 M.Samples are then placed in a water-bath at 80–85 °C Scanning mode Peak hop transient and heated for 90 min. After pre-reduction, the solutions are allowed to cool and transferred to the ICP-MS instrumen- FIAS argon flow rate 0.350 l min-1 tation. There, the samples are mixed on-line with the 0.3 M Times/ms m/v thiourea solution and left to react for 180 s.The sample is then reacted with the NaBH4 to form the gaseous hydrides, Element Replicate Dwell 75As 1200 20 which are swept into the ICP-MS instrument for quantification. 121Sb 1200 20 82Se 1200 20 RESULTS AND DISCUSSION Elemental equations Arsenic-75=arsenic-75 Off-line Pre-reduction Antimony-121=antimony-121 Selenium-82=selenium-82-1.001*krypton-83 The reduction of SeVI to SeIV is carried out by heating the sample with HCl. The parameters found to be critical at this stage were the acid concentration and the temperature of the Table 2 Operating conditions for the FIAS 400 flow injection system water-bath.2 A slight decrease in the acid concentration, or a drop in temperature, can result in a significantly longer heating Injection volume 200 ml time being required.Fig. 2 shows the effect of heating time on Repeat from step 2 through step 2 3 times a SeVI solution (10 ng g-1). As can be seen, the signal levels off after 90 min. The SeIV signal was unaffected by the length Speed (rpm) of the heating period. Best results were achieved when the Start Valve Step read Time/s Pump 1 Pump 2 position Remote sterilin sample containers were capped (but only loosely, to allow any chlorine fumes to escape).A build-up of Cl can Pre-sample 10 100 0 1 result in the back-oxidation of SeIV.6 Recoveries of 100% were 1 5 100 0 1 2 30 0 0 1 obtained for the SeVI solution when compared with equivalent 3 2 0 0 1 X SeIV standards. 4 X 15 0 -70 2 X 5 5 0 0 1 X 6 1 50 0 1 On-line Reduction Post-sample 10 100 0 1 X The most commonly used reducing agents for the conversion of AsV and SbV to AsIII and SbIII are iodide or bromide,7 Lcysteine3 and thiourea.1 The main drawback is that they all reduce tetravalent Se to its elemental state, which is unable to form the hydride. For this study the use of thiourea was investigated as an on-line reductant for AsV and SbV, because although thiourea is able to reduce SeIV to elemental Se to a significant degree, the conversion seemed to be slower than that caused by iodide.For the complete reduction of AsV and SbV the concentration of the thiourea and the reaction time required were varied to establish the best set of compromise conditions. Reaction with 0.5 M thiourea brought about the reduction of AsV and SbV in the shortest time, but the loss of Se signal was proportionally higher. A reaction time of 180 s and 0.3 M thiourea were found to be the ideal conditions for complete reduction of AsV and SbV without losing a substantial amount of the Se signal (Fig. 3). Although the concentrated HCl procedure outlined above Fig. 1 Schematic diagram of the HG–ICP-MS system. from the following salts (BDH AnalaR grade, Merck, Poole, Dorset, UK): sodium selenite (Na2SeO3), sodium selenate (Na2SeO4 10H2O), sodium arsenite (NaAsO2 ), sodium arsenate (Na2HAsO4 7H2O), antimony trichloride (SbCl3), antimony pentachloride (SbCl5).The arsenic and selenium stock solutions were prepared in 0.12 M hydrochloric acid (Baker, Phillipsburg, NJ, USA), whereas the antimony stock solution was prepared in 3 M HCl (Aristar, Merck). Working standards were prepared daily by serial dilution with 0.12 M HCl (Baker). Fig. 2 Variation of Se signal with an increase in heating time. A Sodium borohydride (Aldrich, Gillingham, Dorset, UK) 10 ng g-1 Se standard was placed in a water bath set at 80°C and portions removed and analysed at the different intervals. 