摘要:

Determination of mercury in cosmetics by flow injection–cold vapour generation–atomic fluorescence spectrometry with on-line preconcentration L. Ga�miz-Gracia and M. D. Luque de Castro* Analytical Chemistry Division, Faculty of Sciences, University of Co�rdoba, E-14004 Co�rdoba, Spain. E-mail: qa1lucam@uco.es Received 2nd July 1999, Accepted 22nd July 1999 A method for the determination of trace amounts of mercury in cosmetics is reported. The method involves the acid treatment of the sample in a focused-microwave digester, on-line preconcentration in a C18 column and cold vapour generation–atomic fluorescence detection. The method can be applied to the determination of the target analyte at the picogram level in diVerent eye cosmetics (eye liner, eye shadow and eye pencil ), with recoveries ranging between 89 and 104% and with a precision, expressed as relative standard deviation, of 5.9%.staltic pumps (Gilson, Worthington, OH, USA), two Introduction Rheodyne (Cotati, CA, USA) Model 5041 injection valves Microwaves are one of the energy sources commonly applied and Teflon tubing of 0.5 mm id (Scharlau, Barcelona, Spain) for accelerating either digestion or leaching processes.1–3 were used in order to construct the flow injection (FI ) mani- Microwaves (either monomode or multimode) have been fold.A Knauer (Bad Hamburg, Germany) recorder was used applied to oxidise organic to inorganic mercury for atomic in order to record the signals.A conical laboratory-made spectroscopic detection, in both open and closed focused column of Teflon (3.5×0.5 cm id) was used in the preconcenmicrowave digesters,4 normally followed by AAS. The easy tration step. experimental implementation of ion-exchange or sorbent techniques and the simplicity of the basic unit required make them Reagents aVordable in contrast to other continuous separation–preconcentration continuous alternatives.5,6 These techniques have All reagents were of analytical reagent grade.Ultra-pure water been used for mercury preconcentration using diVerent chelat- obtained from a Milli-Q system (Millipore, Bedford, MA, ing agents, such as sodium diethyldithiocarbamate, ammonium USA) was used throughout. Stock standard solutions of 1 g l-1 pyrrolidine dithiocarbamate and dithizone, followed by on-line (as mercury) of mercury(II ) nitrate (Aldrich, Milwaukee, WI, enrichment on RP C18 columns.7,8 The release of the analyte USA) in 10% nitric acid (Panreac, Barcelona, Spain), and from these columns involves the breaking of covalent phenylmercury acetate (Merck, Darmstradt, Germany) in reagent–Hg bonds formed during metal retention.Other mate- ethanol (Scharlau) were prepared and stored at 4 °C. Working rials have been used for the preconcentration of Hg, such as solutions in 1% nitric acid and acetone (Merck) were prepared functionalized silica columns9,10 or resin incorporating dithi- for inorganic and organic mercury, respectively.Nitric acid, ocarbamate groups.11 hydrogen peroxide and sulfuric acid (all from Panreac) were Mercury is used in the cosmetic industry as a preservative, used in the sample pre-treatment. Hydrochloric acid (0.1%) usually as phenylmercury salts and thiomersal [sodium (Panreac) and 5% tin(II) chloride (Merck) in 10% hydrochloric 2-(ethylmercuriothio)benzoate]. Owing to its high toxicity, its acid were used in the determination step.Ammonium content is strongly restricted by law, the maximum allowed 1-pyrrolidinecarbodithioate (APDC) (Aldrich), 5×10-4% in concentration being 0.007% m/m of Hg. Environmental and 30 mmol l-1 ammonium acetate (Merck)–acetic acid biological samples are the most frequent matrices in mercury (Panreac) buVer (pH 6.5), was used as the chelating agent analyses, although organic mercury has also been determined and ethanol as the eluent in the preconcentration step. in pharmaceutical samples, but rarely in cosmetics.12 In this C18 bonded phase, obtained from a Bond Elut cartridge paper, a method for the determination of total mercury in (Varian, Harbor City, CA, USA), was used in the preconcencosmetics is reported.It involves the treatment of the sample tration column. Argon (Carburos Meta�licos, Barcelona, Spain) in an open-microwave digester, on-line preconcentration of was used to flush the Hg to the detector. the inorganic mercury and subsequent atomic fluorescence detection of the Hg0 removed from a gas–liquid separator.Procedure Sample pre-treatment. A 0.25 g amount of sample was Experimental weighed and transferred into the vessel of the microwave digester; 1 ml of nitric acid, 1 ml of hydrogen peroxide and Instruments and apparatus 3 ml of sulfuric acid were added and the mixture was subjected to 45 min of microwave irradiation (85% power, 170W). After A Merlin atomic fluorescence detector (P.S. Analytical, Orpington, Kent, UK) and a Microdigest 301 focused micro- cooling, the treated sample was filtered, the vessel was washed with water and the filtrate and washings were transferred into wave system (Prolabo, Paris, France), with a maximum irradiation power of 200 W, were used.Two Minipuls-3 peri- a 100 ml calibrated flask. APDC (2.5 ml ) was added, the J. Anal. At. Spectrom., 1999, 14, 1615–1617 1615Table 1 Optimisation of variables mixture was diluted to volume and the contents were subjected to the preconcentration/determination step.Range studied Optimum value Preconcentration/determination step. In the case of samples Determination step— that did not required preconcentration, the manifold used was [SnCl2] (%) 1–10 5 [HCl] to dissolve SnCl2 (%) 5–15 10 that shown in Fig. 1(a); for samples that required preconcen- [HCl] (%) 0.01–10 0.1 tration, the manifold in Fig. 1(b) was used. In the latter, the Flow rate, SnCl2/ml min-1 0.5–3.5 1.0 sample (or standard) was aspirated and allowed to pass Flow rate, HCl/ml min-1 1.5–5 3.5 through the preconcentration column for 15 min.Preconcentration step— Subsequently, the two injection valves were simultaneously [APDC] (%) 5×10-5–0.5 5×10-4 switched to the injection position and the retained complex Flow rate/ml min-1 0.4–3.5 3.0 Conditioning time/min 1–10 3 was eluted. After 3 min, the two valves were switched to the Eluent volume/ml 0.2–1.5 0.75 filling position again, and a new sample was introduced.After Preconcentration time/min 5–60 15 elution, the Hg was reduced by a stream of SnCl2 and swept Sample pre-treatment step— out of the gas–liquid separator by an Ar stream into the Microwave power (%) 25–99 85 atomic fluorescence detector and the signal was recorded. Irradiation time/min 5–60 45 The preconcentration column, placed in the loop of the injection valve, allowed the load step to be carried out in the Preconcentration step. DiVerent FI manifolds were tested opposite direction to elution, thus avoiding the shortcomings for the preconcentration step, and the best results were derived from both increased compactness of the packed mateobtained by placing the preconcentration column in the loop rial and the appearance of parasitic signals due to the sample of an injection valve and including a second valve for injection matrix, which can in this way be sent directly to waste. Higher of the eluent.The manifold in Fig. 1(b) was ultimately selected. eYciency was achieved by using a conical rather than a The diVerent parameters, ranges studied and optimum selected cylindrical column arranged in such a way that sample loading values are given in Table 1. The retention kinetics were very takes place from the narrower to the broader end.13 fast as no significant influence of the flow rate used on the retention was observed when this variable was changed Results and discussion between 0.4 and 3.5 ml min-1, so a high flow rate was selected in order to increase either the sensitivity or the sampling Optimisation frequency.Increased preconcentration times provided higher The optimisation of the overall method was developed in three signals, and a preconcentration time of 15 min was finally steps: (1) the determinationnd the FI system were selected as a compromise between sensitivity and frequency studied; (2) the on-line preconcentration of Hg was optimised; of analysis. and (3) the pre-treatment of the samples was studied, the main purpose being to carry out this step in as short a time as Sample pre-treatment.DiVerent acid mixtures have been possible. Inorganic mercury diluted in nitric acid (10%) was reported for the determination of mercury using microused in steps 1 and 2 and phenylmercury in step 3. Table 1 wave digestion.14,15 Br-–BrO3-,16 HNO3–H2O2 and shows the studied ranges and optimum values for all the HNO3–HCl mixtures were tested.Finally, a 1+1+3 variables. The univariate method was used in all instances. HNO3–H2O2–H2SO4 mixture was selected. Subsequently, the microwave power was studied, and 85% (170 W) was selected, Determination step. The optimisation of the detection step as above this value violent boiling occurred. The final was carried out with a conventional two channel FI system irradiation time was 45 min, as this was the minimum necessary [see Fig. 1(a)]. One channel contained 0.1% HCl and the to complete the oxidation of the organic mercury (peak height injection valve and the other contained 5% SnCl2 in 10% HCl. within±5% of the signal for the same amount of inorganic The selected flow rates were 3.5 and 1.0 ml min-1, respectively.mercury). Features of the method A calibration curve ranging between 0.1 and 10 mg ml-1 was constructed using the manifold shown in Fig. 1(a). The linear equation was y=2.19+17.71x (r2=99.83) where y denotes fluorescence units and x is mg ml-1 of mercury.This manifold was used to determine the mercury content in an ophthalmic preparation, acquired in a local pharmacy, with a nominal content of Hg(CN)2 as mercury of 30 mg ml-1. The result obtained was 30.1±0.5 mg ml-1 (n=6). The preconcentration system shown in Fig. 1(b) was then used to construct a second calibration curve. Table 2 gives the equation for this curve, Table 2 Features of the method Equationa y=22.77+0.154x Linear range/pg ml-1 14.9–1000 Regression coeYcient (r2) 0.992 Detection limit/pg ml-1b 14.9 Fig. 1 FI manifolds used for the determination of mercury in cos- Quantification limit/pg ml-1c 49.7 RSD (%)d 5.9 metics, (a) without and (b) with preconcentration step. PP=peristaltic pump; IV=injection valve; W=waste; PC=preconcentration column; ay denotes fluorescence units and X=pg ml-1. bAs three times s RC=reaction coil; GLS=gas–liquid separator; AFD=atomic fluo- (IUPAC). cAs 10 times s (IUPAC). d n=7. rescence detector; R=recorder. 1616 J. Anal. At. Spectrom., 1999, 14, 1615–1617Table 3 Application of the proposed method (n=3) Sample Hg sample content/ng Hg added/ng Hg found/ng Recovery (%) Eye-shadow 1 10.2±1.7 50.0 44.4±2.3 88.9±4.6 Eye-shadow 2 5.7±0.4 50.0 50.4±1.5 100.7±3.1 Eye-pencil 1 Below detection limit 50.0 48.0±4.6 95.9±9.1 Eye-pencil 2 Below detection limit 50.0 47.9±3.8 95.7±7.7 Eye-liner 1 17.8±0.9 50.0 51.1±1.9 102.2±3.9 Eye-liner 2 7.6±0.2 50.0 52.0±2.6 104.1±5.2 linear range, detection limit and repeatability, expressed as References RSD. 1 J. R. J. Pare� and J. M. R. Be�langer, Trends Anal. Chem., 1994, 13, 176. Application of the method 2 K. Ganzler, A. Salgo and K. Valko, J. Chromatogr., 1986, 371, 299. The method was applied to six diVerent cosmetics acquired in 3 V.Lo� pez-A� vila, R. Young and W. F. Beckert, Anal. Chem., 1994, local shops: two eye shadows, two eye liners and two eye 66, 1097. pencils, the mercury contents of which were unknown.A 4 G. Schnitzer, A. Soubelet, C. Testu and C. Chafey, Mikrochim. Acta, 1995, 119, 199. 0.25 g amount of sample was used. A portion of the sample 5 M. Valca�rcel and M. D. Luque de Castro, Non-Chromatographic without mercury addition was subjected to the proposed Continuous Separation Techniques, Royal Society of Chemistry, method and its mercury concentration was calculated. Then Cambridge, 1991. three analyses of samples spiked with 50 ng of mercury were 6 Z.Fang, Flow-Injection Atomic-Absorption Spectrometry, John carried out (2×10-5% m/m). The results are given in Table 3. Wiley, Chichester, 1995. Good recoveries (ranging between 88.9 and 104.1%) were 7 R. Falter and H. F. Scho� ler, Fresenius’ J. Anal. Chem, 1995, 353, 34. obtained in all instances, with acceptable precision. The lower 8 M. Ferna�ndez Garcý�a, R. Pereiro Garcý�a, N. Bordel Garcý�a and recovery found for eye shadow 1 could be due to either A. Sanz-Medel, Talanta, 1994, 41, 1833.incomplete dissolution or to the presence of some negative 9 P. Canada-Rudner, J. N. Cano-Pavon, F. Sa�nchez-Rojas and interference. A. Garcý�a de Torres, J. Anal. At. Spectrom., 1998, 13, 1167. 10 P. Canada-Rudner, A. Garcý�a de Torres, J. M. Cano-Pavon and E. Rodrý�guez-Castello� n, J. Anal. At. Spectrom., 1998, 13, 243. Conclusions 11 H. Emteborg, D. C. Baxter and W. Frech, Analyst, 1993, 118, 1007. A method for the determination of phenylmercury in cosmetics 12 C. Demanze, L. RugroV and A. Karleskind, Parfums Cosmet. has been developed. The method allows the determination of Aromes, 1984, 5869. amounts of mercury at least 1000 times lower than the 13 M. D. Luque de Castro and L. Ga�miz-Gracia, Advances in Atomic Spectroscopy, JAI Press, Greenwich, CT, 1998, vol. 4, concentration of mercury allowed in this kind of sample, with pp. 177–212. good recoveries. This method could be applied to detect fraud 14 M. Gulmini, V. Zelano and G. Ostacoli, Ann. Chim. (Rome), in the content of mercury. 1997, 87, 457. 15 C. Y. Zhou, M. K. Wong, L. L. Koh and Y. C. Wee, Anal. Sci., 1996, 12, 471. Acknowledgements 16 D. W. Bryce, A. Izquierdo and M. D. Luque de Castro, Anal. Chim. Acta, 1996, 324, 69. The Spanish Comisio�n Interministerial de Ciencia y Tecnologý�a (CICyT) is thanked for financial support (Project No. PB96–1265) Paper 9/05349B J. Anal. At. Spect

ISSN:0267-9477    

DOI:10.1039/a905349b

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

Comparative studies on chemical modification of polytetrafluoroethylene slurry in ETV-ICP-AES and ETAAS Wang Fuyi, Jiang Zucheng,* Hu Bin and Peng Tianyou Department of Chemistry, Wuhan University, Wuhan 430072, China Received 10th June 1999, Accepted 12th August 1999 The chemical modification of polytetrafluoroethylene (PTFE) slurry in electrothermal vaporization inductively coupled plasma atomic emission spectrometry (ETV-ICP-AES) and in electrothermal atomic absorption spectrometry (ETAAS) was investigated and compared systematically.For both cases, the graphite furnace can be used as a chemical reactor in which the fluorinating reagent PTFE can convert the oxides of analytes into their volatile fluorides at high temperature. However, diVerent influences resulting from fluorination in ETV-ICP-AES and ETAAS were observed owing to diVerent functions of the graphite furnace in the two techniques. The formation of fluoride enhanced the emission signals of the refractory elements Mo, Cr and Yb significantly in ETVICP- AES, but only improved the sensitivity of Cr in ETAAS.In both ETV-ICP-AES and ETAAS, the addition of PTFE increased the maximum ashing temperature of volatile analyte cadmium.Moreover, PTFE can obviously reduce or remove the matrix interference and particle size eVects of solid samples; this advantage is a benefit in the analysis of complicated samples, especially for the direct analysis of solid samples. As is well known, electrothermal atomic absorption Experimental spectrometry(ETAAS) has developed into one of the most Apparatus and operating conditions eYcient techniques for the determination of trace elements, and is widely applied in biological, environmental, geological A WFX-IF2 atomic absorption spectrometer with a deuterium and industrial analysis.1,2 As the atomizer, a graphite furnace background corrector and a WF-4A electrothermal graphite is frequently used for ETAAS.Another important application furnace system (Beijing Second Optics, Beijing, China) was of the graphite furnace is concerned with the sample vapor- employed.izer in inductively coupled plasma atomic emission A 2 kW power, 27±3 MHz ICP spectrometer (Beijing spectrometry(ICP-AES). Since the first paper describing the Broadcast Equipment Factory, Beijing, China) and a convenuse of a tantalum filament device as a vaporizer for ICP was tional plasma torch were used with a WF-1 graphite furnace published in 1974,3 electrothermal vaporization (ETV) has system as the sample vaporizer for ETV-ICP-AES.A detailed been considered to be a useful tool for sample introduction in description of ETV-ICP-AES is available elsewhere.5 ICP-AES owing to following advantages: less sample consump- The optimized operating conditions for ICP-AES, AAS and tion, high sample transport eYciency and direct analysis of the electrothermal vaporizer/atomizer are given in Table 1.solid samples. However, this approach has the disadvantage of electrothermal atomization with regard to the formation of Solutions and reagents refractory carbides, which leads to a decrease in and sometimes Stock standard solutions (1 mg mL-1) of molybdenum, complete suppression of analytical signals.4,5 chromium, ytterbium, cadmium, copper, cobalt and nickel In order to prevent analytes from forming refractory were national reference solutions with 5% H2SO4 (Mo), 10% carbides and to improve the volatility of the elements of HNO3 (Yb) or 10% HCl (Cr, Cd, Cu, Co, Ni) supplied by interest, several halogenating reagents6–9 have been proposed the National Center for Analysis and Testing of Steel Materials for the determination of refractory elements by ETV-ICP- (Beijing, China).Test solutions were prepared by diluting AES. Our previous work5,10–16 confirmed that fluorinationthese stock standard solutions. A 60% m/v PTFE slurry assisted electrothermal vaporization (FETV)-ICP-AES using (Shanghai Organic Chemistry Institute, Shanghai, China) was polytetrafluoroethylene (PTFE) slurry as a halogenating commercially available.Nitric acid and other reagents used reagent was an eYcient method for determining trace refracwere Specpure or of analytical-reagent grade (Shanghai tory elements in biological and environmental samples and Reagent Factory, Shanghai, China). Doubly distilled water high-purity materials. To the best of our acknowledge, there was used throughout. have been few reports of the application of PTFE as a chemical modifier in ETAAS.Our recent work17–19 indicated that the Slurry sample preparation addition of PTFE could clearly improve the sensitivity of chromium and increase the tolerable ashing temperature of Slurry samples were prepared by placing 1 mL of 60% PTFE cadmium in ETAAS. Moreover, the modification of PTFE slurry in 10 mL calibrated flasks containing the relevant also reduced or removed matrix interference in ETAAS.amount of analytes and diluting to 10 mL with doubly distilled The aim of this work was to attempt to investigate and water. The resultant mixtures were dispersed with an ultrasonic compare the chemical modification of PTFE slurry in both wave vibrator for 15 min. The flasks were shaken vigorously ETV-ICP-AES and ETAAS, and some conclusions were drawn prior to analysis. A cadmium slurry sample was prepared in from the experimental results. Molybdenum, ytterbium, chro- 4% HNO3 medium.Blank slurries had to be prepared when mium and cadmium, which have diVerent physical and chemi- interfering components were added. The emission signal intencal properties, were selected as the analytes. The mechanism sity or absorbance of analytes was obtained by subtracting the blank values from the total values. of chemical modification of PTFE is also discussed. J. Anal. At. Spectrom., 1999, 14, 1619–1624 1619Table 1 Operating conditions for ETV-ICP-AES ICP-AES— Wavelength/nm Mo 379.325; Yb 328.937; Cr 267.716; Cd 361.051 Incident power/kV 3.5 for Mo, Cr and Cd; 3.2 for Yb Carrier gas (Ar) flow rate mL min-1 0.6 Coolant gas (Ar) flow rate ml min-1 16 Auxiliary gas (Ar) flow rate ml min-1 0.8 Observation height/mm 12 AAS— Wavelength/nm Mo 313.3; Yb 398.8; Cr 357.9; Cd 228.8 Spectral bandwidth/nm 0.4 for Mo and Cd; 0.2 for Cr and Yb HCL current/mA Mo 4.8; Yb 1.5; Cr 2.0; Cd 1.8 Deuterium lamp current/mA 45 for Mo and Cd; self-absorption background correction for Cr and Yb ETV— Dry 110 °C, ramp 20 s, hold 20 s Ashing 400 °C for Mo; 1200 °C for Yb; 900 °C for Cr; 800 °C for Cda.Ramp 15 s, hold 25 s Atomization (AAS)b 2800 °C for Mo; 2600 °C for Yb and Cr; 2400 °C for Cd. Hold 12 s (Mo); 6 s (Yb, Cr, Cd) Vaporization (ICP) 1900 °C for Mo; 2400 °C for Yb, Cr and Cd. Hold 6 s (Mo, Yb, Cr); 3 s (Cd) aNo ashing step was used for Cd without PTFE. bThe flow rate of Ar was 50 ml min-1 (Mo) or stopped (Yb, Cr and Cd).Recommended procedure The general analytical procedures for ETV-ICP-AES and ETAAS were recommeded in earlier work.5,17 Results and discussion Signal profiles of analytes Typical emission and absorption signal profiles of Mo, Yb, Cd and Cr are shown in Figs. 1–4, respectively. Based on the experimental results obtained, the following conclusions could be drawn. (1)The PTFE slurry showed an excellent chemical modifi- cation of refractory elements (Mo and Yb) in ETV-ICP-AES.The presence of PTFE not only enhanced the signal intensity Fig. 2 Typical emission (A) and absorption (B) signal profiles of Yb with (a) and without (b) PTFE. (A) a, 4 ng Yb; b, 20 ng Yb; a¾, b¾, residual signal of first firing. (B) a, 200 ng Yb; b, 1.6 ng Yb. of the analytes, but also eliminated the memory eVect. In contrast, the addition of PTFE made the absorption signals of Mo and Yb decrease significantly in ETAAS. Furthermore, a background signal was observed during Mo atomization.(2) In both ETV-ICP-AES and ETAAS, the addition of PTFE had no obvious influence on the signal intensities of Cd owing to its high volatility and easy atomization. However, the PTFE slurry (with 4% HNO3) could stabilize Cd during the thermal pre-treatment. Therefore, its maximum ashing temperature increased in both ETV-ICP-AES13 and in ETAAS.18 In ETAAS, the integrated absorption signal of Cd was slightly enhanced and appeared after a delay.Fig. 1 Typical emission (A) and absorption (B) signal profiles of Mo (3) No matter whether the PTFE slurry was used in ETV- with (a) and without (b) PTFE. (A) a, 2 ng Mo; b, 20 ng Mo; a¾, b¾, ICP-AES or in ETAAS, it had a positive eVect on the residual signal of first firing. (B) a, 20 ng Mo; b, 4 ng Mo; c, background signal in the presence of PTFE. determination of chromium. In both cases, the Cr emission 1620 J. Anal. At. Spectrom., 1999, 14, 1619–1624and absorption signal intensities were increased significantly compared with no PTFE.Detection limits of analytes The detection limit is defined as the analyte concentration which yields a signal equal to three times the standard deviation of the signal produced from 10 replicates of a 6% PTFE (with 4% HNO3 for Cd) blank slurry. With and without PTFE, the detection limits of Mo, Yb, Cr and Cd determined by ETVICP- AES and ETAAS were calculated and are presented in Table 2.The obtained results indicated that compared with conventional ETV-ICP-AES, the detection limits of Mo, Cr and Yb obtained by FETV-ICP-AES were decreased by one to two orders of magnitude, with the exception of Cd determination. Further, in ETAAS, the PTFE modifier improved the detection limit of Cr by a factor of 3.5, but no improvement in detection limit was observed for Cd. In contrast, the detection limits of Mo and Yb were much worse than those without PTFE.Matrix interference Previous work5,10 demonstrated that a 104–105-fold excess of alkali and alkaline earth elements and transition metals had no obvious eVect in the determination of Cr, Ti, Y, V, La, B and Mo by FETV-ICP-AES. These tolerable amounts of matrix elements in FETV-ICP-AES were superior to those in pneumatic nebulization ICP-AES by one order of magnitude or more. The influences of common matrix elements on Cr and Cd Fig. 3 Typical emission (A) and absorption (B) signal profiles of Cd absorption in ETAAS were also investigated, and the results with (a) and without (b) PTFE.(A) a, 10 ng Cd; b, 10 ng Cd: a¾, b¾, are given in Table 3. It can be seen that in the presence of residual signal of first firing. (B) a, 0.2 ng Cd; b, 0.2 ng Cd. PTFE slurry, the interferences of matrix elements (Ca, Mg, Ba, Sr, Cu, Fe, Al ) on Cr and Cd absorption were reduced substantially, and the background absorption produced by sodium chloride could also be eliminated by using PTFE in 4% HNO3.18 Fig. 5 shows the signal profiles of Cr absorption in CuCl2 solutions with and without PTFE. It is obvious that without the PTFE matrix, copper seriously suppressed the Cr absorption and delayed the appearance of the analytical signal, but this interference was completely removed with the introduction of PTFE. Direct analysis of solid samples Particle size eVect. The particle size eVects of SiO2 powder on the recovery of Cr in ETV-ICP-AES and the recovery of Co in ETAAS were investigated with and without PTFE, and the experimental results are given in Tables 4 and 5, respectively.These results show that the eVects of particle size of solid silicon dioxide powder on both ETV-ICP-AES and ETAAS could be decreased considerably by the fluorination reaction which took place in the graphite furnace during thermal pre-treatment. When the average particle size of sample was <150 mm, the recoveries of the analytes were >90% with RSDs of 4.7 and 5.9%, respectively.In comparison with data in the literature20,21 concerning the tolerable particle Table 2 Detection limits (pg) of the analytes determined by ETVICP- AES and GFAAS ETV-ICP-AES GFAAS FETV-ICP-AES ETV-ICP-AES With Without Elements PTFE PTFE Mo 7.0 100 174 5 Yb 18 150 754 5 Fig. 4 Typical emission (A) and absorption (B) signal profiles of Cr Cr 14 500 2.8 9.6 with (a) and without (b) PTFE. (A) a, 4 ng Cr; b, 20 ng Cr; a¾, b¾, Cd 141 138 0.12 0.12 residual signal of first firing.(B) a, 0.4 ng Cr, b, 0.4 ng Cr. J. Anal. At. Spectrom., 1999, 14, 1619–1624 1621Table 3 Interference of matrix components(as chlorides) on chromium and cadmium absorption Matrix components Parameter None Na K Mg Ca Ba Sr Cu Fe Al Concentration of cations/mg mL-1 — 2.0 2.0 0.08 2.0 1.0 1.0 4.0 2.0 2.0 Relative absorbance of Cr (50 ng mL-1)— Without PTFE 1.00 1.02 1.01 2.02 0.31 0.32 0.29 0.17 3.34 2.91 With PTFE 1.00 1.03 0.97 1.02 1.03 0.87 0.90 0.98 0.99 1.05 Concentration of metal/mg mL-1 1.00 3%a 100 10 10 — — 100 100 10 Relative absorbance of Cdc (10 ng mL-1)— Without PTFE 1.00 —b 1.00 0.18 0.67 — — 0.28 0.24 1.00 With PTFE 1.00 1.00 1.00 1.00 0.97 — — 1.00 1.00 1.00 aConcentration of NaCl.bThe cadmium absorption signal could not be distinguished from the strong background signal of NaCl. cCd slurry samples contain 4% HNO3. Fig. 6 Influence of concentration of silicon dioxide on the signal intensity of Cr emission and Co absorption in SiO2 slurries in the presence of PTFE.dependence of the Cr emission signal and Co absorption signal on the concentration of SiO2 suspension in ETV-ICP-AES and ETAAS. It can be seen that with PTFE as chemical modifier, the tolerable concentrations of the matrix (SiO2) in Cr and Co determinations are 25 and 15 mg mL-1, respectively. This result indicates that in the presence of PTFE, the matrix interferences in ETAAS and ETV-ICP-AES remain at a low level, which is attributed to the eYcient separation of most matrix Si from the analytes by selective vaporization Fig. 5 Typical signal profiles of Cr absorption in the absence (A) and during thermal pre-treatment.14,22 presence (B) of PTFE: a, 1.0 ng Cr; b, 1.0 ng Cr+40 mg Cu. Modification mechanism size for conventional ETV-ICP-AES and ETAAS, the present data indicate that satisfactory recoveries with better RSDs can In order to explore the vaporization mechanism of analytes in ETV with and without PTFE, an X-ray diVraction analysis of be obtained for larger dimensions when the fluorination modification is employed.residues in the graphite tube after the ETV heating cycle was carried out in our previous work.5,10 The results obtained indicated that in ETV-ICP-AES the chemical modification of Matrix interference. The matrix interference is one of the primary factors influencing the direct analysis of solid samples PTFE resulted from the fluorination taking place between its pyrolytic product and the analytes. The reaction mechanism in conventional ETV-ICP-AES and ETAAS.Fig. 6 shows the Table 4 EVect of particle size on the recovery of Cr in SiO2 determined by FETV-ICP-AES Particle size/mm 280–154 154–105 105–90 90–74 <74 Recovery (%)a 74.8±8.9 98.6±4.7 101.5±3.1 100.0±2.1 99.5±1.9 aMean±s (n=3). Table 5 EVect of particle size on the recovery of Co in SiO2 determined by ETAAS with PTFE Particle size/mm >300 300–150 150–125 125–97 97–74 <74 Recovery (%)a 80±19 87±9.5 91±5.9 96±4.1 98±2.8 100±2.5 aMean±s (n=3). 1622 J. Anal. At. Spectrom., 1999, 14, 1619–1624Table 6 Gibb’s free energy (kJ ) changes of chemical reactions Reaction temperature/°C Chemical reaction 500 1000 1500 2000 2500 2800 2MoO3(s)+7C(s)�Mo2C(s)+6CO -42.9 -586.7 -1135.7 — — — 2MoO3(s)+3C2F4(g)�2MoF6(g)+6CO -962.1 -1392.2 -1829.3 — — — Mo2C(s)�2Mo(g)+C(s) — — — 694.8 — 467.9 2MoF6(g)+3C(s)�2Mo(g)+3CF4(g) — — 118.8 -348.5 -824.6 — MoF6(g)�Mo(g)+3F2 — — — 1230.7 — 889.6 Cr2O3(s)+3C(s)�2Cr(s)+3CO(g) 386.1 121.8 -140.2 — — — 3Cr2O3(s)+13C(s)�2Cr3C2(s)+9CO(g) 972.6 152.6 -668.6 — — — 2Cr2O3(s)+C2F4(g)�4CrF3(g)+6CO(g) -1883.8 -2468.1 -3075.3 — — — Cr3C2(s)�3Cr(g)+2C(s) — — 528.6 330.9 136.3 20.8 2CrF3(g)�2Cr(g)+3F2(g) — — 1877.1 786.6 435.2 402.4 4CrF3(g)+3C(s)�4Cr(g)+3CF4(g) — — 132.3 16.1 -106.0 — 2Yb2O3(s)+3C2F4(g)�4YbF3(s)8.4 — — — 2Yb2O3�4Yb(g)+3O2 — — — 2026.5 1569.8 1298.3 2YbF3�2Yb(g)+3F2 — — — 1916.8 1779.5 1697.4 2YbF3(g)+3C(s)�2Yb+3CF2(g) — — — 1023.6 699.8 501.9 2CdO(s)+C2F4(g)�2CdF2(s)+2CO -574.5 -689.4 -819.0 — — — 2CdO(s)�2Cd(g)+O2 — 205.8 3.6 -196.3 — — CdF2(g)�Cd(g)+F2 — 472.9 358.0 248.7 143.7 82.4 CdF2(g)+C(s)�Cd(g)+CF2(g) — — 120.3 -1.5 -124.7 — could be written as follows: The fluorides of Cr and Cd have suitable volatilities and moderate stabilities, and can be eYciently atomized at tempera- PTFEAR–CF2–CF2–CF2VAC2F4 (g) + (at 415 °C) tures lower than 2500 °C.Therefore, the introduction of PTFE as a modifier is a benefit for Cr and Cd in ETAAS MOx(s)+C2F4(g)AMF2x(g or s)+CO(g) determination. where M represents the analyte. The reduction/removal of the matrix interference in ETVThe chemical thermodynamic data for relevant chemical ICP-AES and ETAAS by using PTFE as a chemical modifier reactions and physico-chemical data for the analytes and the is also attributable to the fluorination reaction taking place related compounds23 are presented in Tables 6 and 7, respect- between PTFE pyrolytic products and the analytes or matrix ively.These given data support the above reaction mechanism. in the graphite furnace at high temperatures. The formation In ETV-ICP-AES, since the fluorides of Mo, Yb and Cr are of volatile fluoride not only improves the vaporization and more easily evaporated than their metals, oxides or carbides, transport eYciency of the analytes, but also enhances the the formation of fluorides greatly increases their vaporization diVerence in the volatilities between the analytes and matrix eYciency.On the other hand, when the analytes are trans- components. This diVerence facilitates the separation of ported into the ICP torch as fluoride molecules, their transport interferent matrices by selective evaporation. eYciencies are much higher than those when transported as free atoms, because free atoms are easy to deposit while being transported.As a result, the emission signals of Mo, Yb, Cr are enhanced significantly with the addition of PTFE. Conclusions In contrast, with the addition of PTFE the situation in In both ETV-ICP-AES and ETAAS, the fluorination reagent ETAAS was much more complicated. In ETAAS the improve- PTFE can convert the refractory oxides into volatile fluorides ment of the analytical performance depends not only on the in a high temperature graphite furnace and prevent the forma- volatility of the analyte fluorides, but also on the decompotion of refractory carbides.For ETAAS, the graphite furnace sition energy and atomization eYciency of the fluorides. When is not only a vaporizer but also an atomizer, in which the the fluorides (such as MoF6 and YbF3) are easy to vaporize, analyte should be atomized within a short retention time. but diYcult to decompose in a graphite tube atomizer, it will However, for ETV-ICP-AES, the graphite furnace is only a be diYcult to complete their atomization prior to their vaporizvaporizer, in which the analyte is evaporated and then trans- ation in the graphite tube.Therefore, the formation of fluorides ported into the ICP for atomization and excitement. Therefore, significantly decreases the absorption signals of Mo and Yb. the fluorination reaction taking place in the graphite furnace However, this decrease in analytical signal is not observed in has diVerent eVects on the performance of ETV-ICP-AES ETV-ICP-AES, because the ICP torch can oVer a high enough and ETAAS. temperature to ensure the complete decomposition and atomiz- In ETV-ICP-AES, the formation of fluorides can improve ation of these fluorides.the vaporization eYciency of the analytes Mo, Yb, Cr, and reduce memory and matrix eVects. Compared with those in Table 7 Melting and boiling points (°C) of analytes and their conventional ETV-ICP-AES, the detection limits of Mo, Yb, compounds Cr with FETV-ICP-AES determination are a factor of 10–100 lower.For the volatile element Cd, however, the fluorination Metal Oxide Carbide Fluoride has no obvious influence on the analytical sensitivity. Element Mp Bp Mp Bp Mp Bp Mp Bp In ETAAS, the fluorination modifier PTFE is unfavorable for the atomization of the refractory elements Mo and Yb. Mo 2610 4646 801 1155a 2430 — 17.61 33.89 However, PTFE is an eVective chemical modifier for the Yb 824 1430 2372 3227 — — 1157 2230 volatile element Cd and the refractory element Cr, whose Cr 1857 2682 2330 3000 1895 3800 1100 1200a fluorides have a more favorable thermal stability than their Cd 321 770 1497a — — — 1100 1760 oxides or carbides.Obviously, the improvement in ETAAS aSublimation temperature. performance depends not only on the volatility of the analyte J. Anal. At. Spectrom., 1999, 14, 1619–1624 16237 G. Zaray, T. Kantor, G. WolV, Z. Zadgorska and H. Nickel, fluorides, but also on the decomposition energy and Mikrochim.Acta, 1992, 107, 345. atomization eYciency of these fluorides. 8 J. M. Ren and E. D. Salin, Spectrochim. Acta, Part B, 1994, The formation of fluoride can increase the volatility 49, 555. diVerence between the analyte and matrices. This diVerence in 9 J. M. Ren and E. D. Salin, Spectrochim. Acta, Part B, 1994, volatility makes it easy to separate in situ the interferent 49, 567. 10 Z. C. Jiang, B. Hu, Y. Qin and Y. Zeng, Microchem. J., 1996, matrices (such as Si) in both ETV-ICP-AES and ETAAS, and 53, 326.could be used for the direct analysis of solid samples, especially 11 Y. Qin, Z. C. Jiang, Y. Zeng and B. Hu, J. Anal. At. Spectrom., for refractory materials (such as silicon dioxide). 1995, 10, 455. 12 Z. C. Jiang, B. Hu, M. Huang and Y. Zeng, Xitu, 1993, 15(6), 23. 13 Y. Qin, Z. C. Jiang, B. Hu and Y. Zeng, Fenxi Kexue Xuebao, 1995, 11(2), 14. Acknowledgements 14 T. Y. Peng and Z. C. Jiang, Anal. Sci., 1997, 13, 595. 15 T. Y. Peng and Z. C. Jiang, Fresenius’ J. Anal. Chem., 1998, This work was supported by the National Natural Science 360, 43. Foundation and the Education Ministry Foundation of China. 16 T. Y. Peng, Y. Qin, Z. Y. Fan and Z. C. Jiang, J. Rare Earths, 1997, 15(2), 149. 17 F. Y.Wang and Z. C. Jiang, J. Anal. At. Spectrom., 1998, 13, 539. 18 F. Y.Wang and Z. C. Jiang, Anal. Chim. Acta, 1999, 391, 89. References 19 F. Y. Wang and Z. C. Jiang, Fenxi Kexue Xuebao, 1999, 15(2), 1 S. J. Hill, J. B. Dawson, W. J. Price, I. L. Shuttler, C. M. M. Smith 111. and J. F. Tyson, J. Anal. At. Spectrom., 1997, 12, 327R. 20 B. Raeymaekers, T. Graule and J. A. C. Broekeart, Spectrochim. 2 J. S. Crighton, B. Fairman, J. Haines, M. W. Hinds, S. M. Nelms Acta, Part B, 1988, 43, 932. and D. M. Penny, J. Anal. At. Spectrom., 1997, 12, 509R. 21 C. Bendico and M. T. C. de Loos-Vollebregt, Spectrochim Acta, 3 D. E. Nixon, V. A. Fassel and R. N. Kniseley, Anal. Chem., 1974, Part B, 1990, 45, 695. 46, 210. 22 F. Y. Wang, Z. C. Jiang and T. Y. Peng, J. Anal. At. Spectrom., in the press. 4 M. Huang, Z. C. Jiang and Y. E. Zeng, Guangpuxue Yu 23 Lange’s Handbook of Chemistry, ed. J. A. Dean, McGraw-Hill, Guangpufenxi, 1994, 14(6), 61. New York, 13rd edn., 1985. 5 B. Hu, Z. C. Jiang, Y. Qin and Y. Zeng, Anal. Chim. Acta, 1996, 319, 255. 6 T. Kantor and G. Zaray, Fresenius’ J. Anal. Chem., 1992, 342, 927. Paper 9/04639I 1624 J. Anal. At. Spectrom., 1999, 14, 1

