摘要:

Non-destructive Sampling Method of Metals and Alloys for Laser Ablation= Inductively Coupled Plasma Mass Spectrometry* A. RAITH R. C. HUTTON AND I. D. ABELL Fisons Instruments Elemental Analysis Ion Path Road Three Winsford Cheshire UK CW7 3BX J. CRIGHTON BP Research & Engineering Centre Chertsey Road Sunbury-on-Thames Middlesex UK TW16 7LN A sampling method is described for analysis of metals and alloys using laser ablation-inductively coupled plasma mass spectrometry (LA-ICP-MS). The steel rubbing technique involves polishing the surface of the metal component of interest with a diamond lapping film disc and then analysing the material transferred to the disc using LA-ICP-MS. The method has the advantages that it is essentially non- destructive components can be virtually any shape or size and sampling is simple and safe and requires no electrical equipment.Furthermore samples can be taken by non- specialist staff anywhere in the world by following simple instructions and the samples posted back to the laboratory for analysis. Investigations carried out on steel standards have shown that the precision of the technique is generally better than 10% (relative standard deviation) with accuracy within that experimental error. Limits of detection are in the range Keywords Non-destructive sampling method; steel rubbings; laser ablation; inductively coupled plasma mass spectrometry 1-10 pg g-1. Ferrous and non-ferrous alloys are amongst the most com- monly used construction materials for a wide variety of indus- trial and commercial applications. In many instances such as for petroleum and petrochemical constructions there are strin- gent specifications on the type of materials that can be used e.g.the correct grade of steel and type of welding rods. It is perhaps inevitable therefore that situations arise where it is necessary to check retrospectively whether fabricated compo- nents have been manufactured from the correct grade of material and that the latter is within the required specifications. X-ray fluorescence (XRF) and optical emission spectrometry are commonly applied in the metals industry for identifying and checking alloy specifications. However conventional appli- cation of these techniques generally requires a relatively large well prepared (often polished) sample or in the case of portable equipment can often not be used in restricted environments such as oil rigs and refineries.Solution techniques such as atomic absorption spectrometry or inductively coupled plasma mass spectrometry (ICP-MS) can also be used for alloy analysis but normally these also require destructive sampling of the component to be tested e.g. drillings turnings and filings. For over 15 years a technique has been in use at BP Research and Engineering Centre which circumvents many of these problems.' The method involves rubbing the surface of the metal component of interest with a diamond lapping film ~ ~ * Presented at the 1995 European Winter Conference on Plasma Spectrochemistry Cambridge UK January 8-13 1995. Journal of Analytical Atomic Spectrometry disc and then analysing the material transferred to the disc using XRF.The technique has several advantages for field sampling applications since it is essentially non-destructive. It consists of only polishing the sample surface and can be used with components of virtually any shape size or location. Furthermore sampling can be carried out by non-specialist staff anywhere in the world following simple written instruc- tions and the sample then posted back to the laboratory for analysis. Owing to its unique sampling capabilities the XRF steel rubbing technique has lent itself to a large number of applications within the petroleum and petrochemical industry. Typical examples include checking that the correct alloy grades had been used for fabrication of pipework vessels and engine components; testing welds during fabrication of oil rigs to ensure that correct welding rods had been used; and checking the integrity of isolation valve coatings.One of the main limitations of XRF analysis of the steel rubbings however is that since only a small amount of material is presented to the spectrometer limits of quantifi- cation can be relatively high (typically 0.1% m/m),' restricting application of the technique only to the determination of major components. Laser ablation (LA)-ICP-MS has been widely used over many years for elemental analysis of solid samples down to pg g-' The main application areas include not only the geological and environmental fields,'-* but also quantitative analysis of powders glasses e t ~ . ~ " ' The advantage offered by the use of LA-ICP-MS is principally that it offers improved sensitivity over the XRF approach. A review of the LA technique can be found in refs.11-13. This paper discusses the extension of the steel rubbing technique to the determination of minor and trace components of alloys using LA-ICP-MS. EXPERIMENTAL Non-destructive Sampling Method The steel rubbing technique is a sampling method that permits non-destructive analysis of metals and alloys. It involves gently abrading the sample for a few seconds using a flexible polymer disc impregnated on one side with 15 l m diamond particles (661X; 3M St Paul MN USA). The amount of material transferred to the plastic disc is small approximately 1 mg and in general is representative of the material sampled to a depth of about 1 pn from the surface. It is therefore important to remove all debris and oxidation layers before sampling the bulk.The actual rubbing technique itself is very simple. The disc is placed abrasive-side down on the surface to be sampled. The disc is then rubbed in a rotary fashion for about IOs rotated through about 120" then rubbed for a further IOs followed by another rotation of 120" and a final rubbing of Journal of Analytical Atomic Spectrometry September 1995 Vol. 10 59110s. The disc can then be analysed directly by XRF or LA-ICP-MS. Instrumentation The LA-ICP-MS analyses were carried out using the commer- cial LaserLab coupled to a PlasmaQuad ICP-MS instrument (Fisons Instruments Elemental Analysis Cheshire UK). The LaserLab consists of an infrared Nd:YAG laser operating at a wavelength of 1064 nm.The laser beam is focused onto the target sample located in a quartz sample holder. The ablation cell is maintained slightly above atmospheric pressure aiding rapid sample change-over. The ablation cell is mounted on a motorized x:y:z stage which is fully software-controlled. The sample and ablation processes are displayed in situ on a video monitor. Material is ablated from the sample by a combination of mechanisms including vaporization sputtering and acoustic cloud. Ablated material in the form of a microparticulate cloud is transported to the plasma by an argon carrier gas stream where it is ionized and then analysed in the mass spectrometer. LA-ICP-MS Analysis Parameters The steel rubbing disc was placed on a holder in the laser cell.The sample was then ablated by using a raster of 5 x 5 points to enlarge the area of analysis. In addition the laser was defocused 10 mm beneath the sample surface to achieve larger cratering and lower ablation rate avoiding any damage to the disc itself. The laser craters were 500 pm apart; therefore the 5 x 5 point raster covered a sample area of approximately 3 x 3 mm. The analysis area was rastered once each single crater was ablated with the laser operating in Q-switched mode with five laser shots using a frequency of 4 Hz. It should be noted that no rigorous optimization of ablation parameters was made the manufacturer’s default parameters being used in this case. It might be expected that a further improvement could be expected with a more rigorous approach however this was not the objective of this feasibility study.The param- eters were chosen only to optimize the signal from the steel on the rubbing disc without ablating or damaging the disc itself. The acquisition time to pass the raster once was 60s. For all samples 57Fe was used as the internal standard. Fig. 1 Photograph of the steel rubbing disc with laser ablation rasters Fig. 2 Optical microscope picture of the same laser ablation raster at a magnification of 50 x Fig. 3 at a magnification of 100 x Optical microscope picture of the same laser ablation raster Table 1 LA-ICP-MS detection limits (2s) in steel rubbings* Element P V Cr Mn Ni Cu As Mo Detection limit (% m/m) 0.006 0.001 0.002 0.001 0.004 0.001 0.001 0.0007 * Detection limits were calculated from the standard deviation (2s) of five repeat rubbings of the high-purity iron CRM 097-1.RESULTS The steel rubbing technique followed by LA-ICP-MS analysis was applied to steel standards to prove the capability of this method. Eight steel standards were selected for the studies consisting of the unalloyed British Chemical Standard (BCS) Certified Reference Materials (CRMs) 451 452 453 454 and 455 and the low alloy Bureau of Analysed Samples (BAS) CRMs 456,458 and 460. These standards were used to establish calibration graphs. In addition samples of known concen- tration were analysed as unknowns. All standards and samples were analysed on five different sites using the laser raster of 5 x 5 points each time. These data sets could be used to calculate the precision of LA-ICP-MS and to check the homogeneity of the rubbed powder on the disc.592 Journal of Analytical Atomic Spectrometry September 1995 Vol. 10L I 2,Table 3 Comparison of expected values with LA-ICP-MS data. All values in %. Samples diamond steel rubbings* Element Ti V Cr Mn Co Ni Cu Sn Expected value (0.005) (0.2) (1.3) (0.9) (0.014) (0.14) (0.13) (0.01) LA-ICP-MS 0.006 f 0.0005 0.19 k 0.008 1.33k0.05 0.96 +_ 0.04 0.013 k0.0004 0.1 1 0.008 0.1 1 10.004 0.009 0.0009 Sample B Expected value (0.006) (3.25) (0.64) (0.019) (0.26) (0.006) (0.2) (0.5) Sample J LA-ICP-MS 0.01 k 0.003 0.18+0.003 3.37 +0.008 0.70 _+ 0.02 0.019 +0.0006 0.28 _+ 0.04 0.62 5 0.004 0.0062 0.0004 Sample F1 Sample F2 Expected value LA-ICP-MS (0.003) 0.01 +0.001 (0.007) 0.0065 & 0.0002 (1.21) 0.78 fO.01 (0.7) 0.79 & 0.02 (0.01) 0.015 k0.0003 (1.5) 1.87&0.06 - 0.10 + 0.00 1 (0.003) 0.008 +0.0002 LA-ICP-MS 0.008 k 0.0005 0.0065 k 0.0002 0.78+0.01 0.78 k0.02 0.016 f 0.0005 1.86k0.04 0.10+0.002 0.008 f 0.0002 * Each sample was analysed on five different sites and treated as an unknown sample.57Fe was used as internal standard. Samples B J and F samples of known concentration. The detection limits were calculated from 2s standard deviation of five repeat rubbings of a high-purity iron BCS CRM 097-1 showing detection limits in the range 1-10 Fg g-'. The 2s level i s the accepted confidence level used by the proponents of the technique and indicates the industrial requirements. The eight steel standards were ablated on five different sites.Site 1 of each standard was used to establish the calibration graphs while sites 2-5 were then re-analysed as unknown samples. Figs. 4-6 show the calibration graphs for the elements Cr M n and Co. Table 2 gives the LA-ICP-MS results averaged over the five different ablation sites for the steel standards 455 456 458 and 460 and for sample A and shows a comparison with the reference values. The precision between the different laser rasters is mostly better than 10% (relative standard deviation). The correlation of the data points on the calibration graphs has generated accuracy for most elements sufficient for the purposes of the test. Samples of known concentration B J and F were then re-analysed using the calibration graphs established with the unalloyed and low alloyed steel standards.One of the samples (F) was rubbed with two different discs F1 and F2 and the results are displayed in Table 3. The results for the element Ni for example show an expected value of 1.5% in sample F and the first diamond disc shows an LA-ICP-MS result of 1.87% and the second disc 1.86%. This demonstrates the reproduc- ibility achievable with the steel rubbing approach. CONCLUSIONS The non-destructive steel rubbing technique followed by LA-ICP-MS analysis has been found to be extremely effective for the analysis of metals and alloys. A wide range of elements can be determined and hence the technique could be used for the industrial characterization of materials in oil rig fabri- cations welding engine components and as a tool for fault diagnosis.Comparison of accuracy on a range of steel stan- dards and samples showed good correlation with certified values certainly sufficient for identification purposes. The initial results presented in this paper are extremely encouraging and this sampling method shows considerable promise for rapid non-destructive sampling on metals and alloys. REFERENCES 1 Purdue G. E. and Williams R. W. BP Sunbury Technical Memoranda Nos. 121#223 121#645 and 20#931 1976-1981. 2 Jarvis J. in Handbook of Inductioely Coupled Plasma Mass Spectrometry eds. Jarvis K. E. Gray A. L. and Houk R. S. Blackie New York 1990 vol. 7 p. 172. Abell I. in Applications of Plasma Source Mass Spectrometry eds. Holland G. and Eaton A. G. The Royal Society of Chemistry Cambridge 1990 p. 209. 4 Remond G. Batel A Roques-Cannes C. Wehbi D. Abell I. D. and Seroussi G. Scanning Microsc. 1990 4 249. 5 Pearce N. J. G. Perkins. W. T. Abell I. Duller G. A. T. and Fuge R. J. Anal. At. Spectrom. 1992 7 53. 6 Perkins W. T. Pearce N. J. G. and Jeffries T. E. Geochim. Cosmochim. Acta 1993 57 475. 7 Jackson S. E. Longerich H. P. Dunning G. R. and Fryer B. J. Can. Mineral 1992 30 1049. 8 Jarvis K. E. and Williams J. G. Chem. Geol. 1993 106 251. 9 Raith A and Hutton R. C. Fresenius' J. Anal. Chem. 1994 350 242. 10 Raith A. and Hutton R. C. Fresenius' J. Anal. Chem. in the press. 11 Gray A. L. Analyst 1985 110 551. 12 Moenke-Blankenburg L. Laser Microanalysis Wiley- Interscience New York 1989. 13 Longerich H. P. Jackson S. E. Fryer B. J. and Strong D. F. Geosci. Can. 1993 20 21. 3 Paper 5/00940E Received February 16 1995 Accepted M a y 9 1995 594 Journal of Analytical Atomic Spectrometry September 1995 Vol. 10

ISSN:0267-9477    

DOI:10.1039/JA9951000591

出版商:RSC    

年代:1995

数据来源: RSC

摘要:

Effect of Laser Parameters and Tooth Type on the Ablation of Trace Metals from Mammalian Teeth* Journal of Analytical Atomic Spectrometry PETER M. OUTRIDGET AND R. DOUGLAS EVANS Environmental and Resource Studies Program Trent University Peterborough Ontario Canada K9J 788 This study has shown that the analysis of trace metals in tooth material by laser ablation (LA) ICP-MS is wavelength- beam energy- and matrix-dependant. Ultraviolet (UV) light offers more promise for laser work with teeth because it results in Ca-normalized signals for a number of metals which are greater than or equal to those obtainable with green light and with less mass ablated and so less intra-instrument Ca deposition. Two distinct types of ablation processes appear to occur with tooth material. Bulk ablation of metals indicated by Ca-normalized signals that are generally independant of beam energy occurs at green light energies of >2-3 m J and UV energies of > 3-3.5 mJ.Below these energy levels selective thermal desoprtion (fractional ablation) of metals is indicated by increasing Ca-normalized signals associated with declining rates of matrix removal. If fractional ablation could be achieved reproducibly it may have potential for reducing Ca interference effects during the ionization phase of LA-ICP-MS analysis of calcified tissues. Walrus dentine and beluga cement appear to ablate in fundamentally different ways; these matrix effects are probably species-related because beluga dentine and cement ablate similarly. In beluga teeth increasing amounts of matrix are removed with increasing UV and green beam energies up to the maximum levels of our laser while the mass of walrus dentine removed is constant above UV energies of 3.8 mJ and green energies of 7.0 mJ.This suggests the existence of a bulk ablation power density threshold in walrus dentine which once attained produces an ablating plasma that is not affected by increasing beam energy. Pulse rate is an important variable with evidence of greatly reduced fractional ablation and more consistent Ca-normalized signals of metals at a pulse frequency of 5 Hz compared with 10 Hz. Keywords Laser ablation; inductively coupled plasma mass spectrometry; teeth; trace metals Calcified hard biological structures such as teeth bones and shells have long been recognized as important subjects for study by environmental scientists interested in monitoring the concentrations of heavy metals and other potentially toxic elements in the environment.Elements are accumulated and stored in the calcified tissues of a wide variety of organisms and the hard tissue concentrations of many metals may be related either to ambient environmental concentrations or to the levels in soft tissues such as kidneys or liver,’ which are often the target organs of metal toxicity. The hard tissues are intriguing for another reason viz. in many instances they are composed of sequentially deposited layers of mineralized mate- rial which are laid down according to daily or seasonal cycles in bodily metabolism. Analysis of the chemical composition of these layers may have the potential to reveal the history of *Presented at the 1995 European Winter Conference on Plasma ‘Present address Geological Survey of Canada Analytical Spectrochemistry Cambridge UK January 8-13 1995.Chemistry Laboratories 601 Booth St. Ottawa Canada K1A OE8. element accumulation or exposure throughout an organism’s lifetime. Electron and proton micro-probe analyses of calcified tissues have contributed useful biological archival information for some major elements but not for heavy metals and other trace elements because of relatively poor sensitivity.’ Much better detection limits with spatial resolution comparable to other micro-probe techniques are offered by laser ablation-induc- tively coupled plasma mass spectrometry (LA-ICP-MS) which thus has the potential to give new insights into the accumu- lation of trace metals and other elements in hard biological structures.However LA-ICP-MS has received little attention in the environmental and biological sciences. Several basic questions which have been more thoroughly explored during the ablation of other types of solids (mineral- ogical and metallurgical) await investigation with calcified tissues. These questions include the ablation efficiencies and relative signal strengths obtainable with different laser wave- lengths under varying laser operational parameters and by corollary the effect of subtle variations in matrix character- istics such as organic content inorganic composition density and hardness which occur between the different types of calcified structures.Even within a single structure such as a tooth enamel dentine and cement exhibit significant physical and chemical differences,’ the effects of which are currently unpredictable in terms of laser-sample interactions. Thus the object of this study is to investigate the effects of both laser and matrix parameters on the observed signals obtained by LA-ICP-MS for selected trace elements in mam- malian teeth. Relatively homogeneous sections of the teeth of Arctic walrus and beluga whales were ablated with green and ultraviolet (UV) light wavelengths and the signals of selected trace and major elements studied as a function of tooth type (dentine or cement) and laser energy controlled by sequential adjustments of the Q-switch delay setting.The wavelengths studied offer better potential for controlled ablation of tooth material than infrared wavelengths which are absorbed extremely poorly by apatite crystal^,^ the dominant constituent of teeth and bones.’ Earlier work with infrared ablation of various calcified tissues4 demonstrated a large degree of sample cracking and scorching suggesting that poorly defined uncontrolled ablation was occurring. EXPERIMENTAL Sample Preparation The tooth material studied was collected in 1987-88 by the Canadian Department of Fisheries and Oceans from animals harvested in the Arctic by Inuit hunters. Two sets of tooth sections were prepared and ablated in separate experiments. Experiment 1 examined ablation differences across a range of laser beam energies with teeth from two different species.Beam energy was modulated by incremental adjustments of the Q-switch delay which ranged from 240 to 400 ps. The dentine (the inner core) from a walrus tooth and the cement (the outer Journal of Analytical Atomic Spectrometry September 1995 Vol. 10 595mineralized layer) from a beluga tooth were exposed by the radial thin-sectioning (Isomet Low-speed Saw Buehler) of specimens encased in epoxy resin. Thin-sections were mounted on petrographic slides and ground and polished to a consis- tent thickness ( z 0.75 mm) with a Minimet Grinder/Polisher (Buehler). Analyses of beluga cement were performed within a single annual growth layer at the root of the tooth where the layers tend to become wider owing to vertical tooth displace- ment and consequent increased cement deposition during tooth growth.Walrus dentine contains no visible annual layers and is visually quite homogeneous however there may be physical and chemical variations in the material which are not apparent under transmitted light microscopy. Experiment 2 aimed to explore matrix differences between dentine and cement in the same beluga tooth. A single section of 0.6mm thickness was prepared as above except that the tooth was transversely sectioned. Sampling occurred as far as practicable within annual layers of cement and across layers visible within the dentine of beluga teeth. Discs of a glass standard reference material from the National Institute of Standards and Technology (NIST SRM 611 Glass nominal trace element concentration 500 ppm) were also analysed during the second experiment with the same instrumental conditions.All specimens were washed with 20% trace grade HN03 and distilled water to remove surface contamination prior to analysis. Instrumentation A Perkin-Elmer/SCIEX Elan 5000 ICP-mass spectrometer and modified Model 320 laser sampler (frequency doubled and quadrupled gaussian beam profile) were used during this work. The mass spectrometer and laser settings are outlined in Table 1. To accomodate the petrographic slides used for tooth specimens a flattened rectangular sample chamber was con- structed in the laboratory from perspex with a sloping quartz glass cover. A charged-coupled device camera was mounted axially above the sample stage. Control of laser settings sample position and focus was achieved via the ALAS ~oftware.~ Laser light which had been purified by being passed through a system of three dichroic mirrors was focused onto the sample through a 75 mm planoconvex lens.Light wavelengths were Table 1 Operating parameters Elan SO00 ICP-MS Rf poweriW 1000 Nebulizer 0.89 Ar gas flow/l min-' Outer 14.8 No. of replicates 1 Dwell time/ms 200 Readings per replicate 60 Laser sampler (PE 320) Type Nd YAG Wavelengths/nm Green 532 (2nd harmonic) uv 266 (4th harmonic) Green 70 uv (Experiment 1) 70 (Experiment 2) 48 Lamp power input/J No. of pulses Experipment 1 100 Experipment 2 so Experiment 1 10 Experiment 2 S Read delay/s 3 Pulse frequency/Hz Q-switch delay/ps 240-400 (vaned) 10 - Green - Expt. 1 *-. Green - Expt. 2 (t-o UV - Expt.1 a E . L % 4 m 2 0 240 280 320 360 400 0-Switch delaytps Fig. 1 Laser beam energy at different Q-switch settings in Experiments 1 and 2 266 nm for UV and 532 nm for green. To reduce damage to the optics UV lamp power input was reduced to 48 J in Experiment 2 compared with 70 J in the first which led to significantly lower UV beam energies (see Fig. 1). The green lamp power was 70 J in both experiments. The following isotopes were measured 43Ca "Co 63Cu 66Zn "Sr and z08Pb. Data were collected and viewed in Graphics Mode then taken into Quantitative Analysis where the signal profile was integrated. The resulting numeric data were imported into a spreadsheet where blank correction calculations took place. RESULTS AND DISCUSSION Modulation of beam energies by adjustments of the Q-switch delay showed peak energy for green light at Q-switch delays between 260 and 280 ps (Fig.1). The UV beam energy showed a similar pattern with the differences between experiments being due to the lower lamp energy employed in Experiment 2. Experiment 1 Sample pits resulting from both UV and green light ablation exhibited little or no scorching and cracking which was a feature of IR a b l a t i ~ n . ~ Since Ca is the dominant inorganic component of the tooth matrix 43Ca was used both as an indicator of the mass of tooth matter ablated and as an internal standard for the other elements. In Experiment 1 the relationship between 43Ca intensities and laser beam energy (Fig. 2) showed distinct differences between the two types of tooth material.With beluga cement 43Ca intensity displayed a linear relationship with energy for both green and UV wavelengths up to the maximum energy levels of each type of laser light. Conversely with walrus dentine the 43Ca signal was relatively constant at green light energy of >7 mJ and UV energy of > 3.8 mJ. Below these levels the Ca signal was proportional to energy as it was for beluga cement. There are two possible explanations for these data. First the constant Ca signal from dentine may simply be an artifact related to use of the Q-switch delay to vary the beam energy. It is possible that varying the Q-switch settings affects beam energy and other parameters such as pulse length or beam diameter simultaneously resulting in a confounded yet con- stant signal.As Fig. 2 shows different Q-switch delays with similar beam energies in this case 240 and 320ys produce markedly different Ca intensities. However this explanation does not account for the differences between walrus dentine 596 Journal of Analytical Atomic Spectrometry September 1995 Vol. 10I - UV -Beluga Q-S=240 I 1 w Green - Walrus / I /Q-S-320 ~ 0 2 4 6 8 10 Laser energylmJ Fig. 2 43Ca intensities from walrus dentine and beluga cement as a function of beam energy and light wavelength in Experiment 1 (‘Q-S’ indicates Q-switch delay in p) and beluga cement. A second explanation may be that the data for walrus dentine reflect the existence of a power density threshold (ie. energy per unit area per unit time) that is necessary for the formation of an ablating plasma.Once this threshold is attained the mass of tooth matter removed is not significantly affected by increasing levels of beam energy. The absence of an obvious ablation threshold for beluga cement suggests that a plasma was formed which was enhanced by increasing beam energies in contrast to that formed with walrus dentine. In any case it is apparent that the two types of tooth material respond to laser energy in fundamentally different ways. Beluga cement gave higher 43Ca signals than walrus dentine over most of the range of beam energies for both green and UV lasers. This finding may be related to differences in ablation processes between the tooth types or may be indicative of a higher Ca concentration in beluga cement. The latter sugges- tion is supported by the Sr data and by the observation that the volume of material ablated in cement and dentine appeared similar.The Sr:Ca ratio was virtually constant over most of the energy range [Fig. 3(a)] suggesting that Sr and Ca are chemically associated in the tooth matrix and ablate in an identical manner. Strontium is known to substitute for Ca in tooth apatite crystals,6 a fact which was corroborated by a significant correlation between Sr and Ca intensities resulting from the IR ablation of walrus cement.’ Strontium may thus provide an independant surrogate measure of Ca concen- tration. Like Ca the raw ‘*Sr intensities were 3-5-fold higher from beluga cement than from walrus dentine in this study and displayed a similar relationship with laser energy. Analysis of the signals for various transition metals indicates that there are two generalities discernible [Fig.3(b)-(e)]. First although green light typically results in much greater raw signals for all metals (data not shown) the Ca-normalized signals are usually greater from UV ablation. Second an increase of at least an order of magnitude occurs in the Ca-normalized signals of all metals at relatively low energy levels. For green light the increase typically becomes noticeable below 2-3 mJ for both tooth types while for UV the increase occurs below 3.0-3.5 mJ depending on the metal and tooth type. This increase in the metal signal relative to Ca may be indicative of selective volatilization of metals through thermal desorption a process termed fractional ablation.’ If trace metals were removed exclusively by bulk ablation i.e.by mechanisms involving photochemical bond cleavage within the m a t r i ~ ~ or the thermally-related ejection of particulates and aerosols,” then the metal Ca ratio should remain constant across energy settings as it generally is at green light energy levels above 3 mJ and UV energy levels above 3.5 mJ. Increases in the Ca-normalized signals with declining rates of bulk ablation as shown by the Ca intensity (Fig. 2 ) indicate that fractional ablation increases at energies significantly below the power density threshold for bulk ablation as reported by Cromwell and Arrowsmith.’ The occurrence of fractional ablation is supported by observations that at these low energy levels Zn Pb and other metals appear in the sample stream earlier than Ca and Sr a pattern which is consistent with the existence of gaseous metal phases and particulate Sr and Ca (Outridge and Evans in preparation).Fractional ablation has been demonstrated in a number of metallurgical and this study may be the first documentation of the phenomenon with a biological material. There are few other generalities in the data which apply to all metals. For 66Zn at maximum laser energy levels (8.5 mJ for green and 4.0 mJ for UV) there was no difference between light wavelengths in the Ca-normalized signal [Fig. 3(b)]. At lower energy levels the relative signal obtained by each wavelength was matrix-dependant. The UV light gave a higher Zn Ca ratio in beluga cement than green light over most of the energy range below 4 mJ while green light gave a higher Zn Ca in walrus dentine.The Ca-normalized ”‘Pb signal was markedly higher from beluga cement with UV than with green light as for Zn. However this was not the case with walrus dentine in which UV and green wavelengths gave similar signals at any given energy level. In contrast to both Pb and Zn Ca-normalized 59C0 intensities were 10-fold higher from walrus dentine with UV than with green light across the range of beam energies but not from beluga cement except at energy levels less than 3 mJ [Fig. 3(d)]. For 63Cu the Ca-normalized signals from UV and green light are very similar in walrus dentine however in beluga cement UV produced markedly greater Cu:Ca ratios over most of the energy range [Fig. 3(e)].Experiment 2 In this part of the study cement and dentine layers were ablated in the same beluga tooth. Compared with Experiment 1 green light ablation of beluga cement appeared to be less affected by increasing laser energy at the lower pulse frequency (5 Hz) and lower pulse number ( 5 0 ) used here. For example the 43Ca intensity as an indicator of the mass of tooth material ablated decreased only 4-fold between energies of 8.4 and 2 mJ (Fig. 4) compared with a 10-fold decrease Experiment 1. A possible explanation is that the higher pulse frequency used resulted in a hotter plasma than in Experiment 2 in which the lower pulse frequency allowed for greater heat dissipation at any given energy level. Geertsen et al.’ proposed that hotter plasmas ‘shield’ the sample surface from much of the energy in a laser beam allowing only a small part of that energy to be transferred towards the surface.Cooler plasmas by infer- ence should allow proportionally more energy to be directed towards the sample thereby increasing effective power density at the sample surface. Thus a cooler plasma in this experiment may have resulted in relatively greater ablation at similar or lower total beam energies than in Experiment 1. For both green and UV light the 43Ca data indicate that beluga cement and dentine ablated in an identical fashion in contrast to the pattern in Experiment 1. This suggests that the different responses in 43Ca between cement and dentine were due to species-related matrix variations between beluga and walrus rather than to differences between cement and dentine per se.Journal of Analytical Atomic Spectrometry September 1995 Vol. 10 59710.0 (a ) 0 2 4 6 8 10 I- 0 2 4 6 8 100 - UV - Beluga 0 UV-Walrus - Green - Beluga . Green - Walrus I 1 - 0" a 0.1 1 $ (0 0.01 - 0.1 ' 10 1 m 0 9 t) a 8 0.1 N 0.01 l O 0 r - - T - - - 10 0.01 ' I I I I I I 0 2 4 6 8 10 1 s a 0.1 5 Y) 0.01 0.001 0 2 4 6 8 10 0.001 ' I I I I 0 2 4 6 8 10 Laser energylmJ Fig. 3 43Ca-nomalized signals for (a) "Sr (b) 66Zn (c) zOsPb ( d ) 59C0 and (e) 63Cu from dentine and cement in Experiment 1 [legend as in Fig. 3(a)] As in Experiment 1 there were few common patterns between elements as to the effect of wavelength energy or matrix on the Ca-corrected signals. We have chosen to rep- resent the data for Zn and Pb as illustrative of the complexity of results encountered in this study.There are several notable features in the 66Zn results [Fig. 5(u)]. First the Ca-normalized Zn signal resulting from green light ablation of cement and dentine is independant of laser energy compared with the declining trend seen in Experiment 1. This constant Zn Ca ratio suggests the absence of fractional ablation of Zn. In contrast UV ablation at low energy levels again produced enhanced Zn Ca signals indicative of fractional ablation. Second UV light produces a 10-fold greater Zn Ca ratio from dentine than green light however the difference between wavelengths is much smaller in the case of cement particularly at the maximal energy levels of each type of laser. The Ca-normalized "*Pb data show that UV ablation produced consistently higher signals across the energy range than green light in both types of beluga tooth material [Fig.5(b)]. There was no evidence of fractional ablation with either laser wave- length since the Pb Ca signals increase or remain constant with increasing beam energy in contrast to Experiment 1. 598 Journal of Analytical Atomic Spectrometry September 1995 Vol. 10100 2 10 8 9 2 1 2 J 2 d . .- m 2 0.1 0.01 . 0-0 UV - Cement - UV - Dentine 9 *-.-* Green - Cement m- -... Green - Dentine I I I I I 0 2 4 6 8 10 Laser energy/rnJ Fig. 4 43Ca intensities from beluga cement and dentine as a function of beam energy and wavelength in Experiment 2 (samples were taken from the same tooth) I 1 m . m . , i 0.01 ~1 1 0.1 0.01 $ 99 1 ~ 1 0 - ~ 1310-4 b. i-4 1 ~ 1 0 - ~ 1x104 .3 0 I I I I I 0 2 4 6 8 10 Laser energy/mJ Fig. 5 Ratios of (a) 66Zn 43Ca and (b) 208Pb 43Ca from beluga cement and dentine in Experiment 2 (legend as in Fig. 4) NIST SRM 611 Glass Ablation of a NIST glass reference material (611; nominal concentration of trace elements 500 ppm) indicated that glass ablates in a considerably more consistent fashion and that the 100 I 10 i I 0.1 li c-i! UV-Zn intensity H Green-Zn intensity 0 0 UV-Zn:Ca ratio 0 Green-2n:Ca ratio /;@. * . . 0.01 1 I I I 1000 1 y--+- = 0 100 I D-il UV-Co intensity )-. Green-Co intensitv Green-Co Ca ratio 0 01 0 2 4 6 8 Laser energy/mJ 5 4 3 m 0 9 . E N 2 9 1 0 5 4 3 mz 9. 0 u 2 % 1 0 Fig. 6 (a) 66Zn intensities and 66Zn 43Ca and (b) "Co intensities and "Co 43Ca from NIST SRM 61 1 Glass Standard in Experiment 2 results are more consistent between elements than tooth material.Results for Zn and Co are presented here. As with tooth material 66Zn intensities were considerably higher from green light ablation compared with UV [Fig. 6(a)] however the signals for Zn and for Ca (not shown) from glass were remarkably independant of green light energy compared with teeth. The Zn Ca ratios were again higher with UV than with green light. The 59C0 intensities resulting from green light ablation were also relatively constant across the range of energies and Co Ca ratios generally were similar to Zn Ca for both light wavelengths [Fig. 6(b)]. Wavelength- Matrix- and Energy-dependant Ablation One of the unwelcome features of ablating calcified material is the deposition of a fine white powder presumably Ca on sampler and skimmer cones which may also extend to the ion lenses.One of the goals of this work therefore is to maximize the signal of the elements of interest relative to Ca so that intra-instrument Ca accumulation is minimized. On this basis the data presented here indicate that UV wavelengths are preferred over green light because irrespective of laser energy the Ca-normalized signals are higher than or equal to those obtained with green light in most instances but with a much lower mass of tooth material ablated. Higher Ca-normalized signals are also produced through putative fractional ablation at relatively low beam energies for both UV and green light. If the fractional desorption of metals from teeth could be Journal of Analytical Atomic Spectrometry September 1995 Vol.10 599achieved reproducibly it may offer a preferred mode of laser ablation since it would minimize Ca interference effects during plasma ionization. Data from Experiment 1 indicate that in terms of the measured Ca intensity a bulk ablation threshold exists for walrus dentine which is wavelength-dependant. The dentine ablation threshold appears to lie at a lower beam energy for UV light (3.8 mJ) than green light (7.0 mJ) although we have not measured these differences in terms of power density. These findings are interesting in light of the views of Masters and Sharp” and Geertsen et al.’ that UV lasers are largely associ- ated with a photo-ablative removal of material as opposed to thermal ablation in the case of IR and presumably green wavelengths.Our results suggest that the disruption of chemical bonds in the dentine matrix by UV photo-ablation occurs at lower energy levels than those required to form a thermally ablating plasma by green light. Walrus dentine appears to ablate in a fundamentally different manner to beluga cement based on their respective responses to increasing laser beam energy. Together the results of both experiments indicate that the differences are species-related since cement and dentine from the same species (beluga) generally showed similar responses to variations of energy and wavelength. Matrix-related effects are presumably due to physi- cal or chemical differences which affect the degree of coupling between sample and laser pulse. Yet the limited knowledge concerning the physical and chemical characteristics of cement and dentine indicates that they are relatively similar compared with glass.The largest potential difference appears to be qualitative variations in the type of organic fraction present. While little is known about the physico-chemistry of the teeth of walrus and beluga human dentine possesses only a slightly lower organic content (20% m/m) than human cement (24-26%) but virtually all of the organic matter is collagen. In cement collagen accounts for only 20% of the organic matter.’ There are no significant differences in terms of apatite crystal structure or size between cement and dentine although there are between enamel and other tooth and bone tissues.’V6 In addition to the complex and interacting effects of matrix wavelength and beam energy on signal strengths work with tooth samples is further complicated by a high degree of matrix heterogeneity.Each type of tooth tissue consists of mineralized material deposited in sequential layers although those layers are not always visible.’ Physiological variables related to animal nutrition age sex and tooth growth may affect the physico-chemical characteristics of the matrix and the concen- trations of trace elements deposited in each layer. Intra-layer variations of orders-of-magnitude have also been recorded in a number of calcified tissues for metals such as Pb.’s4 Micro- spatial variations such as these occur to a much smaller degree in mineralogical and metallurgical standard reference material^.*^^.'^ This work was funded by strategic equipment and operating grants from the Natural Sciences and Engineering Research Council (NSERC) of Canada to R.D.E.; P.M.O.was the recipient of an NSERC post-doctoral fellowship. Richard Hughes and Geoff Veinott gave valuable technical assistance. REFERENCES 1 2 3 4 5 6 7 8 9 10 11 12 13 Outridge P. M. Veinott G. and Evans R. D. Enuiron. Rev. in the press. Hillson S. Teeth Cambridge University Press Cambridge 1986 Jackson S. E. Longerich H. P. Dunning G. R. and Fryer B. J. Can. Min. 1992 30 1049. Evans R. D. Richner P. and Outridge P. M. J. Anal. At. Spectrom. 1994 9 985. Richner P. and Evans R. D. At. Spectrosc. 1993 14 157. Driessens F. C. M. and Verbeeck R. M. H. Biominerals CRC Press Boca Raton Florida 1990 pp. 137-142. Evans R. D. Richner P. and Outridge P. M. Arch. Enuiron. Contam. Toxicol. 1995 28 55. Cromwell E. F. and Arrowsmith P. Anal. Chem. 1995 67 131. Geertsen C. Briand A Chartier F. Lacour J.-L. Mauchien P. Sjostrom S. and Mermet J.-M. J. Anal. At. Spectrom. 1994 9 17. Moenke-Blankenberg L. Spectrochim. Actu Rev. 1992 15 1. Chan W.-T. Mao X. L. and Russo R. E. Appl. Spectrosc 1992 46 1025. Masters B. and Sharp B. L. paper presented at the 1995 Winter Conference on Plasma Spectrochemistry Cambridge U.K. January 8-13 1995. Richner P. Evans R. D. Wahrenberger C. and Dietrich V. Fresenius J. Anal. Chem. 1994 350 235. pp. 103-214. Paper 5/01453K Received March 9 1995 Accepted June 5 1995 600 Journal of Analytical Atomic Spectrometry September 1995 Vol. 10