1.2% m/v was prepared weekly in 0.1 M sodium hydroxide 314 Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12hydride generation systems, its effectiveness in the proposed method was investigated. A study of the effect of the most common interferents was carried out. A 10 ng g-1 multielement standard was spiked with a variety of interferents at 1, 10 and 100 mg g-1 (Table 3). The results are expressed as a percentage of the signal without interferent.Precision data, expressed as an RSD of the three replicates performed for each solution, have also been included here. As can be seen, the biggest interference was caused by Cu (100 mg g-1) on Se. The signal has decreased by 53% from that obtained for the Se standard with no added Cu. This may be caused by adsorption of the hydride by metal precipitates present in the solution.1,8 Fig. 3 Effect of reaction time with 0.3 M thiourea on As, Sb and Se Cd and Ni also had an adverse effect on the Se signal, but to signals.Variation of signal of a 10 ng g-1 multi-element standard in a lesser extent. As and Sb were generally unaffected by the 5 M HCl, as the reaction time with thiourea increases. interferents examined. Analytical Performance A typical injection profile for a 200 ml 10 ng g-1 multi-element standard is shown in Fig. 5. Using the present parameter file conditions (Table 2), the linearity of the system was found to be up to 25 ng g-1 for As and 50 ng g-1 for Sb and Se.This linearity can be increased by desensitizing the detector at specific masses by utilizing the Perkin-Elmer OmniRange software option. Reproducibility was measured by analysing 10 consecutive determinations of a 10 ng g-1 multi-element standard, followed Fig. 4 Effect of acid concentration on AsIII and SbIII signals. by several further determinations over a 4 h period. The short Differences in hydride formation of a 10 ng g-1 SbIII standard in 0.12 M and long term stability of the analytes under these conditions, HCl compared with a 10 ng g-1 SbIII standard acidified to 5 M HCl.expressed as RSDs, are 8 and 12% for As, 9 and 12% for Sb, and 7 and 10% for Se, respectively. does not reduce AsV and SbV to their trivalent states, we did One of the requisites of the proposed method is that it find that the high concentration of acid seemed to affect the ability of AsIII and SbIII to form the hydrides in the proposed system.A solution of AsIII and SbIII prepared in 0.12 M HCl and heated for 90 min at 80–85 °C gave a consistent signal regardless of the reaction time with the 0.3 M thiourea (Fig. 4). However, the same solution heated in 5 M HCl gave signals which were dependenton the thiourea reaction time, mimicking the behaviour of the pentavalent species. The signal for the AsIII standard acidified to 5 M HCl equalled that of the standard prepared in 0.12 M HCl. The signal for the 5 M HCl SbIII standard exceeded that of the standard prepared in 0.12 M HCl.This suggests that although the Sb was present at the lower oxidation state in 0.12 M HCl this acid concentration was not ideal for Sb hydride generation (Fig. 4). Interferences Several elements, in particular the transition metals, have been found to interfere with the generation of hydrides.1,8 As Fig. 5 Typical signal profile of a 10 ng g-1 As, Sb and Se multielement standard mixture: 200 ml injection volume.thiourea has been used as an interference suppresser in other Table 3 Effect of interferents on As, Sb and Se signals—10 ng g-1 multi-element standard spiked with the interferents at 1, 10 and 100 mg g-1. Results expressed as a percentage of the result without interferent Interferent concentration/mg g-1 As Sb Se Interferent 1 10 100 1 10 100 1 10 100 Cu2+ 94 95 95 94 95 96 89 95 47 RSD (%) 3.6 3.7 2.4 4.5 2.8 2.5 4.3 3.0 2.4 Cd2+ 94 97 94 91 93 91 90 84 83 RSD (%) 1.0 3.2 3.4 2.2 3.4 3.1 0.8 4.6 2.8 Ni2+ 94 95 99 92 87 92 84 78 83 RSD (%) 7.0 4.9 2.5 4.1 2.9 2.6 5.8 6.7 4.8 Pb2+ 102 96 114 96 96 97 96 97 99 RSD (%) 4.5 2.1 1.1 0.6 0.8 1.3 4.6 2.1 6.6 Fe3+ 94 95 102 87 82 92 89 90 93 RSD (%) 6.6 6.4 6.8 3.9 5.4 3.9 9.4 4.9 6.1 Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12 315Table 4 Accuracy data from Water Research Centre inter-laboratory reference waters. All concentrations are in ng g-1 As Sb Se Sample Obtained Expected Obtained Expected Obtained