ISSN:0267-9477    

DOI:10.1039/a904639i

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

INTER-LABORATORY NOTE Determination of (ultra)trace amounts of lead in biological materials by on-line coupling flow injection microcolumn separation and preconcentration to electrothermal atomic absorption spectrometry using a macrocycle immobilized silica gel sorbent Xiu-Ping Yan,*† Michael Sperling and Bernhard Welz‡ Department of Applied Research, Bodenseewerk Perkin-Elmer GmbH, D-88647 U� berlingen, Germany Received 18th June 1999, Accepted 27th July 1999 A fully automated procedure was developed for the determination of (ultra)trace lead in biological materials by on-line coupling flow injection (FI ) microcolumn separation and preconcentration with electrothermal atomic absorption spectrometry (ETAAS) using a macrocycle immobilized silica gel sorbent (Pb-02).The analyte was selectively and eYciently collected on a conically shaped column (50 ml ) packed with Pb-02 over a wide range of sample acidity (0.08–3 mol l-1 HNO3). Quantitative elution of the retained analyte from the column was achieved with 46 ml of 0.03 mol l-1 ethylenediamine tetraacetic acid (EDTA) solution at pH 10.5.The eluate was driven with an air flow into the graphite tube preheated to 110 °C. No precise timing was needed during analyte elution and eluate introduction. When 0.15 mol l-1 HNO3 was used as the wash medium and the residual solution was removed from the column and connecting tubes by air before elution, the only potential interferents were found to be Ba(II ), Sr(II) and K(I) due to their competition for the cavity of the macrocyclic compound because their ionic radii are similar to that of Pb(II).Nevertheless, these potential interferences were eliminated or minimized by a proper increase of eluent volume and/or EDTA concentration. Under the optimized conditions, the tolerated concentrations of Ba(II), Sr(II) and K(I) were at least 10, 100 and 5000 mg l-1 in the digest, respectively. With a sample loading rate of 3 ml min-1 and a 20-s preconcentration time, an enhancement factor of 23 and a sampling frequency of 23 h-1 with a collection eYciency of 70% were obtained.The detection limit (3s) was found to be 2 ng l-1. The relative standard deviation (n=9) was 2.9% at the 500 ng l-1 Pb level. The results for a number of standard reference materials (rice flour, blood and urine) demonstrated the applicability of the proposed method to the analysis of biological materials with simple aqueous standards for calibration.fully automated sample management and operation in a closed Introduction system. The on-line coupling of FI preconcentration and The lead content in biological materials such as blood and separation technique with ETAAS has been proved to be a urine provides an important basis for the diagnosis of clinical powerful technique for (ultra)trace determination of a variety disorders, of intoxication and for the monitoring of environ- of elements.5–9 Up to now, several techniques have been mental pollution.The low content of lead in biological mate- adapted to FI on-line separation and preconcentration ETAAS rials necessitates the use of highly sensitive techniques for its systems, including solid sorbent extraction,10–19 solvent extracdetermination. ETAAS is one of the most sensitive techniques, tion,20–23 coprecipitation,24–26 electrodeposition27 and sorption with low volumes of sample required for a large number of in knotted reactors (KRs).28–33 elements with detection limits in the mg l-1 to ng l-1 range, The majority of work conducted in this area has been associated with preconcentration on packed columns.5–19 and therefore it is often chosen for such determinations.1–3 In Compared with ion-exchangers as column packing materials, ‘real’ samples of complex composition, however, matrix eVects bonded-silica with octadecyl functional groups (RP C18) as can have a significant influence on the performance of ETAAS.sorbent, and dithiocarbamate (DTC) derivatives as com- While interferences may be overcome, to various degrees, by plexing agents, were shown to be more selective and less prone applying the ‘stabilized-temperature platform furnace’ (STPF) to interferences for the determination of trace heavy metals in concept,4 separation of the analyte element from the matrix is sea-water since the most common ions in sea-water, alkali and undoubtedly even more eVective in avoiding matrix eVects. alkaline earth elements, do not form complexes with DTC and Flow injection (FI ) as a microsample introduction system can be quantitatively separated from the other metals.5–9 oVers some distinct advantages over manual batch-type pro- However, the determination of trace metals using the DTC-RP cedures for analyte preconcentration and separation, such as C18 preconcentration system suVers interferences from high concentrations of other transition metals in the sample due to †Present address: Department of Geological Sciences, University the competition of their DTC complexes for active sites at the of Saskatchewan, Saskatoon, SK, S7N 5E2, Canada.column packing. The use of more selective chelating agents, E-mail: xiu-ping.yan@usask.ca such as diethyldithiophosphate derivatives,11,29 is one way to ‡Present address: Departamento de Quimica, Universidade Federal solve this problem. Another solution is the application of de Santa Catarina, 88040–900 Florianopolis, S.C., Brazil.E-mail: welz@qmc.ufsc.br highly selective sorbents.6,34 J. Anal. At. Spectrom., 1999, 14, 1625–1629 1625Recently, we applied a macrocycle immobilized silica gel Operational procedure sorbent (Pb-02) to FI on-line microcolumn preconcentration The FI manifold and details for the sequence of operation for the ETAAS determination of ultra-trace amounts of lead were the same as given in Fig. 1 and Table 2, respectively, of in high purity reagents.34 The interferences from other heavy the previous report,34 except for the sample loading time of metal ions, which were encountered in DTC-RP C18 systems, 20 s used in this work.A complete cycle of preconcentration were overcome using this highly selective sorbent. Howand elution without a pre-fill stage required 112 s with a ever, interferences from Ba(II), Sr(II) and K(I) were observed sample loading time of 20 s. when their concentrations exceeded 1, 10 and 100 mg l-1, respectively.34 The objectives of this work were to make a further investi- Results and discussion gation into the potential interferences and their elimination or Optimization of the FI on-line preconcentration system reduction in a recently developed FI on-line column preconcentration and separation system for ETAAS,34 and to evaluate The factors aVecting the analyte sorption (sample acidity, the possibility of applying this system for interference-free sample loading time and loading rate), the column wash (wash determination of lead in biological materials.medium and duration, etc.), the analyte elution (EDTA concentration, and pH, etc.), and the ETAAS determination (preheating temperature, drying temperature and pyrolysis Experimental temperature) were investigated for the optimization of the Apparatus present system. The optimum range and the selected conditions are summarized in Table 1. The atomic absorption spectrometer, the graphite furnace Although the specific selectivity of the macrocyclic com- temperature programme for the determination of lead and the pound could achieve eVective separation of the analyte from flow injection system with the conically shaped microcolumn matrix constituents, as for the analysis of high purity (ca. 50 ml ) packed with Pb-02 used in this work were the same reagents,34 a column wash step was also necessary for the as described in the previous publication,34 except for the determination of lead in biological materials, since introduc- preheating temperature of 111°C used in this work.tion of the residual matrix from the column and connecting tubing into the graphite atomizer could cause interferences in Reagents and standard solutions the subsequent ETAAS determination. Fig. 1 shows the eVect The reagents and standardolutions used in this work were of the duration of the column wash with 0.15 mol l-1 HNO3 the same as described in the previous publication,34 except for at a flow rate of 3 ml min-1 on the signal of lead in spiked the preheating temperature of 110 °C used in this work.whole blood, indicating that a minimum column wash time of 15 s was required for quantitative recovery of the spiked Samples analyte. The following biological reference materials were analyzed to demonstrate the capability of the proposed method for Interferences and their elimination the analysis of biological samples: SRM-1568 Rice Flour Macrocyclic compounds are uncharged and contain a cavity (National Bureau of Standards,Washington, USA), Seronorm in which a cation can be encapsulated, which gives them highly Whole Blood I 205052, Seronorm Whole Blood 902, Seronorm selective complexing abilities for metal ions.Their respective Urine 009024 (Nycomed Pharma AS, Oslo, Norway). cation selectivity depends mainly on the following factors:35 (a) relative size of the ion and the cavity of the macrocyclic Sample pretreatment compound; (b) type, number and location of hetero atoms (binding sites); (c) conformational flexibility of the ring in the A 0.5 ml volume of whole blood or urine or 0.5 g of rice flour were digested with 5 ml of 65% m/v nitric acid in a sealed macrocyclic compound; (d) electrical charge of the ion.Change or modification of the above factors, therefore, can result in PTFE vessel using a Model MDS-81D microwave digestion system (Kuerner Analysentechnik, Rosenheim, Germany).All diVerent selectivities. On the other hand, it can be inferred that the most likely interfering ions are those with a size instrumental parameters were chosen according to the manufacturer’s recommendations. The clear digest was transferred similar to that of the analyte ion. Experimental results for the evaluation of potential inter- into a 25-ml calibrated flask and diluted to volume with doubly de-ionized water (18 MV cm-1). ferences are shown in Table 2, which were observed using 36 ml Table 1 Optimum range and selected conditions for the FI on-line column preconcentration and separation system for ETAAS Parameter Optimum range Selected value Sample acidity/mol l-1 HNO3 0.08–3 0.15 Sample loading time/s 10–180a 20 Sample loading rate/ml min-1 1.4–10.8b 3 Concentration of HNO3 for column wash/mol l-1 0.08–0.45 0.15 Duration for column wash/s 15 20 Concentration of EDTA for elution/mol l-1 0.02–0.10 0.03 pH of EDTA for elution 10.3 10.5 Eluent volume/ml 31–56 46 Elution rate/ml min-1 0.6–2.3 1.4 Preheating temperature/°C 100–120 110 Drying temperature/°C 130–170 160 Pyrolysis temperature/°C 500–700 500 aThe range giving linear relationship between sensitivity and sample loading time.bThe range giving linear relationship between sensitivity and sample loading rate. 1626 J. Anal. At. Spectrom., 1999, 14, 1625–1629Fig. 1 Influence of wash time on the recovery of 0.2 mg l-1 Pb spiked Fig. 2 EVect of eluent volume on the recovery of 0.5 mg l-1 Pb in the to Seronorm Whole Blood I 205052. 46 ml of 0.03 mol l-1 EDTA presence of diVerent concentrations of K (g l-1): (a) 0; (b) 0.1; (c) 1; solution (pH 10.5) used as the eluent. (d) 5. Eluent: 0.03 mol l-1 of EDTA solution (pH 10.5). Table 2 EVect of coexisting ions on the recovery of lead at the minimum eluent volume required for quantitative elution of 0.5 mg l-1 level with the use of 36 ml of 0.03 mol l-1 EDTA (pH 10.5) the retained analyte increased with increasing K(I) concen- as the eluent tration.Compared with a simple aqueous solution [Fig. 2(a)], the presence of 5000 mg l-1 of K(I) necessitated a 15 ml larger Interfering ion Concentration/mg l-1 Recovery (%) volume of 0.03 mol l-1 EDTA solution (pH 10.5) for quanti- Fe(III ) 5000 98 tative elution [Fig. 2(d)]. When the volume of 0.03 mol l-1 Cu(II) 5000 99 EDTA solution (pH 10.5) was increased from 36 to 46 ml, the Ca(II) 5000 96 tolerated concentration of K(I) increased from 100 mg l-1 to Ni(II ) 5000 102 at least 5000 mg l-1, the highest concentration investigated.Zn(II) 1000 97 The influence of the EDTA concentration used for elution La(III ) 5000 102 Al(III) 5000 104 on the recovery of lead in the presence of 1000 mg l-1 of K(I) Mg(II) 5000 98 was investigated as a function of eluent volume. As is shown Na(I) 10 000 98 in Fig. 3, the minimum volume of EDTA solution required K(I ) 100 97 for quantitative analyte elution decreased as the EDTA concen- 500 66 tration increased.Similar results were obtained in the presence Sr(II) 10 98 of 10 mg l-1 of Ba(II) and 100 mg l-1 of Sr(II ) (not shown in 100 68 Ba(II ) 1 102 the figures). These results indicate that the interferences in the 10 71 presence of Ba(II), Sr(II) and K(I) resulted from incomplete elution if a too small volume or an insuYcient concentration of EDTA was used, owing to the change of the analyte of 0.03 mol l-1 EDTA (pH 10.5) for elution.As in the previous distribution over the column, rather than from analyte loss study,34 interferences occurred in the presence of more than due to their competition for the cavity of the macrocyclic 100 mg l-1 of K(I), 10 mg l-1 of Sr(II) and 1 mg l-1 of Ba(II ), compound at the sample loading stage. This seems to be in whereas no interferences were found for Fe(III ), Cu(II), Ni(II), disagreement with the results obtained in the FI column Zn(II), Al(III ), Mg(II), La(III ) and Na(I), even at very high preconcentration for a flame AAS system,37 where analyte loss concentrations (5000 mg l-1). Among the coexisting ions due to the competition from 1 mg l-1 of Ba(II), >10 mg l-1 investigated, Ba(II), Sr(II) and K(I) have an ionic radius of Sr(II) or 100 mg l-1 of K(I) was observed with the same similar to that of Pb(II) (about 1.4 A° ).36 It was found that the column as in this work (50 ml, conical ), which was ascribed smaller the diVerence in the radius between Pb(II ) and the to the insuYcient column capacity.37 Such apparent disagreeinterfering ion, the more serious the interference.The ment might, however, be caused by the 400-fold diVerence in interferences from Ba(II), Sr(II) and K(I), therefore, could be attributed to their competition for the cavity of the macrocyclic compound on the column. In a search for ways of eliminating or minimizing the interferences from Ba(II), Sr(II) and K(I), a preliminary attempt was made by changing sample acidity.However, this was not successful because the optimum ranges of sample acidity for the retention of these interfering ions were found to be even wider than that for the retention of Pb(II). In recent work, using the same FI column preconcentration system for flame AAS showed that the presence of Ba(II), Sr(II) and K(I) was shifting in time and broadening the analyte absorbance signal.37 This result encouraged us to investigate how the analyte recovery in the presence of these interfering ions changed with eluent volume and EDTA concentration in order to see if the increase of eluent volume and/or EDTA concentration could improve the tolerated concentrations of Fig. 3 Influence of the EDTA concentration on the eluent volume the interferents. required for complete recovery of 0.5 mg l-1 Pb in the presence Fig. 2 depicts the eVect of the eluent volume on the recovery of 1000 mg l-1 K.EDTA solutions (pH 10.5) used as the eluent (mol l-1): (a) 0.03; (b) 0.06; (c) 0.1. of lead in the presence of diVerent concentrations of K(I). The J. Anal. At. Spectrom., 1999, 14, 1625–1629 1627Table 3 Characteristic performance data for the on-line microcolumn Analysis of biological samples separation and preconcentration ETAAS system To evaluate the feasibility of the present system for inter- Working range/mg l-1 0.01–1.5 ference-free determination of lead in biological materials, an Preconcentration time/s 20 overview of typical concentrations of lead and the potential Sampling frequency/h-1 23 interferents, Ba(II), Sr(II) and K(I) in human body fluids and Enhancement factora 23 tissues is made in Table 4.Considering the concentration levels Sample consumption/ml 2 of lead in body fluids and tissues, lead in a final sample Reagent consumption/ml solution of 25 ml, prepared from the digest of 0.1 g of the tissue 0.15 mol l-1 HNO3 1 0.03 mol l-1 EDTA (pH 10.5) 0.046 or 0.5 ml of the fluid, can easily be detected by the proposed Extraction eYciencyb (%) 70 method under the conditions used in this work.A tolerant Calibration equation (5 standards, Aint=0.000 74+0.084CPb concentration of 100 mg l-1 K(I ) in the digest corresponds to n=3; CPb in mg l-1) 25 mg g-1 K(I) in the original tissues or 5000 mg l-1 K(I) in Correlation coeYcient 0.9987 the original fluids, which is much higher than the actual concen- Detection limit (3s)/ng l-1 2 trations of K(I) in body fluids and tissues.In fact, the tolerant Precision (n=9) (%) RSD 2.9 (0.5 mg l-1) concentration of K(I) in the present system was up to at least aCompared with direct injection of 30 ml aqueous solution. bCompared 5000 mg l-1 in the digests. Clearly, the concentrations of Ba(II ) with the total mass of the analyte transported through the column. and Sr(II) in the digests of body fluids and tissues are much lower than the tolerant concentrations, so that no interferences the analyte concentration used in the FI preconcentration, from Ba(II), Sr(II) and K(I) should be expected.which was 0.5 mg l-1 in the case of ETAAS and 200 mg l-1 in To check the accuracy of the proposed method for the the case of flame AAS.37 Thus, in the FI preconcentration determination of lead in biological samples, a number of ETAAS system, even in the presence of higher concentrations biological reference materials were analyzed.It should be of Ba(II), Sr(II) or K(I), the capacity of a 50 ml column was noted that although the maximum acidity of the dilute digest still suYcient to accommodate the ultra-trace amounts of lead. could be as high as 3 mol l-1 HNO3, it was not necessary to Based on the above results and discussion, interferences in the adjust the dilute digest to the same acidity as that of the presence of at least 10 mg l-1 Ba(II ), 100 mg l-1 Sr(II) or standards (0.15 mol l-1 HNO3) since no significant influence 5000 mg l-1 K(I) could be eliminated by properly increasing of sample acidity on the sensitivity was observed in the range the eluent volume and/or the EDTA concentration.of 0.08 to at least 3 mol l-1 HNO3. The analytical results obtained using simple aqueous standard solutions for cali- Analytical performance bration are given in Table 5. An excellent agreement between the result obtained by the proposed method and the certified Characteristic data for the performance of the on-line column value was achieved for the rice flour. For the whole blood and preconcentration system for ETAAS are given in Table 3.For urine samples, the recommended values were stated to be only a sample loading flow rate of 3 ml min-1, with a preconcenpreliminary, and no control ranges were given. The results tration time of 20 s, an enrichment factor of 23 was obtained obtained can therefore only be used to show the general in comparison with the direct injection of 30 ml of an aqueous capability of the proposed procedure.solution. The relative detection limit, which is defined as the analyte concentration that gives an integrated absorbance equivalent to three times the standard deviation of the blank, Conclusions was 2 ng l-1. Compared with the total analyte mass transported to the column, the collection eYciency was 70%. The The results in this work demonstrated the feasibility of the FI on-line column preconcentration and separation ETAAS precision at the 0.5 mg l-1 level was 2.9% RSD, including the entire process of preconcentration, elution and ETAAS system using the Pb-02 macrocycle immobilized silica gel sorbent for interference-free determination of (ultra)trace determination.Table 4 Concentrations of lead and potentially interfering elements in human body fluids and tissues. Data from refs. 38 and 39a Fluid/tissue Unit Pb Ba Sr Kb Brain mg g-1 0.3±0.1 0.006±0.0003 0.08±0.01 (1.7–3.8) Kidney mg g-1 (0.10–1.40) 0.01±0.001 0.1±0.02 (1.3–2.2) Liver mg g-1 (0.19–3.13) 0.01±0.003 0.1±0.03 (1.7–3) Lung mg g-1 (0.0045–0.59) 0.03±0.008 0.2±0.02 (1.2–2.2) Milk mg l-1 (6–30) (2–170) (17–295) — Muscle mg g-1 (0.01–0.23) 0.02±0.006 0.05±0.02 (1.8–5.5) Blood mg l-1 (30–790) 100±60 (20–31) (1200–2000) Urine mg l-1 (8–40) 4.8±1.4 (<0.01–0.03) — Plasma/serum mg l-1 (0.02–5.7) (30–290) (10–70) (<300) Hair mg g-1 (0.1–110.9) (0.121–29.0) (0.75–10.8) — aWet weight, Mean±s; range in parentheses. bData for K estimated from Fig. 2 in ch. 1 of ref. 39. Table 5 Analytical results for lead in biological reference materials Sample Unit Certified Determined (mean±s, n=5) NBS-SRM-1568 Rice Flour mg g-1 0.045±0.002 0.046±0.002 Seronorm TM Whole Blood I 205052 mg l-1 0.041a 0.040±0.002 Seronorm TM Whole Blood 902 mg l-1 0.062a 0.058±0.002 Seronorm TM Urine 009024 mg l-1 0.090a 0.101±0.004 aRecommended values. 1628 J. Anal. At. Spectrom., 1999, 14, 1625–162917 V.Porta, O. Abollino, E. Mentasti and C. Sarzanini, J. Anal. At. amounts of lead in biological materials. The potential inter- Spectrom., 1991, 6, 119. ferences from Ba(II), Sr(II) and K(I) due to their competition 18 Z.-S. Liu and S.-D. Huang, Anal. Chim. Acta, 1993, 281, 185. for the cavity of the macrocyclic compound, observed in 19 L. C. Azeredo, R. E. Sturgeon and A. J. Curtius, Spectrochim. previous work,34 can be eliminated by properly increasing Acta, Part B, 1993, 48, 91.the eluent volume and/or the EDTA concentration. It is 20 K. Backstrom and L. G. Danielsson, Anal. Chem., 1988, 60, 1354. 21 K. Backstrom and L. G. Danielsson, Anal. Chim. Acta, 1990, expected that the developed system is also applicable to the 232, 301. determination of (ultra)trace levels of lead in a variety of 22 S.-L. Lin, Q.-H. Zhao and G.-L. Yu, Fenxi Kexue Xuebao, 1994, environmental samples. 10(1), 24. 23 G.-H. Tao and Z.-L. Fang, Spectrochim.Acta, Part B, 1995, 50, 1747. 24 Z.-L. Fang and L.-P. Dong, J. Anal. At. Spectrom., 1992, 7, 439. References 25 L.-P. Dong and Z.-L. Fang, Fenxi Shiyanshi, 1992, 11(6), 9. 1 B. Welz and M. Sperling, Atomic Absorption Spectrometry, Wiley- 26 H.-W. Chen, S.-K. Xu and Z.-L. Fang, J. Anal. At. Spectrom., VCH, Weinheim, Germany, 3rd edn; 1999. 1995, 10, 533. 2 D. T. Miller, D. C. Paschal, E. W. Gunter, P. E. Stroud and 27 E. Beinrohr, M. Rapta, M. L. Lee, P. Tschopel and G.To� lg, J. D’Angelo, Analyst, 1987, 112, 1701. Mikrochim. Acta, 1993, 110, 1. 3 J. A. Navarro, V. A. Granadillo, O. E. Parra and R. A. Romero, 28 M. Sperling, X.-P. Yan and B. Welz, Spectrochim. Acta, Part B, J. Anal. At. Spectrom., 1989, 4, 401. 1996, 51, 1891. 4 W. Slavin, D. C. Manning and G. R. Carnick, At. Spectrosc., 29 X.-P. Yan and F. Adams, J. Anal. At. Spectrom., 1997, 12, 459. 1981, 2, 137. 30 X.-P. Yan, W. Van Mol and F. Adams, Lab. Rob. Autom., 1997, 5 Z.-L.Fang, Flow Injection Separation and Preconcentration, VCH, 9, 191. Weinheim, Germany, 1993. 31 X.-P. Yan, W. VanMol and F. Adams, Analyst, 1996, 121, 1061. 6 B. Welz, Microchim. J., 1992, 45, 163. 32 E. Ivanova, X.-P. Yan, W. VanMol and F. Adams, Analyst, 1997, 7 Z.-L. Fang, Flow Injection Atomic Absorption Spectrometry, John 122, 667. Wiley, Chichester, UK, 1995. 33 E. Ivanova, X.-P. Yan, W. Van Mol and F. Adams, Anal. Chim. 8 Z.-L. Fang, Spectrochim. Acta, Part B, 1998, 53, 1371. Acta, 1997, 354, 7. 9 J. L. Burguera and M. Burguera, Analyst, 1998, 123, 561. 34 M. Sperling, X.-P. Yan and B. Welz, Spectrochim. Acta, Part B, 10 Z.-L. Fang, M. Sperling and B.Welz, J. Anal. At. Spectrom., 1990, 1996, 51, 1875. 5, 639. 35 K. Kimura and T. Shono, J. Liq. Chromatogr., 1982, 5 (Suppl. 2), 223. 11 R.-L. Ma, W. Van Mol and F. Adams, Anal. Chim. Acta, 1994, 36 CRC Handbook of Chemistry and Physics, ed. D. R. Lide, CRC 293, 251. Press, Boca Raton, FL, USA, 74 edn., 1993–1994,p. 12–8, 9. 12 B. Welz, M. Sperling and X.-J. Sun, Fresenius’ J. Anal. Chem., 37 X.-P. Yan, M. Sperling and B. Welz, Anal. Chem., in the press. 1993, 346, 550. 38 D. L. Tsalev, Atomic Absorption Spectrometry in Occupational and 13 M. Sperling, X.-F. Yin and B. Welz, J. Anal. At. Spectrom., 1991, Environmental Health Practice, vol. II, CRC Press, Boca Raton, 6, 615. FL, USA, 1984. 14 M. Sperling, X.-F. Yin and B. Welz, Spectrochim. Acta, Part B, 39 G. Venkatesh Iyengar, Elemental Analysis of Biological System, 1991, 46, 1789. Vol. I, CRC Press, Boca Raton, FL, 1989, p. 3. 15 M. Sperling, X.-F. Yin and B. Welz, Analyst, 1992, 117, 629. 16 B. Welz, X.-F. Yin and M. Sperling, Anal. Chim. Acta, 1992, 261, 477. Paper 9/04876F J. Anal. At. Spectrom., 1999, 14, 1625–1629 1629