ISSN:0267-9477    

DOI:10.1039/JA9951000595

出版商:RSC    

年代:1995

数据来源: RSC

摘要:

Extraction of Trace Elements in Coal Fly Ash and Subsequent Speciation by High- performance Liquid Chromatography with Inductively Coupled Plasma Mass Spectrometry* Plenary Lecture JIANSHENG WANG MEDHA J . TOMLINSON AND JOSEPH A. CARUSOt University of Cincinnati Department of Chemistry Cincinnati OH 45221-01 7 2 USA The speciation of arsenic in coal fly ash is reported. The amounts of As" and As' were found and these values released into water from coal fly ash are dependent on the pH of extracted solution and the type of coal fly ash (e.g. location collected type of coal and combustion conditions). A small amount of arsenic conversion from Adn to As" was found especially from the lower pH extractions. The arsenic conversion may be caused by co-existing high oxidation states elements in the extracted solution and may be promoted by the grinding process.Oxygen from the atmosphere was found to be an unlikely source for the arsenic conversion. A preliminary investigation for speciation of vanadium and nickel in coal fly ash shows that V' V'" and Ni" exist in the extracted solutions. This extractionspeciation approach can provide some basic information for predicting the hydrolytic behaviour of trace elements in the coal fly ash-water systems and can be extended to the speciation of other elements. Keywords Inductively coupled plasma mass spectrometry; high-performance liquid chromatography; speciation; arsenic; vanadium; nickel The total amount of potentially hazardous toxic elements in coal fly ash is relatively small however there may be significant contributions to the environment since a large amount of coal fly deposits on the land surface every year.There is an increasing interest in trace element speciation of coal fly ash due to the toxic nature of many of its components and because the toxicological profiles of some of the potentially hazardous compounds depend on their total amount and on the oxidation states or chemical forms of the elements. For example several arsenic compounds such as arsenite (As"') dimethylarsinate (DMA) monomethylarsonate (MMA) and arsenate (As') have various toxicities. Among them As"' is the most toxic form about 50 times more toxic than As' and several hundred times more toxic than MMA or DMA.'.' Chromium(rI1) is an essential element for the maintenance of glucose lipid and protein metabolism.However Cr'' is known to be toxic because of its ability to oxidize other species. Exposure to chromium(v1) can result in various forms of ~ a n c e r . ~ . ~ Vanadium(v) has also been determined to be more toxic than vanadium (rv).' Standard reference materials (SRM) (e.g. Coal Fly Ash 1633a 2689 2690 and 2691) from the National * Presented in part at the 1995 European Winter Conference on t To whom correspondence should be addressed. Plasma Spectrochemistry Cambridge UK January 8-13 1995. Journal of Analytical Atomic Spectrometry L I Institute of Standard and Technology (NIST) have only the total amount for each trace element provided. The speciation of arsenic in various samples has been performed by high-performance liquid chromatography (HPLC) coupled with inductively coupled plasma mass spec- trometry ( ICP-MS)6-'2 or with inductively coupled plasma atomic emission spectrometry (ICP-AES).'3-'8 The speciation of chromium with ICP-AES4*15819 and the speciation of vanadium with ICP-MS has also been reported.A difficulty with speciation lies in the successful removal of these trace elements from coal fly ash solid without changing their oxidation states or chemical forms. Most digestion pro- cedures involve harsh conditions which may change any of the original forms of the element to only one of its species. In this study the procedures for trace element speciation in coal fly ash solid samples have been performed in the following order (a) extraction of coal fly ash with water at various pHs (pH = 1-12 extraction may occasionally be referred to as dissolution); (b) determination of trace elements in solutions of various pH values with ICP-MS; and (c) speciation of trace elements in the extracted solution from coal fly ash solid by HPLC-ICP-MS.The ICP-MS detection is chosen for its ultra- trace level detection capabilities. The extraction strategy is to study the effect on total trace element and individual trace element species by varying pHs grinding and/or sonicating. Because of the relatively mild extraction conditions total amounts of the various elements in the fly ash are not extracted. However these extraction conditions are more representative of the actual environmental risk than the conditions imposed by harsh total digestion methods.A detailed study of arsenic speciation and a preliminary investigation of the speciation of vanadium and nickel are also reported. EXPERIMENTAL Instrumentation A commercial ICP-MS instrument (VG PlasmaQuad I1 STE VG Elemental Winsford Cheshire UK) was used for all data acquisition. The operating conditions are shown in Table 1. A concentric nebulizer (C-Type Precision Glassblowing Colorado Englewood CO USA) and a double-pass Scott- type spray chamber (Precision Glassblowing) cooled to 5 "C by means of a refrigerated chiller (Neslab Instrument Portsmouth NH USA) were used. The spray chamber was maintained at 5°C to reduce the amount of solvent vapour (thereby reducing condensation in the torch elbow) reduce oxide interference and improve plasma stability.Journal of Analytical Atomic Spectrometry September 1995 Vol. 10 601Table 1 ICP-MS operating conditions ~- ICP system Forward power/W Reflected power/W Coolant flow rate/l min-' Auxiliary flow rate/l min-' Nebulizer flow rate/ml min-' Sampling depth*/mm 1350 < 2 12 1.5 750 12 Mass spectrometer Sampler (nickel) orificejmm 1.0 Skimmer (nickel) orifice/mm 1.0 1st stage pressure/mbar 1.9 2nd stage pressurejmbar < I 1 0 - 4 3rd stage pressure/mbar Dwell time/ms 320 Channels/u 20 Sweep time/s 0.5 Scanning time/s 60 *Defined as the distance between the front coil of the load coil and < 2 x 10-6 Mass range (m/z) 6-239 the tip of the sampling cone. A Dionex Model DX-300 HPLC System (Dionex Sunnyvale CA USA) and a Rheodyne Model 9125 injector (Cotati CA USA) were used.Samples were injected using a 100 pl loop. A Wescan Anion/R IC column (250 mm x 4.1 mm id Wescan Instrument Deerfield IL USA) was used for the speciation of arsenic. A 45 cm length Teflon PTFE tubing (0.012 in id and 0.030 in od Cole-Parmer Niles IL USA) was used to connect the column to the concentric nebulizer. The eluent 1 (2% propanol) used in the speciation of arsenic was prepared from propan-1-01 (HPLC grade Fisher Scientific Fair Lawn NJ USA) with distilled de-ionized water (18 MR Barnstead Newton MA USA). Eluent 2 (50 mmol 1-' carbon- ate buffer) was prepared from ammonium carbonate and ammonium hydrogen carbonate (Fisher Scientific) and distilled de-ionized water (adjusted to pH 7.5 by ammonia solution). A gradient programme was used for the speciation of arsenic.At the beginning 70% of eluent 1 (2% propan-1-01) and 30% of eluent 2 (50 mmol dm- carbonate buffer) were run for 3 min. The mobile phase was then stepped to 100% eluent 2 until the separation was ~ompleted.~ The flow rate was kept constant at 1 ml min-'. A mixed mode column (250mmx4mm id HPLC-CS5 Dionex Sunnyvale CA USA) was used for the speciation of vanadium and nickel. The eluent was prepared by dissolving lithium hydroxide followed by the addition of 2,6-pyridine dicarboxylic acid. The pH of the mobile phases were adjusted to 3.6 using concentrated nitric acid for the speciation of vanadium and to 6.8 using ammonium hydroxide for the speciation of nickel as specified by Tomlinson et al." Reagents and standards Nitric acid solution (2% m/m) was prepared from doubly distilled concentrated nitric acid (GFS Chemicals Columbus OH USA) with distilled deionized water.Multi-element stock solutions (10 mg g-') of V As Zn Ni Mn Cr Pb and Cd were prepared from 1000 mg ml-' single element stock solu- tions (Certified Atomic Absorption Standards Fisher Scientific) in 5% nitric acid and 5% hydrochloric acid respectively. Multi-element standard solutions for the external standard calibration were prepared by serial dilution of the multi- element stock solutions with 2% nitric acid. Two sets of standard solutions 5 20 50 and 100 ng g-' as well as 10 100 500 and 1000 ng g-' were used to construct calibration graphs for the analysis of extracted solutions containing low and high levels of analytes. Stock standard solutions of As"' As" DMA and MMA were made from NaAsO (As"') Na,HAsO *7H20 (As') (CH,),AsO*OH (DMA) and CH3As0,Na2.6H,0 (MMA) (Sigma St. Louis MO USA).Standard solutions ranging from 2 to 100 ng g-' were injected into the HPLC-ICP-MS system for preparing the calibration curve. Stock standard solutions were made using vanadyl(1v) sulf- ate trihydrate [VOS04 .3H20] for V'" ammonium metavan- adate (NH,VO,) for V" and nickel(I1) nitrate hexahydrate [Ni(N03)2 .6H20] for Ni". All were purchased from Aldrich Chemical Company (Milwaukee WI USA). Extraction From Coal Fly Ash Effect of varying pH Coal Fly Ash SRM 1633a (NIST Gaithersburg MD USA) was weighed (1kO.001 g) in an HDPE bottle (Nalge Rochester NY USA) using an electronic balance (Sartorius Bonemia NY USA).Then 20 g (kO.001 g) of distilled de-ionized water (DDW pH 7) or water at various known pH values were carefully introduced into the bottle. Magnetic stirring bars (l/2 in long x 1/8 in od Fisher Scientific) and a Thermix stirring hot-plate (Model 610T Fisher Scientific) were utilized to stir the suspensions continuously at room tempera- ture (18 "C) for 7 d. The extractants with known pHs (water at various pHs) were prepared by diluting stock solutions of HNO H2S04 and HC1 with distilled de-ionized water in order to study the effects of different pHs and different types of acid. Concentrations of HN03 and H2S04 were at low levels; even the solution at pH = 1 required only a few drops of stock acid solution. It is assumed that no oxidation- reduction reaction takes place due to HN03 and H2S04 during the process of extraction.The extractants made from diluted HCl solution were used to study the possible polyatomic interferences formed in the ICP on the determination of analytes such as 4oAr35C1+ on 75As+ and 35C1160+ on 51V+. The extractants prepared from NaOH and Na,C03 (certified ACS Fisher Scientific) were to study the effect on the trace element solubility due to different bases. The pH values of extracted solutions were measured by a pH meter (pHTestr 2 Cole-Parmer). No significant changes in pH were observed before and after the neutral water extraction of SRM 1633a. Eflect ofgrinding and ultra-sonication A planetary micro grinding mill (Fritsch GmbH Idar- Oberstein Germany) was used to obtain a small particle size of coal fly ash.The procedure is the following an appropriate amount of Coal Fly Ash SRM 1633a was placed in a grinding bowl equipped with grinding balls (Syalon containing 90% silicon nitride). After carefully inserting the bowl into the recess of the bowl plate the grinding mill was turned on and operated continuously for 24h. No apparent heating was observed during the grinding process nor contamination from the grinding bowl and balls determined by analysis of the solution washed out from the grinding bowl and balls after 3 h grinding with no coal fly ash present. The final particle size of the ground coal fly ash was estimated to be from 3 to 10pm according to the manufacturer's information. The ground coal fly ash in a pH 7 solution was sonicated for 5 h using an ultra- sonic bath (Fisher Scientific) and then the suspension was extracted for 7 d.Effect of extraction time and temperature Coal fly ash solid (1 i 0.001 g) was weighed in an HDPE bottle and then 20 g (kO.001 g) of distilled de-ionized water were carefully introduced into the bottle. Magnetic stirring bars and a Thermix stirring hot-plate were used to stir the suspension 602 Journal of Analytical Atomic Spectrometry September 1995 Vol. 10continuously at room temperature (18 "C) for 3 h 6 h 1 d 2 d 3 d and 7 d. To study the effect of extraction temperature the suspensions were heated and maintained at 55°C by using a hot-plate. This extraction was carried out for 3 d with temperature as the only variable that differed from the previous extraction.Extraction of coalfly ash collected from various locations Coal fly ash samples were collected from different locations (labelled STK CZ1 and SFA and provided by Electric Power Research Institute) and extracted with pH 7 (DDW) and pH 2 (adjusted by diluted HNO,) solution. The distilled de-ionized water for the extraction was purged with He gas to reduce dissolved 02. The extraction process was also kept in a nitrogen atmosphere by introducing nitrogen gas (99.99% purity) and sealing the bottles with parafilm (American National Can Greenwich CT USA). This was done to reduce the possibility of the arsenic oxidation. All suspensions were filtered by 0.2 pn hydrophilic nylon membrane filters (Alltech Deerfield IL USA). After filtration the clear solutions were directly introduced into the concentric nebulizer for elemental analysis or were injected into the HPLC column for elemental speciation.All measurements are based on 3 replicate runs. RESULTS AND DISCUSSION Fig. l(a) shows the concentrations of some trace elements in Coal Fly Ash (NIST SRM 1633a) after stirring for 7 d with water at various pH values. Most of them are released in small amounts in water under normal environmental pH conditions (pH = 5-7). However when the pH is < 2 or > 11 the solubilit- ies of trace elements in coal fly ash are markedly increased. At pH 1 and 12 arsenic and vanadium are the most soluble species among those elements under investigation. Even at the pH 1 extracted solution only about 50% of arsenic and 17% of vanadium (calculation based on their total masses in solution compared with the solid) in coal fly ash solid dissolve in water when compared with the certified values.Fig. l(b) with an expanded concentration scale further demonstrates the pos- sibility for speciating arsenic (As) vanadium (V) zinc (Zn) nickel (Ni) and manganese (Mn) in the neutral water extracted solutions (pH = 7) since the concentrations of those elements ranging from 10 to 20 ng g-' is within the working range of The concentrations of chromium (Cr) lead (Pb) and cad- mium (Cd) in the neutral water extracted solutions may be too low to be speciated. The trace elements in coal fly ash can exist in various forms such as salts silicates and oxides. The coal fly ash-water equilibrium system is complex because of various water-solid interactions.These interactions may involve chemical reactions such as precipitation-desorption complexation adsorption-desorption and oxidation- reduction although precipitationdesorption was found to be a dominant reaction for Cr compounds.22 In a simplest case for example As203 and As205 are assumed to exist in the coal fly ash solid phase. The solubility of these oxides in aqueous media depends mainly on their hydrolytic behaviour. It is known that both As,O and As20 are soluble in water although the solubility of As205 is about 6 times higher than that of A S ~ O . ~ ~ An acidic or basic solution increases the trace element solubility as the addition of H+ or OH- forces the equilibrium to form more water-soluble species. It is understandable that the concentration of As in the extracted solution depends on the total amount of As the As species concentrations in the coal fly ash and the extraction pH.The pH-dependent behaviour is likely for other elements as well although there may be only one chemical form such as HPLC-ICP-MS. Fig. 1 (a) Effect of varying pH on the solubilities of trace elements in Coal Fly Ash (NIST SRM 1633a). The concentration represents an average from 3 replicate runs. (b) An expanded concentration scale oxide in the coal fly ash. In addition as elements are released into aqueous solution chemical reactions such as oxidation- reduction may take place. For example at pH <4.8 only Cr"' was found." This was believed to be due to Fe" in the solution acting as a reductant for CrV1.Furthermore even if no oxidation-reduction takes place the trace elements released from the solid phase to aqueous phase would change their original form to the hydrolysed species. Based on the above discussion and on the environmental concerns only the stable hydrolytic species in the coal fly ash-water system were investi- gated in this work since only the 'mobile species' would give the contamination or pose an environmental or health risk. No attempt has been made to investigate the original (and possibly unavailable) forms in the solid phase or the dissolution reaction mechanism. The effect of extraction time from a 3 h to 7 d equilibration period at pH 7 on solubilities of trace elements in coal fly ash (NIST 1633a) is shown in Fig. 2. As can be seen the concen- trations for most of the trace elements slightly increase with longer equilibration except Cd Cr and Pb suggesting their relative insolubility in the aqueous extracting solutions.For As however a plateau between the 3 d and 7 d extraction indicates that maximum extraction (dissolution) has probably been reached. As a result 3 or 7 d extractions were thought to be reasonable and were used throughout the investigation. The Journal of Analytical Atomic Spectrometry September 1995 Vol. 10 6033 h 6 h 1 day 2days 3 days 7days v) 30- - 2 0 % 20 . - x v1 m .- Extraction time Fig. 2 Effect of dissolution time on the solubilities of trace elements in Coal Fly Ash (NIST SRM 1633a) at pH 7 I I - AS"' effect of extraction temperature on the solubilities of trace elements in Coal Fly Ash (NIST SRM 1633a) also was studied by comparing the trace element concentrations in the extracted solution between 18°C (room temperature) and 55°C.The results (Fig. 3) show no significant increase in solubility even though a higher temperature might result in a higher solubility in some cases. Nickel was the only exception as at 55 "C the Ni concentration was found to be about four times higher than that at 18 "C. No explanation for this is available at this time since the precise species are unknown. Comparisons of trace element solubilities in different types of acidic as well as basic solutions are shown in Fig.4. No significant difference was observed between HN03 and HCl; HN03 and H2S04; and NaOH and Na,C03. This means that the potential interference of 40Ar35C1+ on 75As+ as well as 35Cl'60t on "V' due to chloride can be neglected with these extracted solutions.In addition the complexation reaction by various acids or bases appears insignificant. The trace element solubilities for coal fly ash depend primarily on the pH of the extracted solution and little on the acid or base used for the studied here. Fig. 5 shows a chromatogram of As"' DMA MMA and AsV spiked with 500 ppm NaC1. As can be seen the chloride appearing as 40Ar35Clc after about 6 min can be completely separated from all arsenic species (m/z = 75). Therefore soluble C1- salts commonly should have no effect on arsenic speciation with this HPLC-ICP-MS procedure. The initial neutral water extraction s t ~ d y ~ ~ ~ ~ for arsenic speciation in NIST SRM 1633a by HPLC-ICP-MS shows As' ( Z 20 ng g-') as the predominant species (by comparison with standard retention times) with a small amount ( ~ 0 .2 ng g-l) of As"' [Fig. 6(a)]. Fig. 6(c)-(e) shows the neutral extraction- speciation with grinding and sonication and will be discussed below. By contrast only As' was found at increased concen- 0 J Fig. 3 Effect of dissolution temperature on the solubilities of trace elements in Coal Fly Ash (NIST SRM 1633a) at pH 7 Fig. 4 Comparison of trace element solubilities in different acidic and basic solutions DMA AS" Time/min Fig. 5 Speciation of a mixed standard solution of As"' As" DMA and MMA each at 50 ppb spiked with 500ppm of NaC1. Chromatographic conditions are given in the text tration at pH 2 [Fig. 6(b)].A conversion from As"' to As' might have taken place during the acidic extraction process with oxidizing agents e.g. 0 from the atmosphere or high oxidation state elements in the extracted solution such as Mn or Cr. The reduction potentials of these oxidizing agents are usually higher in acidic than neutral aqueous media. These early observations suggested neutral pH and an 0,-free con- ditions for the extraction of arsenic to minimize conversion from As"' to As'. However when the same extracted solution was analysed again by HPLC-ICP-MS after five months storage there was no significant change intimating that atmos- pheric O2 is not the likely oxidizing source for arsenic or more conversion would have been noted as the solution aged. To study the effect of oxygen an extracted solution from coal fly ash (CZ1) containing significant amounts of As"' and AsV was analysed.It was found that CZ1 released about 12 ng g-l of As"' and 230ng g-' of As' in the neutral water extracted solution without de-gassing or other treatment. By comparing the results from the same extraction period with de-gassing and with N atmosphere no significant difference was found. This further indicated that O2 has little part in As"' to AsV conversion. The extracted solutions from both coal fly ash CZ1 and NIST SRM 1633a at pH 7 were acidified to pH 1 and then injected into the HPLC-ICP-MS system for further study. No obvious decrease in As"' or increase in AsV was found. 604 Journal of Analytical Atomic Spectrometry September 199.5 Vol.10( a ) p H = 7 I 20 15 10 i AsV I 255 ppb /Pn = z 39.7 ppb I ,PH = 9 ( a ) - - r 15 I ( b ) pH = 2 7.5 h AsV 8 1 ( d ) pH = 7 Sonicated t 10 I 1 ( e ) pH = 7 Ground & Sonicated Time/min Fig. 6 Speciation of arsenic in dissolved solutions from Coal Fly Ash (NIST SRM 1633a); Chromatographic conditions same as Fig. 5 . (a) Extraction solution at pH 7; (b) extraction solution at pH 2; (c) extraction solution at pH 7 with grinding; (d) extraction solution at pH 7 with ultra-sonication; and (e) extraction solution at pH 7 with both grinding and ultra-sonication However this finding does not preclude the possibility of As"' oxidation by other elements in high oxidation states since those element concentrations at pH 7 may not be sufficient to convert such As"' to AsV.The overall chemistry differences from extraction at pH 1 compared with extraction from neutral water then acidified to pH 1 would be expected to be different. Further investigation is needed to provide confirmation for As"' oxidation by high oxidation state elements in the fly ash. The effect of pH on the speciation of nickel was also performed using the method developed by Tomlinson et aL20 as shown in Fig. 7. At a pH of 2 at m/z 58 255 ppb of nickel was extracted. At a pH of 9 39.7 ppb was extracted and at a pH of 12 16.9 ppb was seen. When the fly ash sample SFA was speciated at pH 2 and 7 an additional peak appeared at pH2 which could be due to a polyatomic interferent from calcium. At a pH of 7 343 ppb of Ni" was seen and a signal equaling an amount of 289 ppb Ni was determined at a pH of 2.Fig. 8(a) shows the speciation of vanadium at m/z=51 I - 5 u) v) - -_ 5 '0 100 200 300 400 500 600 7 10 5 ' 0 100 200 300 400 500 600 700 Time/s Fig. 7 (a) Speciation of nickel from Coal Fly Ash (NIST SRM 1633a) and (b) speciation of nickel in sample SFA. Mobile phase contained 6 mmol 1-' 2,6-pyridinedicarboxylic acid and 8.6 mmol 1-' lithium hydroxide adjusted to a pH of 6.8 with ammonium hydroxide Flow rate used for the separation was 1.5 ml min-' r I - 7.5 12.5 7.5 12.5 Time/min Fig. 8 (a) Speciation of vanadium extraction solution at pH 1 from Coal Fly Ash (NIST SRM 1633a) and (b) chromatogram of a vanadium standard solution. Mobile phase contained 6 mmol I-' 2,6-pyridinedicarboxylic acid and 8.6 mmol 1-' lithium hydroxide adjusted to a pH of 3.6 with nitric acid solution extracted at pH 1 and adjusted to a pH of 2.6 for the chromatography.'' Fig.8(b) shows the vanadium(Iv) and vanadium(v) standards on the same scale as the unknowns from SFA. The vanadium and nickel results suggest that other elements can be extracted and speciated. Extraction at pH 7 was done with and without grinding- ultra-sonicating. Grinding and ultrasonicating should improve the extraction efficiency since it theoretically provides more efficient contact between fresh solvent and solid surfaces from the greater surface area of finely divided particles. Fig. 9 illustrates the effect of grinding and ultra-sonicating on the solubilities of trace elements in Coal Fly Ash NIST SRM 1633a at pH 7.The solubilities of most of the trace elements under investigation were increased by various degrees with grinding and/or ultra-sonication except with Cr and Pb. Among them Mn showed the greatest improvement in the extraction efficiency. In digestion procedures ideally all solids would dissolve Journal of Analytical Atomic Spectrometry September 1995 Vol. 10 605Fig. 9 Effect of grinding and sonicating on solubilities of trace elements in Coal Fly Ash (NIST SRM 1633a) at pH 7 5 r while in extraction the solubility of the solid compound depends in part on solid surface availability and solubility rather than harsh chemical reaction. If the trace element compounds in the fly ash are not homogeneously distributed on the particle surface different results from one extraction to another will be expected.It is seen from Fig. 6(a)-(e) that the intensity of AsV increases with grinding and sonicating or with a lower pH as might be expected from surface area and solubility considerations. The concentration of As"' however is unpredictable and may depend on the sample batch. Sample batch dependent behaviour for AsV may also have an effect but it is not apparent at the higher AsV concentration. Thus particle surface homogeneity is a concern with aqueous extractions of fly ash as would be expected. Possible inconsistencies in As"' results can be found in Fig. 6(a)-(e). It was previously indicated that grinding and sonicating should improve the extraction efficiency for both AsV and As"' because of more effective contact between solvent enhanced surface areas.Higher concentrations of both AsV and As'" in the extracted solution should be expected. However Fig. 6(a) and (c)-(e) shows increasing amounts of AsV and small but varying amounts of As(II1) from neutral water extracted solutions with grinding sonicating or both although the amounts are small and near the low working range of the technique. Coal fly ash sample STK was used to study further this arsenic conversion since a relatively large amount of As"' was found in this sample. The arsenic speciation chromatograms of extracted solutions at pH 7 from sample STK with and without grinding are illustrated in Fig. 10 and an increase in Asv and a decrease in As"' with 24 h grinding can be seen. This confirms that arsenic conversion did occur.The conver- sion is probably promoted by the grinding process which may result in more high oxidation state elements in the extracted solution. The possibly inconsistent results from Fig. 6(a) and (c)-(e) may be caused by several factors. Grinding and sonicat- ing improves the extraction efficiency for both As" and As"' and more arsenic will be released into the extracted solution; however with grinding and sonicating other elements with high oxidation states are extracted at the same time potentially resulting in more conversion from As" to AsV. The processes are competitive but not easily predictable without much more being known about the sample. Non-homogeneous distribution on the surface of coal fly ash particles will also influence the extraction of differing arsenic species and may therefore cause uncertainty in As"' results.Results at the low working range for the speciation studies also have larger uncertainties. ( b ) pH = 7 24-hour grinding 3.0 ( a ) pH= 7 1.5 I As"' h As" I 0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 Tirnelrnin Fig. 10 Speciation of arsenic in dissolved solution from coal fly ash STK (a) extraction solution at pH 7; and (b) extraction solution pH 7 with grinding Fig. ll(a) shows a small amount of As"' and a dominant amount of AsV extracted at pH 7 from SFA but an increasing amount of As"' with a basically constant amount of AsV at pH 2 [Fig. 11 (b)]. These results differ somewhat from those of NIST SRM 1633a and may be due to the differing nature 3.0 ( a ) pH = 7 1 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 Tirnelrnin Fig 11 Speciation of arsenic in extraction solution from coal fly ash SFA (a) extraction solution at pH 7; and (b) extraction solution at pH 2 606 Journal of Analytical Atomic Spectrometry September 1995 Vol.10of arsenic species in the coal fly ash solid. Sample SFA may contain more As"' and clearly more was extracted at lower pH. Arsenic conversion in these samples may be dependent on the nature and amount of co-existing high oxidation state elements. Furthermore the elemental species profiles for any coal fly ash samples are highly dependent on original coal starting materials combustion conditions etc. and therefore must be determined on a case-by-case basis until a broader data base is assembled. The authors are grateful to the Electric Power Research Institute (EPRI) for providing funding for this research through grant numbered RP-2485-26 the United States Environmental Protection Agency (USEPA) for assisting in the purchasing of the VG PlasmaQuad instrument through grant number CR 818301-02 the National Institute of Environmental Health Sciences (NIEHS) for partial financial support through grants number ES 04908 and ES 03221 and Dr.Jeffrey Giglio for his assistance and helpful comments. REFERENCES Fowler B. A Biological and Environmental Effect of Arsenic Elsevier New York 1983. Hodgson E. Mailman R. B. and Chambers J. E. Dictionary of Toxicology Macmillan London 1988 40-41. Snow E. T. and Xu Li-Sha Boilogical Trace Element Research Krull I. S. Bushee D. Savage R. N. Schleicher R.G. and Smith S . B. Jr. Anal. Lett. 1982 15 267. Willsky G. R. White D. A. and McCabe B. C. J. Biol. Chem. 1984,259 13273. Sheppard B. S. Caruso J. A. Heitkemper D. T. and Wolnik K. A. Analyst 1992 117 971. Sheppard B. S. Shen W. L. Caruso J. A. Heitkemper D. T. and Fricke F. L. J. Anal. At. Spectrom 1990 5 431. 1989 V O ~ 21 61-71. 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Heitkemper D. T. Creek J. Caruso J. A and Frike F. L. J. Anal. At. Spectrom. 1989 4 279. Larsen E. H. Pritzl G. and Hansen S . H. J . Anal. At. Spectrom. 1993 8 557. Hansen S . H. Larsen E. H. Pritzl G. and Cornett C. J. Anal. At. Spectrom. 1992 7 629. Beauchemin D. Siu K. W. M. McLaren J. W. and Berman S. S. J . Anal. At. Spectrom 1989 4 285. Branch. S.. Ebdon L.. and O'Neill P. J . Anal. At. Spectrom 1994 9 33. Morita. M.. Uehiro. T.. and Fuwa. K.. Anal. Chem. 1981.53. 1806. Morital M. and Shibata Y. Anal. Sci. 1987 3 575. Roychowdhury S . B. and Koropchak J. A. Anal. Chem. 1990 62 484. Spall W. D. Lynn J. G. Andersen J. L. Valdez J. G. and Gurley L. R. Anal. Chem. 1986 58 1340. Low G. K.-C. Batlev. G. E.. and Buchanan S . J. J. Chromatogr 1986 386 423. Nisamaneeoonp. W.. Ibrahim. M.. Gilbert. W.. and Caruso. J. A.. . . . J. Chromatogr.,-Sci.,'1984 22,'413. Nakata F. Hara S. Matsuo H. Kumamaru T. and Matsushita S. Anal. Sci. 1985 1 157. Tomlinson M. Wang J. and Caruso J. A. J . Anal. At. Spectrom. 1994 9 957. Hirayama S. Kageyama S. and Unohara N. Analyst 1992 117 13. Rai D. and Szelmeczka R. W. J. Enuiron. Qual. 1990 19 378. CRC handbook of Chemistry and Physics CRC press Cleveland OH. Wang J. Tomlinson M. and Caruso J. A. presented at the 41st ASMS (American Society for Mass Spectrometry) conference San Francisco CA USA June 1993 paper No. 628. Caruso J. A Tomlinson M. and Wang J. presented at the 20th Annual Meeting of the Federation of Analytical Chemistry and Spectroscopy Societies (FACSS) Detroit IL USA October 1993 paper No. 343. Paper 5/01 332A Received March 3 1995 Accepted May 5 1995 Journal of Analytical Atomic Spectrometry September 1995 Vol. 10 607