Expected AQUACHECK 85 24.9 24.1 2.78 2.81 3.21 3.13 AQUACHECK 93 27.9 27.0 3.00 3.10 3.25 3.00 AQUACHECK 101–5 23.9 22.0 8.50 8.40 4.11 4.30 AQUACHECK 101–5A 38.7 36.0 5.35 5.30 4.63 4.20 Table 5 Recovery results from spiked water samples—obtained by with using a single element HG–AFS instrument for the spiking the samples at 10 ng g-1 with As, Sb and Se analysis of different analytes, each with different chemical strategies, is no longer an issue due to a unified sample pre- Recovery (%) treatment methodology.The HG–ICP-MS method produces savings in analysis time of 50% over the single element Sample As Sb Se HG–AFS method. This rapid analysis time, compared with 1 101 97 102 the single element approach, leads to improvements in 2 116 105 108 efficiencies and increased productivity, welcome in any com- 3 105 107 98 4 107 97 88 mercial laboratory environment. The work described in this paper was supported by the would be able to comply with the presentWater Supply (Water Department of Trade and Industry, UK, as part of the National Quality) Regulations 19899 (present limits are 50 ng ml-1 for Measurement System Valid Analytical Measurement As and 10 ng ml-1 for Sb and Se).Although many compro- Programme and the Government Chemist Programme. mises had to be made in the development of the method, the detection limits of 0.08, 0.06 and 0.10 ng g-1 for As, Sb and REFERENCES Se, respectively (calculated as 3s based on 10 determinations of the blank) are well below the required levels. 1 Uggerud, H., and Lund, W., J. Anal. At. Spectrom., 1995, 10, 405. To check on the accuracy of the method a number of Water 2 Hill, S. J., Pitts, L., and Worsfold, P., J. Anal. At. Spectrom., 1995, Research Centre (WRC) spiked Aquacheck reference solutions 10, 409. 3 Welz, B., and S¡ ucmanova�, M., Analyst, 1993, 118, 1417. were analysed (Table 4). As can be seen the results are very 4 Bye, R., T alanta, 1990, 37, 1029. acceptable, with agreement falling within 10% of the WRC 5 Halicz, L., and Russell, G. M., Analyst, 1986, 111, 15. declared spike values. Some typical water samples (drinking 6 Excalibur Methods of Analysis Manual, P S Analytical, and bore waters) submitted to the laboratory were spiked with Orpington, Kent. As, Sb and Se at approximately the 10 ng g-1 level. The 7 Nakahara, T., and Kikui, N., Anal. Chim. Acta, 1985, 172, 127. samples were analysed using the multi-element method and 8 Bye, R., Engvik, L., and Lund, W., Anal. Chem., 1983, 55, 2457. 9 UK Water Supply (Water Quality) Regulations 1989, HM recovery values calculated (given in Table 5). All recovery Stationery Office, London, UK. results were encouraging, with over 90% of the values falling within 10% of the expected amount. Paper 6/05704G Received August 14, 1996 CONCLUSIONS Accepted November 6, 1996 A method has been successfully developed for the simultaneous determination of As, Sb and Se in water. Problems associated 316 Journal of Analytical Atomic Spectrometry, March 1997, Vol.

ISSN:0267-9477    

DOI:10.1039/a605704g

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Use of Flow Injection Cold Vapour Generation and Preconcentration on Coated Graphite Tubes for the Determination of Mercury in Polluted Seawaters by Electrothermal Atomic Absorption Spectrometry† PILAR BERMEJO-BARRERA*, JORGE MOREDA-PIN� EIRO, ANTONIO MOREDA-PIN� EIRO, ADELA BERMEJO-BARRERA Department of Analytical Chemistry, Nutrition and Bromatology, Faculty of Chemistry, University of Santiago de Compostela, Avenida de las Ciencias, s/n E-15706-Santiago de Compostela, Spain Methods have been developed for the trace determination of utilized successfully for hydride forming metals.The use of different coating reagents such a palladium,21–24 platinum,21 total and inorganic mercury in natural waters by cold vapour generation–trapping and atomization in a graphite furnace by rhodium,21 rutenium,21 zirconium,25–27 tungstate,25 thalium25, molybdeno25, silver28 and palladium–iridium29 has been suc- selective reduction with NaBH4 and