ISSN:0267-9477    

DOI:10.1039/a904876f

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

COMMUNICATION Alkylating vaporisation of antimony using tungsten boat furnace–sample cuvette technique for inductively coupled plasma mass spectrometry Yasuaki Okamoto Department of Chemistry, Graduate School of Science, Hiroshima University, Kagamiyama 1-3-1, Higashihiroshima 739-8526, Japan. E-mail: Yasuaki.Okamoto@sci.hiroshima-u.ac.jp Received 5th July 1999, Accepted 6th September 1999 A low temperature vaporisation method based on methylation its application to the ICP mass spectrometric determination of ultra-trace antimony are reported for the first time.of the analyte was proposed for introducing ultra-trace antimony into an inductively coupled plasma for mass spectrometric detection. Following the dosing of the aqueous sample and Experimental application of a drying process to remove all water, the resulting Instruments and reagents residue was reacted with a solution of methyllithium in diethyl ether to permit the release of the volatile methylantimony species A vaporiser head for the alkylmetal generation device, equipped at vaporisation temperatures as low as 650 °C.As a result, the with a TBF (10×70 mm), was made by modifying a Seiko II quantitative generation and introduction of the analyte antimony (Chiba, Japan) Model SAS-705V metal furnace atomiser head, could be achieved, permitting its complete separation from not originally manufactured for an electrothermal AAS. Sample only the matrix but also the methylating agent itself.The cuvettes were shaped by cutting the TBFs. This modification detection limit was estimated to be 0.1 pg of antimony, which was similar to the sample cuvette–TBF device for direct solid corresponds to 1.0 pg ml-1 of the antimony concentration when sampling-ICP-AES, and has been described in the literature.12 a sample injection volume of 100 ml is applied. The relative Electronic current was supplied from a Seiko II Model EV-300 standard deviation for 8 replicates of standards each containing metal furnace vaporiser unit, which was attached to a Seiko II 20 pg of antimony was calculated to be 2.8%.Model SPQ9000 ICPmass spectrometer. Both the vaporiser unit and the ICP mass spectrometer were controlled synchronously with an MS-DOS/Windows 3.1 computer. Introduction A 1.4 mol l-1 solution of methyllithium in diethyl ether solution (Aldrich Chemical, Milwaukee, WI, USA) was used Over the last few years it has become clear that chemical as received.Standard solutions were prepared by diluting a modifiers are as important to electrothermal vaporisation- 1000 mg l-1 antimony(III ) stock solution (Kanto Chemical, inductively coupled plasma mass spectrometry (ETV-ICP-MS) Tokyo, Japan) with 0.14 mol l-1 nitric acid. as they are to conventional graphite furnace atomic absorption spectrometry (AAS).1,2 Many chemical modifiers have been Recommended procedure used in ETV-ICP-MS for a vast array of analytes and matrices.First, up to 100 ml of an aliquot of an aqueous sample solution Regarding antimony, Fairman and Catterick have described containing less than 200 pg of antimony were pipetted into complex interactions between the analyte antimony and varithe sample cuvette. As the drying stage, these cuvettes were ous chemical modifiers. According to their results, a mixed placed for several minutes, typically about 5 min, on a hot modifier of Pd(NO3)2–Mg(NO3)2–ascorbic acid was preferplate kept at 150 °C.After one of the cuvettes had been able.3 It allowed the use of higher ashing temperature, 1000 °C, superposed on the TBF and 10 ml of the methyllithium solution and thus oVered eVective matrix removal prior to the vaporisinjected into it through an insertion port of the glass dome, ation and introduction of the analyte into the plasma. the TBF was maintained at 50 °C for 50 s to expel the solvent This communication describes a modifier based on a new of methyllithium through the insertion port. Then the port concept, i.e., the analyte is selectively transferred to the ICP ion was closed with a rubber stopper and the temperature was source, the matrix being retained throughout the ETV.With an taken up to 650 °C for vaporisation. A momentary cloud ICP-atomic emission spectrometer, the preliminary approach containing the antimony species was generated and immedi- was examined for zinc and gallium, i.e., dibutylzinc was generately transported into the ICP ion source through a PTFE ated by a reaction with butyllithium and ethylgallium was also tube (4 mm i.d.×45 cm long) by argon carrier gas volatilised by ethylation with Grignard reagent.4,5 In this experi- (1.0 l min-1).The transient signal of 121Sb was integrated and ment, a low-temperature sample introduction method for the peak area was estimated with a computer. Fig. 1 shows ICP-MS, based on alkylation of antimony by using methylliththe sampling procedure and schematic diagram of the sample ium, was proposed for the ultra-trace determination of antimony.cuvette–tungsten boat furnace vaporiser. For this purpose, a tungsten boat furnace (TBF)–ETV device, together with sample cuvettes, was utilised as a thermochemical reactor to undertake the methylation of antimony. The analytical Results and discussion performance of the method was also demonstrated. Optimisation of conditions Regarding the alternative generation of volatile antimony species, the use of sodium tetrahydroborate for the production In order to introduce antimony into the ICP at a relatively of antimony hydride and its eVective transport to the ICP ion low temperature, the conversion of inorganic antimony into a source has been reported for the sensitive method of ICP mass methyl derivative was necessary.Methyllithium was selected spectrometric detection.6–11 However, the generation of volatile for this purpose. The amount of methyllithium was examined in the range of 0–16 mmol.When no methyllithium was added, antimony species with the alkylating agent methyllithium and J. Anal. At. Spectrom., 1999, 14, 1631–1632 1631a very small, almost negligible, peak was observed. The peak height increased with increase of added amount up to 10 mmol, and remained nearly constant over the range 10–16 mmol. Regarding the vaporisation temperature, a drying step (50 °C for 50 s) was programmed prior to the vaporisation step.During drying, diethyl ether, the solvent of the methyllithium, was allowed to escape through the open cuvette insertion port. After the port was closed with a silicone rubber stopper, the cuvette was then maintained at the vaporisation temperature to generate the methylantimony species. The antimony signal appeared at 500 °C. The signal increased with increasing the vaporisation temperature up to 650 °C, remained constant at 650–700 °C, and then decreased.Further temperature elevation over 700 °C caused an evaporation of lithium, whose introduction into the torch resulted in a deterioration in the reproducibility of the method. Therefore, 650 °C was selected as the optimum vaporisation temperature. At this temperature no lithium emission was observed with an ICP atomic emission spectrometer. Direct sampling on the furnace was not recommended, because the residue of methyllithium remained on the surface after each firing.The residue could be washed easily by soaking Fig. 1 Sampling procedure and schematic diagram of the sample the sample cuvettes in dilute nitric acid, although it was cuvette–tungsten boat furnace vaporisation method. A, Argon gas diYcult to remove it by conventional bake out. inlet port; B, bakelite plate; C, O-ring; D, tungsten boat furnace; E, sample cuvette; F, furnace electrode; G, quartz dome; H, sample Basic analytical performance characteristics cuvette insertion port; I, silicone rubber stopper; J, port from which argon flows to the plasma torch; and K, digital micropipette. Since the vaporisation temperature was relatively low compared with that of the conventional ETV technique, few foreign anions during passing through skimmers.No lithium species were and cations especially metal ions were vaporised and introduced introduced at that vaporisation temperature. From the point into the ICP. The reduced interference was also attributed to the of view of few co-products, methyllithium is preferable to strongly basic methyllithium medium, in which almost all metal Grignard reagents, which would degrade to produce poly- ions were formed into their hydroxide salts.Excepting Sr2+, atomic species such as ArCl+, ArBr+, ClO+, etc. Further Cr3+, Ni2+ and SiO32-, all the foreign ion species tested were work is now focusing on applications to practical samples. tolerated at a 100051 weight ratio against 40 pg of Sb3+. In the Since the generated organometallic species is immediately presence of Sr2+, Cr3+, Ni2 +, or SiO3 2-, the tolerance limits introduced into the detector, the proposed method could be were restricted to 700–80051 weight ratios. applicable to the other analyte metals that might produce The 3s detection limit of the proposed method was estimated unstable or moisture-sensitive compounds. to be 0.10 pg of antimony, which corresponds to 1.0 pg ml-1 of the antimony concentration when a sample injection volume of 100 ml is applied.With a larger sample volume, a pro- References portionally lower detection limit could be attained. For this 1 R. D. Ediger and S. A. Beres, Spectrochim. Acta, Part B, 1992, 47, purpose, repeated injections would be eVective. The detection 907. limit was better than that of the published paper.3 The relative 2 I. Havezov, A. Detcheva and J. Rendl, Mikrochim. Acta, 1995, 119, standard deviation obtained for 8 replicate measurements of 147. 3 B. Fairman and T. Catterick, J. Anal. At. Spectrom., 1997, 12, 863. 20 pg of antimony(III) was estimated to be 2.8%. A linear 4 S. Tao,Y. Okamoto and T. Kumamaru, Anal. Sci., 1995, 11, 319. calibration graph for antimony intersecting the origin of the 5 T. Kumamaru, S. Tao, M. Uchida and Y. Okamoto, Anal. Lett., coordinate axis and covering absolute amounts of up to 200 pg 1994, 27, 2331. was established. By using a number of exchangeable small 6 P. Smichowski, Y.Madrid, M. Beatriz de la Calle Guntin�as and C. sample cuvettes, sample throughput can be facilitated. Ca�mara, J. Anal. At. Spectrom., 1995, 10, 815. Approximately 40 batches could be vaporisable per hour. 7 L. S. Zhang and S.M. Combs, J. Anal. At. Spectrom., 1996, 11, 1043. 8 G. E.M.Hall and J.-C. Pelchat, J. Anal. At. Spectrom., 1997, 12, 97. In conclusion, it has been demonstrated that the generation 9 G.E.M.Hall andJ.-C. Pelchat, J.Anal.At. Spectrom., 1997, 12, 103. of alkylated volatile species of antimony (probably SbMe3), 10 J. Bowman, B. Fairman and T. Catterick, J. Anal. At. Spectrom., using methyllithium as the alkylating agent, can be applied to 1997, 12, 313. the determination of trace levels of inorganic antimony(III, V) 11 S. J. Santosa, H. Mokutani and S. Tanaka, J. Anal. At. Spectrom., in aqueous samples. Because of strong basicity of the reaction 1997, 12, 409. media and the relatively low vaporisation temperature, matrix 12 Y. Okamoto, N. Takagaki, T. Fujiwara and T. Kumamaru, Anal. Commun., 1997, 34, 283. metal ions are not vaporisable. As for the degradation products from methyllithium, although hydrocarbons accompanied the methylantimony, the molecular components were exhausted Paper 9/05374C 1632 J. Anal. At. Spectrom., 1999, 14, 1631–