ISSN:0267-9477    

DOI:10.1039/JA9951000601

出版商:RSC    

年代:1995

数据来源: RSC

摘要:

Determination of Arsenic Species by High-performance Liquid Chromatography-Ultrasonic Nebulization-Atomic Fluorescence Spectrometry* Journal of Analytical Atomic Spectrometry AGNES WOLLER ZOLTAN MESTER AND PETER F O D O R ~ Department of Chemistry and Biochemistry University of Horticulture and Food Industry 29-35 Villdnyi H-1114 Budapest Hungary A novel technique involving the coupling of high-performance liquid chromatography with atomic fluorescence spectrometry using ultrasonic nebulization as an interface was assessed for the determination of As"' AsV dimethylarsinic and monomethylarsonic acid. The effects of hydrogen and argon gas flow rates nebulization temperature loop volume pump rate and phosphate buffer conditions on signal intensities were investigated. The detection limit for As"' AS" DMAs and MMAs were 35 50 20 and 20 ng respectively (for 250 mm3 volume injected).Linearity ranges were 250-2500 ng for all four arsenic species. Keywords Arsenic speciation; high-perjormance liquid chromatography; ultrasonic nebulization; atomic fluorescence spectrometry Speciation the identification of different chemical forms of elements presents many challenges to modern analytical chem- istry. The most popular analytical methods used for element speciation today seem to be the so called 'hyphenated tech- niques' especially the combination of a powerful separation process with an adequate element-specific Many previous publication^^*^ involve the speciation of arsenic. The different species of arsenic exhibit wide ranging levels of toxicity for example As"' and As" are very toxic whereas arsenobetaine and arsenocholine are virtually non- toxic.Moreover it seems as happened with selenium that arsenic is on the way to being recognized as a life-supporting element. For arsenic speciation ion chromatography ion exchange and ion pair reversed-phase chromatography are most fre- quently used. Anion-exchange chromatography is especially popular because of the widely varying first pK values of arsenic specie^,^^^ however the ion-exchange columns used often have a very limited lifetime. In a recent paper Liu et aL7 presented the separation of a mixture of four toxicologically important arsenic species [As"' As" dimethylarsinic (DMAs) and monomethylarsonic (MMAs) acids]. The separation involved the use of a modified C18 bonded silica column with didodecyldimethylammonium bromide (DDAB) the run time was 10 min.This method was efficient and fairly robust and was therefore used as a starting point for the development of our system. Most of the hydride forming elements can be detected using atomic fluorescence spectrometry (AFS) in the ultraviolet (UV) region below 250 nm. This is a useful spectral region because only very little background emission can be seen when a low energy cool flame is utilized. Other advantages of the system *Presented at the European Winter Conference on Plasma t To whom correspondence should be addressed. Spectrochemistry Cambridge UK January 8-13 1995. are high sensitivity and selectivity with a relatively simple technical solution.To obtain good detection limits a high radiance excitation source with a uniform level of radiance is needed. One possible reason for AFS not being widely used is that until recently there was no reliable high-intensity exci- tation source on the market. Recently boosted-discharge hollow cathode lamps (BDHCL) have become commercially available and they have proved to be very good excitation sources for AFS detection.' Most of the AFS devices in use have been developed for the determination of hydride forming elements and use hydride generation as a method of sample introduction. The atom cell for hydride generation is commonly a hydrogen-argon diffusion flame where a continuous hydrogen stream is pro- vided by the chemical reaction of sodium tetrahydroborate and hydrochloric acid.Since we did not use hydride generation in our system we needed to solve the introduction of external hydrogen gas. The main problem encountered during the development of this method was from interfacing high-performance liquid chromatography (HPLC) with AFS. The separated species arriving with the eluent from HPLC have to be converted into gaseous form. One possibility for example is hydride gener- ation another is to nebulize the sample into the flame. Aerosols from the nebulizer provide an easy means of introducing liquid samples into atom cells such as flames and plasmas. Mortong was probably the first to use nebulization of a solution by an airstream however the development of pneumatic nebulizers is based on the work of Gouy." Ultrasonic nebulizers (USN) were developed" in the 1920s and have been used in recent times to improve the detection capabilities of atomic spectrometry.The USN type CETAC U-5000 (CETAC Omaha NE USA) differs from the previous USN designs in that it has an air- cooled transducer assembly for heat dissipation and regulation of the transducer power and the heating and cooling tempera- tures of the desolvation system. It consists of a temperature controlled heated cell followed by a water cooled condenser to obtain dry aerosol particles. This way a high amount of water vapour is removed from the system thus decreasing the loading effect of the sample on the atom source.12 According to manufacturer's results the detection limits for most elements obtained with an ultrasonic nebulizer are generally 5-50 times better than those obtained with pneumatic nebulizers.Other sources report enhancement of between 6- and 87-fold but only when normal pneumatic nebulization was compared with ultrasonic nebulization with desolvation. When both systems were compared using desolvation the detection limits of USN were better by a factor of 3-8. The improved detection limit in USN systems is not only the result of using desolvation systems. Another important factor is that increased analyte Journal of Analytical Atomic Spectrometry September 1995 Vol. 10 609transport results from the primary aerosol output in the case of USN because of the generally smaller droplet sizes than those obtained from pneumatic neb~lization.'~ The nebulizer systems needed for analytical AFS must meet a set of requirements quite different from those for nebulizers used in inductively coupled plasma (ICP) spectrometry.Many of our problems arose from this fact since the USN we possess has been developed for ICP detection. This paper reports our experience in using an ultrasonic nebulizer as an interface between HPLC and AFS for arsenic speciation. The aim of this study was to investigate this HPLC- USN-AFS system and to evaluate its analytical performance characteristics concerning the toxicologically important arsenic species. EXPERIMENTAL Instrumentation A Shimadzu Model LC-7A HPLC pump was attached to a sample injection valve (six-port Rheodyne system LMIM Hungary). A 250mm3 sample loop was used for sample introduction. The analytical column was a Bio Separation Technologies (BST Budapest Hungary) CI8 Rutin column (25 x 4.6mm id 10 pm particle size).A Cetac Model 5000 Ultrasonic Nebulizer (CETAC Omaha NE USA) was connec- ted to the column. A Cetac Model 2050 Heated and Refrigerated Constant Temperature Bath provided the required conditions for desolvation. Removal of the drain from the nebulizer was carried out with a peristaltic pump (Rainin Instruments Woburn MA USA). The nebulizer was further connected to an AFS (PSA Excalibur PS Analytical Sevenoaks Kent UK) that utilizes an arsenic boosted- discharge hollow cathode lamp (Superlamp Photron Victoria Australia) as an excitation source. Measurements were carried out around the resonance wavelength of arsenic (193.7 nm) using a multi-reflectance filter with a spectral bandpass between 20 and 40 nm.Because the spectral bandpass was so wide in our experiments we detect all the emitted fluorescence lines in this region. Argon functioned as a carrier gas in the USN and mixed with the hydrogen supported the diffusion flame. The constant gas flows were maintained by Cole-Palmer rotameters (Niles IL USA). A Hewlett-Packard 3396 Series I1 Integrator attached to the AFS functioned as a data collection manipu- lation and recording device. All peaks were evaluated by their peak height. A schematic diagram of the system is shown in Fig. 1. Reagents A stock solution of arsenite (1000 mg dm-3 as As) was pre- pared by dissolving 1.320 g of As,03 (Reanal Budapest I Eluent Argon Hydrogen m m fj 1njecti;n port I u 1 I Column U U AFS (excalibur) Integrator P.pump Cooler Drain Fig. 1 arsenic speciation Schematic diagram of the HPLC-USN-AFS system used for Hungary) in 25 cm3 of 0.5 mol dm-3 NaOH solution and then diluting the solution to 1 dm3 with 0.6 mol dm-3 HC1. AsV stock solution (1000 mg dm-3) was obtained from Merck. A stock solution of MMAs (12.75 mg dm-3) was prepared from strychrotonin solution (Chinoin Budapest Hungary) which is used for medical purposes. A 1 cm3 vial of solution contained 1 mg of strychninium chloride and 50mg of methylarsonic acid sodium salt (12.75mg as As). A solution of DMAs (C,H6AsNa02 *3H20) was obtained from Fluka. The stock solutions of MMAs DMAs As"' and AsV were further diluted in de-ionized water. All working solutions were prepared daily.The DDAB solution (0.01 m ~ l d m - ~ ) was prepared by adding 0.23 g of DDAB (Fluka) to 50 cm3 of water and sonicating the solution until all added DDAB was dissolved. This solution was used for preparing all the other DDAB solutions mentioned in this work. The mobile phase was prepared by adding 0.5% v/v meth- anol and 0.1% v/v of 0.01 mol dm-3 DDAB solution to the Na,HPO buffer solution (20 mmol dm-3). The pH was set to 6.0 by the addition of NaH2P04 solution containing the same amount of phosphate methanol and DDAB as the eluent.' Procedures Column modiJcation The CI8 bonded silica column was modified by passing through 500 ml of a DDAB solution (0.01 mol dm-3) in aqueous 50% methanol at a flow rate of 1 ml min-'. De-ionized water was then passed through the column.The modified column was kept in de-ionized water when not in use. Arsenic speciation Typically a 250 mm3 portion of working standard solution is injected directly into the HPLC column via the injection port. (Fig. 1.) After separation the arsenic species are transferred to the air-cooled transducer of the ultrasonic nebulizer by the eluent. A fine aerosol is formed by the acoustic waves desolv- ation takes place by heating and then suddenly condensing the analyte. A continuous stream of argon enters into the system by sweeping the transducer face plate and it carries the dry aerosol from the nebulizer to the AFS. The hydrogen- argon diffusion flame is maintained by externally introduced hydrogen where the hydrogen gas enters the system via a Y-shaped connection tube located about 25 cm from the flame.(Tube lengths were reduced as much as possible to minimize dead volumes.) All chromatographic peaks were evaluated by their peak heights. Experimental conditions were optimized these operating parameters are presented in Table 1. RESULTS AND DISCUSSION Effect of Gas Flow Rates on Net Signal In order to study the effect of gas flow rates on the signal intensities of different arsenic species we investigated the effect of different hydrogen flow rates as a function of all argon gas flow rates. All the other instrumental parameters of separation and nebulization were kept constant. This type of investigation seemed to be essential in the present system because the argon gas flow rate influences not only the parameters of atomization and the geometry of the flame (as can be expected in the case of hydrogen) but the nebulization process as well.It has to be mentioned that the aim of this investigation was to find optimal working conditions first (to obtain the best detection limit) and to see the complex effect of gas flows on the system. The effect of argon and hydrogen gas flow rates on the signal intensity of As"' is shown in Fig. 2. (This effect was investigated for all four species and the shapes of the curves were similar.) 61 0 Journal of Analytical Atomic Spectrometry September 1995 Vol. 10Table 1 Operating conditions for the HPLC-USN-AFS system Chromatography Column Column temperature/"C Sample loop size/mm3 Mobile phase Pump flow rate/ml min-' Heating temperature/"C Cooling temperaturerc Atomic fluorescence detector Primary current/mA Boost current/mA Argon/ml min-' Hydrogen/ml min- Ultrasonic nebulizer Gas flow rates BST C Rutin column (25 x 4.6 mm id 10 pm particle size) 24 250 20 mmol dm-3 Na2HP04-NaH2P0 buffer +0.1% vjv of mol dm-3 +0.5% v/v methanol 1 (PH 6.0) DDAB 140 5 21.5 35 390 140 700 - 1 1 - $ 600 2 500 2 400 - 300 z .- C 3 .$ 200 5 100 c a 0 50 100 150 200 250 300 Hydrogen flow/ml min-' Fig.2 Influence of hydrogen flow rate on As"' signal intensity at argon flow rates (ml min-') of ( 1 ) 230; (2) 390 and (3) 430 The effect of hydrogen can be explained mainly by its role in the excitation processes (which are very complex) and in the mechanisms of atomization (which are not fully understood).As can be seen in Fig.2 the influence of the hydrogen gas flow rate on the signal intensities is smaller than in the case of argon.This is presumably due to the fact that changes in the flame parameters cannot be observed entirely with our AFS system because the observation height of the detector is fixed. Thus a slight deviation from the optimal gas flow rates can cause changes in the flame geometry in such a way that the emission zone may move to either above or below the optimal observing height of the detector. In addition it is important to consider the residence time of analyte in the flame which is also a function of gas flow. An increase in argon flow dilutes the flame thus changes in the composition geometry and temperature of the flame can occur. On the other hand an increase in argon flow improves the efficiency of ultrasonic nebulization owing to the decreased settling of relatively larger sized droplets.Although the nebulization process (production of aerosol) is basically independent of gas flow it is believed that analyte transport is considerably more sensitive to gas fl0w.I3 This assumption is in good agreement with our finding that at very low Ar flow rates severe band broadening occurred. The USN we use was designed for ICP techniques. According to the User's Manual its optimal argon flow rate is around 700-800 ml min-' which proved to be far too much in our case as it blew out the flame. As can be seen in Fig. 2 the best signals were obtained at an argon flow rate of 430 ml min-'. As a working condition we picked 390 min-' because the measured background noise was lower at this rate.Effect of Nebulizer Temperature on Net Signal Aerosol desolvation using a heated chamber followed by a condenser not only increases the analyte transport efficiency relative to transport without desolvation but decreases the diluting effect caused by the eluent flow on the analyte. In our system the evaporation of solution is essential otherwise the large amount (1-2 ml min-') of solution would extinguish the relatively small and cold flame. We expected a positive relation- ship between the heating temperature and the fluorescence signal (higher temperature more efficient desolvation) but the measured effect (Fig. 3) was different from the one we expected. Fig. 3 shows the influence of nebulization temperature on As"' and DMAs signal intensities although it is not shown the same effect was registered in the case of AsV and MMAs.The higher sensitivity at the lower temperature can be explained by the complexity of the system where the signal intensity depends more on the volatilization losses of arsenic compounds than on the effectiveness of solvent evaporation. The heating temperature (see Table 2 ) selected was 140 "C instead of 100 "C because at this low temperature condensation occurred in the non-heated tube leading to the AFS owing to the inadequate desolvation. Presently we are unable to explain why the net signal was optimal at 100 "C (Fig. 3) but further experiments investigating this question will be carried out. Effect of Loop Volume Normally in HPLC an increase in loop volume results in severe band broadening.In our system where aerosol forma- tion takes place in the USN which is located directly after the separating column no significant peak broadening was observed. This means that the detection power of the system can be improved by increasing the volume of the HPLC loop. It is now believed that the main band broadening factor of the system is not the loop volume but the buffer effect of the V 0 50 100 150 200 250 300 350 400 Nebulization temperaturePC Fig. 3 Effect of nebulizer temperature on the fluorescence signals of (1) As"' and (2) DMAs Table2 Figures of merit for the HPLC-USN-AFS method for arsenic speciation As species RT*/min DLt/ng RSD$ (%) As"' 3.0 35 4.1 DMAs 3.7 20 5.5 MMAs 4.8 20 3.7 AsV 9.3 50 4.2 *Retention time.t Detection limit. $Relative standard deviation at 10 pg ml-' (n=9). Journal of Analytical Atomic Spectrometry September 1995 Vol. 10 61 1system which is due to the fact that the volume of the nebulization-evaporation section of the system is 2-3 orders of magnitude higher than the sample loop. - L -1 +2 74-3 _- $74 In Effect of Pump Rate (Mobile Phase Flow Rate) on Net Signal According to general chromatographic rules the effect of eluent pump rate on chromatographic resolution is small. Since our system consists of a nebulization and a desolvation step after the chromatographic separation an investigation into the effect of pump rate on these processes seemed to be a promising idea. Results of these measurements are reported in Fig.4. The figure shows that no significant changes happened even the methanol had no effect on the background and this supports the general chromatographic rule. '5 800 Analytical Conditions The chromatogram of the four separated species of arsenic obtained at optimal working conditions is shown in Fig. 5. The drop in the baseline just before the first peak can be explained by the effect of de-ionized water which functioned as a solvent in our samples. As can be seen the resolution of the first three components is not complete but was as a compromise thought to be acceptable for this system. When the phosphate concentration of the eluent was decreased in order to improve resolution' of As species the retention time of AsV became unacceptably high.To shorten the long retention -f . I 1 min H 0 1 I I I FIOW rate/ml m i d 1 1 2 1 4 1.6 1.8 2 Fig.4 Influence of HPLC pump rate on signal intensities of 10mg dm-j of (1) As"'; (2) DMAs; (3)As"; and (4) 2.55mgdm-3 of MMAs A PI n 0 Fig. 5 HPLC-USN-AFS chromatogram of the four investigated arsenic species A As"'; B DMAs; C MMAs; and D As". The amounts of injected arsenic were 1250 1250 315 and 1250ng respectively - L In .- 5 150 F E - .- 5 100 Y E 01 .- 50 m a O d 250 750 1250 1750 2250 Amount of arsenidng Fig.6 Linear range of four arsenic species (l) (2) DMAs; (3) MMAs; and (4) As" time of AsV the concentration of methanol in the mobile phase was increased. According to the theory of classical reversed- phase chromatography the retention times should have been shortened despite this our results showed no response even to concentrations of methanol several times higher than the original 0.5% v/v.A possible explanation of this phenomenon is that DDAB which functions as an ion-pairing reagent interacts more strongly with the octadecyl chains of the stationary phase than with C4-C6 chains. That is why this separation technique can be classified more as ion exchange than reversed-phase chromatography. After the effects of different parameters on the net signal had been investigated the limits of detection (LOD) for the four arsenic species were determined (the calculated LOD are based on a 3s criterion). The results obtained for As"' DMAs and MMAs are shown in Table 2 (solutions used contained all four arsenic species together.) The relatively high LOD of AsV can be explained by the effect of peak broadening caused by the very long retention of AsV.According to our investigations systems designed for the determination of inorganic arsenic species can only elute As" within 4min using the following parameters phosphate buffer with a mobile phase containing 25 mmol dm-3 (pH 6.5). Considering the complexity of the system the standard deviations are very good (Table 2). The linear ranges obtained for the four arsenic species are presented in Fig. 6. CONCLUSIONS Four toxicologically important arsenic species were separated and detected by the newly developed HPLC-USN-AFS system. The detection limit of the system needs further improvement in order to use it for analytical work that requires higher sensitivity. A promising direction is to adopt this system for hydride generation techniques where we can expect an order of magnitude signal enhancement based on our previous measurements.Another way to improve the analytical perform- ance of the system would be to find a more suitable eluent for AFS because the present system suffers from problems caused by the very high background emission of the phosphate. Since the spectral transmission range of the present instrument is very wide using a filter of specific wavelength instead of the present multi-reflectance filter could improve sensitivity. With regards to the USN (designed for ICP techniques) where the parameters of eddy current are not as important as for separation techniques it would be useful to develop a nebulizer specifically designed for these hyphenated techniques.This research was partly supported by the Hungarian Scientific Research Foundation (OTKA) Grant No. T014329. 61 2 Journal of Analytical Atomic Spectrometry September 1995 Vol. 10REFERENCES Donard 0. F. X. and Martin F. M. Trends Anal. Chem. 1992 11 17. Quevauiviller P. Donard 0. F. X. Maier E. A and Griepink B. Mikrochim Acta 1992 109 169. Irgolic K. J. in Hazardous Metals in the Environment ed. Stoeppler M. Elsevier Science Publishers B.V. Amsterdam 1992 Anke M. in Trace Elements in Human and Animal Nutrition ed. Mertz W. Academic Press New York 1986 pp. 347-372. Hansen S. H. Larsen E. H. Pritzl G. and Cornett C. J . Anal. At. Spectrom. 1992 7 629. Heitkemper D. Creed J. Caruso J. and Fricke F. L. J . Anal. At. Spectrom. 1989 4 279. pp. 288-340. 7 8 9 10 11 12 13 Liu Y. M. Fernandez Sanchez M. L. Gonzalez E. B. and Sanz- Medel A. J. Anal. At. Spectrom. 1993 8 815. Corns W. T. Stockwell P. B. Ebdon L. and Hill S. J. J. Anal. At. Spectrom. 1993 8 71. Morton H. Chem. News 1868 17 231. Gouy C. L. Ann. Chim. Phys. 1879 18 5. Wood W. R. and Loomis A. L. Phil. Mag. Ser. VII. 1927 4,417. Brenner I. B. Bremier P. and Lemarchand A J. Anal. At. Spectrom. 1992 7 819. Tarr M. A Zhu G. and Browner R. F. J. Anal. At. Spectrom. 1992 7 813. Paper 5100988 J Received February 20 1995 Accepted May 15 1995 Journal of Analytical Atomic Spectrometry September 1995 Vol. 10 613