SnCl2. Iridium, tungstate and zirconium coated graphite tubes were investigated for the cessful for metals such as lead,26 arsenic,21,24 antimony,21,24 selenium,21,24,28 bismuth,21,29 germanium,23 tellurium28 and preconcentration of the mercury vapour.The carrier solution was 0.4 m of HCl, and 0.2% NaBH4 and 1.0% SnCl2 were tin.21,22,27 For adsorption of the mercury cold vapour, the use of used as reducing solutions for total and inorganic mercury, respectively. Using iridium coated graphite tubes characteristic gold,30 platinum31 and palladium20 coated graphite tubes has been proposed for preconcentration. Palladium coated graphite masses of 300 and 240 pg for total and inorganic mercury, respectively, and detection limits of 90 and 60 ng l-1 for total tubes have been referred to in the literature as those that present a higher adsorption efficiency for volatile hydrides32 and inorganic mercury were obtained for a 1500 ml sample volume.The precision was between 0.7 and 2.1%.Accuracy of and mercury cold vapour20 than graphite tubes with other coatings. Otherwise, owing to the high volatility of palladium, method was confirmed by the analysis of IAEA/W-4 fresh water reference material. an injection of this species into the furnace is necessary for each cycle. In addition, a change in the graphite furnace Keywords: Electrothermal atomic absorption spectrometry; programme is necessary because of the tube coating and in cold vapour; total mercury; inorganic mercury; coated graphite order to take measurements, thus increasing the time necessary tubes; preconcentration; natural waters per cycle.Therefore, the study of other permanent coating species remains of interest. In the present work, selective mercury cold vapour gener- The concentration of mercury in natural waters is minute, about 0.1 mg l-1 in fresh water,1 and this concentration is ation using SnCl2 or NaBH4 was carried out.33 Preconcentration of mercury cold vapour using iridium, tung- significantly reduced in unpolluted natural waters.In order to monitor levels of mercury in natural waters a detection limit state and zirconium coated graphite tubes was also carried out and the results obtained were compared. at of least of 10 ng l-1 (preferably 1 ng l-1) is required. Therefore, sensitive methods are necessary to detect such low concentrations. Owing to the simplicity, high sensitivity and relative freedom EXPERIMENTAL from interferences, CVAAS, with or without a preconcentration Apparatus step before the mercury cold vapour generation, has generally been used for the determination of mercury in seawater A Perkin-Elmer (Uberlingen, Germany) FIAS-400 system with samples.2–7 Continuous flow systems have also been used for a five-port FI valve was used with the cold vapour accessories.mercury cold vapour generation in natural water samples.8–16 The rotation speed of the two multichannel peristaltic pumps Despite the high sensitivity of CVAAS and owing to the low were programmed and controlled automatically by a separate mercury concentration in un-contaminated waters, variations personal computer. A schematic diagram of the FIAS-400 of conventional CVAAS, (introducing a step preconcentrating system is shown in Fig. 1. The FIAS–furnace sample transfer the mercury cold vapour, such as the amalgamation technique tube consists of a quartz capillary (2 cm long×1.3 mm using different amalgamating media based on copper, silver, od×0.5mm id) attached to a metal mount with a short piece gold and platinum metals17–19) have been proposed in order of silicone rubber tubing, which in turn is attached to a PTFE to increase the sensitivity of the method.However, the use of tube. The metal mount of the FIAS–furnace sample transfer the amalgamation technique presents several disadvantages, tube was loaded into the spring clip on the end of the such as the necessity to clean the amalgamating medium, also autosampler arm (Fig. 2). The FIAS mercury cold vapour the sensitivity is flow rate limited and the reproducibility is generation programme