ISSN:0267-9477    

DOI:10.1039/a905374c

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

ATOMIC SPECTROMETRY UPDATE Atomic mass spectrometry JeVrey R. Bacon,*a JeVrey S. Crain,b Luc Van Vaeckc and John G. Williamsd aThe Macaulay Land Use Research Institute, Craigiebuckler, Aberdeen, UK AB15 8QH. E-mail: j.bacon@mluri.sari.ac.uk bUnion Carbide Corporation, Technical Center, PO Box 8361, South Charleston, WV25303, USA cMicro- and Trace Analysis Centre, Department of Chemistry, University of Antwerp, Universiteitsplein 1, B-2610 Wilrijk, Belgium dNERC ICP-MS Facility, Centre for Earth and Environmental Science Research, School of Geological Sciences, Kingston University, Penrhyn Road, Kingston upon Thames, Surrey, UK KT1 2EE Received 6th July 1999 1 Accelerator mass spectrometry (AMS) 8 Stable isotope ratio mass spectrometry (SIRMS) 8.1 Reviews 1.1 Small accelerators 1.2 Accelerator SIMS 8.2 Instrumentation 8.3 Analytical methodology 1.3 Developments in radiocarbon analysis 1.4 Developments in the analysis of elements other 8.4 Sample preparation 9 Thermal ionization mass spectrometry (TIMS) than carbon 2 Glow discharge mass spectrometry (GDMS) 9.1 Instrumentation 9.2 Negative ionization procedures 2.1 Review and fundamental studies 2.2 Instrumentation 9.3 Positive ionization procedures 10 Other methods 2.3 Analytical methodology 3 Inductively coupled plasma mass spectrometry 10.1 Electrospray mass spectrometry (ESMS) and ion spray mass spectrometry (ISMS) (ICP–MS) 3.1 Fundamental studies 10.2 Fast atom bombardment mass spectrometry (FABMS) 3.2 Instrumentation 3.3 Sample introduction 10.3 Gas chromatography-mass spectrometry (GC-MS) 3.3.1 Introduction 3.3.2 Laser ablation 10.4 Noble gas mass spectrometry 10.5 Spark source mass spectrometry (SSMS) 3.3.3 Thermal vaporization 3.3.4 Chemical vaporization 10.6 New methodologies 11 References 3.3.5 Nebulization 3.3.6 Flow injection 3.3.7 Chromatography and electrophoresis T he format of this year’s Update follows that used in last 3.4 Interferences 3.5 Isotope ratio measurement year’s1 with some minor changes in the section headings.Although an attempt is made to consider all relevant refereed 4 Laser ionization mass spectrometry (LIMS) 4.1 Fundamental studies papers, conference abstracts, reports, book chapters and patents for inclusion, this review does not aim at being 4.2 Instrumentation 4.3 Analytical methodology comprehensive in its coverage. T he selection of papers is based on criteria applied to focus sharply on the most significant 5 Resonance ionization mass spectrometry (RIMS) developments reported during the period (approximately corresponding to 1998) covered by this Update.T he prime 6 Secondary ion mass spectrometry (SIMS) 6.1 Instrumentation consideration is that the reports should present advances in instrumentation and methodology or improved understanding of 6.2 Fundamental studies 6.3 Analytical methodology the fundamental phenomena involved in the MS process. As a general rule, conference abstracts are not included because 6.4 Quantification 6.5 Single and multi-dimensional analysis they rarely provide suYcient information to judge whether or not they meet the criteria.We consider it better to wait for full 6.5.1 Depth profiling 6.5.2 Imaging details to appear in a refereed journal. A similar policy applies 6.5.3 Three-dimensional (3-D) analysis to those papers in a language other than English and unlikely 6.6 Multi-technique approaches to reach a wide readership. 6.7 Static SIMS (S-SIMS) Routine applications of atomic MS are not covered in this 7 Sputtered neutral mass spectrometry (SNMS) Update and readers are referred to the Updates on Industrial 7.1 Reviews analysis: metals, chemicals and advanced materials,2 7.2 Analytical methodology Environmental analysis3 and Clinical and biological materials, 7.3 Depth profiling food and beverages.4 A book that can be thoroughly recommended is that edited by Gill5 on Modern Analytical Geochemistry.Chapters on most of the techniques included in *Review co-ordinator, to whom correspondence should be addressed and from whom reprints may be obtained. this Update gave full yet concise descriptions of the underlying J. Anal. At. Spectrom., 1999, 14, 1633–1659 1633principles of the techniques, with examples of applications in the system that should result in improvements. Contamination could be reduced by redesigning the sample holders and geochemistry and environmental science.T he substantial review (269 references) of Becker and Dietze6 covered in detail the sensitivity enhanced through better calibration and beam line realignment. inorganic MS techniques that are used for inorganic trace analysis. The microanalysis of geological samples has been performed using a microbeam Cs ion source designed by Sie et al.10 for T his Update follows the policy set by JAAS in deciding the scope of papers which can be considered for inclusion. With the the AUSTRALIS instrument.An important feature of the design was a high magnification viewing system for real-time increased use of atomic spectroscopy techniques, in particular MS, in speciation studies the scope has been widened to include observation of the sputtering process and tuning of the ion beam. A Cs+ microbeam 30 mm in diameter could be produced not only elemental but also speciation studies, including those in which molecular species are determined as long as the focus routinely with adequate intensity for a number of applications.The eVects of source instabilities during isotope ratio measure- of the study is the element and its chemical form. With the boundaries between atomic and molecular MS becoming less ments could potentially be eliminated through the use of fast sequential injections of isotopes. well defined, the judgement of the authors of this Update becomes important in achieving a correct balance.In general, studies of unstable nuclei are excluded but the determination of 1.3 Developments in radiocarbon analysis radioactive elements in ‘real’ samples are included. T he trends apparent for the various MS techniques are For an overview of the current status of radiocarbon analysis, readers are referred to the special edition of Radiocarbon highlighted in the appropriate sections. In general, more attention continues to be paid to sample preparation and (vol. 40, 1998) which was the Proceedings of the 16th International 14C Conference held in Groningen in 1997.The introduction rather than instrumentation. A common feature for many of the techniques reviewed is the need for analysis of two volumes contained numerous examples of the current developments in and applications of radiocarbon AMS. The smaller samples. T his has presented challenges of reducing the levels of contamination and background noise and also major thrust in 14C AMS has been the development of methods for the analysis of samples as small as a few mg as well as calibration and standardization. T he use of ICP-MS for the measurement of isotope ratios is noted as an increasingly investigations into the limitations of such methods.Potential applications included the analysis of atmospheric aerosols and important application. deep-sea particulate matter in which only small amounts of carbon are available. 1 Accelerator mass spectrometry (AMS) Although samples as small as 20 mg could be analysed using 1.1 Small accelerators conventional methods of preparing graphite targets, Wills et al.11 considered it necessary to develop an on-line preparation There have been relatively few developments in AMS to report, method in which analytes would flow directly and continuously probably because the papers from the most recent of the series from a GC to an AMS instrument capable of accepting gas of international conferences have yet to be published.Last phase samples. An initial investigation of coupling a microwave year’s ASU review highlighted the perceived need for small plasma ion source to an AMS instrument led to the conclusion accelerators for use in the biomedical sciences, in particular that the arrangement was not particularly suitable for coupling for the determination of 26Al. In his review, Suter7 considered on-line with a GC because of the low sample flow rates from in some detail the fundamental requirements of small accelerthe GC column.The method would, however, be suitable for ators and discussed both the potential and limitations. A analysis of gaseous samples derived from, for example, carbonnumber of design problems were presented and the elements ates, for which a 120 mA negative ion current could be achieved listed that could potentially be determined on small instru- from a 10 mg sample. A limitation of the system was the ments. It was concluded that precise radiocarbon dating would tailing of peaks caused by sorption/desorption of sample gas be possible using gas stripping at 0.5 MV. The group are in the ion source when the molecules were released into currently designing such an instrument, which would have an vacuum.It was proposed to overcome this by reducing the overall space requirement of 6×5 m. volume of the ion source and silver plating its surface. Mous et al.8 have designed a CO2 ion source for a compact Other laboratories have modified the batch production of radiocarbon instrument dedicated to biomedical research.The targets for the analysis of small samples. In their analysis of instrument was even more compact than that reported above carbonaceous aerosols containing 10–100 mg C, Weissenbok with a footprint of 2.25×1.25 m. The new ion source featured et al.12 produced CO2 which was reduced to carbon over direct introduction of CO2 as well as the conventional analysis manganese chunks and iron wool to form Fe–C bead targets.of graphite targets. Gaseous samples were transported into These were considered to be more stable than graphite targets. the source in a stream of He and absorbed on the surface of Although results were obtained for small samples, considerable a titanium target, from which they were sputtered and ionized. improvements in the procedure were required to overcome This arrangement oVered the potential for interfacing with an problems of a large scatter in the data.The analysis of samples elemental analyser or gas chromatograph. The source has yet smaller than 50 mg was considered the aim for the future. to be operated in the instrument and initial tests have shown Verkouteren et al.13 also used Fe–C bead targets produced in that although ion currents of 50–100 mA for graphite samples sealed quartz tubes and found that measurements were reprocould be produced, those for gaseous samples were an order ducible and internally consistent for 10–100 mg samples, even of magnitude lower.though the ion currents produced were two orders of magnitude lower than those from graphite targets. Isotopic fraction- 1.2 Accelerator SIMS ation was observed during preparation of samples of 100 mg C unless a reduction time of 15 h was used to maximize The short paper by McDaniel et al.9 on the use of Trace Element AMS (TEAMS) for the analysis of high purity the yield. The group at the NOSAMS facility14,15 have reported materials contained useful information on the background to the technique and the estimation of sensitivity levels and instrument-induced fractionation eVects for samples with ,100 mg C and have adopted a procedure which matched LODs.Although the authors were able to demonstrate the measurement of Cu depth profiles in As-implanted silicon sample and standard weights. The fractionation was sampleweight dependent and arose from coulomb repulsion in ion wafers at the ppb level, they also identified modifications to 1634 J.Anal. At. Spectrom., 1999, 14, 1633–1659beams which resulted in larger beam diameters and divergence for 129I was 2×10-10 Bq l-1. The same group have used a TOF system, fitted with a microchannel plate detector to angles. It follows that the traditional tuning of instruments with 1 mg standard samples is not necessarily appropriate for provide start signals and a silicon surface-barrier detector to provide stop signals.26 A detection sensitivity of 6×10-13 for samples of150 mg.The reduction of CO2 over cobalt catalyst was temperature-dependent and at temperatures above 605 °C the 129I5127I ratio was achieved. The AMS determination of 99Tc, an abundant product of the yield decreased. A major consideration in the analysis of small samples is the nuclear fuel cycle and widely distributed in the environment, presents a number of challenges. These result from the the increased importance of contamination.In a systematic study, Schleicher et al.16 determined that a significant, yet lack of a stable Tc isotope which could be used as yield monitor, chemical carrier and sample matrix. McAninch unidentified, blank of 10 mg C was present in the combustion of organic samples. This needed to be resolved before the et al.27 used Rh as a proxy in their preliminary investigation, which demonstrated good linearity and a sensitivity of method could be used for small samples.Contamination in the preparation of targets from inorganic carbon was negligible ~500 fg. Some further improvements were required to overcome problems identified in the study. To realize a detection for samples containing 1 mg C but became more significant for samples of 0.1 mg C. Van der Plicht et al.17 established limit of ~10 fg would require the total Ru in the prepared sample, including Ru contained in the rhodium carrier, to be that CuO was the major contributor to contamination in the graphitization process and considered the potential combined ,10 pg, corresponding to a Ru concentration in the carrier of ,10 ppb. Interference by 63Cu36S, observed in this study contamination of CuO and Fe to be up to 9 mg C.This was considerably higher than values reported in other studies. when using the 12+ charge state, could potentially be overcome by measurement of the 10+ charge state. Further work Alderliesten et al.18 reported that fractionation eVects were due to a combination of sample-mass-independent contami- will focus on optimization of X-ray production and ionization eYciency. nation introduced during sample preparation and samplemass- dependent fractionation occurring both during graphitization and negative ion formation.They derived a formula to 2 Glow discharge mass spectrometry (GDMS) describe the eVects and a procedure for normalizing results for both small and large samples. Campajola et al.19 observed 2.1 Reviews and fundamental studies a contamination problem specific to an instrument previously used for Be analysis.The presence of 7Be2- contributed A comprehensive review by Bogaerts and Gijbels28 has covered the fundamental plasma phenomena, their mathematical mod- significantly to the background in measurements of 14C. GriYn and DruVel20 reported that the advantages of using elling and experimental evaluation, as well as instrumental aspects and applications.Another paper from the same group a single apparatus, when compared with the use of separate vessels, for the preparation of inorganic and organic carbon has addressed the agreement between the modelling and experimental results in Ar dc GD.29 The modelling results qualitat- from the same marine sample were smaller sample requirement (few mg), reduced contamination and faster preparation. No ively predicted the observations, for example, the relationship between current and voltage, the localization of high and low significant diVerences in analytical data were found by Elder et al.21 for samples prepared by diVerent water-stripping Ar+ density zones throughout the discharge and the crater profiles.Because experiments readily disturbed the plasma methods in a number of laboratories for the analysis of seawater organic and inorganic carbon. No eVect of storage on conditions, modelling was considered an adequate approach for correlating observed features with the fundamental 14C concentration could be detected.The radiocarbon measurement, by Larsen et al.,22 of processes. Harrison30 has reviewed high power pulsed GD, which poten- carbonyl compounds in the atmosphere required samples of up to 103 l of air in order to obtain about 150 mg C, yet the tially has analytical applications not possible with low power dc sources. The increased current and voltage produce more preparation method was able to maintain sample integrity. Samples were drawn through cartridges of silica gel coated energetic species for enhanced sputter and ionization yield.The time separation between the ionization of the background with 2,4-dinitrophenylhydrazine and the carbonyl hydrazones extracted in acetonitrile and purified over activated silica gel. gas and the analyte can be exploited in subsequent MS analysis. In addition, pulsed sources allow use of TOF analysers. The potential of multiple pulse modes was considered. 1.4 Developments in the analysis of elements other than carbon For instance, the delay between an initial sputtering pulse and a subsequent excitation and ionization pulse allowed the The use of AMS for providing missing information on the natural, pre-nuclear, levels of 129I in soils, plants and soft neutrals to drift to a region from which ions could be more easily extracted into the MS. Lasers could be used to advantage tissues is a new area of study and requires new methodologies.The discussion by Michel and colleagues23 was notable for its for the atomization of non-conducting samples with low or high resolution and for the selective ionization of plasma detailed review of 129I in the environment and the systematic study of extraction methods and the sources of analytical species. The analytical use of ion manipulation techniques in ion trap error. The authors found it necessary to construct a new iodine-free laboratory to reduce the contamination levels.They analysers for GDMS and ICP-MS has been reviewed.31 Ejection of matrix ions circumvented space charging and demonstrated the measurement of 129I5127I ratios as low as (7.0±1.5)×10-12, which are lower than those previously detection specificity was increased upon elimination of element or reagent ions, reponsible for interferences (for example, H+, deduced by radiochemical NAA but higher than those predicted from model calculations. H3O+). Ion-molecule reactions with methane in a FTMS cell could convert Os+ into OsCH2+ to remove the contribu- The determination of 129I in environmental matrices, in particular water samples, has received attention in a number of tion of 187Os+ from the 187Re+ signal in the determination of Re.Quadrupole traps allowed use of collision-activated laboratories. The carrier-free method of Martin et al.24 for the preparation of samples from sea-water was based on the dissociation to convert CuO+ and CuOH+ into Cu+.The increased analytical possibilities were believed to compensate reaction of I with metallic silver which was pressed as a cathode in a Cs sputter ion source. Xie et al.25 extracted 129I for the lower sensitivity of traps in comparison with ion transport analysers. Ion traps reach their ultimate eYciency from 100 l water samples by anion exchange chromatography and precipitation of I as AgI. The minimum detectable limit with pulsed ionization. J.Anal. At. Spectrom., 1999, 14, 1633–1659 16352.2 Instrumentation bottom. The average sputtering rate in a ms-pulsed GD was ten times lower than that in a dc GD but the transient rates A two-stage GD ion source for elemental analysis of liquids were a factor of 10 higher. The erosion rate depended on the used an insulated metal capillary inserted into the discharge element and was 2.9, 1.3 and 5.2 nm s-1 for Cu, Fe and Zn, chamber and biased 3 kV against the chamber.32 A miniature respectively.The high transient rate and low average sputtering hollow cathode was formed in the vapour of the liquid rate combined with the negligible crater eVect provided high aspirated at 0.2 ml s-1. The bombardment of the hollow cath- ion densities in the depth profiling of thin layers and their ode by the discharge plasma released the analyte ions for MS interface. analysis. The ion source, fitted to a double-focusing magnetic MS, has been tested using aqueous salt solutions.33 The GD ionization was stable and no deposits were formed in the 3 Inductively coupled plasma mass spectrometry capillary.The absence of argon in the GD source avoided the (ICP-MS) formation of Ar+, associated cluster ions and gas impurity 3.1 Fundamental studies ions. When the source was used with a quadrupole MS, LODs in the 1 ppm range were achieved. A multi-component quasi-equilibrium model of thermochem- The growing interest in elemental speciation has stimulated ical processes has been used to determine various factors the development of hyphenated techniques.A new rf GD ion relating to the behaviour of ions in the ‘cold’40 and convensource for speciation studies with GC sample introduction has tional41,42 inductively coupled plasma. The results were used to used axial sample introduction to simplify the source alignment formulate recommendations for achieving the optimum degree and eliminate the inhomogeneous sampling by discharge asym- of ionization and selection of internal standard to overcome metry.34 In comparison with sample introduction perpendicu- ionization eVects.The model was also discussed with respect lar to the cell axis, axial introduction increased the sensitivity to the possibility of using negative-ion MS analysis. Park and by a factor of 3 for tetramethyltin. In addition, the precision Noh43 used a Langmuir probe to measure apparent dc potenwas improved from 5 to 3.5%.The source suppressed the tials of an ICP at atmospheric pressure and inside a supersonic dimethyltin fragments formed at lower pressures. jet. Compared with a 27.12 MHz plasma, the potential gener- The discharge pressure inside the GD cell must be taken ated at 40.68 MHz was significantly higher and that of a into account when calculating relative sensitivity factors shielded plasma was much lower. A diVerent trend of potential (RSFs). Born et al.35 have developed a pressure meter based variation from plasma to supersonic jet was exhibited with the on a capacitance manometer placed outside the vacuum but shielded plasma.Sharp et al.44 described the various means connected to the cell by an electrically non-conducting by which information could be derived from the ICP signals. capillary. Specifically relevant was the use of correlated measurements as a means of reducing noise and improving precision of 2.3 Analytical methodology isotope ratio measurements, as well as the coupling of capillary electrophoresis to ICP-MS for speciation of CrIII and CrVI and In TOF dc GDMS the intense Ar+ and ArH+ signals from the working gas can saturate the detector.Su et al.36 have Ni–humic acid complexes. Some studies continue on the all important interface of the observed that a discharge pressure of about 300–500 Pa at discharge currents between 15 and 30 mA optimized the ratio ICP-MS system. Douglas and co-workers45 constructed a versatile interface to investigate gas dynamics and flow profiles.of analyte to gas signals. The eVect was attributed to the increased sputtering at high pressure and the changes in the Skimmer design and sampling cone–skimmer spacing were studied. Contrary to current convention, an interface pressure relative contributions of Penning ionization by the metastable Ar, electron ionization and asymmetric charge transfer from of 0.24 Torr and a 17 and 27.3 mm spacing produced the highest intensity and narrowest gas beam, whereas a larger Ar+.The apparent incompatibility of a high pressure dc GD ion skimmer orifice reduced beam intensities. In an investigation of skimmer design Jarvis et al.46 found that one particular source with a low pressure FTMS can be overcome through use of an external ion source and proper diVerential pumping in-house design significantly reduced polyatomic species without degrading limits of detection. However, some loss of techniques.Barshick et al.37 have applied GD FTMS to the measurement of Pb isotope ratios in a sample of 10% (m/m) sensitivity was reported. Farnsworth and co-workers47 used laser fluorescence to map a portion of the ion beam formed lead oxide in silver. Use of an ‘Infinity Cell’, equipped with segmented trapping plates to mimic the field of an elongated behind the skimmer cone. Axial measurements of Ba and Sc ion densities from 22 to 45 mm behind the tip of the skimmer cell during excitation, improved the measurement precision from 3%, obtained with the standard cell, to 0.4%.Mass bias were reported, together with maps of radial ion density with and without lead and magnesium matrices. The authors con- could be still up to 2.5% even with the improved cell. Mass resolutions were typically above 20 000 with 100 000 feasible, fessed that their earlier ion deposition experiments had dramatically underestimated the extent of the radial spread of the although precision was compromised at higher resolution.Sputtered crater shapes have been investigated as a function ion beam and the influence of matrix on the beam. Prudnikov and Barnes48 investigated the theoretical dep- of the discharge parameters in order to optimize depth profiling by rf GDMS.38 The gas pressure was shown to play a major endence of the standard deviation on ICP-MS instrument parameters and element concentration, and demon- role in shaping the crater. Low pressures yielded flat crater bottoms but high cluster ion intensities.Both crater shape and strated a relationship between the standard deviation and LOD. Theoretical results were compared with experimental data and atomic signals were optimized for conducting and ceramic samples at a pressure of 0.09 Torr. The increase in crater showed good agreement, which encouraged the authors to propose the theory as a mechanism to improve LODs in roughness with depth was similar for all samples.The procedure was used to analyse thin conducting layers on silicon ICP-MS. Mermet and co-workers,49 on the other hand, questioned the relationship between LOD and LOQ. They rec- and a 15 mm polymer multilayer on aluminium. Su et al.39 have investigated the operational parameters in ommended that the best way to determine the LOQ at any desired repeatability was to plot the RSD of the net signal ms-pulsed GD TOF MS depth profiling of electrodeposited Fe on copper and Zn on steel.Minimum crater curvature occurred against concentration up to about 50×LOD, although they did acknowledge the practical diYculties involved in this at a discharge pressure of 400 to 500 Pa. The discharge frequency considerably influenced the topography of the crater procedure. 1636 J. Anal. At. Spectrom., 1999, 14, 1633–16593.2 Instrumentation of trace elements during IR LA of glass and copper. Particulate trapping at various points in the sample introduction train An improved microwave plasma torch was presented by Pack indicated that ablation and transport phenomena could lead and Hieftje.50 The use of a quartz central channel tube to opposing eVects, giving a confusing overall picture of promoted a stable annular He plasma at a power of 70–200 W.fractionation. Elements such as Au, Bi and Te showed greater Introduction of N2 as a sheath gas at the base of the torch than expected fractionation at higher laser fluences. This increased stability and minimized air entrainment, thereby contradicted current ablation models.The authors found that enhancing signal as well as attenuating background peaks. in the ablation of glass an inverse logarithmic relationship A comparison of three simplex procedures, the Leary, existed between elemental oxide melting points and ablative Euclidean and combined ratio methods, was made by Sartoros or overall fractionation. The ablation of copper exhibited et al.51 for ion lens optimization. The procedure using a single higher ablative fractionation than glass, possibly due to zone mid-mass element was only marginally better than the multi- refinement at the ablation site.Transport fractionation was element approach. Optimization to favour particular mass important in both media; however, the process was found to ranges showed that using average mass was better than using be more selective for glass, aVecting only the most volatile a mid-mass procedure. analytes.Given the many notable diVerences in ablation Landais et al.52 described a new scanning mode for quadru- behaviour, the authors concluded that matrix-matching is an pole mass spectrometers, in which only the rf applied to the essential part of accurate IR LA measurements. quadrupole rods was varied, whilst keeping constant the In a related study, Eggins et al.57 studied elemental fractionamplitude of the ac and dc potentials. Marriott et al.53 showed ation during repeated excimer LA of a single point on a glass how the design process used to determine the geometry of a reference material.At shallow ablation depth, volatile element double-focusing mass analyser was often influenced by the signals were enhanced whereas at greater depth the refractory mass range required of the instrument and was particularly element signals increased. Microscopic examination of the significant for systems used in organic analysis. They indicated ablation pit and the surrounding ejected material revealed that the design of an instrument destined for elemental analysis sequential condensation of refractory and volatile phases was constrained by the mass range to be covered.from the cooling plasma plume. Based on these observations, Vanhaecke et al.54 assessed the mass discrimination and the authors suggested that ablation phenomena were primdetector dead time for a quadrupole instrument with a continu- arily photothermal near the surface but changed to plasmaous dynode electron multiplier and a magnetic sector system dominated mechanisms at greater depths.Ablation in a He with a discrete dynode electron multiplier. In the former, dead atmosphere gave a 2- to 4-fold gain in sensitivity as a result time was found to be strongly mass dependent, whereas in the of attenuation of the surface condensation process. latter it was mass independent. The authors believed that the Because of the ongoing problems with elemental fractiondecreasing gain of continuous dynode multipliers with increas- ation, investigators have continued to develop new methods ing ion mass may be the cause of the observed eVects, which for calibration of LA data.Falk and co-workers58 described should be carefully considered when trying to achieve highly a new approach for the calibration of LA data using pneumatic accurate isotope ratio measurements. nebulization of solution standards.Calibration in ‘nebulization mode’ was not novel but this study was unique in that the LA 3.3 Sample introduction data were normalized internally using the leading edge intensity of a major peak from the matrix element (m/z 62.5 for Cu). 3.3.1 Introduction. Sample introduction continues to play a For the particular instrument used, the authors found that the key role in successful development of ICP-MS methods. leading edge intensity could be measured over a 2.5 h period However, over the last few years, there has been a trend away with 3.6% RSD.They also found that the signal was directly from ‘pure’ studies of sample introduction techniques in favour proportional to analyte concentration, indicating that peak of application-oriented studies. This is predictable given the shape was independent of peak height. age of ICP-MS (nearly 20 years) yet many techniques (for In another study of signal normalization, Watling59 example nebulization) and related practices (for example, described an in-line light scattering cell for determination of calibration of transient signals) still require great improvement. particle mass transport with subsequent normalization of LA In this section some of the most recent and novel developments signals.The cell was fabricated from a low-static fluorocarbon are reviewed. copolymer, with quartz windows at either end. Monochromatic light (190 to 300 nm wavelength) was used to illuminate the 3.3.2 Laser ablation.Though the pace of development has slackened over the last few years, work on new LA sources ablated particle stream as it traversed the cell. The scattered light intensity (a measure of particle density) was convoluted continued, particularly for geological applications. Most notable was work by JeVries et al.,55 in which the authors with analyte signal to achieve normalization. Using this normalization scheme, six sulfide minerals were analysed by IR described the first frequency-quintupled (213 nm) Nd5YAG source for LA of minerals.The output energy of the source LA, and precision was found to be better than 5% RSD. The author also found that with the higher precision, exact matrix was 7.5 mJ per pulse, based upon 350 mJ per pulse output at the Nd5YAG fundamental (1064 nm). The pulse energy was matching was not needed to achieve the data quality required for geochemical exploration. controlled using a wave plate–MgF2 polarizer combination.The authors found that 213 nm incident light was better Matrix matching has continued to be the preferred means of overcoming fractionation and accompanying calibration absorbed than 266 nm (frequency-quadrupled Nd5YAG) light. This higher absorption reduced the problem of ‘catastrophic problems. O� degaerd and co-workers60,61 used UV LA and a double sector ICP-MS device to determine the trace element ablation’ of highly-cleaved mineral thin sections and produced a longer, more stable and more intense signal, which was composition of geological samples after fusion of the samples in lithium tetraborate.Calibration curves were developed by presumed to be due to higher ablation eYciency. The authors also noted that elemental fractionation was less of a problem weighted least-squares linear regression of ablation data from a series of RMs. Barium was used as a reference for internal with 213 nm irradiation. Fractionation continued to be an area of intense investi- normalization of the ablation data.Analysis of high-purity fused silica indicated that method LODs were better than gation during this review period and several notable papers have been published. For example, Outridge and co-workers56 0.05 mg kg-1, and additional analyses indicated that measurement bias was less than 25%. The authors noted that repeated examined ablative and transport mechanisms for fractionation J. Anal. At. Spectrom., 1999, 14, 1633–1659 1637ablation of the sample surface did not have a noticeable eVect they also suggested that the argon dimer would be a more appropriate normalization species than ArH+ as it was more upon subsequent XRF analysis, indicating that the fusion pellets were suitable for either analysis method, regardless of ‘robust’ and possibly more indicative of plasma conditions.In a study similar to that of Alary and Salin, Bjo�rn and their status. Neilsen and co-workers62 have reported the novel use of LA co-workers66 developed a quantitative evaluation protocol and subsequently compared the severity of polyatomic ion inter- with two dimensional gel electrophoresis for spatial imaging of separated cobaltoproteins.The authors developed a protein ferences upon As, Cr, Cu, Ni and Se when introduced by ETV with those occurring when introduced by pneumatic nebuliz- (organic moeity) map using the Coomassie Brilliant Blue staining technique. A 59Co map was subsequently developed ation.Elements were determined in an aqueous reference standard and a solution containing various concentrations of by interrogating the plate, point-by-point, with an Nd5YAG (1064 nm) LA system. Comparison of the two maps allowed Ca, Cl, K, Na and P. A matrix-derived spectral interference was deemed ‘intolerable’ when the matrix concentration was the main Co-binding proteins (for example, albumin) to be identified. Ablation signal area was found to be linear with such that the observed analyte isotope ratio deviated more than 10% from the ratio determined in the aqueous reference respect to Co concentration, and the Co LOD was estimated at 0.3 ng.The precision of replicate analyses was shown to be solution. Given this definition, the authors found that when ETV was used in place of nebulization, spectral interfer- 6% RSD. It is interesting to note that this marriage of liquid– solid separation with solid-sample introduction has further ences, with the exception of 31P2+ on 62Ni+, were reduced by 1–2.5 decades.blurred the boundary between inorganic MS and molecular analysis techniques. In another, mechanistic study, Majidi and Miller-Ihli67 examined the influence of the graphite substrate upon the vaporization characteristics of 18 elements determined by 3.3.3 Thermal vaporization. Despite the maturity of thermal vaporization as a sample introduction technique, investigators ETV-ICP-MS.Single elements and multi-element mixtures were tested in pyrolytically coated furnace tubes (using wall have continued to make improvements, thereby opening new avenues of application. For example, in this review period, vaporization for 7 s at ~2900 K). However, some of the tubes were heated in air before use (40 s at 1023 K) to ‘oxygenate’ Mickeley and Amato63 described a mini-tube furnace catalytic combustion system for the determination of Hg in biological the graphite surface. The authors found that the precision of signals arising from polyatomic carbon species (for example, media.A solid sample (5–20 mg) was placed in an aluminium boat, which was subsequently immersed in the furnace. The CO+) were improved when the furnace tubes were heat-treated before use. They therefore concluded that species transported sample was then burned at 1223 K in an oxygen–argon stream, using a cobalt oxide–alumina catalyst to promote complete with carbonaceous matter would also be measured more precisely during analysis.The authors made several recommen- combustion. Liquid samples were handled in much the same fashion using a combustible filter-paper support to accelerate dations for the general practice of ETV-ICP-MS, including heat-treating furnace tubes before use, keeping sample acidity decomposition. For determination of Hg in a hair RM, the method LOD was 1.9 mg kg-1 (with a 10 mg sample), repro- constant, using other matrix modifiers as necessary and using the method of standard additions to compensate for changes ducibility was 20–30% RSD and the results fell within the uncertainty of the certified Hg concentration (0.57 mg kg-1). in furnace behaviour due to ageing or sample matrix variations.The recommendations made by Majidi and Miller-Ihli, Sample turnaround was 3–4 min. Elucidation of vaporization processes continued to occupy though very commendable, also suggested that more robust calibration techniques (for example, standard additions) are many involved in this field.In an interesting study, Galba�cs and co-workers64 used an analytical furnace with thermogravi- needed to overcome the matrix-dependent properties of thermal vaporization. To address this need Lee and colleagues68 metric, calorimetric and MS sensors to examine thermal decomposition products from various environmental RMs. have used ID ETV-ICP-MS for the analysis of urine. In the determination of Hg, a mixture of 2% HCl, Mg(NO3)2 and Pd Bound water was lost at 393 K and organic matter was decomposed at temperatures below 773 K.Up to a temperature was used as a modifier to promote Hg vaporization from the complex urine matrix. The area of the Hg vaporization transi- of 1073 K carbonates were decomposed, whereas at higher temperatures oxides and other volatile compounds were vapor- ent signal peak was used to calculate the necessary isotope ratios. The method was used to determine Hg in a urine RM ized.The authors believed that spattering, a problem at about 800 K, was due to rapid decomposition of organic matter. and several locally-collected urine samples. Data obtained from the ID method agreed closely with the known Hg Spattering that occurred at higher temperatures was thought to be caused by carbonate decomposition. The authors sug- concentration in the RM. The method LOD for Hg was found to be ~20 ng l-1 and method precision was ,8% RSD.For gested that pyrolysis temperatures should be carefully selected to take into account whether the sample was composed the direct determination of Cd and Pb in urine by ID ETVICP- MS,69 a 1% solution of HNO3 was used as chemical primarily of organic or inorganic matter and to avoid analyte losses and dispersion due to spattering. modifierope ratios were calculated using the area of the vaporization transient signal peaks which were acquired at Alary and Salin65 described the use of argon and argon polyatomic ions as probe species for plasma diagnostics in comparatively low vaporization temperature (1273 K).Data obtained from the ID method agreed closely with the certified ETV-ICP-MS. In this study, water ejected from the furnace was monitored using 36ArH+, and a two-level factorial experi- Cd and Pb concentrations in two RMs. Method LODs were 20 ng l-1 for Cd and 5 ng l-1 for Pb, and precision was ,11% ment was designed to examine the eVect of water loading, plasma power, sampler–load coil separation and injector argon RSD.For samples collected locally, concentrations of Cd, Hg and Pb determined by ID agreed well with those obtained by flow rate upon five analytes and several other polyatomic species (for example, argon dimer). The authors found that the standard additions method. the shape and intensity of the vaporization transient signal peak was greatly influenced by slight variations in water 3.3.4 Chemical vaporization.Two groups have described methods for the determination of Hg. Knight and co-workers70 loading. However, these variations were attenuated when the analyte vaporization transient peak signal was convoluted with described the use of CV AAS in combination with quadrupole ICP-MS for the determination of Hg in hair. A microwave- the ArH+ or the argon dimer transient peak signal. The authors concluded that this convolution method could be assisted microchemical digestion procedure (7 ml total volume) was developed to allow the analysis of small hair samples, applied to other transient techniques (for example, LA) and 1638 J.Anal. At. Spectrom., 1999, 14, 1633–1659such as those taken from newborn children. The interface ulizers. The authors described two novel diagnostic devices, the ‘optical patternator’ and the ‘rainbow refractometer.’ The between the AAS and the ICP-MS was relatively simple and allowed the simultaneous determination of Hg concentration former device could elucidate two dimensional spray structure, planar mass distribution and spatially resolved droplet size by AAS and Hg isotopic composition by ICP-MS.Mercury recovery was found to be ca. 100% and precision was of the distributions. The latter device permitted droplet temperature, temperature–size correlations and temperature–velocity corre- order of 3.5% RSD. Some memory eVects were observed in ICP-MS data, but these did not have an adverse eVect upon lations to be determined.Elucidation of these important parameters may open the way to quantum (rather than the determinations of Hg concentration or isotopic composition. A ‘collect and punch’ CV system was developed by incremental ) improvements in nebulizer performance. It is hoped that these tools will increase the number of publications Karunasagar and co-workers71 for the determination of Hg at ng l-1 concentrations. The CV apparatus was unique in in this mature, yet crucial discipline.that the Hg vapour was collected in a receiving vessel before injection into a quadrupole ICP-MS system. Time-resolved 3.3.6 Flow injection. In an interesting departure from prior years’ experience, this review period saw a surge in the acquisition was used to collect and integrate Hg isotope signals Analysis of an oyster tissue RM indicated that Hg recovery development of new reactor systems for FI analysis. For example, Yan and co-workers76 developed a knotted reactor was ~100%.The method LOD for Hg was estimated to be 6 ng l-1. The authors suggested that this method could be FI manifold for on-line preconcentration and isolation of REEs in environmental and geological media. In the system used to determine Hg reliably at ultra-trace concentrations in biological and other environmental media. described, dissolved REEs were precipitated on-line by mixing the sample with ammonium hydroxide (pH 8.3–9.0).The Selenium HG continued to attract interest during this review period. In an interesting study, Hoppstock and co-workers72 precipitate (free of alkaline earths and alkali metals) was collected within the knotted reactor without prior filtration. used sector field ICP-MS with a locally designed and constructed hydride generation device to determine 79Se, a long- A plug of 1 M HNO3 was introduced to dissolve the precipitate and the acid solution analysed by ICP-MS.Using a flow rate lived radioisotope of importance in nuclear waste disposal. Spectral interferences were of particular concern and the design of 5 ml min-1 and preconcentration time of 120 s, the authors demonstrated that the REEs could be isolated and concen- of the HG apparatus included a ‘wash bottle’, which reduced the introduction of Br species into the ICP. The presence of trated 55–75 fold within 5 min. The method LODs fell within the 0.06–0.27 ng l-1 range. Replicate analyses (n=16) of a other interferents (for example 39Ar40ArH+) were checked and minimized.The instrumental LOD for 79Se was 0.1 mg l-1 spiked pore water sample showed that precision was in the range from 1.8 to 4.2% RSD at a REE concentration of and measurement precision ,10% RSD. Zhang and Combs73 exploited the isotopic selectivity of ICP-MS to develop a 0.1 mg l-1. In a related study, Willie and co-workers77 described a new double calibration scheme for the determination of Se in plant tissues.Samples of alfalfa were spiked with 77Se and standards FI manifold using 8-hydroxyquinoline immobilized upon silicone tubing for the determination of trace elements in sea-water. prepared from a separate (presumably selenium-free) source of alfalfa mixed with natural Se and the 77Se spike. The The ligand was bonded to the silicone tubing via the Mannich reaction and a 2 m length of the functionalized tubing used to samples and standards were digested with HNO3 and HClO4 and the residue treated with HCl to ensure that Se was present isolate metals present in coastal and open sea-water samples.Chemical recoveries were in the 35–95% range (200 injection in the tetravalent oxidation state. Selenium was determined using isotopes at m/z 77, 78 and 82 using HG-ICP-MS. cycles) and LODs for Cd, Co, Cu, Mn, Ni, Pb and Zn were in the ng l-1 range. Determination of the specified metals in a Analysis of 207 alfalfa samples showed that external calibration (using natural Se standards) and ID (using the 77Se sea-water RM agreed closely with certified concentrations when calibration was by the standard additions method.spike) gave the same results. The parallel calibration procedure allowed the authors to evaluate the quality of any single result Rapid development of electrochemical reactions during this review period has been noted. For example, an electrochemical in the data set. Feng and co-workers74 used a novel approach to traditional stripping manifold was developed by Pretty and co-workers78 for the determination of U in complex media.In this system, hydride generation protocols by exploiting the kinetics of AsV reduction in the presence of L-cysteine to determine inorganic a glassy carbon working electrode was anodically pretreated in situ and U subsequently accumulated at the treated elec- As speciation in natural waters by sector field ICP-MS. Arsenic, in its trivalent form, was determined after 2 and 8 min reaction trode.Concomitant sodium (2% in solution) decreased the deposition rate for U. However, the lower deposition rate was time, and the diVerence in As concentration was attributed to AsV present in the original water sample. Potential chemical easily overcome by increasing deposition time. For a deposition time of 10 min, a LOD of 0.12 ng l-1 was attained, which interferents (for example, transition metals) were removed using Chelex 100 (iminodiacetate) resin.The authors found allowed determination of minor isotopes at total U concentrations of 5 mg l-1. Precision was in the range of 2–6% RSD. that excess amounts of one As species did not greatly aVect the determination of the other. Results from analysis of spiked Repeated determination of U in a sea-water RM indicated that recoveries were in the 100–120% range. samples and a standard river water sample agreed closely with expected values. Machado et al.79 have studied chemical and electrochemical means of HG for the determination of As in natural waters by FI.The authors developed an FI manifold specifically for the 3.3.5 Nebulization. Nebulization unquestionably remains the most utilized means of sample introduction for plasma purpose of comparing the two HG techniques side-by-side. They found that the chemical HG technique was more prone spectrochemistry, although for many years authors have noted that nebulization is the weak link in the typical ICP sample to interference than the electrochemical one.The FI manifold was capable of processing 60 samples per hour. The authors introduction scheme. Therefore, it was very disappointing to review the past year’s literature only to find no publications found that As LOD were 0.05 mg l-1 for chemical HG and 0.2 mg l-1 for electrochemical HG. Precision was better than describing notable improvements in nebulizer design or performance. This lack of improvement may be due to an absence 3% RSD for an As concentration of 10 mg l-1.Valles Mota and co-workers80 compared vaporization and of advanced diagnostic techniques. To address this need, McLean and co-workers75 examined various techniques for nebulization in developing an on-line ID system for the determination of Cd in biological and environmental media. Two FI nebulizer diagnostics, with particular emphasis upon microneb- J. Anal. At. Spectrom., 1999, 14, 1633–1659 1639approaches were examined.In the first, sample and spike were developing an ID method for the determination of methylmercury by HPLC-ICP-MS with ultrasonic nebulization. In this mixed in advance of nebulization, and in the second the sample (presumed to be previously spiked) was mixed on-line with work, an enriched 200Hg spike was added (as divalent Hg) to the sample and methylmercury extracted by the water vapor sodium tetraethylborate to produce volatile organocadmium species.With careful optimization of data acquisition param- distillation method. Subsequent determination of isotopically enriched methylmercury allowed for correction of biases eters, 111Cd5112Cd and 111Cd5114Cd were determined with a precision of ca. 0.2% RSD. Using both techniques, the authors induced by the sample preparation method. Relative to a previously published method, this method was significantly found that concentrations measured in several RMs (water, plant tissue and animal tissue) agreed closely with certified more stable, allowing detection of methylmercury at Hg concentrations of 15 ng kg.1.concentrations The authors preferred the on-line vaporization method because of the lesser amount of manual sample In a very significant paper, Kingston and co-workers89 described a ¡®speciated¡� ID technique for the determination of preparation. Dressler and co-workers81 described the development of an CrIII and CrVI by HPLC-ICP-MS.In this method, spikes of 50CrIII and 53CrVI were added to samples and, after equili- FI manifold for direct introduction of alcohols into an ICP mass spectrometer. In this system, ~100 ml of alcohol were bration, the Cr species were separated by anion exchange and detected by ICP-MS. The isotope ratios 50Cr552Cr and injected into the carrier stream (~0.2 M HNO3) and the sample diluted on-line with an equal volume of water prior to analysis. 53Cr552Cr were determined for each species.Extensive data reduction allowed quantification of each Cr species with Several criteria were used for instrument optimization. Conditions which promoted CeO+5Ce+ ratios less than 0.03 compensation for interconversion. The authors estimated that their double spiking technique compensated for 80% of CrVI were found to yield analyte signal enhancements, in particular for As, Hg and Se. In general, lower nebulizer gas flow and reduction during sample handling, and they suggest that similar approaches could be developed for other metal species higher incident powers were needed to support the plasma.Under these conditions, the authors found that carbon build- such as organometallic compounds. Taking one step in this direction, Ebdon, Hill and Rivas90 up on the sampler and skimmer cones was insignificant even after 200 injection cycles. Interestingly, the authors also found described an ID HPLC-ICP-MS method for determination of organolead species in rainwater.Reverse phase ion-pairing that no advantages were realized when oxygen was injected into the nebulizer gas. HPLC was used to separate inorganic lead from trimethyllead (TML) chloride and triethyllead (TEL) chloride in a sample spiked with 206Pb-enriched TML chloride. Compared with the 3.3.7 Chromatography and electrophoresis. The marriage of separation techniques with ICP-MS continues to be an area conventional external calibration method, the ID technique was found to improve TML LOD two-fold (to 1.5 mg kg.1, of rapid growth in terms of publication.However, in a recent review �©obin. ski82 pointed out that this sub-discipline has as Pb). Reproducibility was found to be ~4% RSD for TML concentrations of between 50 and 1000 mg kg.1. The authors reached a level of stagnation in terms of research originality and marketability. Crews83 supported the view of �©obin. ski found that the accuracy was not sensitive to the amount of spike added.However, they suggested that to achieve best that without collaboration between users and developers the trend toward stagnation will continue. Both Crews83 and reproducibility, spike volumes should be carefully adjusted such that the measured isotope ratio was close to unity. Welz84 suggested that progress in data quality assurance (i.e., defensibility) is needed to permit ¡®routine¡� application of these New chromatography-ICP-MS interfaces may themselves lead to improved data quality, and in this vein, Mendez et al.91 techniques.�©obin.ski,85 indicated that greater flexibility, in terms of more automation and improved dimensionality, was needed reported on the development of an HPLC-HG-ICP-MS method for the determination of selenomethionine enantiomers. to meet the analytical challenges that end-users are likely to face, particularly in environmental or biomedical applications. The D and L forms of the amino acid were separated on a b-cyclodextrin column.Selenium in the column eZuent was In answer to the need for traceable data, Larsen86 studied parameters which influence data quality and optimization of subsequently vaporized by on-line microwave-assisted HG. Using ICP-MS, LODs for Se were 70 mg l.1 compared to HPLC-ICP-MS methods. For the separation of ionic species, the author found that the HPLC apparatus was most stable 2.7 mg l.1 for pre-column derivatization with fluorometric detection. Both techniques were capable of detecting 0.1.1% and robust when isocratic elution was employed with a buVered aqueous mobile phase.Those elements diYcult to ionize (for D-selenomethionine in commercially pure L-selenomethionine samples. However, the ICP-MS method was found to be more example As and Se) were better detected when several per cent. of methanol were added to the mobile phase. Potential selective than the fluorometric method, making it better suited for analysis of more complex samples such as Se-enriched interferents (for example Cl ) were removed on-column or by prior treatment such as extraction of the liquid sample.Mass yeast. The development of nebulizer-based interfaces appears to balance calculations were found to be the most useful measure of data quality. However, method comparisons and recovery have accelerated during this review period, and this may indicate a general movement away from traditional nebulizer experiments were found to be useful for solid extracts and liquid samples, respectively.The author also pointed out the development eVorts in favour of these ¡®applied¡� nebulization studies. In one such paper, B¡�Hymer and co-workers92 com- urgent need for RMs cal speciation is certified. To bring higher quality to speciation analysis, Heumann pared several combinations of nebulizers and spray chambers as interfaces for the HPLC-ICP-MS determination of inorganic and co-workers87 described some general considerations for ID analysis of eZuents from several diVerent chromatographs.and organoarsenic compounds. The authors evaluated the oscillating capillary (OCN), concentric and Meinhard high Two modes of ID were discussed. In the first, a specific isotopically enriched compound was added to the sample, eYciency (HEN) nebulizers with three spray chambers (single pass, double pass and cyclonic), and found that the HEN. appropriate for known species (such as selenate).In the second, a general isotopic spike was added, appropriate for poorly cyclonic spray chamber combination gave the lowest LODs (1.2.3 mg l.1 As) but the poorest reproducibility among dupli- characterized species, such as metal humates. The authors pointed out that this technique would not work for monoiso- cate injections (14%). The concentric nebulizer gave good LODs (1.6.3.6 mg l.1). However, the OCN LODs were an topic elements and they also noted that for accurate results, isotope ratios must be corrected for mass discrimination.order of magnitude poorer. Duplicate injections with the OCN and the concentric nebulizer were reproducible to within 9%. Wilken and Falter88 reduced these principles to practice by 1640 J. Anal. At. Spectrom., 1999, 14, 1633.1659Kirlew, Castillano and Caruso93 evaluated ultrasonic nebuliz- and quadrupole instruments for the analysis of human urine. The former was used at a resolution of m/Dm=3000 for the ation (USN) as an interface for capillary electrophoresis (CE) of inorganic cations and anions.With hydrodynamic injection, determination of Cr, Ni and V in order to overcome spectral interferences and the quadrupole system was used to determine the LODs for anions using CE-USN-ICP-MS ranged from 80 mg l-1 (for arsenate) to 2 mg l-1 (for selenite). The cation Cd, Co, Pb, Rb, Sb and Se. Sample dilution, UV irradiation and microwave digestion procedures were investigated to LODs fell between 30 ng l-1 (for Cs) and 60 mg l-1 (for Co).The anion LODs were improved 300- to 1000-fold simply by clarify the performance of the two systems under diVerent matrix loads and quantify the analytes in the urine sediment. changing to electrokinetic injection. However, Majidi and Miller-Ihli94 observed that in speciation measurements electro- Memory eVects caused by the ‘adhesive’ properties of elements such as Hf, Mo, Nb, Sn, Ta, Th, U, W and Zr can kinetic injection was more likely to induce bias than was hydrodynamic injection.aVect accurate determination of the elements by ICP-MS. A wash protocol using 10% HNO3+10% HF was developed by Taylor and co-workers95 used a commercial microconcentric nebulizer in combination with a cyclonic spray chamber as an McGinnis et al.102 to minimize these eVects. Detection limits (~5–20 pg g-1) were about a factor of 50 lower than those interface for CE-ICP-MS. Nebulizer suction was found to be an important factor limiting separation resolution.The authors obtained using 5% HNO3 washes (300–800 pg g-1). The method was applied to basaltic samples. overcame this limitation by applying a weak vacuum (130 mbar) to the vial on the inlet side of the CE apparatus. Several workers have attempted to overcome polyatomic interferences that compromise the analysis of biological and Work with the optimized system revealed that retention times for two Cd–metallothionein isoforms were reproducible to environmental samples.McLaren and co-workers103 digested three National Research Council of Canada marine sediment within 9% RSD, peak heights (for 1 g l-1 Cd–metallothionein, hydrodynamically injected for 6 s) were reproducible to within RMs (HISS-1, PACS-2 and BCSS-1) in a closed-vessel microwave digestion using a mixture of HNO3 and HF followed by 12% RSD, and peak areas were reproducible to within 13% RSD. Absolute LODs were between 2 fg for Cd and 10 fg open-beaker reflux with HClO4 and H2SO4.Baking the residue at 335 °C removed all traces of residual HClO4 and H2SO4, for Zn. As noted in last year’s Update, studies of GC ICP-MS were thereby eliminating the potential for polyatomic ion interference. Accuracy was better than 5% and the LOD 0.4 mg g-1. few in number yet occasionally oVered fascinating insights into the future of atomic spectrometry. For instance, Feldmann Violante et al.,104 in their analysis of candidate environmental RMs, adopted the approach of analysing synthetic matrices and co-workers96 used ICP-MS in combination with ion trap MS for the determination of volatile Bi, Sb and Sn compounds.simulating the five materials in order to quantify each interferent. Subsequently, either a mathematical correction was In this study, a gas chromatograph (with cryotrap) was coupled directly to an ICP-MS instrument for the detection of metal applied or a matrix-matched blank was subtracted from each of the samples.The validity of the method was confirmed with species and the determination of the metal isotope ratios. Fractions of the column eZuent were also collected for sub- ICP-AES and GFAAS. Barnes and co-workers105 determined Cr and V in biofluids following high-pressure thermal or sequent molecular analysis by ion trap MS. With this system, the authors obtained the first ever element specific chromatog- moderate-pressure microwave sample preparation, both with vapour-phase acid digestion.This successfully reduced blanks rams with isotope ratio data and structural analysis of the organometallic compounds. This work will almost certainly and carbon and chloride polyatomic species. The 52Cr signals were corrected mathematically for remaining ArC+ inter- lead to the development of truly parallel GC-ICP-MS-MS systems in which metal species are identified and quantified ferences and verification was achieved with RMs. Mathematical procedures, combined with standard addition, were without the need for operator intervention.Such instruments would be of tremendous value as environmental researchers also employed by Van den Broeck and Vandecasteele106 to correct for interferences aVecting the determination of As in seek to understand trace element mobility in ecological and biological systems. percolate waters from an industrial landfill site. Results from the three tested correction equations compared well with each other, as well as with the results from capillary zone 3.4 Interferences electrophoresis. Steady state97 and transient98 acid eVects in ICP-MS have been studied. In the former, atomic and polyatomic ion 3.5 Isotope ratio measurement intensities were found to decrease with increasing HNO3 concentration (2, 10 and 25%), although under certain con- It is an irony in the development of ICP-MS that it has taken until relatively recently for the technique to achieve its original ditions enhancements occurred.The eVect could be minimized through optimization of nebulizer gas flow. Transient acid scientific target. Gray, one of the founding fathers of plasma mass spectrometry, wrote in 1982 ‘The first applications for eVects, such as those following abrupt changes in acid concentration, were believed to be caused by changes in the extent the technique are likely to be in the geological sciences, where there is a need for isotope ratio measurement....’, and ‘The of sample aerosol evaporation in the spray chamber. These changes aVected aerosol transport.technique is also likely to make considerable impact in the life sciences since it is possible to use non-radioactive isotopic Analyte signal suppression in the presence of a rock matrix was found to be attenuated at high forward power of the tracers for studying biological systems.....’.107 With the benefit of hindsight we now know that the performance oVered by ICP.99 The suppression was negligible at dilution factors of 600, 400 and 113 at 1.1, 1.4 and 1.7 kW, respectively.This the sequential data-collecting quadrupole systems was not suYcient for their widespread use as isotope analysers, particu- approach used in conjunction with FI was found to give accurate (to within 2%) and reproducible (,6%) data based larly in the geological sciences. However, the recent development and now general use of magnetic sector systems, in on RMs, with LODs of several ng g-1.De Boer et al.100 reduced sodium chloride matrix eVects in the determination particular those with multiple detectors, has opened the door to the original prophecy. of Cd in urine by instrument optimization and the use of Rh as an internal standard. Additional corrections were made One indication of the widespread use of sector instruments is the appearance of reviews of the subject. Heumann et al.297 based on the 35Cl signal. Data from a RM showed recoveries to be between 99 and 101% with a RSD of 2.3% and LOD of gave an overview of plasma MS for isotope ratio measurements, with a comparison of other types of instrumentation. 8 ng l-1. Krachler et al.101 considered both magnetic sector J. Anal. At. Spectrom., 1999, 14, 1633–1659 1641Limitations on accuracy caused by spectroscopic interferences, tope ratio measurements of Ca, K and Mg for stable isotope tracer studies of plant nutrient solutions, enriched with 25Mg, abundance sensitivity, space charge, and nozzle separation eVects were discussed together with factors limiting precision 26Mg, 41K, 42Ca and 44Ca.A resolution of 9000 was required to separate the 38ArH+ and 40ArH+ ions from the 39K+ and such as counting statistics and ion current stability. Multiplecollector magnetic-sector ICP-MS systems were acknowledged 41K+ ions, respectively. Precision of the K isotope ratio was 0.7% at a K concentration of 100 mg l-1. Isotope ratios of Ca to have the precision capability of TIMS systems.It was, however, conceded that the issue of mass bias, which aVected and Mg (each at 50 mg l-1) were determined at a resolution of 3000 with a precision of 0.5 and 0.4%, respectively. accuracy, still needed to be resolved. Jakubowski et al.,108 in a similar type of review, discussed the peculiarities and per- A method for the determination of Pu isotopes and the 240Pu5239Pu isotope ratio was described by Stu�rup et al.116 formance of sector instruments and presented some selected examples of analytical applications to demonstrate the state Results from the measurement of total Pu (239+240Pu) were in good agreement with data obtained by a-spectrometry. of the art.Halliday and co-workers have presented several pieces of Detection limits of 5, 1 and 1 fg ml-1 were found for 239Pu, 240Pu and 242Pu, respectively. The achieved precision (2%) for work involving the application of a multiple-collector magneticsector ICP-MS system.In a review109 with many references, the measurement of the 240Pu5239Pu ratio in environmental samples was close to the theoretical value. The influence of the potential of the technique was illustrated with applications such as the use of the Lu-Hf isotopic system to study garnet dwell time, number of sweeps and sample uptake rate on the ratio measurement precision was investigated and optimized. geochronology or crustal evolution. The short-life 182Hf–182W chronometer has been applied to dating of the Earth’s core, The results from the ICP-MS technique were in good agreement with those obtained by high resolution a-spectrometry.the Moon and Mars. Considered here and detailed elsewhere110 was the use of high precision ID measurements of Cd, In and Townsend et al.117 determined what they considered ‘precise’ Pb isotope ratios in Australian galena samples. Values of 0.13, Te as tests for models for the accretion of the inner solar system.Other areas discussed in this wide ranging review were 0.11, 0.11, 0.046 and 0.048% RSD (at 1s, n=30) were obtained for 208Pb5204Pb, 207Pb5204Pb, 206Pb5204Pb, 208Pb5206Pb and various applications to stable isotope geochemistry including the study of isotopic variations of Cu and Zn, quaternary 207Pb5206Pb. DiVerences between measured ratios and those obtained by TIMS were generally better than±0.2%. Catterick geochronology from Th and U isotopic compositions and the use of LA sample introduction to make in situ measurements et al.118 investigated a method intended to simplify the use of ID for high accuracy measurements. The method was an of Hf, Pb, Sr and W isotopic composition in natural materials. In another study,111 high precision Tl isotope determinations adaptation of that which utilized an iterative ‘matching’ procedure.Many common errors and necessary corrections were were made on igneous rocks, marine Fe–Mn crusts and a chondrite meteorite.Mass discrimination was eliminated by negated or eliminated with the new procedure, which achieved an accuracy of within 1% at the 95% confidence level on blind the inclusion of Pb with the natural samples. By simultaneously monitoring the isotopic compositions of both elements, accu- trial solutions. Kerl et al.119 determined U concentrations and isotope ratios in radioactive waste samples. A precision of rate correction could be made. A ratio precision of 0.01–0.02% was achieved.The results were used to provide useful limits ,0.1% RSD was achieved for isotope ratios at the ng ml-1 concentration level. To minimize radioactive contamination of on the abundance of 205Pb in the early Solar System. A reversal of this technique was applied to the determination of the instrument, samples were introduced via a FI system which reduced sample volume to a few ml. Pb isotope ratios112 using Tl to account for mass discrimination eVects.In a detailed study the accuracy and precision of such Useful isotope studies using quadrupole instruments are being published. Monna et al.120 used Pb and Sr as model analytes an external correction technique was presented, based upon the results of multiple analyses of mixed standard solutions of in a study of the eVect of counting time and ion fluxes on the precision of isotope ratio measurements. Optimum measure- NIST SRM 981 (Pb) and SRM 997 (Tl ). The application of a linear correction function resulted in Pb isotopic results ment times for each isotope were determined by a mathematical simulation to provide lowest theoretical error.DiVerent algor- significantly lower than the previously published reference values measured by TIMS. A power or exponential law ithms of mass bias correction were also taken into account and evaluated in terms of improvement of overall precision. function provided significantly more accurate and precise results.The authors conceded that the benefit of the approach was greater for instruments which were operating with a precision Further evidence for the increased popularity of multiple collector instruments comes from additional commercial sys- close to that predicted by counting statistics. Menega�rio et al.121 determined the 34S532S isotope ratio in plant material tems entering the market place. O’Nions and co-workers113 reported on the evaluation of a new generation of instruments using measurements at m/z 48 and 50, corresponding to the 32S16O and 34S16O polyatomic ions.The S species in aqueous that incorporated a novel variable-dispersion ion-optics arrangement, which directed all m/z to the centre of the solutions and plant digests were separated and pre-concentrated in an anion exchange resin (AG1-X8) column Faraday collectors of a fixed static array. The performance was illustrated by repetitive Pb isotope measurement and inserted into a flow system.For a solution containing 17 mg l-1 of S, the precision of isotope ratio measurement was ,1%. found to be competitive with the TIMS double-spike method. Turner and co-workers114 described a system for the accurate Sample throughput was 45 h-1. Results were compared with those obtained by electron impact MS and considered to be determination of Hf and Pb isotopes in zircons for dating purposes. The system used a single-focusing magnetic-sector in good agreement at the 95% confidence level.with a hexapole collision cell placed between the ICP ion source and the magnetic analyser. This arrangement reduced 4 Laser ionization mass spectrometry (LIMS) gy spread of the ion beam from 15 eV to ,1 eV and resulted in accurate mass focusing, flat-top peaks and a claimed 4.1 Fundamental studies sensitivity (total ion detected per sample atom eYciency) of 1% for Pb. The kinetic energy (Ekin) and emission angle distributions of both ions and neutrals allow the working conditions in LA Although fundamental limitations on precision exist in the use of single-collector magnetic-sector systems for isotope ratio and laser ionization to be delineated and thereby provide the necessary basis for further analytical developments.For measurements, a number of applications have been reported. Applications described by Becker and Dietze115 included iso- instance, the reduction of the laser-material coupling upon 1642 J. Anal.At. Spectrom., 1999, 14, 1633–1659excimer UV irradiation of solids is not yet fully understood. resolution and the numerous isobars from airborne particle components hampered the identification of the U-related ions. Above 1013 W cm-2, inverse Bremsstrahlung absorption may occur at high gas pressure. In addition, generated particles of For that reason, the conversion of ions such as UO2+ into UO2+ by ion-molecule reaction with the background gas was 5–20 nm diameter may scatter the UV beam but it is uncertain that these particles can live long enough to reduce the mate- used for unambiguous identification of the U species.Kuzuya et al.128 have combined a quadrupole mass filter rial–beam coupling eVectively. Elucidation of such aspects is relevant not only for analytical LIMS (with less focused lasers) with a moderately focused 1.06 mm wavelength laser yielding a spot of 0.2 mm diameter for elemental analysis. The benefits and laser microprobes (LMMS, spot size 1–5 mm) but for all techniques using laser sampling of solids (e.g., LA-ICP-MS). of this approach resided in the low cost and small size of the mass analyser. The ion signals from samples with homogeneous Pinho et al.122 have studied the Ekin distribution of neutrals and ions in the LA of metals and polymers at 308 nm and 1013 composition showed a reproducibility within 1% and isotope ratios agreed within 10% with the natural abundances.W cm-2. Neutral particles from both materials showed Ekin maxima at 150, 115 and 55 eV.This pointed to the formation of ions with Ekin above 100 eV by fast ionization and sub- 4.3 Analytical methodology sequent explosive expansion due to Coulomb repulsion and electron pressure. Dang et al.123 have determined the velocity The complexity of physical processes involved in the one-step laser desorption-ionization (DI ), traditionally used in LMMS, and angular distributions of ions and post-ionized neutrals, ablated by 532 and 355 nm lasers at 108 W cm-2 from an is often seen as the major bottleneck for quantitative analysis.Therefore, Dimov and Chryssoulis129 have fully optimized the La0.67Ca0.33MnO3 target. The data were based on the flight time in a quadrupole, the position of which could be altered experimental conditions in laser ablation with high spatial resolution combined with laser post-ionization and TOFMS for to obtain angle dependent data. The atomic ions and neutrals showed Ekin of 610 and 0.2–0.9 eV, respectively. The ion’s element determination in minerals.Specific attention was given to reducing the laser output variations to 3–4% and the tuning angular distribution followed a cosn function whereas the neutrals exhibited a bimodal distribution. The data pointed to of the ion reflector to discriminate the directly ablated ions from the post-ionized neutrals using their higher Ekin. The involvement of selvedge interactions in the generation of high m/z cluster ions, non-thermal (photochemical ) processes for delay between the ablation and the post-ionization had to be matched to the velocity distribution of the neutrals as well as the direct emission of ions and some neutrals and thermal evaporation for the production of the remaining neutrals.to the distance between the sample surface and the postionization volume. Under these conditions, the precision of The transition between thermal and plasma LA has been delineated experimentally from the Ekin of Cu+ in TOFMS RSFs for Cu using Fe as a reference became less than 14% RSD.using fluences up to 4 mJ at 351 nm on a spot of 0.6×1.2 mm2 on copper targets.124 At fluences below 2.6 mJ only a thermal Poels et al.130 have explored the potential of one-step laser DI at 266 nm in Fourier transform (FT) LMMS for the component was seen with low Ekin and a width of less than 0.3 eV. The sharp increase of the mean Ekin and the increasing elemental speciation analysis of solids.The detection of atomic ions in the presence of the high m/z cluster ions within a distribution width at higher fluence reflected the onset of plasma ablation. From that point onwards, the detected Cu+ dynamic range of typically 100 enabled unambiguous speciation. The improved reproducibility in comparison with TOF yield increased almost linearly with the fluence. Taylor et al.125 have studied the ion yield as a function of LMMS was attributed to an increased time domain for ion collection by the ion trap.Isotope ratios corresponded to the the fluence for yttrium oxide-stabilized zirconia using a 266 nm laser above 106 W cm-2. The rapid increase of the Y+, Zr+, natural abundances within 10–20% and the long-term reproducibility for the measurement of ion intensity ratios was also YO+ and ZrO+ signals reflected the onset of LA at 2.5× 106 W cm-2. Above that fluence, the ion yield increased 20%. The lack of molecular reference samples was seen as a major problem for semi-quantitative work.rapidly with fluence and the mass spectrum resembled that obtained from SIMS, suggesting a similar ion formation by Tsugoshi et al.131 have explored the detection frequency technique in one-step laser DI LMMS for the semi-quantitative surface electronic eVects. Multiple photon excitation was believed to produce electron-hole pairs followed by two hole analysis of heterogeneous samples.Instead of the signal intensity itself, the frequency of occurrence of given signals within localization at surface defects, when the hole–hole Coulomb repulsion energy was lower than the energy gain from lattice a data set was used. Use of an independent technique on the same samples allowed calibration of the detection frequency relaxation. Desorption would occur upon localization of two holes in the same bond. to be made as a function of the sample concentration. Determination of sulfur species in particles at the surface of environmentally exposed green tuV specimens by EDX and 4.2 Instrumentation LMMS showed a good correlation between both methods.Resonant laser irradiation conditions for one-step DI may Sysoev et al.126 have developed a versatile TOFMS instrument allowing laser irradiation of solids and electron ionization of improve the selective generation of atomic ions. Gibson132 has studied this approach for actinide dioxides (AmO2, NpO2 , liquids and gases, introduced through a membrane.Laser spots of 50–500 mm and power densities up to 1011W cm-2 provided PuO2, ThO2 and UO2) pressed together into a pellet using a copper binder. The wavelength dependence of both atomic LODs of 1–10 ng g-1 for various elements in solids. Solutions containing 10 mg l-1 benzene and toluene were analysed. and oxide cluster ion yields could be related to the atomic and/or molecular absorption but in some cases excitation also Isotopic abundances could be measured with an accuracy and precision of 1%.In standardless analysis of a bronze-663 resulted from the collisional energy exchange in the plasma. This made wavelength tuning diYcult. The laser power density material, an improved accuracy of 18% could be achieved by using high power density and taking only the singly charged and delay between the laser irradiation and the ion extraction pulse influenced the relative intensities of M+ and MO+.ions into account. The use of an ion trap to overcome isobaric interferences by Hachimi et al.133 have compared resonant and non-resonant laser conditions for the detection and speciation of Cr, Ni and tandem MS has been demonstrated for the on-line analysis of U and its oxides in airborne particulates.127 The sensitivity Pb in environmental and biological samples. The gain in sensitivity for atomic ion detection was often suYcient to allowed detection of U in particles of 0.5–1 mm diameter and containing about 2 mg g-1 of uranyl nitrate.The low mass allow detection of trace elements in biological tissues. On the J. Anal. At. Spectrom., 1999, 14, 1633–1659 1643other hand, the high m/z clusters to be used for speciation a-spectroscopy to levels of 108 239Pu atoms took 1 day. In addition the latter technique was unable to distinguish between were suppressed under resonant conditions. 