ISSN:0267-9477    

DOI:10.1039/JA9951000609

出版商:RSC    

年代:1995

数据来源: RSC

摘要:

Determination of Trace Amounts of Arsenic Species in Natural Waters by High-performance Liquid Chromatography-Inductively Coupled Plasma Mass Spectrometry P. THOMAS AND K. SNIATECKI Institut Pasteur de LilIe Seruice Eaux Enuironnement 1 rue Calmette BP 245 F-59019 Lille Cedex France Using directly coupled ion-pair reversed-phase high- performance liquid chromatography and inductively coupled plasma mass spectrometry arsenic acid (As') was identified as a major arsenic species in spring-waters and bottled mineral waters. Spring-waters were collected from a volcanic area in the centre of France and bottled waters were purchased from local supermarkets. Two bottled waters also contained traces of arsenious acid (As"'); no monometbylarsonic acid (MMA) or dimethylarsinic acid (DMA) were characterized in the samples.Using this developed method six arsenic species were determined with limits of detection sufficiently low to study the chemical species at their naturally occurring concentration levels. Detection limits were in the range of 1.0-3.0 pg 1-' and a good mass balance was obtained with total As content in samples determined by a hydride generation system. Keywords Arsenic; inductively coupled plasma mass spectrometry; ion-pair; high-performance liquid Chromatography; spring-water and bottled mineral waters; speciation Severe poisoning can arise from ingestion of as little as 100 mg of arsenic trioxide; chronic effects may result from the accumu- lation of arsenic compounds in the body at a low intake Carcinogenic properties have also been recognized for arsenic compounds.The arsenic concentration in most drinking waters seldom exceeds 10 vg l - l although values as high as 100 pg 1-' have been rep~rted.~ Aqueous arsenic in the form of arsenite arsenate and organic arsenicals may result from mineral dissolution industrial discharges or application of herbicides. The chemical form of arsenic depends on its source (inorganic arsenic from minerals industrial discharges and insecticides; organic arsenic from industrial discharges insecticides and biological action on inorganic arsenic) and the toxicity of arsenic on its chemical Methods are available to identify and determine arsenite arsenate monomethylarsonic acid (MMA) dimethylarsinic acid (DMA) arsenocholine arsenobetaine and other organic arsenic Nowadays the determination of these species in an aquatic environment is carried out using several hyphenated technique^.^.'^ The analyst has the choice between gas chromatography high-performance liquid chromatogra- phy hydride generation and cold trapping followed by atomic spectroscopic Unpolluted fresh water normally does not contain organic arsenic compounds but may contain inorganic arsenic compounds in the form of arsenate and arsenite.''4 The purpose of this paper is to present a routine method Presented at the 1995 European Winter Conference on Plasma Spectrochemistry Cambridge UK January 8-13 1995. Journal of Analytical Atomic Spectrometry for the determination of arsenic compounds using ion- pair reversed-phase high-performance liquid chromatography ( HPLC) and inductively coupled plasma mass spectrometry (ICP-MS) in order to produce precise and accurate analytical data for the evaluation of water quality.We also improved HPLC separation with regard to time of analysis separation efficiency and sensitivity of four arsenic species which occur in natural waters. EXPERIMENTAL Instrumentation A PlasmaQuad PQ 11' ICP-MS (VG Elemental Winsford Cheshire UK) was used in its standard configuration with a Meinhard nebulizer (Glass Expension Australia) and a double pass water-cooled spray chamber made of borosilicate glass. The operating conditions are shown in Table 1. The mass spectrometer was set to sample ion intensities using the single ion monitoring mode (SIM) at the analyte mass m/z 75 (75As+) during the coupling measurement. A peristaltic pump was used for conventional sample introduction of the solution at 0.80 ml min-' and the acquisition was carried out using both scanning and peak jumping modes.Before starting the HPLC- ICP-MS system an aqueous arsenic solution of arsenic as arsenite (20 pg 1-') in the selected mobile phase was aspirated in order to optimize the instrument adjustments relative to the ion lens voltages and the nebulizer gas flow. The optimized signal intensity was about 43 000-46000 counts s - l for Table 1 ICP-MS optimal operating conditions Instrument Forward power/kW Reflected power/W Plasma gas flow rate/l min-' Intermediate gas flow rate/] rnin-l Nebulizer gas flow rate/l min-' Nebulizer Spray chamber Ion sampling Sampling cone Skimmer cone Sampling depth Expansion/mbar Intermediate/mbar Analyser/mbar Vacuum Measurement Plasma Quad Turbo + (VG Elemental) 1.35 1 14 1.1 0.80 Meinhard concentric glass nebulizer Scott double pass water cooled (5°C) Nickel 1.0 mm orifice Nickel 0.75 mm orifice 13 mm from load coil 1.8 4.6 Single ion monitoring on m/z = 75 0.0 10-4 Journal of Analytical Atomic Spectrometry September 1995 Vol.10 61820 pg 1-' arsenic standard solution (as arsenite) in the mobile phase without methanol. For chromatographic separations an isocratic LDC/Milton Roy Constrametric I11 with a Hamilton PRPl resin based reversed-phase column particule size (250 mm x 4.6 mm id) was used. A Rheodyne 7125 (Rheodyne Cotati CA USA) injection valve with a 50 pl injection loop was used for sample introduction.Total arsenic determination was carried out using continu- ous flow hydride generation followed by atomic fluorescence detection (Excalibur Plus System PSA Sevenoaks UK). Since arsenite and arsenate have different hydride generation kinetics potassium iodide (30% m/v in demineralized water) was used to reduce arsenate to arsenite before analysis. This solution was used for all sample and standard preparations. In order to determine the accuracy of the method total arsenic was determined in two National Water Samples of simulated fresh water (used for laboratory certifications and performance in an external quality assessment programme). Reagents Arsenic standard solutions (1000 mg 1-') were prepared as follows arsenite As,O (Aldrich) dissolved in NaOH (4 g 1-'); arsenate Na,HAsO 7H,O (Aldrich) dissolved in water; MMA CH,AsO(ONa)*2.6H2O (Carlo Erba) dissolved in water; and DMA (CH3),AsO(ONa)*3.5H,O (Sigma) dissolved in water.Arsenobetaine (AsB) and arsenocholine (AsC) were provided by the Community Bureau of Reference (BCR Brussels Belgium). For hydride generation measurements stan- dard solutions were prepared by the appropriate dilution of stock 1000 mg 1-' arsenic (As'!') chloride solution (Johnson Matthey) using hydrochloric acid solutions. Fresh solutions were prepared daily. HPLC mixtures of arsenic species were prepared in water after appropriate dilution. The mobile phase was prepared as follows counter ion 0.5 mmol 1-' tetrabutlammonium phosphate was dissolved in water the pH was buffered with 4mmol 1-' Na2HP0,*2H20 and adjusted to pH9.0 with diluted ammonia.The chemicals used for the mobile phase were all of Fluka purum pro analysi quality and the water was provided by Charlau (HPLC water). Methanol added to the mobile phase was from Carlo Erba (HPLC grade) and was used to increase the signal sensitivity.'6 All the resulting mobile phase was filtered through a 0.45 pm filter and degassed before use. Procedure and Signal Measurement For total arsenic determinations a reduction using potassium iodide at room temperature for 60 min was carried out on standard solutions and all water samples before running the hydride generation system and fluorescence atomic detector. Coupling of the HPLC system to the ICP-MS was done by simply connecting a PEEK tubing (0.17 mm id) from the exit of the column directly to the Meinhard nebulizer (TR 30 A).We kept the capillary tube as short as possible to avoid too much dead volume. Isocratic elution was carried out with the Hamilton PRPl at a flow rate of 0.90 ml min-'. In a previous paper," we described the optimization of the separation with regard to counter ion concentrations pH values and different phosphate buffer concentrations. Since we used a single ion monitoring mode to record the signal we quantified all the generated chromatograms manually (peak height) after base- line correction. A calibration curve was made for each com- pound up to 100 pg 1-' with the use of this manual method. Results and Discussion In order to test the accuracy of the hydride generation method two simulated test water samples provided by the French Environmental Ministry were analysed.The results obtained given in Table 2 demonstrate the accuracy and precision of the continuous flow hydride generation system which could therefore be used to check the accuracy of the HPLC- ICP-MS analysis. In order to increase the signal intensities of arsenic we used the effect of the addition of carbon as methanol to the mobile phase as described previously by Larsen and Stiirup.'s The signal intensity of arsenic compounds dissolved in a range of methanol additions in the mobile phase is shown in Fig. 1. The results show that in our case the maximum arsenic signal enhancement was obtained when 2% v/v of methanol was added to the solution. The arsenic species signal intensities are enhanced by a factor of 2 for As"' and As' and 1.7 for DMA and MMA.There was little effect of the methanol on the amplitude of the base-line noise; counts increased from 40 to 70 counts s-'. The presence of methanol in the mobile phase caused a slight reduction in the retention times of each peak. The reduction of retention time for AsV and MMA was about 60 s and 10 s for AsC AsB As'" and DMA. This shift of the retention times affected the separation of AsC AsB and As"' but as we were looking at arsenic in water samples we decided to prioritize the signal enhancement rather than the separation; when running routine analysis the reduction of total elution time is also useful. The chromatogram in Fig. 2 shows the separation of the four arsenic species using the selected mobile phase (2% v/v of methanol in the mobile phase). Under these conditions the retention times and the limits of detection (LOD) estimated as three times the base-line noise are shown in Table 3 and a complete chromatogram of the six arsenic standards is shown in Fig.3. In the selected area (Puy de Dome in the Massif Central) concentrations exceeded maximum contaminant levels for drinking water established by the Council Directive 80/68/EEC (Groundwater) adopted by French legislation in 1989. Arsenic in particular is a concern in certain basins (Centre of France). Table 2 Analysis of simulated test solutions (pg 1-' n= 5 ) Arsenic by hydride generation/pg 1-' Test solution Low level High level Certified 4.6 0.8 18+3 Found 4.9 0.1 19.3 i0.3 9 J I I I ' 0 1 2 3 4 [Methanol] ("A v/v) Fig.1 ICP-MS signal intensities for four arsenic species (1 As"'; 2 AsV; 3; MMA; and 4 DMA) versus YO of methanol in the mobile phase. Signal intensities are given as counts s-' for 20 pg 1-' of each compound without their respective counter-ions 61 6 Journal of Analytical Atomic Spectrometry September 1995 Vol. 10i i 4500 3000 'g 2500 7 * c al - .e - 2000 re C 1500 1000 500 I 0 390 780 Time/s Fig. 2 Ion-pair HPLC-ICP-MS signals of arsenic species spiked in aqueous solution 1 As"'; 2 DMA; 3 MMA; and 4 AsV. A 1.0ng amount of each species was injected Table 3 Detection limits based on three times the amplitude of the base-line noise and given for the compounds without their respective counter-ions and retention times with 2% v/v methanol in the mobile phase using HPLC-ICP-MS and 50 pl injections Arsenic species Retention time/s LOD/pg I - ' AsC 118 1.4 AsB 139 1.4 As"' 152 0.9 DMA 207 1.5 MMA 335 3.0 As" 600 2.6 12000 11000 10000 7- 9000 g 8000 2 7000 In .- .5 6000 3 500C $ 4000 300C 2000 1 ooc C C .- 0 i3 390 Timels 71 Fig. 3 Ion-pair HPLC-ICP-MS signals of six arsenic species spiked in aqueous solution 1 AsC; 2 AsB; 3 As"'; 4 DMA; 5 MMA; and 6 As'. A 3.0 ng amount of each species was injected The method was applied to twelve spring-waters and three bottled mineral waters which contained a significant amount of arsenic. The results obtained for the arsenic speciation are shown in Table 4. For all of the spring-waters analysed only AsV was observed even when high concentrations of arsenic were present (samples A,B,G); a chromatogram of spring-water is shown in Fig. 4.As the samples contained large amounts of chloride (210-320 mg 1-I) we checked the isobaric interference from 40Ar35C1; the signal from these ions with our chromatographic system corresponds t o 1.1 pg I-' of arsenic (As"') and was eluted with a retention time of 300 s between DMA and MMA. Table 4 Results for analyses of twelve spring-waters and three bottled mineral waters Hydride Ion-pair HPLC-ICP-MS method As total As,O As,O DMA MMA Sample* as As as As as As as As as As A 120 n d t 110 nd nd B 160 nd 150 nd nd C 35 nd 39 nd nd D 40 nd 35 nd nd E 16 nd 15 nd nd F 16 nd 15 nd nd G 120 nd 108 nd nd H 135 nd 125 nd nd I 40 nd 36 nd nd J 40 nd 34 nd nd K 17 nd 17 nd nd L 17 nd 16 nd nd w 1 120 2.5 110 nd nd w 2 29 0.8 25 nd nd w 3 15 nd 14 nd nd * A to L spring-water samples and W1 W2 W3 bottled mineral t Nd below detection limit shown in Table 3.generation waters samples. 1600 1 1400 ' - ; 1200 g 800 C ._ 'x 1000 L .- v) C .- - 2 600 400 200 cn 3 L 1 0 390 7 Tirnels 0 Fig. 4 Ion-pair HPLC-ICP-MS chromatogram of spring-water in 2% v/v methanol in the mobile phase 1 As" (concentration 16 pg I-' as As Analysis of the three bottled mineral waters from the same region shows that even though we obtained a high level of total arsenic content the predominant form was always AsV but on two of the three samples we found small amount of As"' at a very low level. For the same two samples we also observed an unknown peak (see Fig.5) which may possibly correspond to arsenocholine according to the retention time. However it is surprising to observe that the same concentration of unknown species was obtained (3 kg 1-I) even when the total arsenic content in the two samples was very different (29pg 1-I for sample W2 and 120pg 1-' for sample Wl). Until now arsenocholine was not reported in natural water samples. This unknown peak could also be another arsenic species eluted in our chromatographic system which is difficult to qualify because of the use of only six available standards. When referring to Table4 we can say that the recovery is acceptable with regard to the manual quantitative method used because for all of the samples the difference for the total arsenic content found with both methods does not exceed 10% when carrying out the mass balance.Journal of Analytical Atomic Spectrometry September 1995 Vol. 10 61 75500 5000 4500 'Y) 4000 In 2 3500 '< b '5 3000 - L 2500 .- - 2 2000 ' 1500 1000 500 I- 0 390 780 Timels Fig. 5 Ion pair HPLC-ICP-MS chromatogram of two times diluted mineral bottled water in 2% v/v methanol in the mobile phase 1 As" (concentration 120 fig I-' as As); 2 As"' (concentration 2.5 pg 1-' as As); and 3 unknown peak CONCLUSIONS This routine HPLC-ICP-MS method for the determination of arsenite and arsenate in the presence of other compounds is selective rapid and sensitive and it should be interesting to monitor the behaviour of arsenic in an aquatic environment and the evolution of the water quality through time.Therefore further work using special chromatographic software is required to improve the quantitative measurement at a natu- ral level. REFERENCES 1 2 3 4 5 6 I 8 9 10 11 12 13 14 15 16 17 18 Leonard A in Metals and their Compounds in the Environment ed. Merian E. VCH Publishers Weinheim Germany 1991. Franzblau A. and Lillis R. Arch. Environ. Health 1989 44 385. Levesque L. in Les Micropolluants Minkraux duns les Eaux Superficielles Continentales AFEE Report No. 7 Centre National de Documentation Paris 1994. Arsenic in Environmental Health Criteria 18 World Health Organisation Geneva 1981 174 pp. Howard A. G. and Arbad-Zavar M. H. Analyst 1981,106 213. Aggett J. and Kadwani J. Analyst 1983 108 1495. Branch S. Ebdon L. and O'Neill P. J . Anal. At. Spectrom. 1994 9 33. Hill S. J. Bloxham M. J. and Worstold P. J. J. Anal. At. Spectrom. 1993 8 499. Apte S. C. Howard A. G. and Campbell A. T. Environmental Analysis Using Chromatography Interfaced with Atomic Spectroscopy Ellis Horwood Chichester 1989. Larsen E. H. and Hansen S. H. Mikrochim. Acta 1992 109 47. Larsen E. H. Pritzl G. and Hansen S. H. J. Anal. At. Spectrom. 1993 8 551. Howard A. G. and Comber S. D. W. Mikrochirn. Acta 1992 109 21. Farmer J. G. and Johnson L. R. Environ. Geochem. Health 1985 7 124. Robertson F. N. US Geological Survey Report No 213 1989 p. 171. Korte N. E. and Fernando Q. Crit. Rev. Environ. Control 1991 21 1. Beauchemin D. Siu K. W. M. McLaren J. W. and Berman S. S. J. Anal. At. Spectrom. 1989 4 285. Thomas P. and Sniatecki K. Fresenius' J. Anal. Chem. 1995 351 410. Larsen E. H. and Stiirup S. J. Anal. At. Spectrom. 1994,9 1099. Paper 5/01 149C Received February 24 1995 Accepted M a y 15 1995 618 Journal of Analytical Atomic Spectrometry September 1995 Vol. 10

ISSN:0267-9477    

DOI:10.1039/JA9951000615

出版商:RSC    

年代:1995

数据来源: RSC

摘要:

Measurement of Mercury Methylation in Sediments by Using Enriched Stable Mercury Isotopes Combined with Met h y I m e rcu ry Deter m i na t i on by Gas Chromatography-Inductively Coupled Plasma Mass Spectrometry* Journal of Analytical Atomic Spectrometry HOLGER HINTELMANN R . DOUGLAS EVANS AND JANICE Y . VILLENEUVE Trent University Environmental Science Centre Peterborough Ontario Canada K9J 7B8 A novel technique for the calculation of mercury methylation rates in sediments by using enriched stable mercury isotopes is described. The method takes advantages of the ability of an inductively coupled plasma mass spectrometry (ICP-MS) instrument to measure individual isotopes. An ICP-MS instrument was used as a detector for the determination of methylmercury compounds after separation by gas chromatography (GC).CH3Hg+ was isolated from sediments by distillation converted to methylethylmercury by sodium tetraethylborate and analysed after purge-and-trap pre- collection on a Tenax adsorber and thermodesorption onto the GC column. Detection limits were found to be rc 1 pg (as Hg) absolute or 0.02 ng g-' dry sediment. The precision was 2 4 % relative standard deviation when 250 pg of methylmercury were processed. The accuracy of the GC-ICP-MS technique was demonstrated by analysis of an International Atomic Energy Agency certified reference material (IAEA CRM 356) Harbor Sediment giving a concentration of 5.40 k 0.40 ng g - ' compared with the certified value of 5.46 k 0.38 ng g - '. Mercury methylation was investigated by spiking sediments with stable enriched mercury isotopes at in siru mercury concentrations not perturbing the system.More than 3% of the mercury added to a lake sediment was methylated during a 21 d incubation period. Isotope ratios of total mercury differed significantly from isotope ratios of methylmercury at the end of the experiment suggesting that the system was still not in equilibrium after 21 d. Keywords Inductively coupled plasma mass spectrometry gas chromatography; mercury; speciation; methylmercury; methylation; stable isotope Elevated mercury concentrations in freshwater fish constitute a common problem in Canada the United States and Scandinavia.]" Virtually all of the total mercury in fish4 comprises methylmercury and it is the only mercury species which is biomagnified in the aquatic foodweb.' Although numerous papers and conferences have dealt with the environ- mental health hazard of mercury there are still many open questions in the biogeochemical cycle of mercury.Micro-organisms are known for their potential to methylate inorganic mercury in water and though the contribution of abiotic mercury methylation is still unre- Usually radiotracer methods have been used to investigate the methylation rates in sediment^,'^.'^ believed to be the main site for mercury m e t h y l a t i ~ n . ~ ~ ~ ' ~ However this technique has several limitations. Owing to the low specific * Presented at the 1995 European Winter Conference on Plasma Spectrochernistry Cambridge UK January 8-13 1995. activity of commercially available inorganic "'Hg used in these studies the tracer had to be added at concentrations at least 10 times higher than environmental Hg concentrations.This has an unknown effect on the microbial community and might favour more mercury-tolerant strains of micro-organism and the experiments therefore might not represent natural processes. Today the most commonly used technique for the determi- nation of methylmercury is distillation-ethylation with sub- sequent gas chromatography (GC) separation with atomic fluorescence (AF) d e t e ~ t i o n . ' ~ ' ~ This elaborate method is capable of handling the low environmental methylmercury concentrations encountered in unpolluted sediments.'* There are a few other methods published recently demonstrating the determination of methylmercury in sediments and b i ~ t a ' ~ - ' ~ but these methods cannot be used in methylation studies with either radioactive or enriched stable Hg isotopes owing to their inability to determine individual isotopes or they rely on an operationally defined methylmercury determi- nation24 rather than a chromatographic separation of the various mercury species.The development is described of a GC-inductively coupled plasma mass spectrometry (ICP-MS) method for the determi- nation of methylmercury in sediments. Using this method we were able to investigate the mercury methylating potential of sediments by spiking lake sediments with the stable mercury isotope "'Hg at environmental levels and measuring the produced methylmercury. EXPERIMENTAL Reagents Hg(NO& in 1 % m/v HCl and CH,HgCI in propan-2-01 were obtained from Baker (Phillipsburg NJ USA) and Brooks Rand (Seattle WA USA) respectively.Stocks were diluted to working strength as required. '"Hg ( 5 mg) was obtained from Medgenix Diagnostics (Ratingen Germany) as the metal dissolved in 1 ml of concentrated HN03 (Ultrex 11 Baker) and diluted to 25 ml with de-ionized HzO to a final concentration of 185 mg 1-' lg9Hg. The certified and actual abundances of the different Hg isotopes in the final tracer solution are given in Table 1. The differences between certified and measured values reflect the contamination of the tracer with natural Hg during preparation of the solution. The sodium tetraethylbo- rate solution (Strem Chemicals Newburyport MA USA) was prepared by dissolving 1 g of NaBEt in 100 ml of 1 YO m/v KOH (Fisher Scientific Nepean Ontario Canada).This solu- tion was dispensed in 2 ml storage vials and kept in the freezer until used. Solutions of 9 moll-' sulfuric acid (double sub- Journal of Analytical Atomic Spectrometry September 1995 Vol. 10 619Table 1 Comparison of natural mercury isotope abundances with abundances of enriched mercury used for the methylation study. Values are percent of total and include a correction for mass discrimination Hg isotope 196 198 199 200 20 1 202 204 Natural abundance ("/.I 0.15 10.02 16.84 23.13 13.22 29.8 6.85 Certified abundance of the enriched mercury tracer (To) < 0.01 1.25 97.36 1.05 0.13 0.17 0.03 Measured abundance in the enriched tracer solution (%) < 0.01 1.78 92.57 2.28 0.97 2.03 0.46 boiling distilled Seastar Chemicals Sidney British Columbia Canada) and 20% m/v potassium chloride (Fisher Scientific) were used in the distillation procedure and the solution was adjusted to the optimal pH for the ethylation step by means of a 2 moll-' sodium acetate buffer (Fisher Scientific).For total Hg analysis a 0.05% m/v sodium tetrahydroborate (Fluka Ronkonkoma NY USA) solution in 0.05% m/v NaOH (Fisher Scientific) was used as sample reductant and 3% v/v nitric acid (Baker) was used as the acid carrier. Instrumentation Quartz tubes (100 x 4 mm id) were filled with 150 mg of Tenax TA mesh 20/35 (Alltech Deerfield IL USA) as the adsorbent material. Adsorbed organomercury compounds were released by electrothermal heating using nichrome wire wrapped around the tube and connected to a variable transformer which was remote controlled by the ICP-MS software.The GC column consisted of a 40x0.4cm U-shaped silanized glass column filled with 15% m/m OV-3 on Chromosorb W AW DMCS (Brooks Rand Seattle WA USA). The transfer line (1/16 in id) made from polytetrafluoroethylene (PTFE) was connected to the column outlet with a two-way valve and interfaced to the torch adapter of the spectrometer by means of a T- connector (Cole Parmer Anjou Quebec Canada). Heating of the transfer line was not necessary because of the high vapour pressure of the organomercury compounds investigated. An additional argon gas line was coupled to the T-connector to provide sufficient gas flow to produce a central channel in the plasma and internal mass flow controllers ensured stable gas flow. The ICP-MS instrument used in this work was an ELAN 5000 ( Perkin-Elmer/SCIEX Mississauga Ontario Canada) which was equipped with an FIAS 400 (Bodenseewerk Perkin-Elmer Uberlingen Germany) and AS 90 (Perkin- Elmer) autosampler for on-line Hg vapour generation in total Hg analysis.The masses m/z= 199 200 and 202 were moni- tored in peak transient mode. The data acquisition parameters were chosen to obtain approximately one datapoint per second and per m/z. The instrumental operating parameters and conditions for total and methylmercury determination are summarized in Table 2. Chromatographic raw data were collected with the Elan ICP-MS instrument and then imported into and processed with Chromafile-MS software for further analysis.Peak areas were used for calculations. Instrumental bias was removed by correcting ion intensity ratios of enriched samples (after blank correction) for the bias measured with natural abundance Hg according to the following equation true natural ratio measured natural ratio corrected ratio = uncorrected ratio x Table 2 Operating conditions for the GC-ICP-MS system (settings in parentheses were used for total Hg measurements) Inductively coupled plasma mass spectrometry Argon flow rate/l min-' Outer gas 15 Intermediate gas 0.8 Carrier gas 1.35 Forward power/W 1200 Sampler and skimmer cone Platinum Dwell time/ms 41 (20) Readings per replicate 420 (48) Measured isotopes 199Hg 200Hg "'Hg Argon flow/ml min-' 50 Column Column temperaturePC 105 Adsorption tubes Adsorber material Desorption temperature Sweeps per reading 11 ( 5 ) Gas chromatography OV-3 on carbowax 40 x 0.4 cm Quartz 10 x 0.4 cm 150 mg Tenax TA mesh 20/35 Ramp heating to 200°C within 30 s Mercury Analysis The method for methylmercury isolation preconcentration and GC separation by distillation aqueous phase ethylation adsorbent trap precollection followed by isothermal GC is described in detail elsewhere.18 Briefly the procedure was as follows 1 g of wet sediment was placed into a distillation vial.After the addition of 0.5 ml of H2S0 0.2 ml of KC1 and 10 ml of water methylmercury chloride was isolated from the sedi- ment matrix by steam distillation. An aliquot of the distillate was added to 50 ml of de-ionized water in a 150 ml reaction flask.The sample was buffered to pH 4.9 with a sodium acetate buffer and 50 pl of a NaBEt solution were added. The flask was immediately closed and the mixture allowed to react for 20 min generating volatile CH,HgC2HS from CH,Hg+ and C2H,HgC2H from Hg2+. After the reaction period the solu- tion was purged with N2 and the organomercury species were adsorbed onto a Tenax trap connected to the purge outlet. The organomercurials were released onto the GC column by thermal desorption. Each time a new Tenax trap was connected to the GC column the gas flow to the column had to be interrupted at the column inlet. Prior to this operation the column outlet was disconnected from the ICP-MS instrument with a two-way valve. Otherwise the high back-pressure of the support gas may have disturbed the column packing.Two different blanks were analysed distillation blanks were carried out by distilling and analysing only distilled water without a sediment sample. This blank is identical with conven- tional method blanks and sample results were corrected for this blank. An additional ethylation blank was analysed omit- ting the distillation procedure. This blank takes only contami- nation during ethylation and subsequent sample handling into account and was used to calculate the net standard results (mean peak area of standard minus mean peak area of ethylation blank). Total mercury was determined by vapour generation flow injection (F1)-ICP-MS according to the method described by Stroh and VOllkopf2' with slight modifications for the FIAS programming as outlined in Table 3.Total mercury concen- trations were calculated by external calibration using peak areas. Set-up of Methylation Experiment The method for assessing mercury methylation rates was tested in aerobic and anaerobic sediment slurries. Sediments were taken from Lake Vernon in August 1994 from the deepest point of the lake at a depth of 34 m. The sediment was rich in organic matter indicated by a loss on ignition of 11.0&0.3%. 620 Journal of Analytical Atomic Spectrometry September 1995 VoL 10Table 3 Instrument parameters of the FIAS-400 FI device Step Time/s Pump l/rev min-' Pump 2/rev min-' Valve position Remarks 2 15 100 3 4 100 4 25 0 5 2 100 6 15 100 20 20 80 20 20 Fill Fill Start data acquisition Inject Inject sample Fill Switch to fill position Fill Rinse tubing fill sample loop Rinse tubing and sample loop at wash position Sediment suspensions were prepared by manually shaking 600 g of wet sediment with 11 of original lake water in 2 1 Erlenmeyer flasks.Mercury(11) nitrate (18.8 pg as Hg) enriched in lg9Hg was added immediately before slurrying. This is equivalent to a spike addition of x280ng of added Hg per gram of dry sediment. The original mercury concentration of the sediment was 306f22 ng g-' (dry mass). One flask was purged with N2 to maintain anaerobic conditions the other one with mercury-free laboratory air to keep the system aerobic. Samples for methylmercury determinations were taken after day 1 2 4 8 and 21. Calculation of Methylation Rates To determine the amount of methylated Hg at least two isotopes of mercury must be monitored one representing the newly produced methylmercury from the tracer addition; and the other demonstrating the changes in methylmercury concen- trations derived from the mercury originally present in the sample.For each selected mercury isotope the isotopic methyl- mercury Concentration in the sediment sample is calculated separately by external calibration using the corresponding calibration curves. The total isotopic concentration of the enriched tracer isotope 'MeHg can then be described as the sum of the newly produced and the originally present methylmercury xlMeHg= 'MeHg,+ 'MeHg (1) 'MeHg = Z'MeHg - 'MeHg or ( 1 4 The subscript n refers to the natural methylmercury which was originally present in the sample before tracer addition and the subscript sp is used to distinguish the newly produced methyl- mercury from the spike addition.Eqn. (1) is true for all mercury isotopes in the sample since the tracer solution has impurities of all mercury isotopes as outlined in Table 1. To calculate the concentration of the natural 'MeHg a correction for the methylmercury produced from impurities in the tracer solution is made YMeHg = 2MeHg + 2MeHg (2) It is then necessary to measure the isotope ratios of the two isotopes in the sediment. Defining the isotope ratios R is the isotope ratio (corrected for mass discrimination) of the unspiked sediment sample and R is the isotope ratio (corrected for mass discrimination) of the newly produced methylmercury 2MeHgn R =- 'MeHg sp 'MeHg 'MeHgsp R =- ( 3 ) (4) rearranging gives 2MeHg = R,'MeHg ( 3 4 2MeHg,,=R,,'MeHg ( 44 At this point it must be emphasized that R can be calculated from total mercury measurements as well as from methylmer- cury measurements since the natural isotope ratios for both species are identical.In contrast R is calculated from the total mercury isotope ratio measurement in the tracer solution used for spiking. Therefore eqn. (4) does not represent the final isotope ratio of methylmercury isotopes in the sediment after spiking and microbial mercury transformation reactions. Eqn. (4) illustrates that the relative amounts of methylmercury produced from the major and from one minor isotope in the tracer solution equals the isotope ratio in this solution.Substituting eqn. ( 3 4 and eqn. (4a) into eqn. (2) eliminates 'MeHgi and gives ( 5 ) YMeHg = R,'MeHg + R,p'MeHg substituting eqn. (la) into eqn. ( 5 ) gives X'MeHg = R,( X'MeHg - 'MeHg,,) + R,,'MeHg ( 6 ) Rearranging and solving for 'MeHg the amount of methyl- ated mercury deriving from the enriched isotope during the experiment finally gives 'MeHg,,= YMeHg - R,x'MeHg Rs - Rll (7) To calculate the total concentration of methylmercury pro- duced in the sediment from the spike addition the concen- tration of 'MeHg must be divided by the abundance lAsp of isotope 1 in the tracer solution Eqn. (8) was used in turn to calculate the methylation rate as the amount of Hg methylated per gram of sediment per hour and to calculate the specific rate of methylation ie.the percentage of the added Hg methylated. RESULTS AND DISCUSSION Effect of Carrier Gas Flow Rate and Forward Power on the Mercury Signal For methylmercury determination the analyte being intro- duced into the plasma is of a gaseous nature. Thus the optimization should be carried out on a gaseous species as well rather than aspirating a test solution directly into the plasma. On the other hand it is impractical to perform an optimization on a transient signal like a methylmercury peak. Therefore the ICP-MS operating parameters such as carrier gas flow rate and forward power were optimized with the set- up used for total Hg analysis by slightly rearranging the connections at the FI valve. By replacing the carrier line with the sample line and bypassing the sample loop it was possible to generate elemental mercury continuously from a test solu- tion and to feed it into the plasma.The optimization was then carried out on the continuous Hg signal at m/z=202. The carrier gas was always held constant at values ranging from 800 to 1500 ml min-' and the forward power was scanned from 800 to 1400 W over a period of 1 min and the change in sensitivity was recorded in real time. The resulting three- dimensional response area is plotted in Fig. 1. There was an optimum power setting for each fixed carrier Journal of Analytical Atomic Spectrometry September 1995 Vol. 10 62180000 I I 2500w Carrier flow rate/ml rn1n.l Fig. 1 Influence of forward power and carrier gas flow on the signal intensities of Hg at m/z=202. Hg was introduced into the plasma as gaseous Hg'.flow rate. Generally the sensitivity increased with increasing forward power reached a plateau and decreased again. The power needed to obtain maximum sensitivity shifted to higher power settings with increasing carrier flow rates. At the highest flow rate tested the sensitivity was steadily increasing with higher forward power without reaching a maximum. In order to maintain stable conditions over time it is desirable to work in a region where a broad maximum occurs. Therefore a final setting for the forward power of 1200W and a carrier flow rate of 1350 ml min-' was chosen for subsequent analysis. After the optimization procedure was finished the plasma was turned off the FI system was disconnected and the GC system was coupled to the ICP-MS to begin the methylmercury measurements.Determination of Methylmercury The separation of methylmercury from other mercury species is demonstrated in Fig. 2(A). During sample preparation methylmercury compounds are converted into methylethylmer- cury and inorganic mercury into diethylmercury. Thus impurities of Hgz+ in reagents will show up as diethylmercury in the final chromatogram [see Fig. 2(A) and (B)]. Mercury(O) presumably picked up by the Tenax traps while they were exposed to laboratory air is also present in very small quantit- ies. The possibility that the Hgo was formed during the ethylation reaction by impurities in the ethylating reagent or that it was derived from organomercury degraded to elemental mercury upon heating of the trap can be excluded.The isotope ratios of the Hgo peak always represented natural ratios no matter whether unspiked samples or samples spiked with enriched lg9Hg were processed. In the latter case only the isotope ratios of the peaks representing methylmercury and inorganic mercury changed accordingly. Therefore the Hgo could not originate from the sample. This is illustrated in Fig. 3(B). However impurities of inorganic Hg and Hgo do not interfere with methylmercury determination since the peaks were resolved adequately down to the baseline. Separation was completed in 4 min. For the present work the Hg background at m/z= 199 and 202 was about 100 counts s-l as shown in Fig. 2(C). In contrast to previous w ~ r k ~ ~ * ~ ' no bleeding of mercury species from the column or memory effects from the torch or sampling cones was observed.In addition peak shapes were almost symmetrical and showed no obvious signs of tailing indicating absence of memory effects. Compared with the AF detector widely used today ICP-MS offers some simplifications of the overall procedure firstly drying of the adsorber trap is not necessary since the organo- mercury separation is not affected by traces of water vapour and the ICP-MS can easily handle minor quantities of it while the mercury fluorescence signal is quenched by HzO in AF 60000 40000 20000 0 0 1 2 3 4 5 6 7 - I cn v) 6000 c E 800 I 400 t 200 0 1 2 3 4 5 6 7 Time/min Fig. 2 GC-ICP-MS chromatograms of mercury species obtained after aqueous phase ethylation where CH,Hg+ and Hgz+ are con- verted to CH,HgCzH and C,H,HgC,H respectively A methylmercury standard (equivalent to 250 pg of Hg); B ethylation blank; and C background noise obtained after heating a blanked Tenax adsorber trap.(Note the 100-fold increase in scale from A to C). detection. Secondly the thermal decomposition tube was omit- ted since organomercury species are easily ionized in the plasma. Precision based on six consecutive determinations of methyl- mercury (250 pg as Hg) and measurement of peak area showed a relative standard deviation (RSD) of 4%. Calibration curves for each methylmercury isotope based on determinations of 0.04,0.1,0.25,0.7,2 and 5 ng (as Hg) and peak area calculation were all linear with correlation coefficients of 0.9997 or better. The system is capable of detecting less than 1 pg (as Hg) of methylmercury.Absolute detection limits defined as the amount of Hg required to yield a net peak that was three times the standard deviation of the ethylation blank (n= 5) were 0.9 pg Hg (m/z = 202) and 1.5 pg Hg (m/z = 199) when an ethylation blank is performed as shown in Fig. 2(B). These absolute detection limits are comparable to those obtained by GC-AFS" and superior by factors of 5-150 over those obtained by liquid chromatography ( LC)-ICP-MS26-28 or supercritical fluid chromatography (SFC)-ICP-MSZ9 At the time this paper was prepared the authors were not aware of 622 Journal of Analytical Atomic Spectrometry September 1995 Vol. 1050000 I I 30000 1 0 0 1 2 3 4 5 6 7 0 50000 40000 30000 20000 10000 0 0 1 2 3 4 5 6 7 Time/min Fig.3 GC-ICP-MS chromatograms of sediment extracts A unspiked sediment sample; and B sediment sample after 1 d of aerobic incubation with enriched stable lg9Hg11. Shown are m/z = 199 200 and 202 note the change of the isotope ratio 202 199 in the spiked sample compared with the unspiked sample. a published GC-ICP-MS method used for Hg speciation in sediments. The absolute detection limits (3s of distillation blank n = 3) rise to 5 and 6 pg of Hg respectively when real samples are processed reflecting the additional contamination involved with sample distillation. Relative detection limits for sediment samples were about 20 and 25 pg g-' of Hg respect- ively by using an equivalent of 300 mg of dry sediment and a mean (k s n = 6) distillation recovery of 80 & 7%.The accuracy of the results for methylmercury in sediments was checked by the analysis of an International Atomic Energy Agency (IAEA) certified reference material (CRM) 356 Harbour Sediment certified for methylmercury. Unfortunately IAEA CRM 356 is a polluted harbour sediment showing a high methylmercury c~ncentration.~' The mean (& s n = 3) methylmercury concen- tration of 5.40 If 0.44 ng g-' measured with the described GC-ICP-MS method is statistically no different from the certified value (with 95% confidence interval) of 5.46 If 0.38 ng g-' . Additionally National Research Council of Canada (NRCC) CRM MESS-2 a sediment with a low certi- fied value for total Hg was analysed and a mean (ks n=3) of 0.168 f0.018 ng g-' methylmercury was obtained.This technique was developed to measure mercury methyl- ation rates by using enriched stable mercury isotopes. As the calculation of methylation rates uses isotope ratios the mini- mum observable methylation rate depends on the precision of isotope ratio measurements. The mean ratio m/z = 202 m/z = 199 in unspiked samples based on 28 determinations of different amounts of methylmercury spread over a whole working day and measurement of peak area was 1.772 with an RSD of 1.0%. The isotope ratio calculated from natural isotope abundances is supposed to be 1.770. Any methylation of the mercury tracer which is enriched with 199Hg would lead to a change of the isotope ratio. A deviation from the measured ratio of 1.772 is statistically significant when the value drops below 1.772 - 3s = 1.720.A hypothetical sediment containing 300 pg g-' of CH3Hg+ and 300 ng g-' total Hg contains 89.4 pg g-l 202CH3Hg+ and 50.5 pg g-' 199CH3Hg+. Assuming further that the natural CH3Hg+ level has reached its steady-state concentration the '"CH3Hg+ concentration at the end of the methylation experi- ment would still be 89.4 pg g-'. A measured isotope ratio of 1.720 then translates into a 199CH3Hg+ concentration of 52pgg-'. In conclusion an increase of only 1.5pgg-' of CH3Hg+ or a change of 0.5% of the steady-state concentration is sufficient to detect mercury methylation in sediments by using a stable mercury tracer technique. Therefore we predict that specific mercury methylation rates of less than 0.1% can be observed by spike additions of only 2ngg-' of total mercury which is two to three orders of magnitude lower than the concentrations conventionally used in radiotracer experiments.6J2J3 Mercury Methylation Experiment A set of experiments was conducted to determine the applica- bility of the developed method to methylation studies in real lake sediments.The time courses of methylated mercury orig- inating from the added enriched Hg tracer are illustrated in Fig. 4. The concentrations of methylated I9'Hg increased lin- early over the first 48 h with a further but slower increase on the following two days. For the next 17 days the concentrations remained unchanged in the anaerobic system and decreased slightly in the oxygenated flask. In contrast to other stud- ies,31-33 no statistically significant difference between the aero- bic and anaerobic systems was observed. However this finding could be an experimental artifact owing to the slurry technique used.Samples were taken over the whole sediment column. Samples from the oxygenated flask in particular contained oxic and anoxic sediment. No attempt was made to sample only the oxic top layer which had a thickness of approximately 1 cm at the end of the experiment. Specific methylation rates were constant over the first 48 h. Already after the first day of incubation a quantity far above the detection limit was converted to methylmercury. A chroma- togram of a sediment extract from the original sediment before spiking is shown in Fig. 3(A) and a chromatogram of the same sediment after 1 d of incubation with 199Hg under oxidic conditions is shown in Fig.3(B). In the original sample extract the signal intensities obtained at m/z= 199 200 and 202 are in agreement with the natural abundance of the corresponding T -0- new CH3Hg'(0,) 4 new CH,Hg+(N,) 10 0 t. CH,Hg+ (0,) -a. not. CH,HgC (N2) r '0 8 - F 7 C 2 4 3 1 c = 6 .- U E 5 .- +o =- 2 0 0 4 a 12 16 20 Incubation timejdays Fig. 4 Time courses of the mercury methylation experiments with enriched stable 199Hg11 under anaerobic and aerobic conditions. Dotted lines show the fate of the natural methylmercury originally present in the sediment and solid lines illustrate the amount of newly produced methylmercury. Means of triplicate measurements are plotted error bars represent the standard deviation. Journal of Analytical Atomic Spectrometry September 1995 Vol.10 623Hg isotopes Fig. 3(B) shows a dramatic increase of the inten- sity at m/z = 199 owing to newly produced methylmercury from the spike addition. The intensity ratio at m/z=200 and 202 remained unchanged. Thus only short incubation times are required to detect methylation. Again there was no differ- ence in specific methylation rates between the two systems. In both cases approximately 110 ng of methylmercury were pro- duced per gram of sediment per hour. This is in good agreement with rates obtained by others using radiotracer A maximum of 8.65 ng g-' of the added mercury tracer was methylated under oxidic and 7.36 ng g- under anoxidic conditions. This represents 3.10% and 2.64% of the added inorganic mercury. These conversion rates are also similar to those observed in the l i t e r a t ~ r e .' ~ ? ~ ~ A great advantage of the described method is that the fate of the added tracer as well as the fate of the methylmercury originally present in the sediment can be monitored at the same time in the same sub-sample. This is not possible with conventional radiotracer methods. This simultaneous monitor- ing allows mass balances of the mercury species In the described experiment the mean (ks n = 3 ) isotope ratio for total mercury (m/z = 202 m/z = 199) after spiking was 0.305 f 0.021 whereas the same mean (f s n = 3 ) isotope ratio for methylmercury was calculated to be 0.059+0.011 in the oxidic system and 0.064 _+ 0.006 in the anoxidic system. Under equilibrium conditions there should be no difference between total mercury and methylmercury ratios indicating that the added inorganic mercury was much more available for methylation reactions than the originally present Hg.CONCLUSIONS GC-ICP-MS in conjunction with the described purge-and- trap method enables the analysis of methylmercury species in unpolluted sediment samples with very good precision and accuracy at ultra-trace levels. The ability to measure specific isotopes in combination with the use of stable mercury isotopes provides a powerful technique for the determination of mercury methylation rates. The high sensitivity and precision of isotope ratio determination allows spilce additions at or below in situ mercury concentrations and therefore does not perturb the environmental system.Future work will focus on the simul- taneous addition of multiple mercury species each with a different enriched Hg isotope to determine the availability of individual species for methylation. Additionally de-methylation rates can be measured together with methyl- ation rates by the proposed technique in the same sub-sample if isotope enriched methylmercury is used. This work was supported by a Deutsche Forschungs- gemeinschaft (DFG) research fellowship to H. Hintelmann and NSERC research grants to R. D. Evans. REFERENCES 1 Ontario Ministry of the Environment Guide to Eating Ontario Sport Fish Communication Branch Ontario Ministry of the Environment Toronto Ontario Canada 16th edn. 1992. 2 Bahnick D. Sauer C. Butterworth B. and Kuehl D. W. Chemosphere 1994,29 5!7.3 Hikanson L. Nilsson A. and Anderson T. Environ. Pollut. 1990 49 149. 4 Bloom N. S. Can. J. Fish. Aquat. Sci. 1992 49 1010. 5 Surma-Aho K. Paasivirta J. Rekolainen S. and Verta M. Chemosphere 1986 15 353. 6 Kerry A. Welbourn P. M. Prucha B. and Mierle G. Water Air Soil Pollut. 1991 56 565. 7 Jackson T. A. Appl. Organomet. Chem. 1989,3 1. 8 Compeau G. C. and Bartha R. Appl. Environ. Microbiol. 1987 53 261. 9 Xun L. Campbell N. E. R. and Rudd J. W. M. Can. J . Fish. Aquat. Sci. 1987 44 750. 10 Zhang L. and Planas D. Bull. Environ. Contam. Toxicol. 1994 52 691. 11 Weber J. H. Chemosphere 1993 26 2063. 12 Korthals E. T. and Winfrey M. R. Appl. Environ. Microbiol. 1987 53 2397. 13 Furutani A. and Rudd J. W. M. Appl. Environ. Microbiol. 1980 40 770. 14 Saouter E.Gillman M. Turner R. and Barkay T. Enuiron. Toxicol. Chem. 1995 14 69. 15 Gilmour C. C. Henry E. A. and Mitchell R. Enuiron. Sci. Technol. 1992 26 2281. 16 Saouter E. and Blattmann B. Anal. Chem. 1994 66 2031. 17 Bloom N. S. Can. J. Fish. Aquat. Sci. 1989 46 1131. 18 Horvat M. Bloom N. S. and Liang L. Anal. Chim. Acta 1993 281 135. 19 Baldi F. Parati F. and Filippelli M. Water Air Soil Pollut. 1995 in the press. 20 Hintelmann H. and Wilken R.-D. Appl. Organomet. Chem. 1993 7 173. 21 Lansens P. Meuleman C. Laino C. C. and Baeyens W. Appl. Organomet. Chem. 1993 7 45. 22 Fischer R. Rapsomanikis S. and Andreae M. O. Anal. Chem. 1993 65 763. 23 Bulska E. Emteborg H. Baxter D. C. Frech W. Ellingsen D. and Thomassen Y. Analyst 1992 117 657. 24 Beauchemin D. Siu K. W. M. and Berman S. S. Anal. Chem. 1988 60 2587. 25 Stroh A. and Vollkopf U. J. Anal. At. Spectrom. 1993 8 35. 26 Shum S. C. K. Pang H. and Houk R. S. Anal. Chem. 1992 64 2444. 27 Bushee D. S. Analyst 1988 113 1167. 28 Huang C.-W. and Jiang S.-J. J. Anal. At. Spectrom. 1993,8 681. 29 Carey J. M. Vela N. P. and Caruso J. A. J. Anal. At. Spectrom. 1992 7 1173. 30 Horvat M. Mandic V. Liang L. Bloom N. S. Padberg S. Lee Y.-H. Hintelmann H. and Benoit J. Appl. Organomet. Chem. 1994,8 533. 31 Hintelmann H. and Wilken R.-D. Sci. Total Enuiron. 1995 166 1. 32 Gilmour G. C. and Henry E. A. Environ. Pollut. 1991 71 131. 33 Compeau G. and Bartha R. Appl. Enuiron. Microbiol. 1984 48 1203. 34 Ramlal P. S. Rudd J. W. M. and Hecky R. E. Appl. Enuiron. Microbiol. 1986 51 110. Paper 5/01 526J Received March 13 1995 Accepted June 14 1995 624 Journal of Analytical Atomic Spectrometry September 1995 Vol. 10

ISSN:0267-9477    

DOI:10.1039/JA9951000619

出版商:RSC    

年代:1995

数据来源: RSC

摘要:

Stable Isotope Approach to Fission Product Element Studies of Soil-to-Plant Transfer and in Vitro Modelling of Ruminant Digestion Using Inductively Coupled Plasma Mass Spectrometry* PAUL ROBB LINDA M. W. OWEN AND HELEN M. CREWS CSLFood Science Laboratory Norwich Research Park Colney Norwich UK NR4 7UQ A stable isotope approach has been used to investigate two aspects of the behaviour of fission product elements in the environment and food chain using inductively coupled plasma mass spectrometry (ICP-MS). Limits of detection (dry mass LODs) of 0.053 mg kg-' for Sr 0.011 mg kg-' for Cs and 0.084 mg kg-' for Ce were low enough to allow the determination of soil-to-plant transfer factors for soft fruit and the application of the approach to an in vitru model of ruminant digestion. The multi-element measurement capability of ICP-MS also permitted the analysis of selected nutrients including zinc in in oitro experiments.Keywords Inductively coupled plasma mass spectrometry; caesium; cerium; strontium; soil-to-plant transfer; in vitro ruminant digestion Radioactive fission products are present in the environment primarily as a result of the activities of man. Natural sources of radioactive fission products do exist but they are small by comparison with the activities of the nuclear power and weapons programmes which have operated across the world in the last 55 years.' Although the physical and chemical properties of fission product elements have been studied in depth there still remains a need to obtain a better understand- ing of some aspects of the behaviour of these elements in the environment and their transfer through the food chain.This is important in assessing both the impact of routine discharges into the environment from nuclear installations and the most effective remedial action following a major nuclear incident such as that which occurred at Chernobyl in April 1986.2 Performing large scale experiments with radiotracers can be technically challenging and expensive particularly for beta- emitting nuclides such as "Sr. One method which may help researchers in this field is the use of stable isotope measure- ments using endogenous forms of the elements of interest in true field samples. This approach does not require the adven- titious contamination of test samples by radioactive fallout and can be applied to any field sample containing measurable amounts of the elements of interest.Two applications of this approach are reported here which use inductively coupled plasma mass spectrometry (ICP-MS) for the measurement of soil-to-plant transfer factor^^,^ for 133Cs I4'Ce and *%r and in an in vitro model of part of the ruminant gastrointestinal tract576 used to assess the ability of selected chemicals to reduce fission product element solubility. The multi-element capability of ICP-MS can be used to quantify low levels of many elements in a range of matrices. * Presented at the 1995 European Winter Conference on Plasma Spectrochemistry Cambridge UK January 8-13 1995. Journal of Analytical Atomic Spectrometry Similar quantification procedures were employed in both examples described below but additional elements were included in the in vitro study to assess whether the uptake inhibitors could have an adverse effect upon the mineral status of any animals treated.EXPERIMENTAL Materials and Reagents Acid-cleaned plastic-ware glassware and Milli-Q grade (Millipore) de-ionized water (18.2 MSZ cm) were used in all experiments. Polycarbonate Nunc vials (Life Technologies) were used to store soil and dried fruit samples prior to assay and a Rapide Nutrient Analyser soil test kit (Fisons) was used to measure soil pH values colorimetrically. Sterilin plastic test tubes (Merck) seal-again bags (Transatlantic Plastics) and aluminium foil food cartons purchased from retail outlets in Norwich were also used in the study.PrimaR grade nitric and hydrochloric acids and sodium hydrogencarbonate (HPLC grade) were all purchased from Fisons. Merck SpectrosoL solutions of Ca and K (10 g l-') Sr Cs Zn In and Pb (1 g 1-') were used to prepare working standards together with Ce stock solution (1 g 1-I) obtained from Aldrich. All dilutions were performed using nitric acid (0.79 moll-'). Soil-7 and Hay V-10 [International Atomic Energy Agency (IAEA) Seibersdorf ] were used as reference materials throughout this study. The potential adsorbents examined using the in vitro model were Manox SlOD which is a form of Prussian Blue sodium ammonium ferroferricyan- ide (Degussa) and ammonium dodecamolybdophosphate (Merck laboratory-reagent grade). Lyophilized cellulase (Trichoderma viride 238 104 Boehringer Mannheim UK) anhydrous citric acid (C0759) disodium hydrogen ortho- phosphate heptahydrate (S9390) and porcine pepsin ( 1 10 000 P7000) obtained from Sigma were also used in the in vitro model.Apparatus employed in in vitro experiments included Oak Ridge vials (30 ml); a B&T Unitemp shaker water bath; a Gallenkamp fan oven fitted with shaker tray; a Mettler PC440 balance; Kent EIL 7015 pH meter and plastic bodied combi- nation glass pH electrode. A Centra-4X Centrifuge (IEC) was also used in the study. Procedures Sampling and preparation o f f ruit and associated soils Samples of ripe fruit (strawberry raspberry and blackberry) were taken during the summer and autumn of 1992 together with soil from near the roots of the fruiting plants. Strawberries and raspberries were obtained from commercial Pick-Your- Own farms with the consent of the farmers together with Journal of Analytical Atomic Spectrometry September 1995 Vol.10 625Table 1 ICP-MS operating conditions Parameter Mass range/u No. of channels No. of scan sweeps Dwell time/ps Collector type Outer gas flow rate/l min-l Intermediate gas flow rate/l min-' Nebulizer gas flow rate/l min-' (@ 40 psi) Nebulizer type Rf forward power/W Rf reflected power/W Isotopes monitored PQ 11+ Turbo 44-2 1 2 2048 100 320 Pulse 13.5 1 .o 0.900 DeGalan V-groove 1350 <2 *%r 11% 1 3 3 ~ s 14'Ce PQI 44-212 2048 100 320 Pulse 14.0 0.5 0.800 Cross-flow 1350 < 5 39K 43Ca 66,68Zn '% '151n 133Cs I4OCe 2osPb blackberries from suitable hedgerows in Anglesey Cumbria and Suffolk.Raspberries and blackberries were taken from plants at a height of at least 1 m above ground level wherever possible to minimize soil splash contamination. Sampling of strawberries and raspberries was conducted using two (imagin- ary) W-shaped sampling grids crossing each field and a mini- mum distance of approximately 3 m was maintained between sample points. Seven sets of soil and fruit were taken for each of the two grids i.e. 14 samples of each fruit and 14 soils were taken from each location. At each raspberry or strawberry sampling position a minimum of 1OOg of fruit was taken together with at least 10 g of soil (0-10 cm depth). Soil samples were taken using Sterilin plastic test tubes (13 ml) as scoops and the soil was stored in plastic Nunc vials until required for analysis.Hedgerow blackberries were more difficult to sample in a controlled manner. Several samples of ripe fruit were taken from at least two rows of roadside bushes separated by a distance of several metres. No minimum amount of this fruit was taken because of the lack of ripe fruit in some parts of the bush and the widespread growth of some plants. However wherever a fruit sample was obtained a soil sample was taken from close to the foot of the bush following removal of surface debris. The fruit was double bagged in re-sealable plastic bags and placed in aluminium foil food cartons in cool boxes during transport to the laboratory. Upon receipt at the laboratory all fruit samples were rinsed with de-ionized water dried gently using paper tissues and stored at - 20 "C.Bulk samples were prepared by combining equal amounts of the individual sub- samples from each W-grid (or hedgerow in the case of black- berries). Portions of frozen fruit were allowed to thaw for approximately 30 min at room temperature before homogeniz- ation using a Colworth Stomacher laboratory blender 400. Known masses (approximately 20 g of strawberry or raspberry or l o g of blackberry) were combined to form bulked fruit samples. Each bulk was homogenized once more using the Stomacher and re-frozen to a temperature of -20 "C. Known amounts of the frozen fruit were transferred to acid-cleaned 300ml glass Petri dishes covered with loose fitting lids and freeze-dried using an Edwards Supermodulyo freeze dryer fitted with a bulk tray drying system.The freeze-dried material was reweighed and transferred to plastic Nunc vials fitted with screw-cap lids until required for analysis. Soil samples were passed through a 20 mesh sieve to remove stones twigs or other debris and this process also homogenized the soil. Samples of the sieved soil were stored at + 4 "C until required for bulking and analysis. Portions (2 g) of each of the sieved soils were combined and mixed by shaking. In addition separate portions of the soil bulks were oven-dried at 80°C for 24 h to determine the dry matter content. The pH of these soil samples was determined colorirnetrically. A graduated plastic test-tube was filled to the 2 ml mark with soil followed by barium sulfate powder to the 3ml mark. An aliquot of indicator solution (4ml) was then added to these solids and the mixture shaken well.The mixture was allowed to settle before the colour of the indicator solution was compared with that of a reference chart.7 Preparation of fodder samples Samples of grass from pasture were supplied by staff at the ADAS farm at Boxworth in Cambridgeshire. The grass was not treated prior to use in in vitro experiments. QuantiJication of elements in soil-to-plant transfer samples Known masses (z 250 mg of dried fruit or reference material Soil-7 and x 500mg of Hay V-10 or soil) were digested with 5 ml of concentrated nitric acid in PTFE-lined steel digestion vessels which were sealed and heated to 150°C for 6 h. After cooling the resultant liquor was diluted to produce a nitric acid concentration of 0.79 mol I-' and spiked with indium the ICP-MS internal standard. Trace elements in the diluted digest were measured by ICP-MS using a VG PlasmaQuad PQII Plus Turbo (FI Elemental) using the operating conditions shown in Table 1.Quantification was achieved by automated comparison of the instrument response with that of external calibration standards. Reagent blanks were taken through the analytical procedure and sample data were blank-corrected. Quality control checks were conducted in all analytical batches by assessing elemental recoveries from replicate samples of the certified reference materials IAEA Soil-7 and IAEA Hay V-10. In vitro enzymolyses Enzyme blanks containing no added feed were taken through the complete digestion and quantification procedure in each experiment.All measured concentrations in fodder digests were corrected for these reagent (enzyme) blanks. Duplicate 100 mg samples of fodder and where appropriate 100mg of fodder +100mg of uptake inhibitor were also taken through the following digestion procedures. The in vitro enzymolysis procedure is summarized in Fig. 1 and was adapted from the 1iteratu1-e.~~~ Each test sample was treated with an initial cellulase digestion using 6.25 mg ml-' 2 Blanks 4 Fodder samples 4 Fodder + Uptake Inhibitor Samples 1 Blank 2 Fodder + Inhibitor * 2 Fodder i I Simulated Rumen I Cellulase. pH 6. 1 Blank 2 Fodder + Inhibitor 1 Simulated Abomasum I Pepsin. pH 1 5 . -1 * 2 Fodder Fig. 1 Schematic representation of an in uitro model 626 Journal of Analytical Atomic Spectrometry September 1995 Vol.10lyophilized Trichoderma uiride cellulase in 20 ml of citric acid- disodium hydrogenphosphate buffer which had been adjusted to pH 6.4 with a saturated solution of sodium hydrogen carbonate. These samples were incubated at 40°C for 24 h centrifuged for 30 min at 11 000 rpm and the supernatants decanted and frozen prior to trace element quantification. In order to simulate the next (gastric) stage of ruminant digestion replicate digests were adjusted to approximately pH 1.5 using concentrated hydrochloric acid and 20 ml of pepsin solution C0.2 g pepsin per 100 ml hydrochloric acid (0.1 mol 1-') pH 1.51 were added. These samples were incubated at 40°C for a further 5 h centrifuged for 30 min at 11 000 rpm and the supernatants decanted and frozen prior to trace element quanti- fication.After thawing the test supernatants were diluted 10-fold and spiked with indium as internal standard (25 ng ml-') prior to quantification using a VG PlasmaQuad PQII Plus Turbo or PQI (FI Elemental) using the operating conditions shown in Table 1. RESULTS AND DISCUSSION Soil-to-Fruit Transfer Factors Table 2 shows the measured Sr Cs and Ce data obtained for the reference material IAEA Soil-7. Strontium gave consistently low results (approximately 85% of the reference value) follow- ing pressure bomb digestion with nitric acid indicating possible retention of this element on the insoluble residues observed in soil digest solutions. Mean recoveries (n = 5) for added Sr Cs and Ce were 107 110 and 106% respectively.The Hay V-10 data (Table 2) show close agreement between measured and certified Sr levels and the mean Cs value is within 1 standard deviation of the recommended value. No data were available for Ce in this reference material. Strontium levels in fruit were in the range 0.4-2.6 mg kg-' which is in agreement with the literature values (0.5-9 mg kg-I).' Caesium levels in fruit were approximately three orders of magnitude lower being in the range < 1.1-2.5 pg kg-I (reported levels' are in the range <0.1-2.9 pg kg-'). However Ce levels in all of the fruit were < 14 pg kg-' (fresh mass) with relatively few measurements above the limit of detection (LOD). There is relatively little reported in the literature concerning Ce levels in plants but vegetables have been reported to contain between 2 and 50 pg kg-I dry m a s 2 Measured Sr levels in soils were in the range 11-70 mg kg-' Cs from 1.3 to 5.5 mg kg-' and Ce from 16 to 68 mg kg-'.The results for Cs and Ce are comparable to literature valuesg of 10 and 60 mg kg-' respectively for these elements in the Earth's crust but measured Sr levels are lower than the equivalent value of 375mgkg-' reported in the literat~re.~ Examples of the measured transfer factors are shown in Table 3 which gives the concentration of each element in fruit (fresh mass) divided by the concentration of the relevant element in the soil (dry mass). When fruit was dried mass losses of between 78 and 88% were noted. Transfer factors calculated using dry mass fruit would therefore differ from those shown in Table 3 by almost an order of magnitude.In those cases where a fruit sample contained less than the fresh mass LOD then that value was used in the calculation of the transfer factor to give a 'less than' transfer factor. Variations in the value of such 'less than' transfer factors can generally be attributed to differences in fission product element concen- tration in soils. The transfer factors measured using the stable isotope approach relate to systemic transfer of the elements from the soil into the fruit. It is possible that the higher transfer factors reported in the literat~re,'*-'~ measured after adventitious contamination following a nuclear incident arise from a combi- nation of foliar and systemic uptake. The transfer of surface contamination from leaves and skin into fruit would be a major translocation route and this is known to be a relatively efficient process.'o However in the absence of continued depos- ition of radioactivity other mechanisms including systemic transfer will become increasingly important.Variations in transfer factor over a period of time have been reported for radiocaesium when taken up by mushrooms lichen and higher plants.''.12 Radioactive Cs from fallout is more bioavailable than the endogenous stable Cs but becomes less available over a period of time as the radioactive Cs exchanges with stable Cs bound relatively tightly in the soil. Therefore the results of this work will be of more value when assessing the longer term impact of radionuclide deposition on agricultural land.The transfer of trace elements into plants from soil is influenced by many factors. Soil pH clay content humic and fulvic acid content particle size and the concentrations of other ligands and metals can all affect the transfer proce~s.'~-'~ The proprietary method of determining soil pH was used to identify any large scale variations in pH between sampling locations. The results of analysing the bulked soils showed that all of the soils had a pH in the range 6.0-7.5. Consequently no attempt was made to correlate pH with transfer factor in this limited data set. There is a high degree of uncertainty and variability associ- ated with some reported radiocaesium data within given plant species e.g. mushrooms with quoted transfer factors of from 0.005 f0.006 to 0.183 f0.166.17 Mean soft fruit transfer factors Table 2 Recoveries of elements from reference materials and spiked reference materials (mg kg-l) Sample Value Soil-7 Measured (n = 6) Certified Certified range Measured (n = 6) Certified Certified range Non-certified Non-certified range Limit of detection (dry mass) Hay V-10 Sr 86.5 2 10.2 108 39.6 & 1.2 40 37-44 103-1 14 0.053 c s Ce 5.4 61 5.2 f 0.4 5 1.4 & 4.7 4.9-6.4 50-63 0.02 & 0.002 0.087 & 0.045 0.0 17 0.01 1 0.084 0.01 6-0.019 Table 3 Examples of transfer factors measured using the stable isotope approach Fruit Element Suffolk Cumbria Anglese y Literature Blackberry c s 1.2 x 10-3 <3.8 x 10-4 6.7 x 10-4 1.0 x Strawberry Ce <2.5 x 1 0 - 4 <2.2 x 10-4 2.1 x 10-5 2.5 x 10-3 Raspberry Sr 4.0 x lo-' 3.2 x lo-' 1.9 x 1.7 x Journal of Analytical Atomic Spectrometry September 1995 Vol.10 627measured using the stable isotope approach (ignoring those values below detection limits) were (3.6k2.2) x for Sr (4.9 _+ 4.4) x for Ce. Bunzl and KrackeI8 reported Sr and Cs transfer factors for wheat flour which differed by an order of magnitude with a Sr transfer factor of 0.12+_0.006 and a Cs transfer factor of 0.019f0.001. Similar differences have also been noted for rye barley and oats in both the flour and bran.18 The trends shown by the transfer factors reported here therefore are not at odds with those found in the literature for other plant species. Another aspect of possible contamination of plants by radio- nuclides which can be investigated using a multi-element stable isotope approach is the behaviour of fission product elements under the conditions found in the ruminant gastro- intestinal tract.for Cs and (8.9 f 8.6) x In Vitro Model These experiments were designed to investigate metal solubility under the conditions found in the gastric phases of ruminant digestion. In vitro models of ruminant digestion have been developed in the past to study cellulose digestion’ or for use with radiocaesium contaminated fodder when screening new uptake inhibitors.6 The current model however enables studies into both nutrient and fission product element behaviour without the need to prepare labelled samples. Presenting the results of such an investigation poses prob- lems because of the wide range of element concentrations present in fodder.The concept of an inhibition factor (Fi) has therefore been developed to permit the ready display of the effect of adding an inhibitor to in vitro enzymolyses. The inhibition factor for a specified set of conditions is given by Fi=Amount of element released from fodder + inhibitor assuming an equal mass of fodder is employed in each case. A value for Fi which is equal to unity means that the addition of inhibitor has had no net effect upon the amount of element present in the digest solution. A value of Fi which is greater than 1 means that there is less element present after the addition of inhibitor and uice versa for values of Fi of less than unity. The effect of added inhibitor on several elements can therefore be displayed in a single figure.Thus in Fig. 2 Manox 510D acts as an effective inhibitor of Ce solubility under simulated rumen conditions whereas ammonium molybdophosphate has little inhibitory effect. In Fig. 3 Ce solubility is shown to be not greatly affected by either potential in hi bi t or under simulated a bomasum (gastric) conditions. The in vitro model has the benefit of being able to consider both nutrient elements (e.g. zinc) and contaminants simultaneously and in Figs. 2 and 3 results are given for Ca Ce K Pb Sr and Zn. An additional problem occurred in studies using Manox 510D as residues were found to precipitate out from supernates and the indium ICP-MS internal standard became adsorbed onto the inhibitor. The results for Manox must therefore be considered as being approximate and represent an underesti- mate of the inhibitory effects of this material.Future studies using this approach should therefore employ an alternative ICP-MS internal standard. Figs. 2 and 3 show the change in solubility which occurs upon moving from the simulated rumen (pH 6.4) to the abomasum (pH 1.5). From these figures it is clear that ammonium molybdophosphate is likely to be less effective in reducing Ce solubility than Manox 510D but it may reduce Sr solubility more effectively than Manox 510D in the abomasum. Ammonium mofybdophosphate was chosen because of its chemical properties alone and no attempt was made to assess the physiological or toxicological behaviour of this reagent in animals. The object of the exercise was to assess Amount of element released from fodder alone Ammonium molybdophosphate Zn Sr Ce K Ca 0.8 0.75 I 0.7 - Manox 5 IOD 100 Zn Sr Ce Pb K Ca Fig.2 Behaviour of uptake inhibitors in simulated rumen Ammonium molyWophosphate Zn Sr CO K Ca 1.2 T 1 0.8 L- 0.6 0.4 0.2 I Manox 5 1 OD Zn Sr Ce Pb K Ca 0.01 1 Fig. 3 Behaviour of uptake inhibitors in simulated abomasum whether a stable isotope approach could provide useful infor- mation regarding a novel uptake inhibitor and this has been demonstrated. Sodium ammonium ferroferricyanide is a well established uptake inhibitor and is known to be effective both in vivo and in ~ i t r 0 . l ~ It is of note that both ammonium molybdophosphate and Manox bind zinc effectively in the simulated abomasum and for Manox this effect is also found in the simulated rumen.This may have implications when considering the health of animals treated with Prussian Blue following a major nuclear incident. The enzymes employed in this study are not totally soluble and a solid residue (enzyme + undigested fodder) always remains in the digestion vial at the end of the experiment. Some metals can be scavenged out of the test solutions by this solid and this was observed for Cs in all of these examples. Although Cs levels were often below the LODs instrument 628 Journal of Analytical Atomic Spectrometry September 1995 Vol. 10Table 4 LODs (pg 1-I) for the in uitro ruminant digestion model Zn Sr Cs Ce Pb K Ca Rumen 160 14 6.4 3.9 8.9 19000 16000 Abomasum 160 17 1.6 0.3 1.6 15000 10000 Analytical 7.8 1.1 0.4 0.3 3.5 160 1300 Table 5 Levels of elements of interest (pg 1-') present in the in uitro ruminant digestion model ~~~ Zn Sr Cs Ce Pb K Ca Rumen 260 60 1.0 0.4 <3.5 55000 <16000 Abomasum 170 30 0.4 <0.3 <3.5 27000 <10000 response corrected by the internal standard was consistently lower in enzyme digests than in reagent blanks. The current in vitro model is not without limitations.It is a crude approximation of the physical and biochemical processes which occur in vivo and strictly speaking only addresses part of the ruminant digestive system. Digestion in vivo is a dynamic process with continual secretion and reabsorption of enzymes associated trace elements and water. The changes in equilibria which occur cannot be duplicated in a simple in uitro model.Improvements to the current in vitro model could be made by the introduction of dialysis to approximate removal of soluble trace elements from digest solutions. There are also analytical problems with this type of study. Many of the elements of interest are also present in the enzymes used and levels have been found to vary from supplier to supplier. Table 4 shows three sets of LOD. The first two have been calculated as three times the standard deviation of the measured levels (Table 5) in enzyme blanks for the model rumen and abomasum respectively. Instrumental limits of detection [three times the standard deviation in analytical blanks (acid diluent)] are also shown. The variation in detec- tion limit between the two stages of the in uitro model arises because of the additional dilution that occurs when pepsin solutions are added to cellulase digests.The pepsin itself contains trace elements and these contribute to the variation observed in enzyme blank solutions. An additional limitation of the model is that no account has been made of intestinal absorption in these experiments. Whilst this is important in man the absorption sites in ruminants are less clear2' although uptake in the rumen and abomasum may be important for the fission product elements investigated in this study. CONCLUSIONS A stable isotope approach to studying fission product element behaviour in the food chain has been demonstrated using ICP-MS to measure soil-to-plant transfer factors for soft fruit and in an in vitro model of ruminant digestion.The simul- taneous determination of stable forms of the fission product elements Sr Cs and Ce as well as selected nutrient elements were measured using ICP-MS. Although Cs and Ce levels in soft fruit were found to be below detection limits in some samples site specific data can be obtained for several elements simultaneously and the approach could be applied to any type of agricultural produce. This could be extremely useful in filling gaps in empirical databases used with computer models to assess in particular the long term impact of fission product element release on the food chain. The results of in vitro enzymolyses can be used to study both fission product and nutrient elements simultaneously allowing investigators to identify potential animal nutrition problems prior to starting relatively expensive and lengthy in vivo experiments.Such an approach can be used to screen potential inhibitors of fission product solubility prior to field trials. The authors would like to acknowledge the financial support of the Ministry of Agriculture Fisheries and Food who funded this work and staff at the ADAS farm at Boxworth in Cambridgeshire who supplied fodder for use in the in vitro model. REFERENCES 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 1s 19 20 Choppin G. R. and Rydberg J. Nuclear Chemistry. Theory and Applications Pergamon Press Oxford 1980. International Nuclear Safety Advisory Group. Report Safety Series No.75-INSAG- 1. IAEA Vienna 1986. Kubik M. Bern H. Kusmierek E. and Michalczuk L. Fruit Science Reports.1991 XVIII 17-24. Roca V. Napolitano M. Speranza P. R. and Gialanella G. J . Enuiron. Radioact. 1989 9 117. Jones D. I. H. and Hayward M. V. J. Sci. Food Agric. 1973 24 1419. Howard B. J. and Beresford N. A. Health Phys. 1991 61 715. Rapide Nutrient Analysis pH Indicator Instruction Sheet Agribotanic Ltd. 1992. Kabata-Pendias A. and Pendias H. Trace Elements in Soils and Plants. CRC Press Boca Raton FL 1984. CRC Handbook of Chemistry and Physics ed. Weast R. C. 1st Student Edn. CRC Press Boca Raton FL 1988. Monte L. Quagga S. Pompei F. and Fratarcangelli S. J. Enuiron. Radioact. 1990 11 207. Horyna J. and Randa Z. J. Radioanal. Nucl. Chem. Lett. 1988 127 107. Varskog P. Naeumann R. and Steinnes E. J. Enuiron. Radioact. 1994 22 43. Koranda J. J. and Robinson W. L. Enuiron. Health Perspect. 1978 27 165. Bunzl K. in Low-Leuel Measurements of Man-Made Radionuclides in the Enuironment eds. Garcia-Leh M. and Madurga G. World Scientific Publishing 1991 p. 328. Gerzabek M. H. Mohamad S. A. and Muck K. Commun. Soil Sci. Plant Anal. 1992 23 321. Oughton D. H. Salbu B. Riise G. Lien H. lilstby G. and Narren A. Analyst 1992 117 481. Mascanzoni D. in Transfer of Radionuclides in Natural and Semi- natural Enuironments eds. Desmet G. Nassimbeni P. and Belli M. Elsevier Amsterdam 1990 pp. 459-467. Bunzl K. and Kracke W. Sci. Total Enuiron. 1987. 63 111. Arnaud M. J. Clement C. Getaz F. Tannhauser F. Schoenegge R. Blum J. and Giese W.J. Dairy Res. 1988 55 1. Khoransi G. R. and Armstrong D. G. Livest. Prod. Sci. 1992 31 271. Paper 5/01 389E Received March 7 1995 Accepted May 24 1995 Journal of Analytical Atomic Spectrometry September 1995 Vol. 10 629

ISSN:0267-9477    

DOI:10.1039/JA9951000625

出版商:RSC    

年代:1995

数据来源: RSC

摘要:

Basic Investigations of Nanosecond Laser-induced Plasma Emission Kinetics for Quantitative Elemental Microanalysis of High Alloys a c. - BELA NEMET AND LASZLO KOZMA Department of Physics Janus Pannonius University 7624 Pies Hungary n 431 .I 1 The new application of time-resolved laser-induced breakdown spectrometry (TRELIBS) to the direct undamaged qualitative and quantitative analysis of tertiary high-alloys (gold jewels) has been studied. A plasma was formed by using a low energy (10 mJ 10 ns) Q-switched Nd YAG laser in atmospheric air. The plasma emission was measured by a 0.6 m polychromator and a computer controlled gated optical multichannel detector system. The TRELIBS system greatly improved the line-to- background ratios. This investigation was carried out over a wide spectral range in the ultraviolet and visible region for the analytical line pairs of copper-gold silver-gold and copper- silver.Keywords Time-resolved laser-induced breakdown spectrometry; laser ablation; time evolution of microplasma emission; high alloys; gold jewels 2200 421.09 421.27 A d I/ 0 420.0 422.5 425.0 427.5 430.0 432.5 Journal of Analytical Atomic Spectrometry arrav I Controller Polychro- mator Nd:YAG laser ' / Fig. 1 Block diagram of the TRELIB spectrometer 2000 1 1 517.59 517.95 0 512 514 516 518 520 522 Wavelengttdnm Fig. 2 Time-resolved laser-induced plasma emission of Ag I around the transitions of 421.09 431.11 and 520.91 nm 7,=50 and 100 ns; and 7d =40 80 160,320 500 lo00 and 1700 ns Journal of Analytical Atomic Spectrometry September 1995 Vol.10 63112000 4 200 I 8000 1 I 300 442.5 445.0 447.5 450.0 452.5 455.0 457.5 460.0 462.5 465.0 467.5 470.0 Wavelengt h/nm Fig. 3 Time-resolved laser-induced plasma emission of Ag I around the transitions of 447.61 and 466.85 nm The laser-produced plasmas (LPP) are transient signals which can be changed by pressure temperature volume and composi- tion over a wide range. For binary or tertiary samples the figures of merit of the laser-induced breakdown spectrometry (LIBS) (accuracy precision reproducibility linearity dynamic range and detection limits) for quantitative elemental analysis have not been satisfactory.'" If the content of a binary or tertiary alloy varies over a wide range (e.g. from 10 to 60%) the amount of ablated material per laser shot from this sample also varies by 50-loo% because of the changing thermodynamic parameters of the target.For this reason the correlation between line intensity and concentration of the components is not linear and fre- quently there is no simple correlation In general the amounts of the laser ablated material vary considerably. In the case of atomic emission spectrometry of a laser-produced plasma the solution to this problem is the use of internal standardization by choosing moderately intense line-pairs of analytes with approximately the same excitation energy.'.' The basic requirement for internal standardization is the complete atomization of the ablated material in the micro- which can be achieved later on in the plasma lifetime i.e. by time-resolved laser-induced breakdown spec- trometry (TRELIBS).The time-integrated laser-induced plasma spectrum has a very high background the application of TRELIBS has greatly improved the line-to-background ratios (L/B) and the precision compared with time-integrated In this paper the temporal evolution of the Nd YAG laser- produced microplasmas formed in air at atmospheric pressure from copper silver and gold as well as tertiary high-alloy ~ 1 ~ ~ . 4 . 5 . 7 . 1 0 (Au-Cu-Ag) targets are studied by time-resolved multichannel atomic emission spectrometry. The optimum time delay (zd) and time gate (zg) for the intensity measurement for every analyte transition line was found for optima line-to- background ratios. A number of analytical line-pairs of cop- per-gold silver-gold and copper-silver were chosen (in many cases their upper electron energy levels multiplicities and the change of the inner quantum number are the same) and used to determine the concentration. The aim of the emission kinetic studies was to increase the number of analytically useful lines and to develop the TRELIBS technique in order to achieve good analytical performance for high-alloys.EXPERIMENTAL Instrumentation A Nd YAG laser was used and operated in a self Q-switched mode.6 (Fig. 1). The laser output had a pulse duration of 10 ns and the pulse energy at the surface of the target was 10 mJ. The laser beam was focused by an objective lens (f= 200 mm) and directed by a prism at an angle of 25" onto the solid sample surface. The diameter of the focal point was 0.4mm so the power density was 1 GW cm-2.The radiation of the laser-produced plasma was observed perpendicular to the surface. The imaging ratio of the plasma was 2 1 which was collimated by a quartz condenser lens (f= 120 mm) onto an entrance slit of a Czerny-Turner polych- romator (k two 0.3 m monochromators LOMO-MDR6 Russia) having gratings of 1200 grooves mm-' and a reciprocal 632 Journal of Analytical Atomic Spectrometry September 1995 Vol. 1012000 * y 5 0 n s N" 444.70 rg=50 ns 463.05 A 0 sg=10O ns F 448.83 4 443.73 A d 442.5 445.0 447.0 450.0 452.5 455.0 N" I1 400 500 01. I " . ' . . . I . ~ . . I ' " . I . . . . ' . 457.5 460.0 462.5 465.0 467.5 470.0 Wavelengt h/nm Fig. 4 Time-resolved laser-induced plasma emission of Au I around the transitions of 443.73 448.83 and 460.73 nm linear dispersion of 1.3 nm mm-'.The spectra were detected by a photodiode array (PDA) detector of 5 12 channels (element size 25 pm x 2.5 mm) with a microchannel plate (MCP) image intensifier (Princeton Instruments USA). The combination of a polychromator with this detector system resulted in a practi- cal resolution of typically 0.12 nm (when an entrance slit-width of 30 pm was used) and a spectral window x 15 nm. The grating-drive mechanism rotated the optical grate until the required wavelength window covered the detector elements. Various spectral regions could be observed setting the grating in the range 260-640 nm. Gating the intensified MCP enables the spectra to be temporally resolved (40-1700 ns). Our detector was time-gated with a gate-width range of 50-2500 ns (using an FG-100 pulse generator).Spectral data acquisition and processing were carried out with an optical processing system (type S1-180). A detailed description of the detector system is given in a previous paper.6 Samples and Calibration The investigated samples were high-purity copper silver gold metal and gold jewel etalons from the Hungarian Assay Office. The compounds of these etalons are shown in Table 1. The Cu I Ag I and Au I line selection covered a wide spectral range in the UV and a number of analytical lines (295-525 nm) in the visible spectrum but their upper levels did not cover a large range of excitation energies (45000-69000 cm-I 4.5-7.0 eV) from the most important transiti~ns.''-'~ (The selected lines are described in italics in Table 1 Investigated gold-silver-copper alloys gold jewel etalons [Au] (Yo) [Ag] (%) [CU] (%) [CU]:[AU] [Ag]:[Au] [Cu]:[Ag] 58.5 0 41.5 0.7 1 0 0 58.5 10.0 31.5 0.54 0.17 3.15 58.5 20.5 21.0 0.36 0.35 1.03 58.5 31.0 10.5 0.18 0.53 0.34 58.5 41.5 0 0 0.71 0 Table 2.) The levels belong to excited electron energy states of several multiplicities (singlet doublet quartet) and inner quan- tum numbers (s1,2 p3/2 PI/ D5/2 D3/2 F7p FsI p5/2,3/2,1/2 Line-pairs of copper-gold silver-gold and copper-silver were applied for the determination of the ratio and the concentration of the components of tertiary high-alloys (Au-Cu-Ag) (internal standardization). The basic requirements in our choice of line-pairs were not only the approximately same excitation energy but the same multiplicity.(Table 3). For the Cu-Au and the Cu-Ag line-pairs these requirements could be obtained while for Ag-Au line-pairs they could not. No special sample preparation was necessary for the measure- ments. Before making the spectral measurement 5-6 laser shots were used for cleaning the surface of the metals. Typically ten shots were accumulated and averaged with laser pulses delivered to the sample at a rate of 1 Hz. The mass per laser shot ablated from the metal surface was 30-40ng the depth and the diameter of the crater was 1 Irn x 300 pm. The dark current of the detector was measured separately during the same exposure time and was automatically sub- D7/2.5/2,3/2,1/29 F9/2,7/2,5/2,3/2). Journal of Analytical Atomic Spectrometry September 1995 Vol.10 633Table 2 Analyte transition lines (nm); selected lines are in italics ,ool17 - c u ' 324.75 327.40 453.08 448.04 296.1 2 327.98 359.40 345.79 4p'-4s' 6 ~ l - 4 ~ ' 4s4p-42 - 4p1-42 510.55 570.02 578.21 515.32 521.82 522.01 402.27 406.27 406.33 427.51 437.82 441.56 467.48 458.70 470.47 465.1 1 4dl-4~' 5dl-4~' 4s4p-4s4p 300 Ag' 5p'-5s' 7s1-5p' 5s5p-5sz 328.07 338.29 466.85 447.60 282.44 354.26 353.83 - - 5dl-5~' 520.91 546.55 547.16 6dl-5~' 405.53 421.09 421.27 431 .I 0 5s5p-5s5p - - - - - 487.42 100 Au' 6pl-6~' 242.80 267.60 424.18 365.18 6s6p-6s' 268.87 302.92 264.15 389.79 6p'-6s' 3 12.28 506.46 627.82 6d'-6p1 406.51 479.26 481.16 7dl-6~' 332.15 379.59 380.30 6 ~ 6 p - 6 ~ 6 ~ 389.79 460.73 443.73 431.51 448.83 523.03 8 ~ l - 6 ~ ' - - D7P x 460.73 nm D-F 81443.73 nm +431.51 run 6448.83 nrn A d - - 500 I I 1 Table 3 Analyte line-pairs Copper-gold line-pairs 296.12(44963 cm- ')2F7/2-2D5/2 302.92(42163 ~ m - ' ) 4 P ~ / ~ - 2 D / ~ 441.56( 63585 ~ m - ' ) 4 D ~ / ~ - 4 P ~ ~ 443.73( 68705 cm-')4D5/2-4F5/2 Silver-gold line-pairs 421.09( 5421 3 cm- ')2D5/2-2P3/2 431.51(68705 cm-')4D512-4F7,2 466.85 ( 5 1887 cm - ' ) 2S1/2-2P3/2 460.73 ( 68 70 5 cm - ' ) 4D ,/,-4F 512 Copper-silver line-pairs 427.31(62403 cm-')4D,/,-4P5/ 421.09( 5421 3 cm - ')2D5/,-2P3/ 453.08( 52849 CIII-')~S,/~-~P~~~ 447.61 (5 1887 cm - ')2S1/2-2P1,2 427.31 (62403 cm-')4D7/2-4P5/2 431.51 (68705 cm-1)4D5/2-4F7/2 458.70( 62948 ~ r n - ' ) 4 D ~ / ~ - 4 F ~ / ~ 460.73(68705 Cm-1)4D5/2-4P3/2 447.61( 51887 ~m-')2S,/~-2P, 443.73(68705 cm-1)4D5/2-4F5/2 520.91 (48744 cm-')2D3~,-2P,~ 523.03(67811 cm- ')4D7/2-4F9/2 448.04( 52849 cm-')2S,~,-2P1~ 447.61(51887 ~m-')2S,/~-2P,~ 5 15.32( 49935 ~m-')2D~/~-2P,~ 520.91 (48744 cm-')2D3~,-2P1~ tracted from the previous emission spectrum to yield a net signal from the plasma.The blank spectra for line intensities determination were measured from plasmas of other metals which had no lines at the given range. Prior to each measure- ment the wavelength axis of the monochromator-detector system was calibrated with the use of lines from a mercury emission lamp and an iron plasma. The determination of the concentration from the inten- sity ratios of the line-pairs were made using the following calculations u = [CU] [Au] b = [Ag] [Au] c = [CU] [Ag] and [CU] + [Ag] + [Au] = 100% therefore ~[Au] + ~ [ A u ] + [Au] = 100% S-P 0448.04 nm 0453.08 nm D-PA515.82 tlnl cu' I x427.51 tun I S-PO447.60 run Ag' 0466.85 nm Dp v421.09 nm 5 1uur LL 1 ox c I I(- I I I I Fig.5 Analyte line intensities of copper silver and gold as the function of time for Cu I 515.82 nm Ag-I 520.91 nm andAu I 448.94 the right hand scales are valid and [Au] =- x100% a + b + l RESULTS AND DISCUSSION Time-resolved Spectra Firstly we measured the emission spectra of the Q-switched Nd YAG laser-induced microplasmas in air at atmospheric pressure at different delay and gate times.These time-resolved plasma spectra can be seen in Figs. 2-4 where the target metals were silver and gold. Using the copper target results from TRELIBS measurements have been published in a previous paper.6 The delay times zd were 40 80 160 320 500 1000 and 1700 ns and the gate times zg were 50 and 100 ns.These spectra represent the most typical effects that can be observed during the lifetime of the plasma. In general the spectra corresponding to zd < 200 ns indicate a very high intensity of the continuum emission and only a few neutral atomic lines can be observed (Ag'; 431.11 and 520.91 nm; Au' 448.83 nm). Lines or bands from air compo- nents (nitrogen oxygen) were also observed under the operating conditions because the discharge was formed in air (0 I1 431.72 and 441.70nm; N I1 517.59 517.95 444.70 and 463.05 nm). However above 200 ns the continuum emission and these bands were diminished. In the absence of a target laser-induced breakdown of air was not observed.Although there was a fast expansion of the plasma in the presence of the target the air diffused into the discharge. Several line broadenings can be observed at the different 634 Journal of Analytical Atomic Spectrometry September 1995 Vol. 10448.83 I j rd=1000ns . - cu' ~d=1000ns rg=2500ns 427.51 I cu' Ad Au'l 431 5 1 421.09 I 1 444 446 448 450 452 454 c. *2 12000 Q) L rd=looo ns 465.1 1 - - rg=2500ns I cu' Au' 460.73 Fig.6 Time-integrated spectra of Cu I (450.75 450.94 453.08 and 453.94 nm) Ag I (447.61 nm) and Au I (443.73 and 448.83 nm); and time-integrated spectra of Cu I (458.70 465.11 467.48 469.75 and 470.47 nm) Ag I (455.85 nm) and Au I (460.75 nm). transitions. A similar line broadening effect (minimum homo- geneous broadening and symmetric line profile) can be seen e.g.at four copper lines (282.43 510.55 570.02 and 578.21 nm) and at four gold lines (302.92 312.28 506.46 and 627.82 nm) which are considered to be due to having the same lower energy level and similar transition probabilities (P-D transitions). A significant stronger homogeneous line broadening can be seen at the copper (515.32 and 521.82nm) silver (520.91 and 546.55 nm) (Fig. 2) and gold (406.51 and 479.26 nm) lines which belong to the same transitions (D-P transitions). An asymmetric broad line profile can be observed at the following lines copper (402.27 and 406.27 nm) silver (405.53 and 421.09 nm) (Fig. 2) and gold (332.15 and 379.59 nm) (also D-P transitions but from higher upper levels). The inhomogeneous broadening effect can be seen at the S-P transitions (Cu I 448.04 and 453.08 nm; Ag I 447.61 and 466.85 nm; Au I 365.18 and 424.18 nm) (Fig.3) too. These homogeneous and inhomogeneous line broadenings can be explained by Stark and pressure effects. The ionic form of copper silver and gold (as the ionic Cu 11 Ag I1 and Au I1 emissions show) are present up to 1000 ns and not only during the laser interaction (Cu I1 276.94,283.76 287.67 404.35 and 455.59 nm; Ag I1 276.74 nm; Au I1 280.55 380.40 401.61 and 405.28 nm).6 One can therefore assume that the Stark effect would be significant for longer periods during several hundred nanoseconds or one microsecond. In addition the plasma could be at a relatively high pressure because its expansion was limited by the surrounding air v) 3 c 430 521.83 522.00 I Au' 523.03 512 514 516 518 520 522 Wavelengthhm Fig.7 Time-integrated spectra of Cu I (427.51 nm) Ag I (421.09 and 431.11 nm) and Au I(431.51 nm) and time-integrated spectra of Cu I (510.55 515.32 and 520.91 nm) Ag I(521.83 and 522.00nm) and Au I (523.03 nm) atmosphere. Therefore pressure effects would also play a significant role in the line broadening. Line intensities versus time are shown in Fig. 5. The copper and the silver atoms have transitions between singlet and doublet levels where the intensities don't decrease up to 1.5 ps and moreover permanently increase (Cu I 515.82nm; Ag I 520.91 nm; Cu I 448.04 and 453.08 nm; Ag I 421.09 nm). But there are transitions between quartet and doublet levels where the intensities have maxima about 0.5 ps (Cu I 427.51 and 458.70 nm; Ag I 431.51 and 448.83 nm).This seems to be the result of a dependence of the emission intensity evolution on the multiplicities (singlet doublet quartet) and inner quantum numbers. In Figs. 6 and 7 the 'time integrated spectra' can be seen when in every case the delay time was 1 . 0 p and the gate time was 2.5 ps. The measured ratios of the line intensities are compared with the literature data whether spark discharge or arc discharge was Quantitative Analysis High-alloys The RSD of the background was <4% and the RSD of the line-pairs was <7-8%. The intensity ratios of analytical line-pairs of copper-gold silver-gold and copper-silver (Icu IAu IAg IAu Icu IAg) as a function of the ratios of several concentrations ([Cu] [Au] [Ag] [Au] [Cu] [Ag]) are shown in Fig.8. With most of the line-pairs a linear correlation Journal of Analytical Atomic Spectrometry September 1995 Vol. 10 6352 c 296.1 2 427.31 441.56 458.70 302.92431.51443.73460.73 a 1 I 0 /u 0 - l t I7 "0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 - 0 0.8 0.4 . n f o x + 421.09 447.61 466.85 52/ 431.51 443.73 460.73 523.03 I I I 1 I I l o "0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 [Cu]:[Au] Fig. 8 Intensity ratios of four-four line pairs of silver gold cop- per gold and copper silver as the function of ratio of the component concentration in the jewel etalons could be obtained when the concentration of copper and silver changed between 0-45%. After obtaining the calibration curves the accuracy of the method was tested by measurement of gold jewels.The Icu IAu and ICu:IAg ratios at four line-pairs for two samples were measured and the concentration ratios calculated. From these data we calculated the amount of gold and after that the silver and copper amounts. The deviation between our measurement and the conventional value was <3%. Low-alloys Using aluminium samples of low copper concentration (up to 1.1 YO) the detection limit was determined. The calibration curve was linear when we used the line at 521.83nm while the line at 324.75 nm showed some curvature since this transition is resonant with the ground state and therefore intense. (The self-absorption is not negligible even at a copper concentration of 0.5y0.)~ The detection limit (cL) was calculated according to the conventional definition (which applies when the calibration curves are linear) ZSB CL =- S where z is a constant sB is the standard deviation of the background and s is the slope of the calibration curve.If z= 2 the detection limits calculated are 15 ppm for the 324.75 nm line and 160 ppm for the 521.83 nm. These results are similar to those previously p~blished.~.~ In conclusion TRELIBS has given new possibilities for the quick direct undamaged qualitative and quantitative multiele- ment analysis of tertiary high-alloys (gold jewels). It can be applied for quantitative analysis of several other tertiary and binary high-alloys ( Fe-Cr-Ni Cu-Zn-Ni Cu-Zn Cu-Sn and Cu-Pb). The authors thank the Hungarian Assay Office for gold jewel etalons. REFERENCES 1 2 3 4 5 6 7 8 9 10 11 12 13 Radziemski L.J. and Cremers D. A. in Laser-induced Plasmas and Applications eds. Radziemski L. J. and Cremers D. A. Marcel Dekker New York 1989 ch. 7. pp. 295-334. Moenke-Blankenburg L. in Laser Microanalysis eds. Moenke- Blankenburg L. Winefordner J. D. and Kolthoff I. M. Wiley New York 1989. Marshall J. Carroll J. Crighton J. S. and Barnard C. L. R. J. Anal. At. Spectrom. 1994 9 319R. Grant K. J. Paul G. L. and O'Neill J. A. Appl. Spectrosc. 1990 44 1711. Leis F. Sdorra W. KO J. B. and Niemax K. Mikrochim. Acta [Wien] 1989,II 185. NCmet B. and Kozma L. Spectrochim. Acta Part B 1995 50 in the press. Quentmeier A. Sdorra W. and Niemax K. Spectrochim. Acta Part B 1990 45 537. Kuzuya M. Matsumoto H. Takechi H. and Mikami O. Appl. Spectrosc. 1993 47 1659. Thiem T. L. Salter R. H. Gardner J. A. Lee Y. I. and Sneddon J. Appl. Spectrosc. 1994 48 58. Lupkovics G. NCmet B. and Kozma L. in Current Trends and Problems in Optics ed. Dainty J. C. Academic Press London Harrison G. R. Wavelength Tables Wiley New York 1980. Wiese W. L. and Martin G. A. NSRDS-NBS 68 Part I I . Transition Probabilities National Bureau of Standards Washington DC 1980. Moore C. E. Atomic Energy Leuels Circular of NBS 467 vol. II. 1952 and vol. 111. 1958 US Government Printing Office Washington DC. 1994. V O ~ . 2 ch. 5 pp. 69-81. Paper 5/01 235 J Received Accepted June 1 1995 636 Journal of Analytical Atomic Spectrometry September 1995 Vol. 10

ISSN:0267-9477    

DOI:10.1039/JA9951000631

出版商:RSC    

年代:1995

数据来源: RSC

摘要:

Cluster Formation Processes in Laser and Spark Plasmas of Rare Earth Oxide-G rap hi te Mixtures" JOHANNA SABINE BECKER AND HANS-JOACHIM DIETZE Zentralabteilung fur Chemische Analysen Forschungszentrum Jiilich GmbH D-52425 Jiilich Germany The abundance distribution of cluster ions (Cn+ MC,' and MO * ; M = rare earth element) of a rare earth oxidegraphite mixture in laser ionization mass spectrometry (LIMS) and spark source mass spectrometry (SSMS) was investigated. In LIMS and SSMS a typical alternating abundance distribution for positively charged carbon and carbide cluster ions of rare earth elements was found. Clusters with an odd number of atoms were observed with a higher intensity in comparison with clusters with an even number of atoms. A maximum cluster formation rate of carbon clusters (C,,') and metal carbide clusters (MC,' ) using a rare earth oxidegraphite mixture target was observed at a laser power density of about lo8 W cm-'.The intensities of positively singly charged monoxide and dicarbide ions of rare earth elements correlate with the dissociation energy of the molecules which depends directly on the transition energy from the 4f"-'5d6s2 electronic state to the 4f "6s2 electronic state for rare earth elements. Additionally measurements using high resolution inductively coupled plasma mass spectrometry (ICP-MS) have shown that the typical distribution of oxide ions of rare earth elements with increasing atomic number is comparable to that found in LIMS and SSMS. Keywords Cluster formation; laser plasma; rare earth oxide-graphite mixture; spark plasma The existence of polyatomic and cluster ions in plasma mass spectrometry'-6 is well known and the clusters mostly play a secondary role in the analysis of matrix elements because of their minor intensities. Knowledge of the abundance distri- bution and intensities of cluster -ions is of particular interest for the mass spectrometric trace analysis of inorganic mate- r i a l ~ .~ ~ ~ The cluster ions formed in spark laser or inductively coupled plasmas disturb the atomic ions of the analyte by isobaric interferences. These interferences induce an increase in the detection limits of chemical elements. The determination of rare earth elements (REE) in geological or technical samples is often difficult owing to their very similar physical and chemical properties.Inductively coupled plasma mass spectrometry (ICP-MS) requires a chemical dis- solution of the solid sample and the trace analysis of REE is limited by many interferences of oxide hydroxide and chloride polyatomic ions with analyte ions.' Therefore solid-state mass spectrometric methods are suit- able for the direct analysis of REE in non-conducting materials. The oldest mass spectrometric technique for the sensitive determination of trace amounts of impurities is spark source mass spectrometry (SSMS). For SSMS measurements the insulating material can be mixed with high-purity graphite or silver to prepare conducting electrodes. In general high-purity graphite is mixed with insulating powder samples because in *Presented at the 1995 European Winter Conference on Plasma Spectrochemistry Cambridge UK January 8-13 1995.I Journal of I Analytical I Atomic Spectrometry this way more homogeneous electrodes are prepared in com- parison with silver as the conducting powder. The determination of REE in the ppm or ppb concentration range by SSMS8 is made difficult in samples with relatively high concentrations of Ba and light lanthanides by isobaric interferences of oxides and carbide ions with atomic ions (Fig. 1). In spite of many observed interferences in the mass range of REE,1° the determination of trace concentrations is possible using a high resolution solid-state mass spectrometer. A mass resolution of up to loo00 requires the separation of cluster ions from the atomic ions. In order to overcome the interference problems in mass spectrometric trace analysis the following basic approaches are possible (a) the analysis of interference-free isotopes of the analyte; (b) analysis under experimental conditions where cluster formation is minimal; (c) the correction of interferences using measured cluster intensities; or (d) the use of a mass spectrometer with high mass resolution (e.g.double-focusing sector field instruments). In this work the formation of polyatomic and cluster ions of a rare earth oxide-graphite mixture in a spark plasma was investigated and compared with the formation of cluster ions in a laser plasma. Such investigations can be helpful for understanding cluster formation processes in plasmas. Further a systematic study of the types of clusters formed in laser and spark plasmas is useful for estimating the mass spectral inter- ferences of cluster and atomic ions at the same mass number.EXPERIMENTAL For comparative measurements of the abundance distribution of cluster ions in spark and laser plasmas a double-focusing mass spectrometer with Mattauch-Herzog geometry (Fig. 2) was used. In this high resolution instrument the laser and spark sources are interchangeable. The ions formed in the laser or spark ion source were separated in the Mattauch-Herzog instrument. Positively charged ions are extracted and acceler- ated into this double-focusing mass spectrometer. Ilford 4 2 ion-sensitive photoplates were used for the ion detection. The laser ion source is described in ref. 11. The laser beam is focused on the sample surface in the reflection mode.For the evaporation of solid samples and ionization of evaporated species in laser ionization mass spectrometry (LIMS) a Q-switched Nd YAG laser (J. K. Lasers) was used at a wavelength of 1064 nm and a pulse width of 15 ns. With an experimental mass resolution of approximately m/Am z 12 000 all line interferences in the mass range of the REE could be resolved. The experimental parameters of LIMS and SSMS are summarized in Table 1. A homogeneous mixture of lanthanide oxides (Specpure; Johnson Matthey) and high-purity graphite (RWS Ringsdorff) (1 1) was pressed into electrodes for SSMS (size 2 x 2 x 20 mm3) or targets for LIMS (size 5 x 5 x 2 mm3). The measurements of oxide intensities of lanthanides were carried out using the high resolution inductively coupled plasma mass spectrometer 'ELEMENT' from Finnigan-MAT Journal of Analytical Atomic Spectrometry September 1995 Vol.10 637135 145 150 155 1160 165 1170 175 180 mlr U Ha Tm Lu Tb E" u La - =d - I u Er (HO Ce DY Yb Fig. 1 Scheme of interferences for REE trace analysis of a rare earth oxide-graphite mixture (with BaO) by SSMS laser plasma Fig. 2 Double-focusing mass spectrometer with laser or spark ion source Table 1 Experimental parameters of LIMS and SSMS Laser ion source laser wavelength pulse width repetition frequency Spark ion source spark voltage pulse width repetition frequency acceleration voltage Ion separation main slit electrostatic analysator magnetic analysator registrable mass range ion-sensitive photoplate range of exposition mass resolution (m/Am) ion source analyser Analytical results.detection limit reproducibility lon detection Vacuum conditions Nd YAG 1064 nm 15 ns 100 Hz 60 kV lo00 Hz 18 kV 100 p 10 Cun 31,8" sector field 90" sector field 1 36 e.g. 8-288 u Ilford 4 2 photoplate 10-14-3 x C ~ 1 2 0 0 0 mbar mbar 1 ppm-10 ppb 10-20% at a mass resolution of 7500. For the ICP-MS studies a mixed standard solution of REE in nitric acid [from the National Institute of Standards and Technology (NIST)] was used. The concentration of each REE in the mixed standard solution was 10 pg1-l. The reproducibility of oxide ion intensities measured by ICP-MS (five determinations) was about 5%. RESULTS AND DISCUSSION A multiplicity of cluster ions were observed in the laser and spark mass spectra for the rare earth oxide-graphite mixtures investigated e.g.carbon cluster ions C,' up to the end of the mass range in our experiments-with up to 25 atoms in SSMS and smaller clusters in LIMS (as a function of laser power density and exposure of photoplate); oxide ions of REE with one or two oxygen atoms; and carbide ions MC,' with n= 1-8. Distribution of Carbon Cluster Ions The abundance distribution of carbon cluster ions in laser and spark plasmas using graphite is well known. The first meas- urements using SSMS were published by Franzen and Hintenberger12 in 1961. In Fig. 3 our mass spectrometric results for the formation of carbon cluster ions in spark and laser plasmas at different power densities are compared. In general in spark and laser mass spectra a typical alternating abundance 638 Journal of Analytical Atomic Spectrometry September 1995 Vol.10103 102 10' 100 fl 10'' ,o 3 1 0 2 10-3 1 0 4 lo-' 10-6 1 5 10 15 20 25 n Fig. 3 Abundance distribution of Cn* cluster ions in laser and spark plasmas where A Ic-+:Ic+ at @x2 x lo8 W cm-2 (LIMS); B Icm- I,- at @x2 x lo8 W cm-2 (LIMS); C Icn+/Ic+ at @x5 x lo9 W cm-* (LIMS); and D Icn+:Ic+ (SSMS) distribution was measured with double periodicity and local maxima at an odd number of atoms for cluster ions with less than nine atoms and 4-fold periodicity for cluster ions of more than nine atoms. In our experiments the atomic carbon ions with the highest intensity were observed at a laser power density of 5 x lo9 W cm-'. This is in agreement with the results of SSMS at a power density of 109-1010 W cm-'.For positively charged cluster ions at a laser power density of 2 x 10' W cm-' Cll+ cluster ions have a marked maximum (Ic,+:Ic+ in Fig. 3 curve A). In comparison with positively charged cluster ions negatively charged carbon cluster ions appear with local maxima for even clusters. The distribution of negatively charged cluster ions in LIMS (Icn-:Ic- in Fig. 3 curve B) is in agreement with measurements by Fiirstenau and Hillenkarn~.'~ From theoretical investigations by Seifert et al.,14 it can be concluded that the higher intensity of cluster ions correlates with a higher electronic stability. Our experimental results for the formation of carbon clusters as a function of laser power density are discussed in this theoretical work.14 At low laser power density small clusters with a maximum intensity for trimers were observed.This result is in agreement with investigations by ordinary thermal evaporation or laser evapor- ation at 4000K.15 With increasing laser power density the formation of larger clusters of C5 C and Cll in agreement with other workers (e.g. Rohlfing et a l l 6 ) was observed. At a laser power density of 10' W cm-2 a maximum of cluster formation with a maximum intensity of Cll+ cluster ions was measured. A further increase in laser power density causes a decrease in cluster formation and cluster size owing to dis- sociation processes. Experimental conditions under which the cluster formation rate is minimal are of interest for inorganic trace analysis by LIMS.At high laser power density @> 1O'O W ern-' cluster formation is negligible because of the high dissociation rate or low formation rate of possible clusters. Distribution of Carbide Cluster Ions and Oxide Ions of Lanthanides The abundance distribution of carbide clusters in laser and spark plasmas is very similar for all the lanthanides. Local maxima at an odd number of atoms and a general decrease in intensities were measured e.g. as demonstrated for carbide cluster ions of Tm (TmC,+:Tm+) in Fig. 4. A similar distri- bution of carbide ions (MC,+:M+) was measured in laser and spark mass spectra for all other lanthanides and also B W Si Th and U.17918 A comparison of this measured abundance distribution with the well-known carbon cluster ion formation shows a similar abundance distribution with typical alternation and local maxima at an odd number of atoms.Possible mechanisms of formation processes of carbide cluster ionslg can be explained by different association or substitution reactions of a metallic ion and a neutral carbon cluster (e.g. M++C,-+MC,+ or M++C,-+MC,-l++C) in a spark or laser plasma. The cluster formation of carbide and oxide ions of a rare earth oxide-graphite mixture was investigated as a function of laser power density (Fig. 5 ) . Two different curves for oxide ions were observed. Oxide ions were formed at low laser power density (see the left-hand side of Fig. 5 oxide-1) by direct evaporation of initial oxides thermal dissociation to the monoxides and post-ionization for example by electron bom- bardment [e.g.M,O,(s)+e- -+2MO+ +O(g)+3e-]. At a laser power density higher than lo7 W cm-2 plasma chemical reactions [such as association substitution or ion molecule 1 2 3 4 5 6 7 8 fl Fig. 4 Distribution of carbide cluster ions of Tm (ITmc;:ITm+) as a function of cluster size (n) where A LIMS and B SSMS 1 degree of dissociation 4 I oxide-I ~ - \ degree of ionization ' 1 I I I I I 1 06 10' 1 08 1 o9 10'0 +W cm4 Fig.5 Scheme of oxide and carbide formation of rare earth oxide-graphite mixture as a function of laser power density. I lo6 W cm-2<@< lo7 W cm-2 (direct evaporation ofclusters; partial dissociation of clusters; ionization of clusters by electron bombard- ment); 11 @> lo' W cm-l (cluster formation by plasma chemical reactions association substitution and ion molecule reactions or by expansion of plasma in the vacuum; plasma ionization) Journal of Analytical Atomic Spectrometry September 1995 Vol.10 639reactions; e.g. M' + O(g)-MO'] dominate (see Fig. 5 oxide-2). A further increase in laser power density causes an increase in the degree of dissociation of oxide ions. A maximum cluster ion formation rate was observed for carbide ions (MC' MC2+) at 10' W cm-' using a rare earth oxide-graphite mixture target. The relative intensities of oxide ions of REE (MO+:M+) in spark and laser plasmas for REE with increasing atomic number are given in Fig. 6. A similar abundance distribution to that measured for oxide ions was found for carbide and dicarbide' ions of REE. The highest ion intensities were measured for La Gd and Lu; the minimum oxide intensities were observed for Eu and Yb. The oxide ion intensities for the REE vary over about 3 orders of magnitude.Variations of about 14 orders of magni- tude with a similar dependence of oxide ion intensities on atomic number were measured by Nguyen and de Saint- Simon" by thermal surface ionization mass spectrometry at about 2000-2600 K. The very large difference in the variation of oxide ion intensities for REE in the thermal surface ioniz- ation mass spectrometric method in comparison with LIMS or SSMS could be explained by the dominating thermal effects in the surface ionization mechanism at low temperature. The distribution of oxide ion intensities for REE in an inductively coupled plasma (Fig.7) is similar to the abundance distribution in LIMS and SSMS. The results for oxide ion formation of REE in an inductively coupled plasma using high resolution ICP-MS for the separation and detection of ions are in agreement with measurements made by Longerich et al." using a low resolution quadrupole ICP-QMS instrument. Maximum ion intensities of oxides of La Gd and Lu correlate with the filling of 4f orbitals; i.e. an empty 4f orbital for La and half-filled and filled 4f orbitals for Gd and Lu respectively. The electron configuration for the transition from the solid to the gaseous state is the same for these elements. The transition energies of a 5d electron from a 4f"-'5d6sz configuration to the 4f "6s' electronic configuration from the solid to the gaseous state are shown for REE in Fig.7 curve C. The oxide ion intensities observed in laser spark and induc- tively coupled plasmas also correlate with the bond dissociation energies of oxides [Dz( MO)] determined by Knudsen effusion mass spectrometry with electron bombardment at 2000 K;" these are given in Fig. 7. By analogy the cluster intensities of dicarbide ions of REE La Pr (Prn) Eu Tb Ho Trn Lu Ce Nd Sm Gd Dy Er Yb Rare earth element Fig. 6 Relative intensities of oxide ions (IM,,+:IM+) of REE in laser and spark plasmas as a function of atomic number where A LIMS and B SSMS 1 € 3 1 I - :\'"" -3 La Pr (Pm) Eu Tb Ho Tm Lu Ce Nd Sm Gd Dy Er Yb Rare earih element Fig. 7 Distribution of oxide ions of REE in an inductively coupled plasma measured by high resolution ICP-MS dissociation energies of lanthanide oxides and properties of REE where A (IMo+:IM+) high resolution ICP-MS; B bond dissociation energies (0;) of lanthanide oxides (MO); and C transition energies for the transition of the 5d electron from the 4f"-' 5d6s2 configuration to the 4f shell of the 4f"6s2 configuration./ I I I I I I I I I I I I I I in laser and spark mass spectra correlate with the bond dissociation energies of dicarbides (Fig. 8). These results are in agreement with measurements by Michels and Gijbel~.~ These workers found a correlation for oxide ions of chemical elements in laser-induced mass spectra with bond dissociation energies by laser microprobe mass analysis (LAMMA 500). By using the following empirical approximation (from Nguyen and de Saint-Simon") an estimation of plasma temperatures is possible log [I,,+:I,+] =0.4343[Dz(MO):kT+aZ+b] where Dz(M0) is the bond dissociation energy of the lanthanide oxide k is the Boltzmann constant T is the absolute tempera- ture and Z is the atomic number; a and b are empirical constants and are determined from the straight line through log,,[I,,+:I,+] of La Gd and Lu for which the energy to displace the 5d electron from the 4f "-'5d6sZ configuration in the solid form to the 4f shell of the 4f"6s2 configuration of the gaseous form is zero.Ion temperatures of about 25 000 K in the laser plasma and 30 000 K in the spark plasma were estimated for the dicarbide ions. For the oxide ions temperatures of 58 000 K in the laser plasma and 65 000 K in the spark plasma were estimated.The reason for the lower ion temperatures in the laser plasma in comparison with the spark plasma is the lower laser power density in our experiments uiz. 5 x 10' W cm-'. For experi- ments using SSMS a power density of 109-1010 W cm-2 can be assumed. CONCLUSIONS In LIMS and SSMS a similar abundance distribution was found for positively charged oxide and carbide cluster ions of REE. A maximum cluster formation rate of carbon clusters ((2"') and metal carbide clusters (MC,') using a rare earth 640 Journal of Analytical Atomic Spectrometry September 1995 Vol. 1010 0 10'' + 10.2 0" 3 1 D 3 10" Pro '"% ' 'g /XGd .-L / Lu Smx I I I I I 4 5 6 7 8 D (MCAleV Fig. 8 Correlation of cluster ion intensities (lMc;:lM+) in LIMS and SSMS as a function of dissociation energiesz3 [D (MC,)] where A LIMS and B SSMS oxide-graphite mixture target was observed at a power density of about 10' W cm-2 in LIMS. Clusters with an odd number of atoms were observed with a higher intensity in comparison with clusters with an even number of atoms.Our experiments using LIMS SSMS and ICP-MS in addition to experimental results from the literature obtained by thermal ionization mass spectrometry and Knudsen effusion mass spectrometry with electron bombardment ionization at different temperatures from 2000 to about 60 000 K prove that the abundance distribution of oxide and dicarbide ions is determined by the electronic structure of the lanthanides. The systematic study of the types of cluster ions formed in a laser or spark plasma can be used for estimating mass spectral interferences of cluster and atomic ions of the same nominal mass.Therefore a knowledge of cluster formation and abundance distributions is of considerable importance for mass spectrometric analysis and for the understanding of chemical and physical processes in plasmas. REFERENCES 1 Martin T. P. Angew. Chem. 1986,98 197. 2 Mark T. D. and Castleman A. W. Jr. Adu. At. Mol. Phys. 1985 20 65. 3 Michels E. and Gijbels R. Anal. Chem. 1984 56 1115. 4 Becker J. S. and Dietze H.-J. Int. J. Mass Spectrom. Ion Phys. 1983 51 325. 5 Stamatovic A. S. and Maerk T. D. Rapid Commun. Mass Spectrom. 1991 5 51. 6 Dietze H.-J. and Becker J. S. in Laser Ionization Mass Analysis ed. Vertes A Gijbels R. and Adams F. John Wiley & Sons Inc.Chemical Analysis Series 1993 vol. 124 p. 453. 7 Becker J. S. and Dietze H.-J. Int. J . Mass Spectrom. Ion Process. 1985 67 57. 8 Becker J. S. and Dietze H.-J. Z . Angew. Geol. 1983 29 136. 9 Dulski P. Fresenius J. Anal. Chem. 1994 350 194. 10 Dietze H.-J. and Becker J. S. Int. J. Mass Spectrom. Ion Process. 1984 56 243. 11 Dietze H.-J. and Becker J. S. Fresenius Z . Anal. Chem. 1985 321 492. 12 Franzen J. and Hintenberger H. Z . Narurforsch. Teil A 1961 16 535. 13 Fiirstenau N. and Hillenkamp F. Int. J. Mass Spectrorn. Ion Phys. 1981 31 85. 14 Seifert G. Becker J. S. and Dietze H.-J. Int. J. Mass Spectrom. Ion Process. 1988 84 121. 15 Berkowitz J. and Chupka W. A J. Chem. Phys. 1964 40 2735. 16 Rohlfing E. A Cox D. M. and Kaldor A. J. Chem. Phys. 1984 81 3322. 17 Dietze H.-J. Becker J. S. Opauszky I. Matus L. Nyary I. and Frecska J. Mikrochim. Acta 1983 111 263. 18 Becker J. S. and Dietze H.-J. Int. J. Mass Spectrom. Ion Process. 1988 82 287. 19 Dietze H.-J. and Becker J. S. Zfl-Mitteilungen 1985 101 3. 20 Nguyen L. and de Saint-Simon M. Int. J. Mass Spectrom. Ion Phys. 1972 9 299. 21 Longerich H. P. Fryer B. J. Strong D. F. and Kantipuly C. J. Spectrochim. Acta Part B 1987 42 75. 22 Ames L. L. Walsh P. N. and White D. Phys. Chem. 1967 71 2707. 23 Filby E. E. and Ames L. L. High Temp. Sci. 1972 4 160. Paper 5/01 2921 Received March 2 1995 Accepted May 15 1995 Journal of Analytical Atomic Spectrometry September 1995 Vol 10 641