is given in Table 1. affected by slight changes in the flow rate.20 Mercury absorbance was measured with a Perkin-Elmer Preconcentration on to coated graphite tubes has been Zeeman 4100 ZL atomic absorption spectrometer equipped with a THGA (transversely heated graphite atomizer) furnace and an AS-71 autosampler.The source of radiation was a † Presented at the Eighth Biennial National Atomic Spectroscopy Symposium (BNASS), Norwich, UK, July 17–19, 1996. mercury electrodeless discharge lamp operated at 4W, which Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12 (317–321) 317Table 2 Furnace temperature programme for mercury determination Step Temperature/°C Ramp/s Hold/s Ar flow rate/ml min-1 1 20 0 3 250 2 1300 0 3 0 (read) 3 1500 1 3 250 Reagents All solutions were prepared from analytical-reagent grade chemicals using ultrapure water, with a resistivity of 18MV cm, which was obtained from a Milli-Q water purification system (Millipore, Milford, MA, USA).Mercury(II) nitrate stock standard solution (1.000 g l-1) was of analytical-reagent grade and supplied by Panreac (Barcelona, Spain). Methylmercury stock standard solution (1.000 g l-1) was prepared by dissolving 58.22 mg of reference material from the Commission of the European Communities Community Bureau of Reference (BCR) in 50 ml of water.Iridium chloride (Aldrich, Milwaukee, WI, USA) stock standard solution (5 g l-1), zirconium oxychloride stock standard solution (10 g l-1) prepared from ZrOCl 8H2O (Merck, Darmstadt, Germany) and sodium tungstate stock standard solution (10 g l-1) prepared from Na2WO4 2H2O (Merck) were used as trapping solutions. The reducing agents used were 1.0% m/v tin chloride dihydrate (Merck) dissolved in 0.4 M HCl and 0.2% m/v sodium tetrahydroborate (Aldrich) dissolved in 0.5% m/v sodium hydroxide solution (Carlo-Erba, Milan, Italy).These solutions were prepared daily and the solutions were filtered before use. The carrier solution used was 0.4 M HCl prepared from 37% HCl (Merck). Fig. 1 Schematic diagram of FIAS-400 system; (a) sampling step Synthetic seawater samples of high 72.8‰ (SSWI) and low (valve in fill position) and (b) cold vapour generation and trapping 34.2‰ (SSWII) salinity were prepared according to the step (valve in injection position).literature31. Argon of N-50 purity (99.999%) was used as the sheath gas for the atomizer and as the internal purge gas. This was obtained from SEO (Madrid, Spain). Coating the Graphite Tubes The coating method used involved the injection of five 100 ml aliquots of iridium (1 g l-1), zirconium (10 g l-1) or tungstate (10 g l-1) trapping solutions. Each injection of solution was dried slowly by heating the atomizer at 150 °C with ramp rate and hold times of 25 and 40 s, respectively.A second dryng step at 200 °C with ramp rate and hold times of 20 and 30 s were then used. This was followed by a reduction step at 2000°C, applied over 5 s. A tube treated in this manner can could pass through about 500 firing cycles. Fig. 2 Schematic diagram of the FIAS–furnace sample transtube. Mercury Cold Vapour Generation and Preconcentration provided a 253.7 nm line. The spectral bandwidth used was 0.7 nm.A THGA with an integrated platform was used. Volatile species of mercury were generated from a 1500 ml sample prepared as follows: 9 ml of a natural water sample Integrated absorbance signals were used throughout. The graphite furnace programme used is given in Table 2. were placed in a 10 ml calibrated flask, with the appropriate Table 1 FIAS mercury cold vapour generation programme: argon flow rate, 50 ml min-1; and trapping temperature, 20 °C Flow rate/ml min-1 Pump 2 Step Pump 1 Carrier solution Reducing solution Time/s Valve Function 1 3 2 2 10 Fill Sampling step 2 0 0 0 8 Inject Quartz capillary is placed into the graphite tube 3 0 — — — Inject Hg cold vapour generation and trapping step 4 0 0 0 8 Inject Quartz capillary moves out of the graphite tube 5 0 0 0 5 Fill Start of furnace programme 318 Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12amount