239Pu and 240Pu. The selectivity of the ionization allowed use of a simple reflectron type TOF and, in addition, sample 5 Resonance ionization mass spectrometry (RIMS) preparation could be simplified. About 106–107 Pu atoms were suYcient for RIMS analysis of a single isotope.Full isotope Optimization of a quadrupole mass filter for RIMS has been based on ion trajectory calculations to assess the eVects of information with an accuracy of 10% required 107 atoms of the least abundant Pu isotope. The practicability of the method imperfect fields due to the use of circular rods.134 The simulations predicted that additional grounded rods would allow was shown for the analysis of soil samples from the Chernobyl area, Pacific Ocean sediments and urine samples from human a six-fold reduction of the field deviations and thereby a tenfold increase in sensitivity.The calculated influence of the ion volunteers injected with 244Pu. entrance energy, rf phase and the slope of the mass scan on the mass resolution and sensitivity was experimentally verified. A mass resolution of over 1500 at m/z 28 could be achieved 6 Secondary ion mass spectrometry (SIMS) but the transmission was reduced by a factor of 100 in Two, essentially diVerent, modes of operation are covered in comparison to that at a mass resolution of 100.Correction of this section. Dynamic SIMS provides fast erosion under high all mass discrimination factors allowed measurement of isotope primary ion current density and generates mainly elemental ratios to within 0.5% of the expected results.A dynamic range ions and small inorganic clusters. No organic structural ions of 108 was achieved. are formed. Static SIMS (S-SIMS) restricts the primary ion The measurement of isotope ratios in RIMS is hampered by dose so that, in practice, the molecular environment is only the fact that ionization yield for each isotope depends on the hit once. Elemental and molecular information is obtained. laser wavelength, line width and Doppler width.A spectral The use of TOF analysers in S-SIMS has led to a common simulation method could be used to compute the isotope ratio confusion between TOF S-SIMS and TOF SIMS, the latter enhancement factor for 6Li57Li in a two photon resonance referring to dynamic experiments with a TOF instrument. For three photon resonance experiment.135 The isotope ratios the sake of clarity, S-SIMS work is discussed separately. It closely agreed with those obtained using ICP-MS.should be noted that the primary beam incidence angles The pursuit of improved LODs by new photoionization mentioned in this and the next section denote the angle schemes in RIMS has continued. Song et al.136 have elaborated between the beam and the normal to the sample surface. optimal 1- and 2-colour schemes for the TOF RIMS determination of Sm on residues obtained from solution. A LOD of 200 fg from a solution of 10 pg l-1 was achieved. The feasibility 6.1 Instrumentation of a copper-vapour pumped dye-laser for Cd analysis has been explored by Erdmann et al.137 The high ionization potential De Laeter and Kennedy142 have reported the analytical performances of a purpose built high-transmission double-focusing of Cd required frequency tripling for the first step at 228.8 nm.An average power of 2 mW was obtained after tripling. Using magnetic SIMS for isotope ratio measurements in U- and/or Th-bearing minerals with a spatial resolution in the order of a second non-resonant step at 643.85 nm, isotope ratios agreed with the expected values. 20 mm. The transmission reached 95 and 17% at mass resolutions of 5000 and 33 000, respectively. This provided about In addition to strictly analytical applications, the RIMS method is ideally suited for the determination of physical 20 counts per ppm of Pb and per 1 nA of O2- at the mass resolution of 5000 required to separate 176HfO2+ from 208Pb. parameters, such as the ionization potential (IP) of actinide elements on extremely small amounts of sample.An experimen- The deviation found for the 208Pb5206Pb ratio as a result of mass discrimination ranged from 0.5 to 0.1% in feldspar (4% tal method for accurate measurement of the first IP of several actinide elements has been based on the ionization of atoms Pb) and zircon (20 ppm Pb), respectively. The dependence of the U+5Pb+ and UO+5U+ on the experimental conditions by multiple resonant laser excitation in an electrical field and subsequent TOFMS analysis.138,139 The wavelength of the last required alternate analysis of unknown and reference samples.The use of the secondary ion coincidence technique in TOF excitation step was scanned across the ionization threshold at various electrical field strengths. Extrapolation to zero field SIMS for nanoscale information on chemical composition has been explored by Hamza et al.143 Specifically, a W layer on yielded the first IP.Samples of only 1012 atoms (400 pg) were needed for analysis. a patterned silicon–silica wafer was bombarded by Th75+ at 250–790 keV and the coincidence of SiO+ with, for example, The production of an atomic beam from actinide elements is usually achieved through heating of the oxides deposited on a W+ pinpointed the generation of a W+ secondary ion from the oxidized sample region. In addition to a TOF analyser, rhenium filament and covered with thin Re or Pb–Pt coatings. However, by considering the thermal evaporation, diVusion, high secondary ion yields were needed and therefore, highlycharged high-energy primary ions were used.This approach desorption and other processes, Eichler et al.140 have showed that evaporation from a metal surface was inadequate for oVers opportunities for other applications. A new method for characterization of ion counting systems actinide elements. Instead, a sandwich type source was proposed, in which the actinide oxide was deposited on a substrate has been based on an element containing three isotopes of which two were a factor 10 less abundant than the other.144 foil and covered with a highly reductible thin metallic coating. The experimental results for a Ta–actinide oxide–Ti sample In comparison with the two isotope technique, this new approach allowed more consistent and faster evaluation of the were reproducible to within 20–40%, whereas the traditional sandwich sources yielded filament-to-filament variations of a mass fractionation as a function of the ion count rate.A low-cost modification of the primary beam rastering elec- factor 100. Several elements (Am, Bk, Cf, Cm) were determined with LODs in the range of 1012 atoms (equivalent to 0.4 fg) tronics for the Cameca 3F SIMS instrument has been intended to overcome the problem of the original design of the elec- and overall eYciencies of about 10-6. Nunnemann et al.141 have compared RIMS and tronics with high frequency components.145 The new electronics, described in detail, avoided the primary beam spending a-spectroscopy for the trace analysis of Pu in environmental samples.Scanning the first and third laser over the reso- more time at the edges of the crater rather than in the middle. Stylus profilometry confirmed the significantly improved flat- nance range in a three-laser RIMS scheme yielded full isotopic composition within 1–2 h, whereas application of ness of the crater bottom. 1644 J.Anal. At. Spectrom., 1999, 14, 1633–16596.2 Fundamental studies lished the relationship between the fundamental ion-sample interaction and the primary beam incidence angle, oxygen Use of highly-charged primary ions such as 69Au+, 70Th+ and flooding, surface transients and depth resolution. 44Xe+ of low energy changes the ionization process completely A predictive approach to simulate shallow depth profile data and results in more secondary ions per incident ion than with from SIMS has been based on Monte-Carlo calculations of the use of conventional beams.Highly charged primary ion the ion beam mixing in doped d-layers and on the assumption bombardment has allowed more accurate estimates of the B that the real profiles could be described by Pearson profiles.150 concentration (1.8×1021 atoms cm-3) in silica films than was Application of the obtained d-layer response function to As possible with conventional O- SIMS.146 The low doses (5×105 and Sb, implanted at low energies in silicon, showed a good ions) of highly charged primary ions eliminated the need for agreement between the convoluted and experimental data with bright sources.The reduced matrix eVects were attributed to O2+ or Ar+ primary ions. The DRF was inadequate for highthe emission of species independent of their IP as a result of energy high-dose implanted As layers, analysed with a Cs+ the Coulomb explosion, which in turn was driven by the beam.Montandon et al.151 have studied the depth dependence release of numerous electrons from the local microvolume of the DRF by Monte Carlo simulations of the sputtering and upon the incident ion impact. The secondary ion yield the atomic collisional mixing, the theoretical implantation and decreased with increasing thickness of the silica layer on silicon the SIMS depth profile. The reduced atomic mixing near the as a result of the diVerent quenching of the primary and surface was based on Pearson functions. The simulated and secondary Coulomb repulsion explosion eVects by the diVerent experimental SIMS data agreed well except for the tail of the materials. Highly charged primary ions also eliminated the distribution.eVects of the surface transients in conventional SIMS. In addition to collisional mixing, radiation-assisted diVusion A fundamental description of dependence of the RSFs on may broaden the depth profiles of multilayers.Paour152 has the physical and chemical matrix features has been based on modelled this process on GaAs–AlAs multilayers analysed the ‘matrix factor eVect’ (MFE), defined as the slope of the with low primary ion energies. The layer thickness was measfunction relating the logarithm of RSF to the IP.147 The ured by transmission electron microscopy (TEM) using crossexperimental MFE in diVerent matrices could be explained sectioned samples. The agreement between the experimental using local thermal equilibrium and bond breaking models.and simulated Al data (3 keV primary ions) supported the The MFE could be predicted from the heat of oxide formation, assumption that the irradiation-assisted diVusion played a average atomic number and electronegativity of the matrix. predominant role in the broadening of the SIMS depth profile. A simplified quantification by the inverse velocity (IV) The ion yield enhancement caused by the presence of oxygen approach was based on the alternate measurement of the in the sample hampers depth profiling through layers with elemental ion ratios (analyte and internal standard) at two diVerent oxygen concentrations.Wittmaack153 has developed energies instead of covering the entire range for each a quantitative correction procedure by monitoring the B+5Si+ element.148 This ratioing compensated for the eVects of the ratios during bombardment of uniformly B-doped silicon with energy-dependent transmission and changing conditions (for 1.9 keV O2+ (0° incidence) and subsequent sputtering of the example, ion current).The influence of the emission angle on synthesised oxide with the same beam at an incidence angle the transmission was negligible for ions with characteristic of 75°. The B mixing was distinguished from the oxygenenergies up to 100 eV and small for those with 1000 eV. The induced yield enhancement using several assumptions. The results for analysis on SRMs approached the certified values SiO+ intensity (normalized to its maximum intensity) was used to within a factor 2–3.to assess the local oxygen content. Polynomial fitting of the Franzreb et al.149 identified analytically useful low energy ion B+ yield and the B5Si sensitivity allowed quantification of the scattering (LEIS) contributions in a SIMS instrument. The depth profiles from 0.5 and 2 keV 11B implanted silicon, LEIS information is helpful for characterizing the undamaged covered with 6 nm of thermal SiO2.Accurate profiling required surface and the matrix eVects. Specifically, the Ar+ and Cs+ O2+ energies less than half the implantation energies. Ekin distributions upon bombardment of Ag and Pb, respect- An extensive study of the surface roughening, erosion rate ively, showed peaks from the ion–surface backscattering and change and depth resolution for B and Ge d-layers in Si under the direct binary collision event (Ar–Ag around 2300 eV, 1 keV O2+ bombardment at 60°, with and without oxygen Cs–Pb 350 eV).The Ekin distribution of 16O+-sputtered or flooding, has been published by Jiang and Alkemade.154 The backscattered Al,Mg and Si under O- and O2+ bombardment B d-layers were 6 nm thick and 15 nm apart and the Ge layers showed signals from the O–Al, O–Mg and O–Si scattering 0.8 nm thick and 12 nm apart. Atomic force microscopy collisions. The energy resolution was suYcient to resolve (AFM) revealed that the surface roughness at a given depth oxygen backscattering from the 27Al and 28Si peaks.The peak (160 nm) increased as the oxygen pressure was increased from energy of the backscattered O+ correlated well with the 10-9 to 10-6 Pa but dropped thereafter. The roughness of the calculated energy. Although SIMS provided less resolution crater bottom at a given pressure increased with depth but did and sensitivity than dedicated LEIS instruments, 1% of S not depend on the dopant concentrations. The erosion rate could be detected by LEIS in SIMS.was fairly constant with depth at O2 pressures up to 10-6 Pa, The increasing demand for analysis of ultra-shallow junc- but showed a fast decay in the first 20 nm above 10-4 Pa tions in micro-electronics has stimulated research on improving corresponding to the onset of surface roughening. Comparison the depth resolution through understanding of the fundamental of the B and Ge layers showed that chemically induced processes involved.The SIMS primary ion bombardment may relocation was less important. The full width at half maximum induce mass transport of, for example, dopants in silicon by (FWHM) of both B and Ge peaks went through a maximum collisional mixing and radiation-enhanced diVusion. This with increasing oxygen pressure but at high pressure the decay atomic transport is a function of the beam (ion mass, energy, length of Ge increased whereas that of B decreased as a result incidence angle) and of the sample material itself.It leads to of the sample swelling due to surface oxidation under flooding broader profiles with a maximum shifted to a greater, ‘appar- conditions. As a result, it follows that the FWHM and the ent’, depth. The depth resolution function (DRF) transforms decay lengths should not be used to define the quality of the the orginal distribution into the actual experimental sputter depth profile. depth profile. Several papers have focused on the development The depth resolution as a function of the incidence angle with of improved DRFs because this allowed identification of the and without oxygen flooding has been determined for alternating Co and Cu layers of 30, 50 and 100 nm thickness on glass important mechanisms.However, other studies have estab- J. Anal. At. Spectrom., 1999, 14, 1633–1659 1645substrate using 3 keV O2+ at 0° to 70° angles of incidence.155 accuracy of typically twice the precision.However, the applicability of these models to other minerals proved to be a problem. Without oxygen flooding, the FWHM of the interfaces varied from 20 nm to over 100 nm with incident beams of 0° and Dynamic SIMS is widely used for the Pb–Th–U dating of Th-and/or U-rich minerals since it is generally assumed that 45°, respectively. A mean depth resolution of 10 nm was achieved with an incidence angle of 52° and O2 flooding. Use no matrix eVects influence the Th5Pb and U5Pb ratios.Nevertheless, Zhu et al.162 have shown significant matrix of AFM showed reduced topographic development for oxygen implantation at 0° incidence and for oxygen flooding under eVects dependent on the ThO2 content of the mineral. The energy distribution of 206Pb was closer to that of 238UO2+ glancing primary ion incidence. Arnoldbik et al.156 studied the changes in the depth profile than to that of 238U+ so that the signal intensity ratio 238UO2+5238U could be used to calibrate the working con- area, position and shape for a Ge d-layer (13.2±0.5 nm below the surface) in silicon during sputtering with a 2 keV O2+ ion ditions (energy filter setting).A linear relationship between 206Pb5238U and 238UO2+5238U+ was found for monazites with beam. A second, marker, d-layer was at a depth of 23.5±0.5 nm. Various doses of 2 keV O2+ ions at 53° incidence low ThO2 content (1–6%) but deviations occurred for monazites with higher concentrations of ThO2 (15%).were applied before analysis by high resolution Rutherford Backscattering Spectroscopy (RBS) and SIMS (700 eV O2+ Accurate isotopic distributions with high spatial resolution are needed to trace isotopically distinct reservoirs in minerals at 77° incidence). The bombardment with 2 keV O2+ ions did not aVect the results for the upper Ge d-layer until the receding or meteorites. Riciputi and Greenwood163 have addressed the problem of mass discrimination which occurred when surface was within about 5 nm of the d-layer.When the surface was 3 nm from the d-layer, removal of Ge became apparent. inclusions of a size of the beam diameter contained diVerent phases. The measurement of 13C and 34S in fine grain mixtures At the same time, some Ge atoms were driven deeper into the layer so that the profile shifted inwards by about 0.7 nm at of carbonates, iron sulfide and iron oxide showed no additional mass discrimination when the co-existing matrix did not the plane of maximum Ge concentration. contain the element of interest.Hence, large grain or homogeneous standards could be used for calibration. However, if 6.3 Analytical methodology the element was present in the grain and the matrix, its distribution over the two phases had to be known. Hiyagon164 Although reactive primary ions from oxygen are widely used to convert the surface layer into oxides and thereby improve studied the oxygen isotope distribution and Fe5Mg ratio in 18 mm spots on olivine grains from the Allende chrondrite and the ion yield, little attention has been given to exploit other reactive species.However, filling the source with neutral N2O, terrestrial samples. Careful alignment of the slit axis, detector dead time corrections and use of a similar count rate were NO and NO2 reactant gases during the bombardment of 3d, 4d and 5d transition metals showed that some metals induced required to minimize the mass discrimination.The reproducibility attained was 0.3% for the 17O516O ratio and 0.2% for a dissociative adsorption with the result that the nitrogen oxide molecule acted as an oxygen source and increased the the 18O516O ratio. Coakley and Simons165 have elaborated a statistical method to identify the source of isotopic ratio ion yield significantly.157 However, some metals underwent molecular adsorption and therefore yielded only additional heterogeneity as being instrument- or sample-generated. The approach was tested on Cr data from a homogeneous RM metal nitride ions. Use of electron-assisted O2+ bombardment with oxygen back- and a simulated data set for two isotope ratios.A depth profile correction for shallow B implants in silicon fill has increased the depth resolution as well as the dynamic range for shallow depth profiling of B in silicon.158 was based on collisional mixing being the main reason for profile broadening and asymmetry under O2+ bombardment Measurement of the 30Si+ and 44SiO+ signals led to the conclusion that the accelerated surface oxidation by the elec- at incidence angles above 30°.166 The correction restored the original distribution with a second moment deviation of less tron beam was responsible for the reduction of the transient width by 0.3–0.5 nm.than 2 nm unless theFWHMof the original profile approached the FWHM of the DRF. The sputter and ionization yield in round particles with a diameter of a few tens of mm are usually dependent on the incidence angle of the primary beam.Tomiyasu et al.159 have 6.4 Quantification circumvented this dependence by cleaving the particle using line scans with a Ga+ liquid metal ion gun in the high current The validity of the IV method for quantification of Au in pyrite has been investigated by Pratt et al.167 The Au levels from mode. The flat surface of the repositioned particle was then analysed with the Ga+ gun at high lateral resolution.Elemental application of the IV method to a Au implanted reference sample agreed with the calculated implanted dose within 3%. ion maps with a lateral resolution of 50 nm were obtained from a 3 mm TiO particle surrounded by a 0.15 mm SiO2 shell. Quantification of Au in individual pyrite grains showed good agreement between the IV approach and the matrix-dependent Schvachko160 has observed that the measurement of H- as opposed to the H+ signal allowed the interferences from RSF method for samples with finely distributed and structurally bound Au, but large deviations for colloid-size inclusions.hydrocarbon contaminants in the H depth profiling of hydrogenated steel to be avoided. Plotting the accumulated H- Although TOF SIMS is primarily used for analysis under static conditions, dynamic experiments at low erosion rates signal against the primary ion current density gave a pulseshaped profile in the case of a surface-adsorbed contamination may allow quantification of trace elements in the upper monolayer( s) of the sample.Schro�der-Oeynhausen et al.168 have layer, whereas a thick layer yielded a plateau and bulk hydrogen gave a steadily rising function. Using this simple determined Al, Ca, Co, Cr, Cu, Fe, Mg, Ni, Si and Zn in silicon with the aid of GaAs standards. Before TOF SIMS approach, consistent correlation between the SIMS data and the hydrogen exposure time of the steel could be achieved. analysis, surface cleaning by UV–ozone was applied to remove contaminants and to convert the surface into a reproducible Riciputi et al.161 have systematically investigated the mass bias of light isotopes in geological materials.In addition to the oxidation state for improved ion yield. The atomic signals, normalized to Ga+, correlated linearly with surface concen- primary ion beam, secondary ion energy and polarity, the matrix composition also played a major role. Various corre- trations between 1010 and 1.5×1012 atoms cm-2.With a 50 mm beam diameter, LODs were 2–4×109 atoms cm-2 for lations between the mass bias for C and O and the matrix composition were proposed. Empirical models allowed predic- all elements except Zn, for which the LOD was 1.3×1010 atoms cm-2. tion of the mass bias for O isotopes in some silicates with an 1646 J. Anal. At. Spectrom., 1999, 14, 1633–1659Nicolescu et al.169 have elaborated the quantification of rare bombardment of the multi-layer films of YBa2Cu3O7-d and 10% Co-doped YBa2Cu3O7-d on LaAlO3.The depth scale was earth elements (REEs) in minerals at a lateral resolution of 60 mm. Energy filtering and peak corrections eliminated the calibrated by surface profilometry and the topography development was assessed by AFM. The primary beam’s incidence interferences of the light REE oxide ions on the heavy REE atomic ions at a mass resolution of 300–500. The REE angle (0–60&de) did not aVect the length of the leading edge in the Co depth profile.However, the trailing edge decay length determinations in four Ca–Al silicate glasses by SIMS and LA ICP-MS (requiring about 1 g of sample) agreed within 10%. increased drastically above 30° incidence angle for Xe+ and O2+ ion beams. The topography development with O2+ was The accuracy for a NIST SRM 613 glass was within 10% for 8 out of the 10 REEs at a concentration between 0.2 and 14%. slower than with Xe+, but the decay length of the leading and trailing edge did not shorten as expected.The counting precision was within 5% and LODs as low as 20–100 ppb. Fahey170 has developed a modified multi-pass A Ta2O5–SiO2 multi-layer film on silicon has been proposed as a new reference material for SIMS depth profiling to replace algorithm to deconvolute the energy-filtered ion signals into atomic and oxide ion contributions through the use of energy- the NIST SRM 2136 with Cr and Cr2O3 layers.177 The identification of the oxide markers by the enhanced Cr signal insensitive ratios.However, RSFs were matrix dependent and varied up to 20% for diVerent materials. prevented use of oxygen sputtering. Argon sputtering induced considerable topography problems. The new standard con- A sensitive high mass resolution (6000) ion microprobe for the isotopic analysis of Pb in 100 mm3 of solid samples has sisted of seven Ta2O5 layers of 18 nm thickness, separated by 1 nm d-layers of SiO2 grown by sputter deposition under given a within-session RSD of 0.1% and a between-session RSD of 0.5%.171 The ratios agreed, within experimental error, oxygen.The choice of Ta2O5 avoided significant surface topography development during sputtering and the SiO2 layers with results obtained by TIMS and LA-ICP-MS. The mass fractionation was below the uncertainty due to the counting eliminated the matrix eVects in metal–metal oxide multi-layers. High resolution TEM revealed sharp interfaces (,0.5 nm) and statistics.Application to microstructures was demonstrated by the measurement of isotope ratios in pyrite inclusions in large AFM showed that the surface roughness only increased from 0.22 to 0.28 nm (Ar+ 7 keV at 60° incidence) when all films depside crystals with an RSD of 0.12%. The use of SIMS for U-contaminated urine samples as a were sputtered. In SIMS, a constant FWHM of 3–4 nm was obtained for all layers. faster alternative to nuclear counting has been based on sampling with films of polypyrrole with alkylammonium Du�mmler et al.178 have warned against the use of crater bottom roughness to assess topography development in multi- groups.172 Dipping of the film into the solution yielded a homogeneous loading ( laterally and in depth) of Pu and U.layers. Manually polished uncoated and silver-coated silicon nitride ceramics were bombarded with a neutral beam to avoid The LODs were 10-11 and 10-10 g l-1 for Pu and U, respectively, and the isotopic ratios were accurate within the exper- sample charging.The roughness of uncoated silicon nitride decreased after erosion of the amorphous silicon nitride layer imental RSD of 1–6%.173 Signal subtraction was used to correct for 238UH and 239Pu interferences, the separation of formed by polishing in the first 100 nm. Below this layer, surface roughness increased until a constant value was reached which would require a mass resolution of 35 000.at a depth of 1.5 mm. In contrast, the roughness of the silvercoated sample reached a maximum at the interface and 6.5 Single and multi-dimensional analysis decreased again to a constant level. In order to determine the interface layer thickness, diVusion and other parameters, 6.5.1 Depth profiling. Hofmann and Schubert174 have developed a model for the DRF in SIMS and Auger electron the topography during the depth profiling, not reflected by the roughness of the crater after analysis, must be known.spectroscopy, taking into account the mixing influence, surface roughness and information depth (MRI). Parameters used in Detection of MCs+ has been used in GaAs and InP molecular beam epitaxial films to determine the excess of the group the model were estimated from the analysis of well-defined samples under constant conditions. The MRI assumptions V element.179 The MCs+ intensity was found to be linear with concentration and independent of matrix eVect so that the were checked on B d-layers in silicon, GaAs–AlAs and SiO2–Ta2O5 multi-layers.The achieved precision of 0.2 nm, non-stoichiometry could be measured whether the element excess was present as point defects or as precipitates. These about one atomic layer, showed that non linear eVects such as preferential sputtering were well accounted for. However, advantages compensated for the poorer sensitivity (0.1 and 0.2% for As and P, respectively) that was achieved in compari- segregation and radiation enhanced diVusion were not incorporated into the model and could limit the applicability son with atomic ion detection.Horita et al.180 have developed a new experimental approach of MRI. Svensson et al.175 have surveyed the possibilities and the to measure the grain boundary diVusion of Sr in a (La,Ca)CrO3 perovskite-type oxide. The traditional experiment involves limitations of dynamic SIMS for doped epitaxial layers of silicon and 6H-SiC used in power and micro-electronics.application of a Sr(NO3)2 layer on the perovskite and depth profiling to yield a profile that can be resolved in the bulk and Adequate LODs made it possible to trace contamination levels of, for example, Ti of the order of 5×1012 atoms cm-3. The grain boundary diVusion eVects by models. Cleaving the sample perpendicularly to the Sr(NO3)2 layer and applying diVusion of H into the material was another source of electrically active defects where SIMS could be used, especially in SIMS with low energy primary ions along the cross section exposed the deeper grains for improved depth profiling of the those cases in which D instead of H implantation was used. This improved the LOD in depth profiling to 1015 atoms cm-3 shallow layers on each grain.The accurate depth profiling of Na in SiOx layers, crucial in from the 1018 atoms cm-3 possible with H. Depth profiling of a Ge d-layer at 36 nm below the surface of silicon using the manufacture of semiconductor devices, is hampered in SIMS by the migration of Na under the action of electrical 5.7 keV Ar+ at 40° incidence yielded a distribution which was shifted and broadened by recoil mixing.The shorter decay fields. Charge compensation is therefore critical. Saito and Kudo181 have proposed the chemical etching of layers less length at the trailing edge of the depth profile using O2+ ions at 40° incidence was due to the swelling of the silicon matrix than 3 nm thick using diVerent concentrations of HF followed by silver coating.Comparison of the etching method with the upon oxygen incorporation. However, O2+ bombardment induced segregation and broadening the profiles of, for traditional use of sample cooling or self charge compensation proved that the new procedure completely eliminated the example, Au, Cu and Ni. Montgomery et al.176 have optimized the Xe+ and O2+ migration towards the SiO2–Si interface so that the data J.Anal. At. Spectrom., 1999, 14, 1633–1659 1647closely agreed with the simulated implantation. Disadvantages position at grain boundaries with a sensitivity down to the sub-ppm level and a lateral resolution of 1 mm. The matrix are the need for large number of samples and the requirement of Na-free solutions for etching. eVect was not a major problem in alloy analysis. Brunner et al.190 have imaged non-metallic carbide impurities of the Lundgren et al.182 have performed depth profiling in biological samples with a lateral resolution of 50 nm.This approached type MC and M2C in high speed steel. Stigmatic experiments provided suYcient lateral resolution (2 mm) to visualize the the limits of modern SIMS in a traditionally diYcult application. The elemental composition in rat incisor dental enamel carbides. A scanned, focused beam was required to image the oxide layer around prepitates of 2–3 mm with a lateral during diVerent stages of the tooth development was studied.The mineralized material circumvented the sample preparation resolution of 300 nm using the 16O- signal. problems encountered with soft tissues. The lateral resolution 6.5.3 Three dimensional (3-D) analysis. The lateral reso- proved adequate to characterize the biological microstructures lution of SIMS is rather poor (1 mm) in comparison with its so the data could be related to various physiological processes. depth resolution in the nm range.Image cross-correlation links Simultaneous multi-element detection in TOF SIMS has been image features at identical lateral coordinates in planes at exploited in the depth profiling of chromium-based conversion diVerent depths. The viability of the approach was evident layers on aluminium.183 The indirect speciation resulting from from the data for D-enriched continuous pores in a D2O- the panoramic elemental analysis made it possible to confirm exposed zirconium oxide film.191 Clustering of Au in the ‘raw’ the duplex nature of the surface layer with CrxPyOz at the top images from arsenopyrite could be attributed to the ‘blooming’ and aluminium oxide and F at the interface. The sensitivity of of the microchannel plate.TOF SIMS for F- proved superior to that of Auger electron spectroscopy. 6.6 Multi-technique approaches 6.5.2 Imaging. Verlinden et al.184 have achieved a lateral Hoskin192 has reviewed the use of dynamic SIMS and resolution of 70 nm in TOF SIMS imaging of halides in the LA-ICP-MS for the determination of minor and trace elements surface layer (,4 nm) of silver halide microcrystals of 700 nm.in natural zircon (ZrSiO4), which is of interest in Pb–U Iodide could be imaged in layers with a global concentration geochronology. The results obtained by the two methods of 0.4% iodide. Imaging TOF SIMS also allowed the rapid agreed within 20% for homogeneous samples but large devisurvey (within minutes) of large collections of particles. ations occurred in the analysis of heterogeneous materials The combination of imaging and indirect speciation through because the two methods provided information over diVerent the measurement of atomic and small cluster ions has been depth ranges.Least squares regression analysis and rejection demonstrated in a study of thin carbon nitride films produced of the outliers in LA-ICP-MS somewhat reduced the amount by a graphite hollow cathode.185 The CCs+ and CNCs+ ions of analytical information but improved the agreement becould be used to distinguish between pure graphite and carbon tween the mean concentration values.Although analysis with nitride. Application toMxCyNz films showed a close agreement LA-ICP-MS is much faster than that with SIMS, and oVers between the CNCs+5CCs+ ratios and the stoichiometric similar sensitivity, precision and repeatability, SIMS is more values. suitable for the analysis of heterogeneous samples.Isotopic analysis in biological tissue at high mass resolution The study of microscopic structures is complex and makes and sensitivity has allowed the study of the transport of tracer the complementary use of diVerent techniques of increasing ions, such as 26Mg, in cryosections of killifish.186 The images importance. Pe�re�z-Rodrý�guez et al.193 have reviewed combiwith a resolution of 0.5 mm were digitally corrected for the nations of physical techniques for the characterization of layers, natural 26Mg contribution and the interfering ions, identified structures and processes in Si technology.In addition to describat a mass resolution of 4000. The resulting uncertainty of less ing the fundamental aspects of SIMS, RBS, X-ray photothan 10% was acceptable in biological imaging since the electron spectroscopy, Auger electron spectroscopy, X-ray variations within one treatment could be as high as 25%. diVraction, AFM, Raman spectroscopy and FT-IR spec- Determination of the isotopic composition of uranium- troscopy, attention was focused on their suitability to study containing microparticles in swipe samples stemming from one or more features, such as amorphous, polycrystalline and environments with nuclear activities or lung tissue sections is single crystal layers, grain size and crystal orientation, disorder, a challenging task in micro-analysis.Tamborini et al.187 have defects, strain and composition.This information is required optimized the instrumental SIMS conditions for fast analysis not only for surfaces but also for interfaces in multi-layers. of carbon-coated swipe samples. A mass resolution of 1000 Hull and Dunn194 have surveyed the combination of focused proved to be suYcient, except for the separation of 235UH ion beam SIMS and TEM to get detailed information on from the artificial isotope 236U. Experiments on certified micro-electronic materials. A resolution of the order of 10 nm3 control particles showed accuracies of 1.2 and 3% for per pixel made 3-D mapping in SIMS profitable in these 235U5238U and 234U5238U, respectively.Images at m/z 238 applications. Under these conditions, determinations of conwere recorded for swipe samples over the entire area and the centration at the 0.1% level are feasible. Examples of analysis actual isotope ratio was derived from data taken within a involved determination of dopant concentrations of 1017 smaller raster. Results obtained by SIMS for environmental atoms cm-3 with a resolution of 30 nm.samples agreed, within 5%, with those obtained by traditional methods (fission track; TIMS). Amaral et al.188 have studied 6.7 Static SIMS (S-SIMS) 1 mm sections of embedded lung tissue of rats instilled with UO4 suspensions and of macrophages from bronchoalveolar Vickerman195 has published a review with 69 references on the use of S-SIMS in the surface analysis of the upper monolayers lavage, spiked with UO2.No sample charging was noticed. The Ca+ images characterized the tissue morphology, whereas of a solid. The processes of sputtering and ionization under static conditions were extensively treated, as well as the recent the UO+ images localized the particles. The 235U+5236U+ ratios were determined under focused primary ion bombard- developments with respect to polyatomic primary ions and post-ionization by lasers under resonant and non-resonant ment and agreed within 1% with the expected abundances.A large niche for dynamic SIMS has been forged in the conditions. Interesting experiments in the field of inorganic catalysts and atmospheric chemistry were described. A major semiconductor industry but great potential exists in many other fields. McMahon189 has discussed the use of SIMS in advantage of the S-SIMS information was the formation and measurement of higher m/z ions, which were needed to specify metallurgy.Ion imaging displayed directly the chemical com- 1648 J. Anal. At. Spectrom., 1999, 14, 1633–1659the original chemical form of the element in the solid sample plasma ionization was adequate for depth resolution but suVered from low lateral resolution. (as opposed to, for example, dissolution of the sample and application of hyphenated techniques). 