ISSN:0267-9477    

DOI:10.1039/JA9951000637

出版商:RSC    

年代:1995

数据来源: RSC

摘要:

Quantitative Analysis of Copper Alloys by Laser-produced Plasma Spectrometry* Journal of Analytical Atomic Spectrometry M. SABSABI AND P. CIELO Industrial Materials Institute National Research Council Canada 75 de Mortagne Boucheruille Quebec Canada J4B 6Y4 Laser-produced plasma spectrometry was applied to the determination of low iron nickel and silver concentrations in copper alloys. A focused Nd YAG laser beam was used to generate a plasma from copper targets in air at atmospheric pressure. The plasma temperature was obtained by using a Boltzmann plot of neutral iron lines. Calibration curves for iron nickel and silver were produced. The precision ranged from 2 to 10% of the analyte concentration. The detection limits were 20 10 and 1 pg g-’ for iron nickel and silver respectively.Keywords Laser-produced plasma spectrometry; copper; plasma temperature; atomic emission spectrometry When a high-powered laser beam is focused onto a target the temperature of the locally heated region rises rapidly to the vaporization temperature of the material whereupon small amounts are ablated from the surface in the form of a vapour plume. For fluxes exceeding the value necessary for the break- down the gas becomes strongly ionized even during the initial stage of the laser pulse and the resultant plasma absorbs the laser beam to a considerable extent particularly at long wavelengths. Once atomized the hot plasma radiates with the characteristic spectral emission of the material and this emis- sion spectrum can be used for elemental analysis.The method based on this analysis is known as laser-produced plasma spectrometry (LPPS) (better known in the literature as laser microanalysis or laser-induced breakdown spectrometry). In recent years a few review papers devoted to this technique and its applications have been published.’-’ It has been used for the quantitative analysis of aerosols,* liquid^,'.'^ liquid metals,I3 iron,14-16 inorganic solids,20 impurities in metals and electronic materials,’I as well as organic materials.” As regards the analysis of copper by LPPS a few studies have been carried out in air at atmospheric In particular Autin et al.” studied the plasma generated from a copper target in air at atmospheric pressure using 3 mJ pulses from a 337 nm ultraviolet (UV) nitrogen laser. Their results indicated that the limitations for LPPS in air under atmos- pheric pressure conditions are self-absorption line broadening and the high intensity of the continuum. These limitations can be minimized or avoided by using time-resolved spectrometry. Lee et aLZ3 reported time-integrated space-resolved studies of copper and lead sample plumes that showed differences in appearance and volume. The 193 nm beam of an ArF excimer laser was used in their experiments.Owing to the different plasma absorption characteristics at various laser wavelengths these characteristics are strongly dependent on the type of laser used for the excitation. In the present work the plasma formed by focusing a YAG laser on the surface of copper samples in air at atmospheric pressure was studied.In order to improve the selection of the temporal * Presented at the 1995 European Winter Conference on Plasma Spectrochemistry Cambridge UK January 8-13 1995. window observed the time-resolved excitation temperature was determined by using Boltzmann plots of the neutral iron lines. The optimum period in the plasma lifetime for spectro- chemical determinations of low concentrations of iron nickel and silver in copper alloys was investigated. Based on these results quantitative analysis of copper alloys by LPPS was carried out and calibration curves for iron nickel and silver were produced. EXPERIMENTAL The experimental apparatus (see Fig. 1) has been described in a previous paper from this laboratory based on aluminium alloys18 and will be described only briefly here.Laser pulses of 8 ns duration and 1064 nm wavelength (Surelite 110 Continuum) were focused by a 25 cm focal lens [Herasil for the infrared (IR)] from a 6 to a 0.6mm diameter. The laser energy incident on the sample surface was 60 mJ per pulse resulting in a laser-power density at the focus of approximately 2.5 x lo9 W cm-’. The spectra emitted were analysed by means of a 0.66 m spectrometer (McPherson 207) which has a linear dispersion at the outlet slit of 0.62nm mm-’ and a gated intensified optical multichannel analyser (OMA Princeton Instruments IRY-7OOSIB). The OMA has 700 active elements (25 pm x 2.5 mm) which allow the acquisition of data over approximately a 10 nm range in the spectra simultaneously. The spectrometer slit was imaged at 1 1 magnification onto the plasma by a He-Ne laser.This configuration allowed sampling of a region of 25 pm x 2.5 mm integrated over the whole volume of the plasma perpendicular to the sample surface. Wavelength calibration was carried out through a mercury spectral lamp operating at reduced pressure. The laser energy was controlled with a power meter (Gentec Sun Series EM1) and it was adjusted by using attenuators (Newport 935-10) in the path of the beam and not by changing the voltage of the laser power supply. Fig. 1 Schematic diagram of the experimental set-up Journal of Analytical Atomic Spectrometry September 1995 Vol. 10 643DISCUSSION Plasma Characterization Excitation temperature The relative emission intensities of lines from a given state of excitation can be used to evaluate the plasma temperature.The lines must be optically narrow well resolved for accurately evaluating their emissions without interference from other lines and the transition probabilities must be known. If the plasma is in local thermal equilibrium (LTE) the populations of the excited states are given by a Boltzmann distribution. Therefore the excitation temperature can be obtained by using a Boltzmann plot. In this particular case the temperatures were determined from the emission of eight neutral iron lines in the 374nm region observed in the spectrum of a laser-produced plasma of copper containing 0.22% iron. The transition probabilities the relevant statistical weights the energies of excited levels and the degeneracies for these lines were taken from Fuhr et aLZ5 A typical spectrum of the plasma obtained showing these lines is presented in Fig.2. Details of the procedures for temperature evaluation for aluminium alloys are described in a previous paper.” In the present study the temperature was measured for a copper target and compared with an aluminium target as described in ref. 18 and under the same conditions. The temporal variation of the temperatures obtained is shown in Fig. 3. The gate width was 5 ps for t< 10 ps and 10 ps for longer delay times. It can be seen that the temperature drops very quickly in the first few microseconds. Conversely only small temperature changes can be observed over a microsecond timescale at later times. The plasma temperature depends on many characteristics of the laser (laser energy wavelength focusing conditions pulse duration spatial and temporal fluctuations) the solid target (mechanical physical and chemical properties) and the sur- rounding atmosphere (pressure and chemical composition).’ Furthermore the laser-produced plasma depends upon the amount of ablated material which in turn is related to the interaction between a laser beam and a metal via various processes including melting erosion vaporization and subli- mation.As discussed in a previous paper,18 the temperatures were calculated from spectra that were measured by integrating the intensity over the line of sight in the plasma perpendicular to the target at the centre of the laser spot. Therefore the 5 1 0 370 372 374 376 378 WavelengtMnm Fig.2 Emission spectrum of a laser-produced plasma on copper alloy containing 0.2% Fe. Time delay and gate width were both 5 ps. The observed lines were mostly assigned to neutral atomic iron 8000 1 0 cu -I 2000 0 10 20 30 40 50 Timdp Fig. 3 Time-resolved plasma temperature for two targets aluminium and copper. The laser energy was 20 J cm-* and the gate width was 5 ps for delay times of t,< 10 ps and 10 ps for the other values plasma is not spatially resolved and no Abel inversion was applied. The laser-produced plasma is space and time dependent and as a consequence it is difficult to compare the time resolved temperature results obtained with those reported in the literature with a copper The results obtained by Lee et a1.23 for a plasma produced by a 193nm excimer ultraviolet (UV) laser are time integrated while the tempera- ture is space resolved.Furthermore plasmas produced by UV and IR lasers are essentially different as they interact in fundamentally different ways with the incoming laser radi- ation.’6 On the other hand Sdorra and N i e m a ~ ~ ~ reported time-resolved temperatures for plasmas produced by focusing 12 mJ YAG laser pulses on a copper sample in different surrounding gases. In their case however the gas pressure was 14 x lo3 Pa. Moreover the line of sight in their plasma was 1 mm above the target surface rather than perpendicular to the surface. Nevertheless in general terms the temporal evol- ution of the values obtained in the present experiments appears to be similar to the results shown by Sdorra and Niema~.’~ In view of the observed temperature evolution a typical time delay i t d ) of 5 ps and gate width of 25 ps were chosen.Such a temporal window resulted in a good signal-to-noise ratioI8 while avoiding the initial strong continuum emission. Time-resolved spectrometry The spectra emitted by a laser-produced plasma vary with observation time after the impact of the laser pulse as illus- trated in Fig. 4. At first the plasma emission consists of an intense radiation continuum that decreases with time. Between 50 ns and 1 ps the emission lines appear to be strongly broadened by the Stark effect owing to the high electron density within the plasma during this period. They are also slightly shifted and superimposed on the intense continuum. At later times (td > 1 ps) emissions from neutral atoms predomi- nate in the spectrum. These atomic lines are well resolved because their linewidths are narrower as the plasma cools and the electron density decreases.For analytical applications a time delay is required to gate off the early part of the signa1,6.8-19,21,27-30 in order to avoid the intense initial con- tinuum emission improve the line resolution and increase the precision. A typical time delay of 5 ps and gate width of 20 ps were chosen in order to obtain a good compromise between the signal-to-noise ratio and the line intensity. It should also be mentioned that the optimum delay time is dependent upon the species and related to the energy of the 1aser.6~8-19,21,27-30 644 Journal of Analytical Atomic Spectrometry September 1995 Vol.10330 332 334 336 338 340 Wavelengt Mnm Fig. 4 Spectra of a copper plasma taken at different delay times after the impact of the YAG laser pulse. The sample was copper alloy (PC-l) the laser energy density was 20 J cm-* and the gate widths were 100 ns 0.5 ps and 20 ps for delay times of 50 ns 0.5 ps and 5 ps respectively 7 - Spectrochemical analysis The experimental conditions suitable for analysis of a copper alloy by LPPS (number of laser shots laser energy focusing light collection etc.) are in general terms similar to those described in a previous paper'* for aluminium alloys. They were applied to the quantification of low concentrations of iron nickel and silver in a copper matrix. The composition of the samples used in the present work is given in Table 1.National Institute of Standards and Technology (NIST) cali- brated standards as provided by Alacan were used as samples for this work. The analytical characteristics of the lines selected for the LPPS analysis which are of neutral species are taken from Reader et aL3' and given in Table 2. The intensity measurements of the analytical lines for the standards were performed by integrating over 20 laser shots at one spot on the sample after cleaning the surface by a few preliminary shots to avoid any effect of the initial surface state. The intensities were calculated (in arbitrary units) by software which numerically integrated the area under the line distri- bution curve after correcting for the background. For cali- bration the mean values were calculated from ten measurements at different locations separated by 1 mm.Calibration curves were then produced by plotting the intensity I Table 1 alloy samples used in the present work Iron nickel and silver concentrations (%) in the SRM copper Sample Fe Ni Ag c u PC- 1 0.2160 0.368 0.0395 99.2 PC-2 0.1120 0.205 0.0190 99.4 PC-3 0.0500 0.078 0.0075 99.6 PC-4 0.0240 0.027 0.0036 99.7 PC-5 0.0115 0.009 0.0020 99.8 PC-6 0.0085 0.004 0.0017 99.9 Table 2 Analytical lines used for LPPS analysis of copper alloys. The data were taken from ref. 31 Element Line/nm E,/cm-' EJcm-' A2,/1O8 s - l g g2 Fe 358.12 6928 34844 1.02 11 13 Ni 341.48 204.8 29481 0.55 7 9 Ni 349.30 879.8 29501 0.98 5 3 Ag 338.29 0 29552 1.31 2 4 of the analyte line of each element uersus the known concen- tration of this element in copper.Systematic investigations of the optical plasma emission were performed in the wavelength region from 180 to 600 nm. All measured spectral features in this range could be identified and assigned to appropriate elements usually present in copper alloys (copper aluminium arsenic bismuth lead manganese selenium sulfur tellurium tin and zinc). The spectra of these elements have been analysed in order to detect possible inter- ference between their lines and the lines given in Table2. In 6 5 4 3 2 1 n Fe " 350 352 354 356 358 360 362 364 18 I I I 1 16 - 14 12 .- 2 l o $ 8 f 6 c 4 2 .- - 15 10 5 n 330 332 334 336 338 340 342 Wavelengt Wnrn Fig.5 Spectra of a laser-produced plasma on copper alloy with increasing (a) iron content; (b) nickel content; and (c) silver content.Laser-energy density was 20 J cm-' and delay time and gate were both 5 ps Journal of Analytical Atomic Spectrometry September 1995 Vol. 10 645330 332 334 336 338 340 Wavelengt Mnm Fig. 4 Spectra of a copper plasma taken at different delay times after the impact of the YAG laser pulse. The sample was copper alloy (PC-l) the laser energy density was 20 J cm-* and the gate widths were 100 ns 0.5 ps and 20 ps for delay times of 50 ns 0.5 ps and 5 ps respectively 7 - Spectrochemical analysis The experimental conditions suitable for analysis of a copper alloy by LPPS (number of laser shots laser energy focusing light collection etc.) are in general terms similar to those described in a previous paper'* for aluminium alloys.They were applied to the quantification of low concentrations of iron nickel and silver in a copper matrix. The composition of the samples used in the present work is given in Table 1. National Institute of Standards and Technology (NIST) cali- brated standards as provided by Alacan were used as samples for this work. The analytical characteristics of the lines selected for the LPPS analysis which are of neutral species are taken from Reader et aL3' and given in Table 2. The intensity measurements of the analytical lines for the standards were performed by integrating over 20 laser shots at one spot on the sample after cleaning the surface by a few preliminary shots to avoid any effect of the initial surface state. The intensities were calculated (in arbitrary units) by software which numerically integrated the area under the line distri- bution curve after correcting for the background.For cali- bration the mean values were calculated from ten measurements at different locations separated by 1 mm. Calibration curves were then produced by plotting the intensity I Table 1 alloy samples used in the present work Iron nickel and silver concentrations (%) in the SRM copper Sample Fe Ni Ag c u PC- 1 0.2160 0.368 0.0395 99.2 PC-2 0.1120 0.205 0.0190 99.4 PC-3 0.0500 0.078 0.0075 99.6 PC-4 0.0240 0.027 0.0036 99.7 PC-5 0.0115 0.009 0.0020 99.8 PC-6 0.0085 0.004 0.0017 99.9 Table 2 Analytical lines used for LPPS analysis of copper alloys. The data were taken from ref. 31 Element Line/nm E,/cm-' EJcm-' A2,/1O8 s - l g g2 Fe 358.12 6928 34844 1.02 11 13 Ni 341.48 204.8 29481 0.55 7 9 Ni 349.30 879.8 29501 0.98 5 3 Ag 338.29 0 29552 1.31 2 4 of the analyte line of each element uersus the known concen- tration of this element in copper.Systematic investigations of the optical plasma emission were performed in the wavelength region from 180 to 600 nm. All measured spectral features in this range could be identified and assigned to appropriate elements usually present in copper alloys (copper aluminium arsenic bismuth lead manganese selenium sulfur tellurium tin and zinc). The spectra of these elements have been analysed in order to detect possible inter- ference between their lines and the lines given in Table2. In 6 5 4 3 2 1 n Fe " 350 352 354 356 358 360 362 364 18 I I I 1 16 - 14 12 .- 2 l o $ 8 f 6 c 4 2 .- - 15 10 5 n 330 332 334 336 338 340 342 Wavelengt Wnrn Fig.5 Spectra of a laser-produced plasma on copper alloy with increasing (a) iron content; (b) nickel content; and (c) silver content.Laser-energy density was 20 J cm-' and delay time and gate were both 5 ps Journal of Analytical Atomic Spectrometry September 1995 Vol. 10 6458 Cremers D. A. Radziemski L. J. and Loree T. R. Anal. Chem. 1983 55 1246. 9 Cremers D. A Radziemski L. J. and Loree T. R. Appl. Spectrosc 1984 38 949. 10 Wachter J. R. and Cremers D. A Appl. Spectrosc. 1987,41,1042. 11 Lorenzen C. J. Carlhoff C. Hahn U. and Jogwich M. J. Anal. At. Spectrom. 1992 I 1029. 12 Aragon C. Aguilera J. A and Compos J. Appl. Spectrosc 1993 47 606. 13 Cremers D.A. Appl. Spectrosc 1987 41 572. 14 Grant K. J. Paul G. L. and O’Neill J. A. Appl. Spectrosc. 1991 45 701. 15 Ieis F. Sdorra W. KO J. B. and Niemax K. Mikrochim. Acta 1989. 2. 185. ~1 16 Aragon C. Aguilera J. A and Compos J. Appl. Spectrosc. 1992 46 1382. 17 Sabsabi. M.. Cielo. P.. Boilv. S.. and Chaker. M.. SPIE. Proc. Optical h e t h . Chem. ProcesiControl 1993 2069 191. . 18 Sabsabi M. and Cielo P. Appl. Spectrosc. 1995 49 499. 19 Autin M. Briand A Mauchien P. and Mermet J. M. Spectrochim. Acta Part B 1993 48 851. 20 Franzke D. Klos H. and Wokaun A Appl. Spectrosc 1992 46 587. 21 Ottesen D. K. Appl. Spectrosc 1992 46 593. 22 23 24 25 26 27 28 29 30 31 Kagawa K. Deguchi Y. Ogata A Kurniawan H. Ikeda N. and Tagaki Y. Jpn. J. Appl. Phys. 1991 39 L1889. Lee Y.-I. Sawan S. P. Thiem T. L. Teng Y.-Y. and Sneddon J. Appl. Spectrosc. 1992 46 436. Sdorra W. and Niemax K. Mikrochim. Acta 1992 107 319. Fuhr J. R. Martin G. A. Wiese W. L. and Younger S. M. in Spectroscopic Data for Iron ed. Wiese W. L. Oak Ridge National Laboratory Oak Ridge 1985 ORNL-6089/V4. Geertsen G. Briand A. Chartier F. Lacour J.-L. Mauchien P. Sjdstrom S. and Mermet J. M. J. Anal. At. Spectrom. 1994,9 17. Iida Y. Appl. Spectrosc. 1989 43 229. Iida Y. Spectrochim. Acta. Part B 1990 45 1353. Grant K. J. Paul G. L. and O’Neill J. A. Appl. Spectrosc. 1990 44 1711. Owens M. and Majidi V. Appl. Spectrosc 1991 45 1463. Reader J. Corliss C. H. Wiese W. L. and Martin G. A. Wavelength and Transition Probabilities for Atomic and Ions National Standard Reference Data System-NBS 68 US Government Printing Office Washington DC 1980. Paper 5/008371 Received February 13 1995 Accepted April 28 1995 Journal of Analytical Atomic Spectrometry September 1995 Vol. 10 647

ISSN:0267-9477    

DOI:10.1039/JA9951000643

出版商:RSC    

年代:1995

数据来源: RSC