of HCl to obtain a concentration of 0.4 M. The tip of addition, it is shown in Fig. 3 that a loss of sensitivity is obtained when using NaBH4, this is due to differences in the the quartz capillary tube was inserted, automatically, from the outlet of the gas–liquid separator at the centre of the graphite mechamism of the reduction reaction in comparison with the use of SnCl2. tube. The carrier, 0.4 M HCl, and reducing solutions, 1.0% SnCl2 or 0.2% NaBH4, were pumped into the generator for 60 s at a rate of 3 and 2 ml min-1, respectively, and the evolved Sample and reaction loops mercury cold vapour was transferred, using an argon flow rate of 50 ml min-1, to the furnace tube where it was sequestered Experiments were carried out using different combinations of onto the coated graphite tube.Mercury was then atomized at sample and reaction loops (200, 500, 1000 and 1500 ml). Results 1300 °C over 3 s using maximum power heating and internal showed that an increase in the sample volume involved an argon gas stop. The FIAS mercury cold vapour generation and equivalent increase in the mercury absorbance signal. An the graphite furnace programme are show in Tables 1 and 2.increase in the length (110 mm, 100 cm, 200 cm and 300 cm) of the reaction loops did not produce an increase or decrease in the mercury absorbance signal. Thus, 1500 ml and 200 cm RESULTS AND DISCUSSION as sample and reaction loops, respectively, were selected. Various experiments using fresh water and seawaters samples and aqueous standard solutions spiked with Hg2+ (10 mg l-1) Flow rate parameters for carrier and reducing solutions, and for and CH3Hg+ (10 mg l-1) were carried out to optimize the carrier gas parameters affecting the efficiency of mercury cold vapour generation and adsorption onto coated graphite tubes, using Generation of the mercury cold vapour was obtained using SnCl2 and NaBH4 as reducing solutions.different pump speeds, and the results (Fig. 4) showed an increase in the integrated absorbance with increasing pumping speed up to 50–60 rev min-1.For higher pump speeds a Optimization of Mercury Cold Vapour Generation Conditions decrease in the mercury absorbance signal occurred. Thus, flow rates of 3 and 2 ml min-1 for the carrier and reducing Carrier and reducing solutions solutions, respectively, were selected for further experiments. The optimum concentration of HCl, used as the carrier solution The optimum flow rate of the carrier gas was 50 ml min-1.and NaBH4 and SnCl2 concentrations, used as reducing solutions for inorganic mercury, and total mercury, respectively, obtained for the cold vapour generation are shown in Table 3. Optimization of Mercury Cold Vapour Adsorption Onto Coated The selective cold vapour generation of mercury species, Graphite Tubes Hg2+ and CH3Hg+, by NaBH4 and SnCl2 at the concen- Reaction and trapping time trations referred to above are shown in Fig. 3. As can be seen, organic mercury could only be reduced by NaBH4.33 In The reaction time for mercury cold vapour generation and trapping time onto the coated graphite tubes was investigated.Results (Fig. 5) showed that a reaction and trapping time of Table 3 Optimum conditions affecting the generation of mercury 60 s is necessary for zirconium, tungstate and iridium coated cold vapour graphite tubes. Parameter Optimum condition [HCl] 0.4 M NaBH4 0.2% m/v SnCl2 1.0% m/v Sample loop 1500 ml Reaction loop 110 mm Carrier flow 3 ml min-1 Reducing solution flow 2 ml min-1 Argon flow 50 ml min-1 Fig. 4 Effect of pump speed on the integrated absorbance of 5 mg l-1 of Hg with 0.4 M of HCl and 1.0% SnCl2 for Ir (&), Zr (2) and W ($) coated graphite tubes. The loop sample was 1500 ml. Fig. 3 Effect of SnCl2 (1.0% m/v) (dotted lines) and NaBH4 (0.2% Fig. 5 Effect of reaction time on the integrated absorbance of 5 mg l-1 of Hg with 0.4 M of HCl and 1.0% m/v of SnCl2 for Ir (&), Zr m/v) (solid lines) on mercury cold vapour generation from solution with various amounts of inorganic and organic mercury.