7.2 Analytical methodology The use of S-SIMS for inorganic speciation analysis has been compared to other micro-analytical techniques for direct Post-ionization by the electron component of a special low- characterization of solids at high spatial resolution.196 The pressure rf-plasma oVers high ionization eYciencies of the ratio between the signals for elemental ions and ions at higher order of 1%. In addition, the plasma ions can be used for m/z, needed for speciation, was found to be higher in S-SIMS sputtering at low energies (for example, 100 eV) to minimize than in FT-LMMS.Hence, the molecular information in the atomic relocation in the sample.Gnaser et al.201 have S-SIMS tended to be lost in applications close to the detection determined detection sensitivities for several elements in alloy limit. On the other hand, S-SIMS provided monolayer infor- RMs and a GaAs wafer doped with 1.3×1017 Te cm-3 using mation depth, whereas FT-LMMS generated ions from a a double-focusing magnetic MS at lolution (300) and 10 nm surface layer. Quantification in both S-SIMS and high transmission (1–10%).The relative signal intensities FT-LMMS was hampered by the diYculty in preparing suitreflected the relative abundance in the sample with a dynamic able standards because of the method’s sensitivity to the range of 9 orders of magnitude. The LODs were in the low (homogeneous) molecular surface composition. ppb range and even the minor 120Te isotope yielded a S/N of Matrix eVects in S-SIMS for adsorbates on metals could be 20 at a Te concentration of 2.85 ng g-1.reasonably well described by the change in work function The absence of matrix eVects in rf plasma SNMS of noncaused by the adsorbate.197 As a result, the peak intensity conducting samples is often quoted as a major advantage. ratios of, for example, RhNO+5Rh+ could be used directly Analysis of standard soil samples has given an agreement to quantify the elemental surface concentration during the within 20–30% with the certified concentrations down to 0.01% reaction of nitrogen-containing gases at the surface of m/m without matrix correction of the elemental sensitivities.202 Rh(100).Temperature programming allowed kinetic infor- However, significant matrix eVects were noticed for sintered mation to be obtained. ceramic samples of Si0.05(AlxTi1-x)0.95Oy (with x=0–1).203 The advantage of panoramic detection in TOF S-SIMS for The RSFs of Si, using Al or Ti as reference element, depended ion imaging has been exploited in a study on the local strongly on the sputtering frequency and the reference element heterogeneities in freshly cleaved calcite surfaces exposed to concentration in the sample.Anomalies in the Ekin distriair. 197 Atomic ion images were recorded from specific areas butions were attributed to the strong emission of positive by TOF S-SIMS and scanning force microscopy was used to secondary ions in the negative half of the rf cycle. complement the composition maps with detailed morphologi- Jenett and Hodoroaba204 have studied the interference from cal information.The Ga+ beam provided an elegant way to the tertiary B+ ions in the rf plasma SNMS analysis of a mark the areas of interest on the surface. The spontaneous H3BO3–Cu powder pellet. Tertiary ions arose from the breakformation of salt crystallites along intersecting cleavage direc- down of stable secondary ions (BO- and BO2-) by interaction tions on the calcite surfaces could be observed as well as the with high energy plasma electrons. The tertiary B+ could diVusion of monovalent ions (Cl, F, K, Na) from the bulk to disturb the B quantification in SNMS using the common the surface.This application has important implications for energy window of 50–60 eV. The energy distributions of the the common geological practice of using trace element and diVerent species revealed a ratio of tertiary5sputtered ions in isotope ratios in calcite for dating. the range of 10-4 to 1, depending on the energy window used.Chakraborty et al.199 have stressed the importance of S-SIMS as an integrated method for elemental and organic 7.3 Depth profiling analysis to explain the weight increase observed following the cleaning of a ‘standard’ kilogram. The build up with time of Higashi et al.205 have made a comparison of grazing-incidence SNMS with laser post-ionization, Auger electron spectroscopy oxygenated species at the surface could be demonstrated but the majority of ions immediately after cleaning were derived and dynamic SIMS for the depth profiling of an In0.53Ga0.47As–InP multi-layer sample.The In0.53Ga0.47As from saturated hydrocarbons. A semi-quantitative assessment pointed to a weight increase of 3.3 mg. interlayers were 12 nm thick and the four InP layers were 50 or 100 nm thick. The 10 keV Ar+ primary ions at a grazing angle of 77° used in SNMS provided better depth resolution than the 1 keV Ar+ beam (at an angle of 70.3°) used in Auger 7 Sputtered neutral mass spectrometry (SNMS) electron spectroscopy, even though sample rotation was used in the latter.Matrix eVects caused ‘peaks’ and ‘dips’ in the 7.1 Reviews SIMS (2 keV O2+ primary ions at 81° incidence) profile of In at the interfaces of the In0.53Ga0.47As interlayer. In contrast, The analytical techniques of SNMS and dynamic SIMS share the primary ion bombardment of the sample under high grazing-incidence sputtering and laser post-ionization SNMS allowed depth profiling with little matrix eVect and a resolution current density conditions.Whereas dynamic SIMS detects the directly emitted secondary ions, SNMS uses post-ionization of 2 nm (instead of 8 nm for normal incidence) up to a depth of about 400 nm. of the sputtered neutrals by electron ionization, electron gas or plasma ionization, resonant or non-resonant laser ioniz- SNMS with electron post-ionization has been useful for depth profiling multilayer systems such as resistive contacts consisting ation. Mathieu and Le�onard200 have reviewed the diVerent post-ionization methods and the analytical use of SNMS in of Pd–Ge–Au–Pd–Au layers on n-type GaSb grown by molecular beam epitaxy.206 The individual layer thicknesses (10 and comparison with SIMS.The absence of matrix eVects and the more uniform ion yield across the periodic table in SNMS 100 nm) were verified by scanning transmission electron microscopy. The depth resolution of SNMS was about 3 nm compensated for the higher sensitivity and dynamic range of SIMS.Although resonant laser post-ionization limited the and the absence of matrix eVects simplified the study of interdiVusion of elements between diVerent layers. information to only one isotope, it provided a dynamic range in SNMS of 106 and a LOD of 2×1013 atoms cm-3 of Co in Electron gas SNMS in the hf mode has been applied to the analysis of lithium and barium disilicate (Li2O 2SiO2 and silicon.Non-resonant conditions provided superior ion yields in comparison with electron and plasma ionization but photo- BaO 2SiO2) coatings on silica glass substrates.207 Use of AFM revealed that the LiO2 2SiO2 coating was mostly crystalline fragmentation could limit the quantification. Electron gas J. Anal. At. Spectrom., 1999, 14, 1633–1659 1649and the SNMS depth profiles showed the complementary achieved for test sample concentrations .500 ng ml-1. This had yet to be achieved for samples separated on an LC fluctuations for SiO2 and Li2O.The depth profiles of the BaO 2SiO2 coating showed that some Al2O3 remnants from column. Some limitations were apparent in the system. Memory eVect resulted in measured d values being dependent the substrate grinding had diVused into the films. The erosion rate of 0.2–0.3 nm s-1 allowed analysis of a 500 nm layer to on sample size, with some drift with time. In addition, the yield for the test solutions was only 55% following possible be made in 30–40 min with a depth resolution of a few nm and no need for matrix corrections or charge compensation.loss of sample in the coating device. The possibility of isotopic fractionation during this process was not considered. Two papers have stressed that SNMS should be considered more often for material analysis as an alternative to SIMS. In Abramson and co-workers,221 pioneers of the LC-SIRMS technique, have developed a LC introduction system that contrast to the complications from the largely diVerent ion yields in SIMS for oxide–metal systems, SNMS has been very allowed continuous monitoring of the 13C512C isotope ratio in non-volatile materials without derivatization.well suited to the study of the addition of an alloying element such as Nb to increase the oxidation resistance of Ti–Al based Goodman222 has built a combustion interface optimized to preserve chromatographic integrity while maintaining the alloys.208 By heating the alloy first in air (preoxidation) and then in an atmosphere of 15N–18O, diVusion from the bulk required functions of solvent diversion, combustion and water removal.The peak shapes were preserved by minimizing into the oxide scales could be distinguished from the diVusi of atmospheric oxygen and nitrogen into the surface. Bock changes in tubing diameter and dead volumes. A single piece of fused silica capillary connected the GC to the MS, thereby et al.209 considered the advantages of hf mode SNMS depth profiling of Ag colloid-containing sol–gel films on glass to be eliminating most fittings.In comparison with the original commercial design, the new system was robust, required less the absence of matrix eVects and the capability to analyse dielectric samples readily. maintenance and reduced leaks. The impressive performance of LA-SIRMS has been demonstrated by Hoefs and colleagues223–225 for the measurement of 8 Stable isotope ratio mass spectrometry (SIRMS) O2 isotope ratios in geological materials.The sample was 8.1 Reviews vaporized with a KrF laser in a stainless steel reaction chamber filled with gaseous fluorine. The lateral resolution was 400– The review of Brenna et al.210 on high precision continuous- 500 mm, corresponding to about 0.1 mg of sample. flow isotope ratio mass spectrometry provided the reader with a wealth of information on all aspects of the technique, 8.3 Analytical methodology including the fundamental details of SIRMS, dual inlet analysis, continuous flow analysis, compound specific analysis and The analytical system of Takahata et al.226 was designed for position specific isotope analysis.Although the review of the isotopic measurement of nitrogen at the sub-nanomole level Burlingame et al.211 predominantly dealt with organic MS, a in rock samples, some three orders of magnitude lower than substantial section on SIRMS presented a comprehensive conventional techniques.A key feature was adoption of a survey of developments mainly in the period 1995–1997. crushing technique carried out under vacuum for nitrogen Reviews restricted to more specific subject areas have been extraction instead of the conventional stepped heating. In those on 15N methodologies,212 the analysis of atmospheric addition, a low background vacuum line was used for the gases213 and applications in forensic science214 and pharmaseparation and purification of nitrogen.Use of a magnetic ceutical research.215 Of particular note was the 141 reference sector instrument allowed determination of Ar, He, CO2, H2O review of Krueger,216 which covered in some detail the importand N2 concentrations and Ar and N2 isotopic compositions ant application of SIRMS to the study of food authenticity. to be made in the same sample. Petzke et al.227 have used GC-combustion-SIRMS for the 8.2 Instrumentation first reported measurement of 15N at the natural abundance level in plasma protein amino acids.Use of N-pivaloyl isopro- The design and operation of MS instruments for use in space missions has been receiving the attention of Pillenger and pyl ester derivatives was important for minimizing the total gas load generated during combustion. This extended the life colleagues. Although only concepts, the designs were discussed in detail with particular relevance to the applications that of the oxidation and reduction catalysts and reduced the retention times.The mean precision for all analysed amino would be desirable within the limitations placed by the need for low weight and in situ operation by remote control. acids was ±0.71‰ for d15N. It should be noted that results could not be presented for all the amino acids as some were Modulus217 was the name given to a philosophy for the measurement of isotope ratios to a high level of accuracy destroyed in the process.Leckrone and Hayes228 examined in some detail water- within a self-contained and space-compatible experiment. It has been designed for a mission to study cometary materials. induced errors in continuous-flow carbon isotope SIRMS. The formation of HCO2+ from CO2 and background H2O resulted The general philosophy could also be used for instruments operating on planets, for example on a Mars lander.218 A in enhancement of the m/z 45 ion current and introduced a systematic error.Failure to maintain constant H2O5CO2 ratios remarkable feature of these designs was the very low weight of 3 kg, with even lighter instruments considered possible for during measurement of ion-current ratios of samples and standards caused a significant error. A general relationship more specialized applications. A key to achieving these low weights would be use of an ion trap instead of the permanent between observed error in d and H2O5CO2 ratio could be derived. The errors could be reduced five-fold by shortening magnet of a conventional SIRMS instrument. Preliminary investigations have been reported219 for reducing the mass, the residence time of CO2 in the ion source by use of stronger ion-extraction fields. The disadvantage of this was a 60% power and volume requirements of an ion trap system designed specifically for the measurement of stable isotope ratios. decrease in sensitivity.The recommended solution to the problem was to remove water before analysis through use of, Brand and Dobberstein220 have tested the first LC-combustion- SIRMS system for the measurement of carbon isotope for example, Nafion dryers.OV-line techniques for the preparation of hydrogen are ratios at natural abundance. The system was based on a moving wire technique in which the LC eZuent was deposited generally diYcult to perform, time consuming and, to some extent, operator dependent. Kelly et al.229 have tested a robust onto a wire and transported through evaporation and combustion ovens. Accuracy and precision of d13CPDB ,1‰ was on-line system based on an elemental analyser coupled to an 1650 J.Anal. At. Spectrom., 1999, 14, 1633–1659instrument capable of resolving the tail of 4He+ from 1HD+. toNH4 +, which was transferred to the acid trap as in the first step. Quantitative diVusion could be achieved in three days Samples were pyrolysed in a helium stream at 1080 °C over an inert form of carbon and the H was separated from other when the containers were shaken continuously in a vertical rotary shaker.A feature of the method was that no intermedi- gases by GC. The precision of on-line dD measurement was ,3.3‰. No memory eVect was observed when measuring ate step of ammonia volatilization was needed before NO3- conversion. A limitation of the method was the need to know natural abundance samples. The accuracy of enrichment determination by SIRMS total NH4+ and NO3- contents so that the optimum sample volume could be taken.Lauf and Gebauer237 converted depends heavily on the calibration of the working standard. Lee et al.230 used highly purified (.99.99% enriched) D2O for trace amounts of NH3, NO and NO2 to N2 in an on-line GC-combustion-SIRMS method for the study of nitrogen calibration of the working standard in D51H measurement. Serial dilution of the D2O was used to construct a standard cycling in ecosystems. A major limitation of the method was the large sample of 100 l of ambient air needed to produce the calibration curve.An error analysis of the whole method showed that most of the variation came from sample handling, minimum sample size for N isotope analysis. The discovery of mass-independent fractionation of oxygen including the serial dilution of the D2O. Mak and Yang231 have reported the measurement of both in atmospheric gases has prompted studies into this new category of isotopic eVect. Brenninkmeijer and Ro�ckmann238 13C and 18O in atmospheric CO using continuous flow SIRMS. Atmospheric CO2 and H2O were removed from samples used a rapid method for the conversion of CO2 to O2 for isotopic analysis as isobaric interferences prevented direct measurement cryogenically and the CO oxidized to CO2 for purification by GC.Analytical precisions of±0.2 and 0.4‰ could be achieved of 17O516O in CO2. The method was based on conversion of CO2 to H2O and CH4 by reaction with H2 over a nickel for d13C and d18O, respectively, in 5 nl samples of CO.These precisions were slightly poorer than those achievable by con- powder catalyst at its Curie point, followed by reaction of the H2O with F2 to produce O2. Novel aspects of the procedure ventional du-inlet SIRMS but the sample requirement was more than two orders of magnitude lower. included the use of 5% F2 in He to reduce health and safety risks and the use of a sapphire single crystal reaction tube for both reaction steps.The method allowed d17O values to be 8.4 Sample preparation measured with a precision and accuracy better than 0.2‰. A potential problem was caused by additional N arising from The need for hydrogen isotope analysis of small water samples has been receiving attention. The eVectiveness of the hydrogen– contamination in the F2–He mixture. Both the CuO and V2O5–SiO2 methods of SO2 production deuterium–oxygen equilibration technique is limited for sample volumes of ,5 ml.Thielecke et al.232 used a fully automated from sulfide and sulfate minerals for sulfur isotope measurements were shown by Park and Ripley239 to be suitable for sample preparation device for sample volumes in the range 0.25–4 ml and a simple mathematical procedure to compensate samples containing as little as 3 mmol of S. The first method was, however, preferred in that the copper metal pellets were for the dependence of measured ratio on the sample volume. Although considered unimportant, fractionation due to water less contaminated than the copper turnings used in the V2O5 method.Preheating of reagents to reduce contamination was loss was detected and the authors advised that caution was still necessary to avoid fractionation. Working on a smaller necessary for the analysis of samples that produced less than 10 mmol of SO2. Accuracy was ±0.1‰ for samples containing scale, Tanweer and Han233 obtained quantitative conversion of 8 ml samples of water to H2 by using manganese as a 10 mmol and ±0.4‰ for smaller samples.Papers reporting the d37Cl analysis of rock and mineral reducing agent. Each sample was reduced in an individual reaction vessel, thereby eliminating memory eVect. A strange samples have been few in number. Although a modified method for the extraction of Cl from rock samples (Musashi et al.240) phenomenon of the study was that whereas the manganese from one supplier gave excellent ease and reproduciblity gave better precisions for d37Cl (±0.12‰, 2s) than previous methods, several g of silicate rocks were still needed for (±1‰) of analysis, that from another supplier suVered from a number of severe drawbacks which made it completely analysis.The method was based on dissolution of rocks in HF, concentration of Cl on anion exchange resin and conver- unsuitable. The sample size requirement for a novel microcombustion sion of Cl to CH3Cl for analysis. The method for the determination of iron isotope ratios, technique, described by Hofmann and Brand234 for the carbon isotopic analysis of ng amounts of C in non-volatile samples, described by Taylor et al.241 was designed to give very high analytical precision of ,3×10-4 for the study of natural was an order of magnitude lower than that for conventional methods.Liquid (0.5–10 ml ) or solid samples placed in a fractionation. The staggering sample requirement of 10 mg of iron would restrict possible applications of the procedure quartz sleeve were combusted at 1000 °C in a continuous stream of He and O2.Reproducibilities were ,1‰ for which was based on down-scaling of a co-condensation method to produce the volatile pentakis(trifluorophosphine)iron(O). .25 nmol of C. Problems associated with contamination of unknown origin led to variable and sample-dependent accu- The yield from this reaction was very low at 10–15%. Problems of interference were identified in the few analyses performed racy.In addition, the yield from small samples was variable. It remained an experimental challenge to reduce contamination and further improvements would be required before the method could become viable. to very low levels yet maintain constant 13C512C ratios in the blank. An intricate series of acid attacks was involved in the Although care should be taken to avoid rearrangements or introduction of reagent carbon during derivatization, the analytical protocol described by Pillinger and colleagues235 for high-resolution, stepped-combustion MS designed for the GC-combustion-SIRMS technique is becoming more widely used for the study of carbon isotopic composition in organic extraction, purification, quantification and isotope analysis of light elements in fine-grained reduced components, chemically compounds.Macko et al.242 found that isotopic fractionation occurred during the preparation of alditol acetate derivatives extracted from natural samples (for example, meteorites).An improved diVusion technique has been used by Georges of monosaccharides. Further limitations of the method included the production of an asymmetric carbon centre during and Dittert236 for the determination of nitrogen isotope ratios in aqueous ammonium and nitrate samples. The first of two the reduction steps. In their characterization of olive oil, Spangenberg et al.243 needed to correct for the isotopic shift steps involved shaking of samples in containers to collect NH4+ in KHSO4 acid buVer traps which were held behind due to carbon introduced in the methylation of fatty acids. Brenna and colleagues,244 on the other hand, have demon- PTFE membranes.In the second step, NO3- was converted J. Anal. At. Spectrom., 1999, 14, 1633–1659 1651strated that the preparation of fatty alcohols from correspond- ation TIMS are commonly reported. A serious problem found in negative ionization TIMS, yet almost unknown in positive ing fatty acid methyl esters, which does not introduce additional carbon, was feasible for isotope analysis.The on-line ionization TIMS, is that of memory eVect. Other observed problems have been the dependence of measured isotope ratios pyrolysis of the fatty alcohols in a position-specific isotope analysis system produced single-bond breakage, stabilization on filament temperatures, sloping peak heights and high and asymmetrical baselines.These eVects would be explained by and isotopically representative fragments for carbon isotope analysis. the processes of negative ionization for Os proposed by Hattori et al.250 Although the prime process involved electron capture very close to the hot filament, a more significant if minor 9 Thermal ionization mass spectrometry process involved electron exchange with O- away from the 9.1 Instrumentation filament. Ions could be formed with low potential energies far from the filament because thermal electrons escaped from the The accurate measurement of large isotope ratios challenges the filament and were accelerated in the ion source.Improved dynamic measurement range of modern mass spectrometers. focusing would prevent the low energy ions entering the flight Although highly sensitive detectors (for example, secondary tube, thereby reducing the problems associated with them. electron multipliers operated in the pulse counting mode for These findings are not in accord with the generally held view measurement of extremely low ion currents) can be used in that both positive and negative ionization only take place at combination with a second detector (for example, Faraday the filament surface with a low energy spread.cup for the measurement of higher ion currents), considerable Improvements to high-precision Re–Os analyses continue to care is required in the calibration of the system. Ovaskainen be reported. Liu et al.251 found that the oxygen isotope ratios et al.245 presented an overview of the problems encountered in the measured OsO3- and ReO4- ions were dependent on in calibrating a multi-detector instrument and suggested an the oxygen pressure in the ion source. Use of an in situ oxygen approach for the determination of 234U and 236U in natural isotope correction together with a stabilized oxygen pressure uranium samples which gave high abundance sensitivity and improved considerably the precision of measured isotope ratios wide dynamic range.to 0.01% for Os (0.1% for 184Os5188Os ratio). Reduction of Multicollector analysis is now commonly used to provide loading blanks through an extensive clean-up of the platinum simultaneous measurement of ion beams but size limitations filaments commonly used in Re–Os analysis allowed Markey can present problems in analysis at high mass. Only Faraday et al.252 to achieve a precision of 0.4% for the dating of cups have hitherto found wide application in multicollector individual molybdenite samples.Sample loadings of only 10 systems but now more sensitive alternatives are becoming and 1 ng were necessary for Os and Re, respectively. An available. Wallenius et al.246 have reported an initial investi- alternative approach to overcoming the problem of impurities gation into the use of a multi-micro-channeltron detection in the platinum filaments, as reported by Hattori et al.,253 was system paying particular attention to stability, linearity and to use tantalum filaments.These had not previously been used operational parameters. A newly developed double collector for negative ionization because of their high electron emission Faraday cup system, installed at the low mass end of a and high reactivity with O2. These problems were solved by traditional Faraday cup multi-collector array, was used by pre-baking filaments before sample loading, by using relatively Nakano and Nakamura247 for the precise determination of B high filament temperatures and high O2 pressures and by by measurement of Cs2BO2+ ions at m/z=308 and 309.The reducing the potential diVerence between the filament and use of static multi-collection in place of the conventional peak draw-out plate. Accurate isotope measurements were possible switching method allowed a reduction in acquisition time to for Os, Pt and Re at the ng level using single tantalum filaments 5 min and sample size to as little as 0.1 mg while giving the with Ba(NO3)2 emitter.Loadings of 1 ng Os, Pt or Re gave same analytical performance. Reproducibilities of measured strong (.10 V) and stable peaks for .24, .24 and .6 h, 11B510B ratios in NIST SRM 951 (boric acid) were±0.012% respectively. No blank correction was required. and±0.023% (2s) for loadings of 1 and 0.1 mg, respectively. The study of Xiao and Wang298 demonstrated an interference 9.3 Positive ionization procedures (Cs2CNO) at m/z=308 and 309 due to the presence of NO3-, in particular in the mannitol used in sample loading.A clean- Correction for mass fractionation is essential for accurate up procedure was developed to remove the interference. determination of isotope ratios but the fundamental processes The high resolution and fast magnetic peak switching system are still not fully understood. Commonly used fractionation built by Nieto et al.248 gave a magnetic resolution of better laws are based on empirical evidence and rarely describe than 1‰ and allowed magnet settling time to be reduced to a observed eVects exactly.Habfast254 combined the Rayleigh quarter of that required by conventional equipment. A mag- distillation equation with the Langmuir equation to produce netic-field-sensitive resistor measured field position and a the ‘Rayleigh–Langmuir’ fractionation model which is the only temperature-sensitive resistor allowed temperature drift to be currently available fundamental description of isotope evapor- corrected by normalization of all measurements to a constant ation in a vacuum.A detailed description of the sources of temperature. fractionation was given and the various fractionation laws The production of stable ion beams in a thermal source is compared. It was concluded that to achieve data with high dependent on precise control of filament temperatures. Halas accuracy required improvement of the stability and reproducand Durakiewicz249 have achieved stabilization of the filament iblity of the evaporation process and stabilization or elimintemperature through use of a Halas–Kaminski bridge in which ation of all static discriminations.the reference resistance of one leg is directly proportional to Sample loading is critical in achieving the reproducible and the filament temperature. They demonstrated that this resulted stable evaporation required to control fractionation.Sahoo in a significantly lower variance of the ion current than and Masuda255 used Li3PO4 loaded on to triple Re filaments stabilization of filament current or voltage. to produce a stable ion beam from ng quantities of Li. The optimum loading was 0.5 mg of Li as LiOH and 2 ml H3PO4 9.2 Negative ionization procedures (100 mg ml-1). Previously reported methods which used polyatomic ions (Li2BO2+ or Li2F+) were much more prone to Recent years have seen a large upsurge in the use of negative ionization TIMS, in particular for the study of the Re–Os fractionation.Taylor et al.256 found that the main criterion for controlling fractionation in the determination of 238U5235U isotope system, yet diYculties experienced with negative ioniz- 1652 J. Anal. At. Spectrom., 1999, 14, 1633–1659in soil samples was the amount of U loaded on to the fila- in elemental speciation. A number of discourses and a large ment. Samples of .350 ng U produced minimal fractionation, number of conference presentations, many of them highly whereas smaller samples required use of a double spike method repetitive, have appeared extolling the potential of ESMS.to achieve acceptable precision. A precision of 0.2% (2s) made However, there have also been some words of caution. The it possible to identify addition of 0.4 ng of enriched uranium review of Olesik et al.264 is recommended for a critical and (93 atomic% 235U) per g of contaminated soil.realistic assessment of the capabilities of ESMS for metal Nowell et al.257 found that the optimum separation, speciation. The need for a fundamental understanding of ion 50.100 mm, between ionization and evaporation filaments for formation in ESMS was emphasized in view of the evidence the eYcient ionization of Hf was only possible with double Re that all the processes of initial droplet generation, droplet filament assemblies and not with the commonly used triple fissioning and gas phase ion formation could be dependent filament assemblies.Achievement of internal precisions of on both sample matrix and species. The use of a modulated 0.002.0.006% for the 176Hf5177Hf ratio required very strict ES source to obtain both elemental and species information adherence to clean operating conditions. Steps taken included from the same solution has been considered by Hieftje265 the use of dedicated mass spectrometer parts for the analysis in his review of atomic spectrometry used for speciation.A of Hf, the running of standards prior to sample analysis to 61-reference review has been given by Bruins266 of the electrosestablish the point at which isobaric interferences caused pray ionization process in general although not specifically for by contaminants were minimized, and ensuring that Hf was elemental speciation. only analysed in dedicated batches. Although a considerable There has been almost an obsession with comparisons of improvement in precision was achieved, a complex three- ESMS with ICP-MS as though the two were competing column procedure was still needed to isolate the Hf from rock techniques.The paper of Houk,267 however, proposed that the samples and a relatively large amount (1.3 mg) of separated two techniques should be complementary to the extent that Hf was loaded on to the filaments. Fietzke et al.258 also used twin instruments could be configured with common sample double filament assemblies in their high precision measurement introduction and a single vacuum system.The ICP-MS would of 231Pa in a manganese crust. In comparison with the carbon- provide elemental analysis of the separated species, the ESMS coated single filament technique used previously there was molecular information. The author reported diYculties in much less contamination of the ion source and multiplier, getting atomic spectra of elements that form strong oxides resulting in improved vacuum and fewer scattered ions.There using ESMS. Declustering into atomic ions was only possible was, however, a reduction in the ionization eYciency of 2.3 with massive losses in total ion signal. Doubt was expressed orders of magnitude. Measurement precisions were about 4% about the feasibility of modulated sources to switch between and the determination limit about 10 fg of 231Pa. atomic and molecular conditions in ESMS. Chassaigne and Other reports on continued improvements in analytical per�©obin .ski268 have in fact used dual ISMS and ICP-MS detection formance have emphasised the advantages of TIMS over other in their study of cobalamins and cobinamides using microbore techniques. Improvements in chemical separation and reversed-phase HPLC. operating procedures were used by JoVroy et al.259 to measure A key question to answer is whether ESMS spectra are a 226Ra, 234U and 238U in water samples smaller than those true reflection of the ion species present in solution.Most of required for conventional counting techniques. Measurements the evidence appears to be that they are, but extreme caution were made routinely on sub-pg and pg quantities of Ra and needs to be taken to get the operating conditions right. Wang U, respectively, equivalent to200 ml and a few ml of sample, and Cole269 used alkali metal halides and other salts in their respectively. Peng et al.260 considered the TIMS U-series study of ionization processes in ESMS, in particular the exact method to be superior to a-spectrometry as sample sizes as mechanism by which gas-phase ions are formed from rapidly low as hundreds of mg could be handled, precisions in ages of shrinking charged droplets.They found that all the salts gave ,1% could be achieved, analysis time was only 3.4 h and the cluster ions when the initial salt concentrations were above dating range was greater. Graphite-coated single Re filaments 10.3 mol l.1.The relative abundance and number of diVerent were used for the analyses. cluster ions depended on several factors intimately related to There have been few new isotope dilution procedures reported the ES ionization process. In a study of the association of in the period covered by this Update. Vanhaecke et al.261 have pyrocyanine with MnII in aqueous media, Vukomanovic developed a multi-element method for the determination of et al.270 concluded that the appearance of complexes in the Cd, Cr and Pb in photographic AgCl emulsions.The AgCl mass spectra were the result of their formation in the ES was removed by precipitation and the determinands isolated process itself and not a reflection of their presence in solution. by electrolytic deposition on to platinum electrodes. A feature In a study on dissolved metal speciation in natural waters, of the procedure was the use of both H3BO3 (for Cr and Pb) Ross et al.271 concluded that accurate determination of dis- and H3PO4 (for Cd and Pb) as ionization aids in addition to solved metal ion speciation required mild ES conditions to a silica gel suspension containing AlCl3.All three determinands were evaporated from the same filament to give LODs in the preserve solution-phase interactions and that optimum ES range 0.4 ng (Cd) to 4 ng (Pb). A technique described by conditions depended on the thermodynamic stabilities of the Smoliar and Ondov262 allowed determination of 50 pg of Ir in species involved.The low voltages required for mild ES soot samples with an accuracy of 0.2%. conditions resulted in lower ion yields. For the purposes of Hinners et al.263 have reported the determination of lead metal speciation, it is important that organic co-solvents do isotope ratios in NIST SRM 1400, Bone Ash, as part of an not interfere significantly with the equilibria established in interlaboratory comparison exercise. The results agreed within aqueous solution.It was found, however, that the tendency of 0.09% with previously reported, uncertifed, values, so this acetonitrile to interact strongly with metal cations could material could eVectively be used as an RM for validating aVect speciation in cases where metal.ligand interactions were isotopic analyses. relatively weak. Experimental results for CuII and 8-hydroxyquinoline in a water.methanol mixture (151) showed good agreement with aqueous speciation predicted 10 Other methods using the thermodynamic equilibrium model MINEQL.A 10.1 Electrospray mass spectrometry (ESMS) and ion spray similar finding was made by Schramel et al.272 in a study of mass spectrometry (ISMS) capillary electrophoresis (CE) ESMS for metal speciation. Whereas strong complexes (for example, Cu.EDTA and Se The period covered by this Update has been one of reflection on the current status of these techniques and their true place compounds) were unaVected by the ionization process, ligands J.Anal. At. Spectrom., 1999, 14, 1633.1659 1653in weak complexes were replaced by solvent (acetic acid– samples there have been a number of reports on the use of GC-MS for the determination of organometallic species. Most water–methanol, 152517) molecules. They reported also that the position of the end of the capillary relative to the ES inlet of the developments reported were in sample preparation and separation, with few advances in the MS.The requirement for capillary was critical and a device was required to optimize the position. the preparation of volatile derivatives was a common factor in all the applications and can place limitations on acceptability The relatively large number of conference abstracts is testimony to the growing interest in the use of ESMS for metal of the method. Pons et al.280 were able to detect ionic alkyllead compounds in human urine with LODs of about 19 pg ml-1. speciation and elemental analysis.The majority of the abstracts contain little information but some have provided an indication Barshick et al.281 determined inorganic Hg with LODs of 7 ppb in water samples and 2 ppm in soils. The soil samples of the direction that studies are taking. Pretty and Van Berkel273 have used an on-line electrochemical flow cell for gave a much higher background and much reduced signal intensity than water samples due to the more complex soil preconcentration of the analytes and removal of sample matrix in the elemental analysis of water samples.It was possible to chemistry. Four organic tin compounds in marine mussels were determined (Binato et al.,282) with a detection limit of transfer inorganic analytes from media unsuitable for ESMS (for example, acidic aqueous solutions) to suitable solvents. 