(2) and W ($) coated graphite tubes. The loop sample was 1500 ml. Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12 319T rapping and atomization temperatures The efficiency of trapping mercury cold vapour onto the coated graphite tubes was found to be optimum at room temperature for all coated graphite tubes. As can be seen in Fig. 6, the sensitivity decreased when the trapping temperature was increased.Thus, an adsorption temperature of 20°C was selected for further experiments. The atomization temperature (using a hold time of 3 s and maximum power heating) of trace amounts of mercury adsorbed onto the different coated graphite tubes was also investigated. The results (Fig. 7) showed optimum atomization temperatures of 1300°C for the use of iridium and zirconium coated graphite tubes and 1700 °C for tungstate coated graphite tubes.For atomization temperatures higher than the above temperatures a decrease in the integrated absorbance was obtained (Fig. 7) whilst the peak height was greatly increased. The mercury cold vapour trapping mechanism could be due to an intermetallic compound or alloy that is probably formed between Hg0 and iridium, tungstate or zirconium, the same as occurs with palladium.20 Analytical Figures of Merit The calibration and standard additions equations obtained for aqueous standard solutions, synthethic seawater of high (SSWII) and low (SSWI) salinity and real freshwater and seawaters samples, all spiked with 5, 10 and 20 mg l-1 of Hg2+ for inorganic mercury determinations; and 2.5, 5 and 10 of Hg2+ and 2.5, 5 and 10 of CH3Hg+ (as Hg) for total mercury determinations, are shown in Table 4.These calibration and standard additions equations were obtained for iridium, zirconium and tungstate coated graphite tubes. For each coated graphite tube, the slopes of the graphs are identical (t-test for a confidence level of 95%34).Therefore, the method is not dependent on the salinity of the natural water. The characteristic mass (i.e., the mass of analyte which provides a defined peak with an absorbance of 0.0044) for a Fig. 6 Effect of trapping temperature on the integrated absorbance of 5 mg l-1 of Hg with 0.4 M of HCl and 1.0% m/v SnCl2 for Ir ($), Zr (2) andW ($) coated graphite tubes. The loop sample was 1500 ml.Fig. 7 Effect of atomization temperatureon the integratedabsorbance Table 4 Calibration and standard addition slopes obtained for inorganic mercury and total mercury determinations with different coated graphite tubes Calibration Standard additions Tube Aqueous standard SSWII SSWI Seawater Fresh water Inorganic mercury— Zr coated graphite QA=0.010+0.014[Hg] QA=0.012+0.013[Hg] QA=0.013+0.014[Hg] QA=0.015+0.014[Hg] QA=0.018+0.015[Hg] W coated graphite QA=0.008+0.016[Hg] QA=0.008+0.016[Hg] QA=0.010+0.015[Hg] QA=0.011+0.015[Hg] QA=0.017+0.014[Hg] Ir coated graphite QA=0.012+0.028[Hg] QA=0.013+0.027[Hg] QA=0.011+0.029[Hg] QA=0.015+0.027[Hg] QA=0.021+0.029[Hg] T otal mercury— Zr coated graphite QA=0.009+0.011[Hg] QA=0.010+0.010[Hg] QA=0.008+0.012[Hg] QA=0.011+0.012[Hg] QA=0.012+0.011[Hg] W coated graphite QA=0.010+0.009[Hg] QA=0.010+0.011[Hg] QA=0.008+0.010[Hg] QA=0.009+0.011[Hg] QA=0.013+0.010[Hg] Ir coated graphite QA=0.012+0.020[Hg] QA=0.013+0.019[Hg] QA=0.011+0.021[Hg] QA=0.013+0.021[Hg] QA=0.018+0.020[Hg] of 5 mg l-1 Hg with 0.4 M of HCl and 1.0% m/v SnCl2 for Ir (&), Zr (2) and W ($) coated graphite tubes.The loop sample was 500 ml. 320 Journal of Analytical Atomic Spectrometry, March 1997, Vol. 12Table 5 Sensitivity for different coated graphite tubes 5 Simeonov, V., and Andreev, G., Fresenius’ Z. Anal. Chem., 1983, 314, 761. 6 Schmidt, D., and Freiman, P., Freseniuz’ Z. Anal. Chem., 1984, m0* (pg) LOD/ng l-1 317, 385. 7 Schintu, M., Kauri, T., Contu, A., and Kudo A., Ecotoxicol. Tube Hg2+ Hg Hg2+ Hg Environ. Saf., 1987, 14, 208. Zr coated graphite 700 900 150 200 8 Goto, M., Munaf, E., and Ishii, D., Fresenius’ Z. Anal Chem., W coated graphite 450 600 90 120 1988, 332, 745. Ir coated graphite 240 300 60 90 9 Munaf, E., Haraguchi, H., Ishii, D., Takeuchi, T., and Goto, M., Sci. T otal Environ., 1990, 99, 205. * m0=characteristic mass. 10 Munaf, E., Takeuchi, T., Ishii D., and Haraguchi, H., Anal.Sci., 1991, 7, 605. 11 Munaf, E., Takeuchi, T., and Haraguchi, H., Fresenius’ J. Anal sample loop of 500 ml and the detection limits obtained for Chem., 1992, 342, 154. each coated graphite tube using both reducing media are 12 Zhan, L., Lu, J., and Le, X., Microchim. Acta. 1993, 111, 207. shown in Table 5. The use of an iridium coated graphite tube 13 Hanna, C., Tyson, J., and McIntosh, S., Anal. Chem., 1993, 65, 653. gives better sensitivity than the other coated graphite tubes. 14 McIntosh, S., Baasner, J., Grosser, Z., and Hanna, C., At. In addition, the use of NaBH4 gives poorer sensitivity than Spectrosc., 1994, 15, 161. the use of SnCl2. 15 Garcý�a, M., Garcý�a, R., Garcý�a, N., and Sanz-Medel, A., T alanta, 1994, 41, 1833. The within-batch precision (RSD for 11 replicate measure- 16 Madrid, Y., Cabrera, C., Perez-Corona, T., and Camara C., Anal ments at different levels) obtained for the various coated Chem., 1995, 67, 750. graphite tubes is good between 0.7 and 4.0% for all concen- 17 Yamamotto, J., Keneda, Y., and Hikasa, Y., Int.J. Environ. Anal. trations investigated. Chem., 1983, 16, 1. The accuracy of the method was studied by analysing an 18 Bloom, N., and Crecelius, E., Mar. Chem., 1983, 14, 49. IAEA/W-4 (simulated fresh water) reference material with a 19 Mertens, H., and Althaus, A., Fresenius’ Z. Anal. Chem., 1983, 316, 696. certified total mercury concentration of 1.8–2.7 mg l-1. The 20 Yan, X.-p., Ni, Z.-m., and Guo, Q.-l., Anal.Chim. Acta, 1993, results obtained were 2.2±0.3 and 2.3±0.3 mg l-1 when using 272, 105. SnCl2 and NaBH4 as the reducing solutions, respectively, 21 Sturgeon, R., Willie, S, Sproule, G., Robinson, P., and Berman, showing that the value for inorganic mercury determined in S., Spectrochim. Acta, Part B, 1989, 44, 667. the reference material is in good agreement with the certified 22 Tao, G., and Fang, Z., T alanta, 1995, 42, 375. value. 23 Tao, G., and Fang, Z., J. Anal. At. Spectrom., 1993, 8, 577. 24 Zhang, L., Ni, Z.-m., and Shan, X.-Q., Spectrochim. Acat, Part B, 1989, 44, 339. CONCLUSION 25 Iwamoto, E., Shimazu, H., Yokota, K., and Kumamaru, T., J. Anal. At. Spectrom., 1992, 7, 421. Mercury cold vapour can be effectively adsorbed onto the 26 Yan, X.-P., and Ni, Z.-m., J. Anal. At. Spectrom., 1991, 6, 483. surface of iridium, zirconium and tungstate-coated graphite 27 Ni, Z.-m., Hang, H.-B., Li, A., He, B., and Xu, F.-Z., J. Anal. At. tubes at room temperature. The iridium coated graphite tube Spectrom., 1991, 6, 385. 28 Ni, Z.-m., He, B., Hang, H.-B., and Li, A., J. Anal. At. Spectrom., gives the best analytical performance. Thus, this permanent 1993, 8, 995. chemical modifier could be used for the preconcentration of 29 Matusiewicz, H., Kopras, M., and Suszka, A., Microchim. Acta, mercury vapour avoiding the problems related to the use of 1995, 52, 282. palladium. The inorganic and total mercury concentration of 30 Lee, S., Jung, K., and Lee, D., T alanta, 1989, 36, 999. 15 samples can be determined per hour. 31 Baxter, D., and Frech, W., Anal Chim. Acta, 1989, 225, 175. 32 Tsalev, D. L., D’Ulivo, A., Lampugnani, L., Di Marco, M., and Zamboni, R., J. Anal. At. Spectrom., 1995, 10, 1003. REFERENCES 33 Oda, C. E., and Ingle, J. D., Anal. Chem., 1981, 53, 2305. 34 Miller, J. C., and Miller, J. N., Statistics for Analytical Chemistry, 1 Metals and their Compounds in the Environment, ed. Merian, E., Wiley, New York, 1984. VCH, New York, 1991. 2 Sanemasa, I., Takagi, E., Deguchi, T., and Nagai, H., Anal Chim. Paper 6/05079D Acta, 1981, 130, 149. 3 Mandal, S., and Das, A., At. Spectrosc., 1982, 3, 56. Received July 22, 1996 4 Freiman, P., and Schmidt, D., Fresenius’ Z Anal. Chem., 1982, Accepted October 21, 1996 313, 200. Journal of Analytical Atomic Spectrometry, March 1997, Vol.

ISSN:0267-9477    

DOI:10.1039/a605079d

出版商:RSC    

年代:1997

数据来源: RSC