80 mg g-1. The use of GC-MS for the determination of isotope ratios Barnett and Horlick274 have used negative ion ESMS to measure both anions and cations in solution.Determination in non-metallic elements has been reported by several workers. The first reported use of negative chemical ionization (NCI) of metal cations was accomplished by complexation with EDTA to form negatively charged metal–EDTA complexes. for the determination of Se has been presented by Van Dael et al.283 for the measurement of stable isotope ratios. All Se The same authors have also demonstrated275 that negative ion ESMS could be used to determine all five Se species in their species in the sample were converted to selenite by acid digestion before derivatization as the volatile piazselenole. sample, whereas positive ion ESMS only detected the three organic species. Selenite reacted with methanol, the ES solvent, Corrections for the inclusion of 13C, 15N and 18O in the measured molecules were required.The NCI technique was to form selenomethyl ester, which caused serious spectral overlap with biselenate. Momplaisir et al.276 were able to chosen because of its greater specificity than EIMS.The detection limit of 1 pg for any Se isotope compared favourably detect nine organoselenium compounds in the sub-ng range in food hydrolysates by using HPLC coupled with ESMS. with that (90 pg) of EIMS. Chen et al.284 have developed an accurate and precise (0.2–0.5%) method for the determination 10.2 Fast atom bombardment mass spectrometry (FABMS) of 15N in uric acid and its oxidation product allantoin in urine samples.Bolz et al.285 found that GC-MS compared well with The only development of note has been the use by Westaway ICP-MS for the determination of B isope ratios in steels et al.277 of FABMS for the measurement of Cl isotope ratios. enriched in 10B. An important observation was the lack of Samples were prepared as AgCl, which was deposited on a matrix eVects in GC-MS. silver plate mounted on the probe tip. The plate was heated to melt the sample, which was bombarded with 6 keV Xe 10.4 Noble gas mass spectrometry atoms.Although the method had poorer precision than SIRMS analysis, this was compensated for by the simplicity of analysis A specific facility for the extraction of Ar, He, Kr, Ne and Xe from petroleum and precise measurement of their isotopic which did not require time-consuming sample preparation and purification. ratios has been described by Pinti and Marty.286 More than 98% of the dissolved noble gases were extracted within 30 min 10.3 Gas chromatography-mass spectrometry (GC-MS) using the procedure which was based on flash release into an evacuated container.Following several stages of purification, The ionization process most associated with traditional the single noble gases were separated cryogenically for analysis. GC-MS is electron ionization, but some recent developments Most of the isotopes measured (36–38Ar, 3He, Kr, 21–22Ne and have also used GC in combination with other MS techniques, Xe) required the use of a Daly collector.Corrections were for example with ICP-MS and GD-MS. Another plasma required for interferences on the Ne isotopes. No interfering technique, microplasma MS, has been developed by Brede and hydrocarbon ions were observed and there was no indication colleagues278 specifically as a detector for capillary GC. A of contamination of the flight tube by organic compounds. miniaturized He discharge was sustained at low pressure within Balogh and Simonits287 have reported improvements in the the end of the capillary GC column inside the ion source method used for K/Ar–Ar/Ar dating. Steps taken to decrease housing of a quadrupole mass spectrometer.Ions were introthe background included surrounding the ion source with a duced directly from the plasma to the mass analyser using cylinder of St707 NEG strip, rotating the sample during only a repeller and electrostatic lenses to focus the ions. The radiation in the reactor and construction of a new furnace for plasma was sustained in a low He flow of only 25 ml min-1, de-gassing samples.which was not only acceptable to the analyser but also A primary Xe isotopic gas standard has been made available enhanced the energy density of the discharge. The maximum following certification by Taylor and colleagues.288 The meassignal was obtained at a power of 1.9 W, which is much lower ured isotope amount ratios were in good agreement with than that required by MIP-MS (50–100 W) or GDMS (30W).previous measurements of xenon, but the relative uncertainties The method was limited by the need to add traces of O2 to have been reduced considerably. avoid deposition of carbon and by the number of polyatomic species reported in the low m/z region. The detection limit for 10.5 Spark source mass spectrometry (SSMS) Cl was 2.2 pg s-1. In a further development by the same group279 for the simultaneous detection of the halogens and There have been few developments in SSMS of note in the period covered by this Update.Jochum289 has reviewed the carbon, only the GC carrier gas (He at 2.3 ml min-1) was used for plasma generation. Both H2 and O2 were introduced technique with particular reference to his multi-ion counting detector referred to in previous Updates. The new detector as reagent gases to suppress analyte reactions with the silica walls of the ion source. In addition, the power level needed to required development of a system for measurement of the ion flux and control of the ion beam reaching the detector.290 The be controlled carefully to avoid formation of H3O+ which interfered with F+ at m/z 19.system, which gave the operator choice of pulse length and repetition rate, was stable to within 0.05%. Ramendik291 has With the increased interest in metal speciation in natural 1654 J. Anal. At. Spectrom., 1999, 14, 1633–16595 R. Gill (Ed.),Modern Analytical Geochemistry, AddisonWesley further expounded his method of standardless analysis with Longman, Harlow, Essex, UK, 1997, 1–316.particular reference to SSMS. The procedure is severely limited 6 J. S. Becker and H.-J. Dietze, Spectrochim. Acta, Part B, 1998, in a practical sense in that five or six internal standards with 53, 1475. widely diVering properties were necessary to achieve quantitat- 7 M. Suter, Nucl. Instrum. Methods Phys. Res., Sect. B, 1998, ive multi-element analysis.Furthermore, only relatively poor 139(1–4), 150. accuracies of 15–21% could be demonstrated. A comparison 8 D. J. W. Mous, W. Fokker, R. Van Dan Broek, R. Koopmans, C. B. Ramsey and R. E. M. Hedges, Radiocarbon, 1998, 40(1), of several MS techniques for the ultra-trace analysis of gallium 283. arsenide by Becker et al.292 concluded that solid state methods, 9 F. D. McDaniel, S. A. Datar, B. N. Guo, S. N. Renfrow, Z. Y. including SSMS, were preferable to methods that required Zhao and J.M. Anthony, Appl. Phys. Lett., 1998, 72(23), 3008. sample dissolution because of the lower detection limits and 10 S. H. Sie, T. R. Niklaus, G. F. Suter and F. Bruhn, Rev. Sci. reduced risk of contamination. Although the authors con- Instrum., 1998, 69(3), 1353. sidered the agreement between methods for the analysis of the 11 J. S. C. Wills, R. J. Schneider, J. M. Hayes, K. F. Von Reden, A. P. McNichol and T. I. Eglinton, Radiocarbon, 1998, 40(1), 95.same sample to be good, the results for several determinands 12 R. Weissenbok, S. R. Biegalski, L. A. Currie, D. B. Klinedinst, diVered by an order of magnitude and emphasized the diYculty R. Golser, G. A. Klouda, W. Kutschera, A. Priller, W. Rom, in getting accurate data, even if the precisions seemed to be P. Steier and E. Wild, Radiocarbon, 1998, 40(1), 265. acceptable. 13 R. M. Verkouteren, D. B. Klinedinst and L. A. Currie, Radiocarbon, 1997, 39(3), 269. 14 Y. Zheng, A.Pearson, P. McNichol, R. J. Schneider and 10.6 New methodologies K. F. Von Reden, Radiocarbon, 1998, 40(1), 61. New interfaces for LC-MS present the prospect of elemental 15 K. F. Von Reden, A. P. McNichol, A. Pearson and R. J. Schneider, Radiocarbon, 1998, 40(1), 247. speciation studies in aqueous solution. Hiraoka et al.293 evalu- 16 M. Schleicher, P. M. Grootes, M.-J. Nadeau and A. Schoon, ated a laser spray source in which a laser beam, focused to Radiocarbon, 1998, 40(1), 85.approximately 0.1 mm in diameter, irradiated aqueous analyte 17 J. Van der Plicht, K. Vandeputte, L. Moens and R. Dams, solution eVusing from the tip of the inner capillary of two Radiocarbon, 1998, 40(1), 103. concentric stainless steel capillaries. Nebulizer gas (N2) was 18 C. Alderliesten, K. Van der Borg and A. F. M. De Jong, supplied through the outer capillary to reduce the angular Radiocarbon, 1998, 40(1), 215. 19 L. Campajola, A. D’Onofrio, A. Feoli, L. Gialanella, divergence of the plume and to entrap and transport the M.C. Morone, G. Oliviero, V. Roca, M. Romano, F. Terrasi, sample to the mass spectrometer orifice. The laser spray C. Rolfs, U. Greife, F. Strieder and H. P. Trautvetter, Nucl. interface yielded ion intensities that were more than 30 times Instrum. Methods Phys. Res., Sect. B, 1998, 140(1–2), 258. stronger than those produced by ES or IS for the same 20 S. GriYn and E. R. M. DruVel, Radiocarbon, 1998, 40(1), 29. solutions. A particle beam interface was used by Magi and 21 K.L. Elder, A. P. McNichol and A. T. Gagnon, Radiocarbon, Ianni294 in the determination of dibutyltin and tributyltin in 1998, 40(1), 223. marine samples with detection limits of 0.95 and 0.65 ng g-1, 22 B. R. Larsen, C. Brussol, D. Kotzias, T. Veltkamp, O. Zwaagstra and J. Slanina, Atmos. Environ., 1998, 32(9), 1485. respectively. Monobutyltin was not eluted from the LC column 23 A. Schmidt, C. Schnabel, J. Handl, D. Jakob, R.Michel, under the conditions used. H.-A. Synal, J. M. Lopez and M. Suter, Sci. Total Environ., Decomposition with F combined with EI-MS for the analysis 1998, 223(2–3), 131. of refractory substances was shown by Broekaert and col- 24 F. X. Martin, G. M. Raisbeck and F. Yiou, Mineral. Mag., 1998, leagues295 to present many analytical problems. Blank contri- 62A(FPt. 2), 953. butions from the system components proved to be significant. 25 Y. Xie, B. Ying, S. Jiang, M. He, J.Shan, R. Baker (Ed.), S. Slate (Ed.) and G. Benda (Ed.), Determination of 129I in envir- In addition, interpretation of the mass spectra proved to be onmental water. Proc. Int. Conf. Radioact. Waste Manage. diYcult because of the presence of many polyatomic ions. Environ. Rem., 6th, American Society of Mechanical Engineers, Heavy absorption phenomena required conditioning of the New York, NY, USA, 1997, 707–708. combustion vessels and the detection system. In-line separation 26 M.He, S. Jiang, S. Jiang, S. Wu and Y. Xie, Yuanzineng Kexue of the combustion products reduced some of these problems Jishu, 1997, 31(4), 301. but severe diYculties remained. 27 J. E. McAninch, A. A. Marchetti, B. A. Bergquist, N. J. Stoyer, G. J. Mimz,M.W. CaVee, R. C. Finkel, K. J. Moody, E. Sideras- A special mass spectrometer, developed by Sysoev et al.296 Haddad, B. A. Buchholz, B. K. Esser and I. D. Proctor, for the chemical analysis of cosmic dust, was based on plasma J.Radioanal. Nucl. Chem., 1998, 234(1–2), 125. formation during impact of the dust particles (10-8–10-15 g, 28 A. Bogaerts and R. Gijbels, Spectrochim. Acta, Part B, 1998, 1–80 km s-1) on a plate. Experiments have proved, however, 53, 1. that there is no simple correlation between the ion composition 29 A. Bogaerts and R. Gijbels, J. Anal. At. Spectrom., 1998, 13(9), of the plasma and the chemical composition of the dust 945. 30 W. W. Harrison, J. Anal. At. Spectrom., 1998, 13(9), 1051. particles.A better understanding of the ionization process is 31 D. C. Duckworth and C. M. Barshick, Anal. Chem., 1998, needed before spectra can be interpreted satisfactorily. Various 70(21), 709A. models have been tested to explain the ionization process 32 G. G. Sikharulidze and A. E. Lezhnev, J. Anal. At. Spectrom., under diVerent conditions. 1999, 14(1), 45. 33 G. G. Sikharulidze and A. E. Lezhnev, J. Anal. Chem. (Transl. of Zh. Anal. Khim.), 1998, 53(4), 375.L. Van Vaeck is indebted to the Flemish National Science 34 M. Belkin, J. W. Waggoner and J. A. Caruso, Anal. Commun., Foundation (FWO, Brussels) for financial support. 1998, 35(9), 281. 35 S. Born, C. Venzago and B. Stojanik, Appl. Spectrosc., 1998, 52(10), 1358. 11 References 36 Y. Su, P. Yang, Z. Zhou, F. Li, X. Wang and B. Huang, J. Anal. At. Spectrom., 1998, 13(11), 1271. 1 J. R. Bacon, J. S. Crain, L. Van Vaeck and J. G. Williams, 37 C. M. Barshick, K. L. Goodner, C.J. Watson and J. R. Eyler, J. Anal. At. Spectrom., 1998, 13(10), 171R. Int. J. Mass Spectrom., 1998, 178(1–2), 73. 2 B. Fairman, M. W. Hinds, S. M. Nelms, D. M. Penny and 38 T. E. Gibeau and R. K. Marcus, J. Anal. At. Spectrom., 1998, P. Goodall, J. Anal. At. Spectrom., 1998, 13(12), 233R. 13(12), 1303. 3 M. R. Cave, O. Butler, J. M. Cook, M. S. Cresser, L. M. Garden, 39 Y. Su, P. Yang, Z. Zhou, X. Wang, F. Li, B. Huang, J. Ren, A. J. Holden and D. L. Miles, J. Anal.At. Spectrom., 1999, M. Chen, H. Ma and G. Zhang, Spectrochim. Acta, Part B, 1998, 14(2), 279. 53, 1413. 4 A. Taylor, S. Branch, D. J. Halls, L. M. W. Owen and M. White, J. Anal. At. Spectrom., 1999, 14(4), 717. 40 A. Pupyshev and A. Lutsak, (Ed.), in Proc.—Semin. At. J. Anal. At. Spectrom., 1999, 14, 1633–1659 1655Spectrochem., 14th, ed. E. Krakovska and S. Ruzickova, StroVek E. A. G. Zagatto and M. F. Gine, J. Anal. At. Spectrom., 1998, 13(12), 1343. Publishing, Kosice, Slovakia, 1998, pp. 325–329. 41 A. A. Pupyshev, A. K. Lutsak and V. N. Muzgin, J. Anal. Chem. 80 J. P. Valles Mota, M. R. Fernandez de la Campa, J. I. Garcia Alonso and A. Sanz-Medel, J. Anal. At. Spectrom., 1999, 14(2), (Transl. of Zh. Anal. Khim.), 1998, 53(7), 627. 42 A. A. Pupyshev and A. K. Lutsak, J. Anal. Chem. (Transl. of Zh. 113. 81 V. L. Dressler, D. Pozebon and A. J. Curtius, Anal. Chim. Acta, Anal. Khim.), 1998, 53(11), 987. 43 C. J. Park and S. M. Noh, J. Anal. At. Spectrom., 1998, 13(8), 1998, 379(1–2), 175. 82 R. Lobinski, Spectrochim. Acta, Part B, 1998, 53, 177. 715. 44 B. L. Sharp, J. Batey, I. S. Begley, D. Gregson, J. Skilling, 83 H. M. Crews, Spectrochim. Acta, Part B, 1998, 53, 213. 84 B. Welz, Spectrochim. Acta, Part B, 1998, 53(2), 169. A. B. Sulaiman and G. Verbogt, J. Anal. At. Spectrom., 1999, 14(2), 99. 85 R. Lobinski, I. R. Pereiro, H. Chassaigne, A. Wasik and J. Szpunar, J. Anal. At. Spectrom., 1998, 13(9), 859. 45 T. N. Olney, W.Chen and D. J. Douglas, J. Anal. At. Spectrom., 1999, 14(1), 9. 86 E. H. Larsen, Spectrochim. Acta, Part B, 1998, 53(2), 253. 87 K. G. Heumann, S. M. Gallus, G. Radlinger and J. Vogl, 46 K. E. Jarvis, P. Mason, T. Platzner and J. G. Williams, J. Anal. At. Spectrom., 1998, 13(8), 689. Spectrochim. Acta, Part B, 1998, 53, 273. 88 R.-D. Wilken and R. Falter, Appl. Organomet. Chem., 1998, 47 B. S. Duersch, Y. Chen, A. Ciocan and P. B. Farnsworth, Spectrochim. Acta, Part B, 1998, 53(4), 569. 12(8–9), 551. 89 H. M. Kingston, D. Huo, Y. Lu and S. Chalk, Spectrochim. Acta, 48 E. D. Prudnikov and R. M. Barnes, J. Anal. At. Spectrom., 1999, 14(1), 27. Part B, 1998, 53(2), 299. 90 L. Ebdon, S. J. Hill and C. Rivas, Spectrochim. Acta, Part B, 49 M. Carre, S. ExcoYer and J. M. Mermet, Spectrochim. Acta, Part B, 1997, 52(14), 2043. 1998, 53(2), 289. 91 S. Perez Mendez, E. Blanco Gonzalez, M. L. Fernandez Sanchez 50 B. W. Pack and G. M. Hieftje, Spectrochim. Acta, Part B, 1997, 52(14), 2163.and A. Sanz Medel, J. Anal. At. Spectrom., 1998, 13(9), 893. 92 C. B’Hymer, K. L. Sutton and J. A. Caruso, J. Anal. At. 51 C. Sartoros, D. M. Goltz and E. D. Salin, Appl. Spectrosc., 1998, 52(5), 643. Spectrom., 1998, 13(9), 855. 93 P. W. Kirlew, M. T. M. Castillano and J. A. Caruso, 52 B. Landais, C. Beaugrand, L. Capron-Dukan, M. Sablier, G. Simonneau and C. Rolando, Rapid Commun. Mass Spectrom., Spectrochim. Acta, Part B, 1998, 53(2), 221. 94 V. Majidi and N.J. Miller-Ihli, Analyst (Cambridge, U.K.), 1998, 1998, 12(6), 302. 53 P. Marriott, R. Fletcher, A. Cole, I. Beaumont, J. Lofthouse, 123(5), 809. 95 K. A. Taylor, B. L. Sharp, D. J. Lewis and H. M. Crews, J. Anal. S. Bloomfield and P. Miller, J. Anal. At. Spectrom., 1998, 13(9), 1021. At. Spectrom., 1998, 13(10), 1095. 96 J. Feldmann, I. Koch and W. R. Cullen, Analyst (Cambridge, 54 F. Vanhaecke, G. de Wannemacker, L. Moens, R. Dams, C. Latkoczy, T. Prohaska and G.Stingeder, J. Anal. At. U.K.), 1998, 123(5), 815. 97 I. I. Stewart and J. W. Olesik, J. Anal. At. Spectrom., 1998, Spectrom., 1998, 13(6), 567. 55 T. E. JeVries, S. E. Jackson and H. P. Longerich, J. Anol. At. 13(12), 1313. 98 I. I. Stewart and J. W. Olesik, J. Anal. At. Spectrom., 1998, Spectrum., 1998, 13(9), 935. 56 P. M. Outridge, W. Doherty and D. C. Gregoire, Spectrochim. 13(9), 843. 99 A. Makishima and E. Nakamura, Geostand. Newsl., 1997, Acta, Part B, 1997, 52(14), 2093. 57 S. M. Eggins, L. P. J. Kinsley and J. M. G. Shelley, Appl. Surf. 21(2), 307. 100 J. L. M. de Boer, P. de Joode and R. Ritsema, J. Anal. At. Sci., 1998, 127, 278. 58 H. F. Falk, B. Hattendorf, K. Krengel-Rothensee, N. Wieberneit Spectrom., 1998, 13(9), 971. 101 M. Krachler, A. Alimonti, F. Petrucci, F. Forastiere and and S. L. Dannen, Fresenius’ J. Anal. Chem., 1998, 362(5), 468. 59 R. J.Watling, J. Anal. At. Spectrom., 1998, 13(9), 927. S. Caroli, J. Anal. At. Spectrom., 1998, 13(8), 701. 102 C. E. McGinnis, J. C. Jain and C. R. Neal, Geostand. Newsl., 60 M. Odegaerd and M. Hamester, Geostand. Newsl., 1997, 21(2), 245. 1997, 21(2), 289. 103 L. Yang, J. W. H. Lam, R. E. Sturgeon and J. W. McLaren, 61 M. Oedegaerd, S. J. Dundas, B. Flem and A. Grimstvedt, Fresenius’ J. Anal. Chem., 1998, 362(5), 477. J. Anal. At. Spectrom., 1998, 13(11), 1245. 104 N. Violante, F. Petrucci, P. D. Femmine and S. Caroli, 62 J. L. Neilsen, A. Abildtrup, J. Christensen, P. Watson, A.Cox and C. W. McLeod, Spectrochim. Acta, Part B, 1998, 53(2), 339. Microchem. J., 1998, 59(2), 269. 105 A. Krushevska, S. Waheed, J. Nobrega, D. Amarasiriwardena 63 N. Mickeley and M. O. Amato, At. Spectrosc., 1997, 18(6), 186. 64 G. Galbacs, T. Kantor, L. Moens and R. Dams, Spectrochim. and R. M. Barnes, Appl. Spectrosc., 1998, 52(2), 205. 106 K. Van den Broeck and C. Vandecasteele, Anal. Lett., 1998, Acta, Part B, 1998, 53, 1335. 65 J.-F. Alary and E. D. Salin, Spectrochim.Acta, Part B, 1998, 31(11), 1891. 107 A. L. Gray, Eur. Spectrosc. News, 1982, (43), 13. 53, 1705. 66 E. Bjorn, W. Frech, E. HoVmann and C. Luedke, Spectrochim. 108 N. Jakubowski, L. Moens and F. Vanhaecke, Spectrochim. Acta, Part B, 1998, 53, 1739. Acta, Part B, 1998, 53, 1765. 67 V. Majidi and N. J. Miller-Ihli, Spectrochim. Acta, Part B, 1998, 109 A. N. Halliday, D.-C. Lee, J. N. Christensen, M. Rehkamper, W. Yi, X. Luo, C. M. Hall, C. J. Ballentine, T. Pettke and C.Stirling, 53, 965. 68 K.-H. Lee, S.-J. Jiang and H.-W. Liu, J. Anal. At. Spectrom., Geochim. Cosmochim. Acta, 1998, 62(6), 919. 110 W. Yi, A. N. Halliday, D.-C. Lee and M. Rehkaemper, Geostand. 1998, 13(11), 1227. 69 K.-H. Lee, S.-H. Liu and S.-J. Jiang, Analyst (Cambridge, U.K.), Newsl., 1998, 22(2), 173. 111 M. Rehkamper and A. N. Halliday, Mineral. Mag., 1998, 1998, 123(7), 1557. 70 R. Knight, S. J. Haswell, S. W. Lindow and J. Batty, J. Anal. At. 62A(Pt. 2), 1241. 112 M.Rehkamper and A. N. Halliday, Int. J. Mass Spectrom., 1998, Spectrom., 1999, 14(2), 127. 71 D. Karunasagar, J. Arunachalam and S. Gangadharan, J. Anal. 181, 123. 113 N. S. Belshaw, P. A. Freedman, R. K. O’Nions, M. Frank and At. Spectrom., 1998, 13(7), 679. 72 K. Hoppstock, J. S. Becker and H.-J. Dietze, At. Spectrosc., Y. Guo, Int. J. Mass Spectrom., 1998, 181, 51. 114 Z. Palacz, P. J. Turner, C. Haines, F. Abou-Shakra and 1997, 18(6), 180. 73 L.-S. Zhang and S. M. Combs, J.AOAC Int., 1998, 81(5), 1060. A. N. Eaton, Mineral. Mag., 1998, 62A(Pt. 2), 1126. 115 J. S. Becker and H.-J. Dietze, J. Anal. At. Spectrom., 1998, 74 Y.-L. Feng, H.-Y. Chen, L.-C. Tian and H. Narasaki, Anal. Chim. Acta, 1998, 375(1), 167. 13(9), 1057. 116 S. Sturup, H. Dahlgaard and S. C. Nielsen, J. Anal. At. 75 J. A. McLean, M. G. Minnich, L. A. Iacone, H. Liu and A. Montaser, J. Anal. At. Spectrom., 1998, 13(9), 829. Spectrom., 1998, 13(12), 1321. 117 A. T. Townsend, Z. Yu, P.McGoldrick and J.A. Hutton, J. Anal. 76 X.-P. Yan, R. Kerrich and M. J. Hendry, J. Anal. At. Spectrom., 1999, 14(2), 215. At. Spectrom., 1998, 13(8), 809. 118 T. Catterick, B. Fairman and C. Harrington, J. Anal. At. 77 S. N. Willie, H. Tekgul and R. E. Sturgeon, Talanta, 1998, 47(2), 439. Spectrom., 1998, 13(9), 1009. 119 W. Kerl, J. S. Becker, H.-J. Dietze and W. Dannecker, Schr. 78 J. R. Pretty, D. C. Duckworth and G. J. Van Berkel, Anal. Chem., 1998, 70, 1141. Forschungszent.Juelich, Reihe Energietech./Energy Technol., 1998, 2(RadioactiveWaste Products 1997), 323. 79 L. F. R. Machado, A. O. Jacintho, A. A. Menegario, 1656 J. Anal. At. Spectrom., 1999, 14, 1633–1659120 F. Monna, J.-L. Loizeau, B. A. Thomas, C. Geuguen and 157 W. Rybczyncki, W. Seidel and H. D. von Zerssen, Fresenius’ P.-Y. Favarger, Spectrochim. Acta, Part B, 1998, 53, 1317. J. Anal. Chem., 1998, 362(3), 249. 121 A. A. Menegario, M. F. Gine, J. A. Bendassolli, A. C.S. Bellato 158 M. Puga-Lambers and P. H. Holloway, Surf. Interface Anal., and P. C. O. Trivelin, J. Anal. At. Spectrom., 1998, 13(9), 1065. 1998, 26(11), 851. 122 G. P. Pinho, H. Schittenhelm, W. W. Duley, S. A. Schlueter, 159 B. Tomiyasu, I. Fukuju, H. Komatsubara, M. Owari and H. R. Jahani and R. E. Mueller, Appl. Surf. Sci., 1998, 127, 983. Y. Nihei, Nucl. Instrum. Methods Phys. Res., Sect. B, 1998, 123 H. J. Dang, Z. H. Han, Z. G. Dai and Q. Z. Qin, Int. J. Mass 136(1–4), 1028.Spectrom., 1998, 178(3), 205. 160 V. I. Shvachko, Surf. Sci., 1998, 411(3), L882. 124 S. Amoruso, V. Berardi, R. Bruzzese, N. Spinelli and X. Wang, 161 L. R. Riciputi, B. A. Paterson and R. L. Ripperdan, Int. J. Mass Appl. Surf. Sci., 1998, 127, 953. Spectrom., 1998, 178(1–2), 81. 125 D. P. Taylor, W. C. Simpson, K. Knutsen, M. A. Henderson and 162 X.-K. Zhu, R. K. O’Nions and A. J. Gibb, Chem. Geol., 1998, T. M. Orlando, Appl. Surf. Sci., 1998, 127, 101. 144(3–4), 305. 126 A. A. Sysoev, S. S. Poteshin, V. I. Pyatakhin, I. V. Shchekina 163 L. R. Riciputi and J. P. Greenwood, Int. J. Mass Spectrom., and A. S. Trofimov, Fresenius’ J. Anal. Chem., 1998, 361(3), 261. 1998, 178(1–2), 65. 127 R. A. Gleray, P. T. A. Reilly, M. Yang, W. B. Whitten and J. M. 164 H. Hiyagon, Antarct. Meteorite Res., 1997, 10, 249. Ramsey, Anal. Chem., 1998, 70(1), 117. 165 K. J. Coakley and D. S. Simons, Chemom. Intell. Lab. Syst., 128 M. Kuzuya, Y. Ohoka, H. Katoh and H. Sakanashi, 1998, 41(2), 209.Spectrochim. Acta, Part B, 1998, 53(1), 123. 166 B. Gautier, J. C. Dupuy, B. Semmache and G. Prudon, Nucl. 129 S. S. Dimov and S. L. Chryssoulis, Spectrochim. Acta, Part B, Instrum. Methods Phys. Res., Sect. B, 1998, B142(3), 361. 1998, 53(3), 399. 167 A. R. Pratt, C. M. Huctwith, P. A. W. van der Heide and 130 K. Poels, L. Van Vaeck and R. Gijbels, Anal. Chem., 1998, N. S. McIntyre, J. Geochem. Explor., 1998, 60(3), 241. 70(3), 504. 168 F. Schroder-Oeynhausen, B.Burkhardt, T. Fladung, F. Koetter, 131 T. Tsugoshi, H. Chiba, T. Yokoyama, K. Furuya and A. Schneiders, L. Wiedmann and A. Benninghoven, J. Vac. Sci. T. Kikuchi, J. Trace Microprobe Tech., 1998, 16(1), 47. Technol., B, 1998, 16(3), 1002. 132 J. K. Gibson, Radiochim. Acta, 1998, 81(2), 83. 169 S. Nicolescu, D. H. Cornell, U. Soedervall and H. Odelius, Eur. 133 A. Hachimi, L. Van Vaeck, K. Poels, F. Adams and J. F. Muller, J. Mineral., 1998, 10(2), 251. Spectrochim.Acta, Part B, 1998, 53(2), 347. 170 A. J. Fahey, Int. J. Mass Spectrom. Ion Processes, 1998, 134 K. Blaum, C. Geppert, P. Muller, W. Nortershauser, 176(1–2), 63. E. W. Otten, A. Schmitt, N. Trautmann, K. Wendt and 171 P. W. O. Hoskin and R. J. Wysoczanski, J. Anal. At. Spectrom., B. A. Bushaw, Int. J. Mass Spectrom., 1998, 181, 67. 1998, 13(7), 597. 135 M. V. Suryanarayana, M. Sankari and S. Gangadharan, Int. J. 172 A. Amaral, P. Galle, C. Cossonnet, D. Franck, P. Pihet, Mass Spectrom.Ion Processes, 1998, 173(3), 177. M. Carrier and O. Stephan, J. Radioanal. Nucl. Chem., 1997, 136 K. Song, H. Cha and J. Lee, J. Anal. At. Spectrom., 1998, 226(1–2), 41. 13(10), 1207. 173 A. Amaral, C. Cossonnet and P. Galle, Radiat. Prot. Dosim., 137 N. Erdmann, V. Sebastian, V. N. Fedoseyev, M. Hannawald, 1998, 79(1–4, Intakes of Radionuclides), 137. G. Huber, T. Kautzsch, K. L. Kratz, V. I. Mishin, M. 174 S. Hofmann and J. Schubert, J. Vac. Sci. Technol., A, 1998, 16(3, Nunnemann, G. Passler and N.Trautmann, Appl. Phys. B: Pt.1), 1096. Lasers Opt., 1998, B66(4), 431. 175 B. G. Svensson, M. K. Linnarsson, J. Cardemas and M. Petravic, 138 N. Erdmann, M. Nunnemann, K. Eberhardt, G. Herrmann, Nucl. Instrum. Methods Phys. Res., Sect. B, 1998, 136, 1034. G. Huber, S. Kohler, J. V. Kratz, G. Passler, J. R. Peterson, 176 N. J. Montgomery, J. L. MacManus-Driscoll, D. S. McPhail, N. Trautmann and A. Waldek, J. Alloys Compd., 1998, R. J. Chater, B.Moeckly and K. Char, Thin Solid Films, 1998, 271(1–2), 837. 317(1–2), 237. 139 J. R. Peterson, N. Erdmann, M. Nunnemann, K. Eberhardt, 177 K. J. Kim and D. W. Moon, Surf. Interface Anal., 1998, 26(1), 9. G. Huber, J. V. Kratz, G. Passler, O. Stetzer, P. Thorle, 178 W. Duemmler, S. Weber, N. Maloufi, C. Tete and S. Scherrer, N. Trautmann and A. Waldek, J. Alloys Compd., 1998, Int. J. Mass Spectrom. Ion Processes, 1998, 176(1–2), 125. 271(1–2), 876. 179 D. P. Docter, J. P.Ibbetson, Y. Gao, U. K. Mishra, T. Liu and 140 B. Eichler, S. Huebener, N. Erdmann, K. Eberhardt, H. Funk, D. E. Grider, J. Electron. Mater., 1998, 27(5), 479. G. Herrmann, S. Koehler, N. Trautmann, G. Passler and 180 T. Horita, N. Sakai, T. Kawada, H. Yokokawa and M. Dokiya, F. J. Urban, Radiochim. Acta, 1997, 79(4), 221. J. Am. Ceram. Soc., 1998, 81(2), 315. 141 M. Nunnemann, N. Erdmann, H.-U. Hasse, G. Huber, 181 R. Saito and M. Kudo, Jpn. J. Appl. Phys., Part 1, 1998, 37(2), J.V. Kratz, P. Kunz, A. Mansel, G. Passler, O. Stetzer, N. 690. Trautmann and A. Waldek, J. Alloys Compd., 1998, 271(1–2), 182 T. Lundgren, L. G. Persson, E. U. Engstroem, J. Chabala, 45. R. Levi-Setti and J. G. Noren, Arch. Oral Biol., 1998, 43(11), 142 J. R. De Laeter and A. K. Kennedy, Int. J. Mass Spectrom., 841. 1998, 178(1–2), 43. 183 E. Cuynen, P. Van Espen, G. Goeminne and J. Terryn, J. Anal. 143 A. V. Hamza, T. Schenkel, A. V. Barnes and D. H. Schneider, At.Spectrom., 1999, 14(3), 483. J. Vol. Sci. Technol., A, 1999, 17(1), 303. 184 G. Verlinden, R. Gijbels, I. Geuens and R. De Keyzer, J. Anal. 144 A. J. Fahey, Rev. Sci. Instrum., 1998, 69(3), 1282. At. Spectrom., 1999, 14(3), 429. 145 R. C. Reedy, M. R. Young and S. E. Asher, J. Vac. Sci. Technol., 185 S. Muhl, A. Gaona-Couto, J. M. Mendez, S. Rodil, G. Gonzalez, A, 1999, 17(1), 317. A. Merkulov and R. Asomoza, Thin Solid Films, 1997, 308, 228. 146 T. Schenkel, A. V. Hamza, A. V.Barnes, D. H. Schneider, 186 S. Chandra and G. H. Morrison, J. Microsc. (Oxford), 1997, D. S.Walsh and B. L. Doyle, J. Vac. Sci. Technol., A, 1998, 16(3, 188(2), 182. Pt.1), 1384. 187 G. Tamborini, M. Betti, V. Forcina, T. Hiernaut, B. Giovannone 147 D.-Q. Yang and C.-Z. Fan, Chin. Phys. Lett., 1998, 15(9), 697. and L. Koch, Spectrochim. Acta, Part B, 1998, 53, 1289. 148 Y. Higashi and Y. Homma, Anal. Sci., 1998, 14(2), 281. 188 A. Amaral, P. Galle, F. Escaig, C. Cossonnet, M.H. Henge- 149 K. Franzreb, A. Pratt, S. Splinter and P. Van Der Heide, Surf. Napoli, E. Ansoborlo and L. Zhang, J. Alloys Compd., 1998, Interface Anal., 1998, 26(8), 597. 271, 19. 150 M. Saggio, C. Montandon, A. Bourenkov, L. Frey and 189 G. McMahon, Microstruct. Sci., 1997, 24 (Understanding P. Pichler, Radiat. EV. Defects Solids, 1997, 141(1–4), 37. Microstructure: Key to Advances in Materials), 149. 151 C. Montandon, A. Bourenkov, L. Frey, P. Pichler and 190 C. Brunner, H.Hutter, K. Piplits, M. Gritsch, G. Poeckl and J. P. Biersack, Radiat. EV. Defects Solids, 1998, 145(3), 213. M. Grasserbauer, Fresenius’ J. Anal. Chem., 1998, 361(6–7), 667. 152 N. Paour, Radiat. EV. Defects Solids, 1997, 142(1–4), 885. 191 M. Srivastava, N. O. Petersen, G. R. Mount, D. M. Kingston 153 K. Wittmaack, Surf. Interface Anal., 1998, 26(4), 290. and N. S. McIntyre, Surf. Interface Anal., 1998, 26(3), 188. 154 Z. X. Jiang and P. F. A. Alkemade, J. Vac. Sci. Technol., B, 1998, 192 P.W. O. Hoskin, J. Trace Microprobe Tech., 1998, 16(3), 301. 16(4), 1971. 193 A. Perez-Rodriguez, A. Cornet and J. R. Morante, 155 F. Jomard and M. Perdereau, Microsc., Microanal., Microstruct., Microelectron. Eng., 1998, 40(3–4), 223. 1997, 8(4–5), 273. 194 R. Hull and D. Dunn, Mater. Res. Soc. Symp. Proc., 1998, 156 W. M. Arnoldbik, Z. X. Jiang, P. F. A. Alkemade and 523(Electron Microscopy of Semiconducting Materials and D. O. Boerma, Nucl. Instrum. Methods Phys.Res., Sect. B, 1998, 136, 540. ULSI Devices), 141. J. Anal. At. Spectrom., 1999, 14, 1633–1659 1657195 J. C. Vickerman, Adv. Spectrosc. (Chichester, U.K.), 1998, 231 J. E. Mak and W. Yang, Anal. Chem., 1998, 70(24), 5159. 232 F. Thielecke, W. Brand and R. Noack, J. Mass Spectrom., 1998, 26(Spectroscopy for Surface Science), 71. 196 L. Van Vaeck, A. Adriaens and F. Adams, Spectrochim. Acta, 33(4), 342. 233 A. Tanweer and L.-F. Han, Isot. Environ. Health Stud., 1996, Part B, 1998, 53(2), 367. 197 R. M. Van Hardeveld, H. J. Borg and J. W. Niemantsverdriet, 32(1), 97. 234 D. Hofmann and W. A. Brand, Isot. Environ. Health Stud., 1996, J. Mol. Catal. A: Chem., 1998, 131(1–3), 199. 198 S. L. S. Stipp, A. J. Kulik, K. Franzreb, W. Benoit and 32(2–3), 255. 235 S. R. Boyd, I. P. Wright, C. M. O. Alexander and C. T. Pillinger, H. J. Mathieu, Surf. Interface Anal., 1997, 25(13), 959. 199 B. R. Chakraborty, D. E. Lehman and N. Winograd, Rapid Geostand.Newsl., 1998, 22(1), 71. 236 T. Goerges and K. Dittert, Commun. Soil Sci. Plant Anal., 1998, Commun. Mass Spectrom., 1998, 12(18), 1261. 200 H. J. Mathieu and D. Leonard, High Temp. Mater. Processes, 29(3–4), 361. 237 J. Lauf and G. Gebauer, Anal. Chem., 1998, 70(13), 2750. 1998, 17(1–2), 29. 201 H. Gnaser, H.-J. Schneider and H. Oechsner, Nucl. Instrum. 238 C. A. M. Brenninkmeijer and T. Rockmann, Rapid Commun. Mass Spectrom., 1998, 12(8), 479. Methods Phys. Res., Sect. B, 1998, 136, 1023. 202 J. Goschnick and M. Sommer, Fresenius’ J. Anal. Chem., 1998, 239 Y.-R. Park and E. M. Ripley, Chem. Geol., 1998, 150(1–2), 191. 240 M. Musashi, G. Markl and R. Kreulen, Anal. Chim. Acta, 1998, 361(6–7), 704. 203 H. Boerner, H. Jenett and V.-D. Hodoroaba, Fresenius’ J. Anal. 362(2–3), 261. 241 P. D. P. Taylor, S. Lehto, S. Valkiers, P. de Bievre, O. Selgrad, Chem., 1998, 361(6–7), 590. 204 H. Jenett and V.-D. Hodoroaba, Fresenius’ J. Anal. Chem., 1998, U. Flegel and T.Kruck, Anal. Chem., 1998, 70(5), 1033. 242 S. A. Macko, M. Ryan and M. H. Engel, Chem. Geol., 1998, 361(6–7), 737. 205 Y. Higashi, T. Maruo and Y. Homma, Surf. Interface Anal., 152(1–2), 205. 243 J. E. Spangenberg, S. A. Macko and J. Hunziker, J. Agric. Food 1997, 26(3), 220. 206 A. Vogt, A. Simon, H. L. Hartnagel, J. Schikora, V. Buschmann, Chem., 1998, 46(10), 4179. 244 T. N. Corso, B. A. Lewis and J. T. Brenna, Anal. Chem., 1998, M. Rodewald, H. Fuess, S. Fascko, C.Koerdt and H. Kurz, J. Appl. Phys., 1998, 83(12), 7715. 70(18), 3752. 245 R. Ovaskainen, K. Mayer, W. De Bolle and P. De Bievre, Adv. 207 R. Schmitz and G. H. Firschat, Glass Sci. Technol. (Frankfurt/Main), 1998, 71(4), 92. Mass Spectrom., 1998, 14, F055500/1. 246 M. Wallenius, H. Kuhn, K. Mayer and L. Koch, Adv. Mass 208 J. D. Sunderkoetter, D. D. Gilliland, H. Jenett, V. A. C. Haanappel and M. V. Stroosnijder, Fresenius’ J. Anal. Spectrom., 1998, 14, F055490/1. 247 T. Nakano and E. Nakamura, Int. J. Mass Spectrom., 1998, Chem., 1998, 361(6–7), 659. 209 W. Bock, B. Kutsch, M. Kopnarski and H. Oechsner, Mater. Sci. 176(1–2), 13. 248 O. M. Nieto, J. Vogel and E. Jagoutz, Adv. Mass Spectrom., Forum, 1998, 287(Trends and New Applications of Thin Films), 407. 1998, 14, F016950/1. 249 S. Halas and T. Durakiewicz, Int. J. Mass Spectrom., 1998, 181, 210 J. T. Brenna, T. N. Corso, H. J. Tobias and R. J. Caimi, Mass Spectrom. Rev., 1997, 16(5), 227. 167. 250 K. Hattori, T. J. S. Cole and D. P. Menagh, Int. J. Mass 211 A. L. Burlingame, R. K. Boyd and S. J. Gaskell, Anal. Chem., 1998, 70(16), 647R. Spectrom., 1998, 176(3), 189. 251 Y. Liu, M. Huang, A. Masuda and M. Inoue, Int. J. Mass 212 E. Deleens, J.-F. Morot-Gaudry, F. Martin, A. Thoreux and A. Gojon, 15N methodology. Assimilation Azote Plant., Inst. Spectrom. Ion Processes, 1998, 173(3), 163. 252 R. Markey, H. Stein and J. Morgan, Talanta, 1998, 45(5), 935. Natl. Recherche Agronomique, Paris, France, 1997, 265. 213 J. Oesselmann, C. B. Douthitt, W. Brand and A. Hilkert, Isot. 253 K. Hattori, D. P. Menagh and T. J. S. Cole, Anal. Chem., 1998, 70(19), 4100. Tech. Study Environ. Change, Proc. Int. Symp. Isot. Tech. Study Past Curr. Environ. Changes Hydrosphere Atmos.1997, 254 K. Habfast, Int. J. Mass Spectrom., 1998, 176(1–2), 133. 255 S. K. Sahoo and A. Masuda, Anal. Chim. Acta, 1998, 370(2–3), International Atomic Energy Agency, Vienna, Austria, 1998, pp. 839–842. 215. 256 R. N. Taylor, I. W. Croudace, P. E. Warwick and S. J. Dee, 214 H. Avak, A. Hilkert and R. Pesch, Isot. Environ. Health Stud., 1996, 32(2–3), 285. Chem. Geol., 1998, 144(1–2), 73. 257 G. M. Nowell, P. D. Kempton, S. R. Noble, J. G. Fitton, 215 T. R. Browne, G. K. Szabo and A. Ajami, Pharmacochem. Libr., 1997, 26(Stable Isotopes in Pharmaceutical Research), 119. A. D. Saunders, J. J. Mahoney and R. N. Taylor, Chem. Geol., 1998, 149(3–4), 211. 216 D. A. Krueger, Anal. Methods Food Authentication, Blackie, London, UK, 1998, pp. 14–35. 258 J. Fietzke, A. Bollhofer, N. Frank and A. Mangini, Nucl. Instrum. Methods Phys. Res., Sect. B, 1999, 149(3), 353. 217 I. P. Wright and C. T. Pillinger, Adv. Space Res., 1998, 21(11, Space Exploration of the Outer Space Solar System and 259 M. E. JoVroy, C. Hillaire-Marcel, B. Ghaleb and L. Dever, Isot. Tech. Study Environ. Change, Proc. Int. Symp. Isot. Tech. Study Cometary Nuclei), 1537. 218 I. P. Wright and C. T. Pillinger, Planet. Space Sci., 1998, Past Curr. Environ. Changes Hydrosphere Atmos., International Atomic Energy Agency, Vienna, Austria, 1998, pp. 727–743. 46(6–7), 813. 219 S. J. Barber, A. D. Morse, I. P. Wright, B. J. Kent, 260 Z. Peng, Z. Wang, W. Sun, Z. Ma, M. Xia, C. Zhang, W. Chen, Z. Zhang and Z. An, Chin. Sci. Bull., 1998, 43(4), 333. N. R. Waltham, J. F. J. Todd and C. T. Pillinger, Adv. Mass Spectrom., 1998, 14, B037640/1. 261 F. Vanhaecke, J. Diemer, K. G. Heumann, L. Moens and R. Dams, Fresenius’ J. Anal. Chem., 1998, 362(7–8), 553. 220 W. A. Brand and P. Dobberstein, Isot. Environ. Health Stud., 1996, 32(2–3), 275. 262 M. Smoliar and J. M. Ondov, Prepr. Ext. Abstr. ACS Natl. Meet., Am. Chem. Soc., Div. Environ. Chem., 1998, 38(2), 222. 221 P. Chen, B. L. Osborn and F. P. Abramson, J. AOAC Int., 1998, 81(1), 40. 263 T. A. Hinners, R. Hughes, P. M. Outridge, W. J. Davis, K. Simon and D. R. Woolard, J. Anal. At. Spectrom., 1998, 13(9), 222 K. J. Goodman, Anal. Chem., 1998, 70(5), 833. 223 K. Simon, U. Wiechert, J. Hoefs and B. Grote, Fresenius’ J. Anal. 963. 264 J. W. Olesik, J. A. Kinzer, E. J. Grunwald, K. K. Thaxton and Chem., 1997, 359, 458. 224 P. Barnekow, J. Hoefs and A. Peccerillo, Mineral. Mag., 1998, S. V. Olesik, Spectrochim. Acta, Part B, 1998, 53(2), 239. 265 G. M. Hieftje, Spectrochim. Acta, Part B, 1998, 53, 165. 62A(Pt. 1), 120. 225 J. Fiebig, U. Wiechert and J. Hoefs, Mineral. Mag., 1998, 266 A. P. Bruins, J. Chromatogr., A, 1998, 794(1–2), 345. 267 R. S. Houk, Spectrochim. Acta, Part B, 1998, 53(2), 267. 62A(Pt. 1), 454. 226 N. Takahata, Y. Nishio, N. Yoshida and Y. Sano, Anal. Sci., 268 H. Chassaigne and R. Lobinski, Anal. Chim. Acta, 1998, 359(3), 227. 1998, 14(3), 485. 227 K. J. Petzke, O. V. Korkushko, T. M. Semesko and C. C. Metges, 269 G. Wang and R. B. Cole, Anal. Chem., 1998, 70(5), 873. 270 D. V. Vukomanovic, J. A. Stone, G. W. van Loon, K. Nakatsu Isot. Environ. Health Stud., 1997, 33(3), 267. 228 K. J. Leckrone and J. M. Hayes, Anal. Chem., 1998, 70(13), and D. E. Zoutman, Spectrochim. Acta, Part B, 1998, 53, 893. 271 A. R. S. Ross, M. G. Ikonomou, J. A. J. Thompson and 2737. 229 S. D. Kelly, I. G. Parker, M. Sharman and M. J. Dennis, J. Mass K. J. Orians, Anal. Chem., 1998, 70(11), 2225. 272 O. Schramel, B. Michalke and A. Kettrup, J. Chromatogr., A, Spectrom., 1998, 33(8), 735. 230 W.-N. P. Lee, S. Lim, S. Bassilian and E. A. Bergner, J. Mass 1998, 819(1–2), 231. 273 J. R. Pretty and G. J. Van Berkel, (Oak Ridge Natl. Lab., Oak Spectrom., 1998, 33(7), 627. 1658 J. Anal. At. Spectrom., 1999, 14, 1633–1659Ridge, TN, USA). Presented at 46th ASMS Conference on Mass 10B11B isotope ratio, VGB Tech. Ver. Grosskraftwerksbetr., [Tagungsber.] VGB-TB, VGB-TB 433, VGB-Konferenz ‘Chemie Spectrometry, Orlando, FL, USA, 31 May–5 June, 1998. im Kraftwerk 1997’, 1997, N3/1-N3/7. 274 D. A. Barnett and G. Horlick, (Dept. Chem., Univ. Alberta, 286 D. L. Pinti and B. Marty, J. Chromatogr., A, 1998, 824(1), 109. Edmonton, AB, Canada T6G 2G2). Presented at 1998 Winter 287 K. Balogh and A. Simonits, Rapid Commun. Mass Spectrom., Conference on Plasma Spectrochemistry, Scottsdale, AZ, USA, 1998, 12(22), 1769. January 5–10, 1998. 288 S. Valkiers, Y. Aregbe, P. D. P. Taylor and P. De Bievre, Int. J. 275 D. A. Barnett and G. Horlick, (Dept. Chem., Univ. Alberta, Mass Spectrom. Ion Processes, 1998, 173(1–2), 55. Edmonton, AB, Canada T6G 2G2). Presented at 1998 Winter 289 K. P. Jochum, PTB-Ber. ThEx—Phys.-Tech.-Bundesanst., Conference on Plasma Spectrochemistry, Scottsdale, AZ, USA, 1997(PTB-ThEx-3), 36. January 5–10, 1998. 290 M. Flanz, H. Achtermann, B. Stoll and K. P. Jochum, PTB-Ber. 276 G.-M. Momplaisir, L. Betowski, W. Winnik and E. Heithmar, ThEx—Phys.-Tech. Bundesanst., PTB-ThEx-3, 1997, pp. 60–67. (US Environmental Protection Agency, Las Vegas, NV, USA). 291 G. I. Ramendik, Int. J. Mass Spectrom. Ion Processes, 1998, Presented at 46th ASMS Conference on Mass Spectrometry, 172(3), 221. Orlando, FL, USA, 31 May–5 June, 1998. 292 J. S. Becker, R. S. Soman, T. Becker, V. K. Panday and 277 K. C. Westaway, T. Koerner, Y.-R. Fang, J. Rudzinski and H.-J. Dietze, J. Anal. At. Spectrom., 1998, 13(9), 983. P. Paneth, Anal. Chem., 1998, 70(17), 3548. 293 K. Hiraoka, S. Saito, J. Katsuragawa and I. Kudaka, Rapid 278 C. Brede, S. Pedersen-Bjergaard, E. Lundanes and T. Greibrokk, Commun. Mass Spectrom., 1998, 12(17), 1170. Anal. Chem., 1998, 70(3), 513. 294 E. Magi and C. Ianni, Anal. Chim. Acta, 1998, 359(3), 237. 279 C. Brede, E. Lundanes, T. Greibrokk and S. Pedersen-Bjergaard, 295 K. Russe, H. Kipphardt and J. A. C. Broekaert, (Dept. Chem., J. High Resolut. Chromatogr., 1998, 21(2), 633. Univ. Dortmund, 44221 Dortmund, Germany). Presented 280 B. Pons, A. Carrera and C. Nerin, J. Chromatogr., B: Biomed. at 1998 Winter Conference on Plasma Spectrochemistry, Appl., 1998, 716(1–2), 139. Scottsdale, AZ, USA, January 5–10, 1998. 281 C. M. Barshick, S.-A. Barshick, P. F. Britt, D. A. Lake, 296 A. A. Sysoev, V. P. Ivanov, T. V. Barinova and A. I. Grishin, M. A. Vance and E. B. Walsh, Int. J. Mass Spectrom., 1998, [Dept. Mol. Phys., Moscow State Eng. Phys. Inst. (Tech. Univ.), 178(1–2), 31. Moscow 115409, Russia]. Presented at 1998 Winter Conference 282 G. Binato, G. Biancotto, R. Piro and R. Angeletti, Fresenius’ on Plasma Spectrochemistry, Scottsdale, AZ, USA, January J. Anal. Chem., 1998, 361(4), 333. 5–10, 1998. 283 P. Van Dael, D. Barclay, K. Longet, S. Metairon and L. B. Fay, 297 K. G. Heumann, S. M. Gallus, G. Radlinger and J. Vogl, J. Anal. At. Spectrom., 1998, 13(9), 1001. J. Chromatogr., B: Biomed. Appl., 1998, 715(2), 341. 298 Y. K. Xiao and L. Wang, Int. J. Mass Spectrom., 178(3), 213. 284 X. B. Chen, A. G. Calder, P. Prasitkusol, D. J. Kyle and M. C. N. Jayasuriyat, J. Mass Spectrom., 1998, 33(2), 130. 285 M. Bolz, W. HoVmann and W. Ruhle, Determination of the Paper 9/05419G J. Anal. At. Spectrom., 1999, 14, 1633–1659 1659

ISSN:0267-9477    

DOI:10.1039/a905419g

出版商:RSC    

年代:1999

数据来源: RSC