Atomic Spectrometry Update. Atomic mass spectrometry
Jeffrey R. Bacon*a, Jeffrey S. Crainb, Luc Van Vaeckc and John G. Williamsd
aThe Macaulay Land Use Research Institute, Craigiebuckler, Aberdeen, UK AB15 8QH
bUnion Carbide Corporation, a subsidiary of the Dow Chemical Company, Technical Center, PO Box 8361, South Charleston, WV 25303, 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
First published on 20th July 2001
Abstract
This Update follows on from last year's1 but covers a longer period. The system we use to access abstracts has been improved with the result that abstracts have become available more quickly and the review more timely. This review effectively covers a period of about 15 months up to the beginning of 2001. Although an attempt is made to consider all relevant refereed papers, conference abstracts, reports, book chapters and patents for inclusion, this review does not aim at being comprehensive in its coverage. The selection of papers is based on criteria applied to focus sharply on the most significant developments in instrumentation and methodology or improved understanding of the fundamental phenomena involved in the MS process. With the boundaries between atomic and inorganic molecular MS, and indeed organic MS, becoming less well defined, the judgement of the authors of this Update becomes important in considering papers for inclusion. The main ruling criterion for all papers is that the work should involve or be intended for the study of natural systems. For example, the study of synthetic metal clusters is generally not included whereas the determination of organometallic compounds in environmental samples is.
Applications of atomic MS are not covered in this Update and readers are referred to the Updates on Industrial Analysis: Metals, Chemicals and Advanced Materials,2 Environmental analysis3 and Clinical and Biological Materials, Food and Beverages.4 Becker and Dietze5 have produced yet another excellent review (210 references) covering in detail the inorganic MS techniques that are used for trace, isotope and surface analysis. Barshick et al.6 discussed the fundamentals and applications of inorganic MS in their substantial review. Although not focused specifically on MS techniques, the review (484 references) of Rao and Biju7 on the determination of REEs was interesting in that it presented most of its data in tabular form.
The trends noted over the last few years have continued. In particular, the growing interest in speciation studies has been matched by the number of papers on the subject. The renaissance of GC-MS has continued and the niche of ESMS for species identification has been confirmed, in particular to complement the elemental information provided by ICP-MS. In all of these studies, sample preparation and introduction have generally received most attention with the aim of further improving analysis. The very wide range of applications of AMS was made apparent by the papers presented at a major international conference.8 The encroachment by ICP-MS into areas of traditional TIMS studies has continued with some success but ultimate levels of precision and accuracy can often only be achieved by the latter technique even if procedures are complex and time-consuming. A number of interesting trends were noted for ICP-MS in this review period. The use of LA, particularly for geological and environmental samples, continued to grow and has almost become a routine application. The use of collision or reaction cells for reduction/removal of polyatomic species has gained momentum and the use of both single- and multiple-collector magnetic sector MS systems has also increased significantly. There has been some interest in ICP-TOFMS systems, but it is not yet clear whether they offer any significant advantages over existing systems.
1 Accelerator mass spectrometry (AMS)
1.1 Reviews
For a complete overview of the current state of AMS and its applications, the reader is referred to a special issue of Nuclear Instruments & Methods in Physics Research, Section B,8 which contained the Proceedings of the Eighth International Conference on Accelerator Mass Spectrometry held in Vienna in 1999. The substantial volume (over 1000 pages) included articles from all AMS facilities and covered all aspects of AMS analysis. Topics included facility reports with details of instrumental upgrades and modifications, improved methods of sample preparation, new instrumentation and new areas of application. The article by Fifield9 gave a detailed review of developments in AMS since the Seventh International Conference in 1996 and, as an indication of the more widespread availability of AMS, listed the 16 new facilities. The most significant advances highlighted in the article were the development of small instruments, the extended list of target isotopes, developments in radiocarbon analysis to give high sample throughputs and the reliable calibration of production rates of 26Al, 10Be and 36Cl in the determination of exposure ages.A large number of review papers has appeared in the period covered by this Update, perhaps reflecting a desire to take stock of the current capabilities of AMS as well as consider the direction of developments. The 55-reference review by Kutschera10 presented a broad view of the principles and potential of AMS in addition to selected applications. Other excellent general reviews were those of Gove,11 which not only covered the historical development of AMS but also presented some thoughts on future developments, and of Hotchkis et al.12 on recent developments and the need for traceable RMs. The recent development of compact instruments for radiocarbon analysis using terminal voltages below 1 MV and ions in +1 or +2 charge states has been notable. Suter et al.13 discussed the current status of the technique and considered its potential for determining other radionuclides. The increased use of such instruments in biochemical and biomedical research was reflected by the number of reviews on these subjects including those of Garner and Leong14 and Turteltaub and Vogel15 on drug research and that of Vogel16 on biochemical research. More specific reviews of note included those of Braucher et al.17 on the use of in-situ produced cosmogenic 26Al and 10Be in the study of soil development, Currie18 on critical events in the development of radiocarbon aerosol science and Kretschmer19 on sample preparation for radiocarbon analysis and applications in archaeology, geology and environmental research.
1.2 Instrumentation
Active development of small AMS systems has continued with several new instruments being reported. Whereas the first generation of small instruments, for example the commercial High Voltage Engineering HVEE design, used ions in the +2 charge state, the operational system at Zürich20,21 was the first to use the +1 charge state. Use of the +2 charge state was considered to be seriously limited by the 7Li+ background. The short accelerator tube with a maximum terminal voltage of 550 kV allowed a compact design with a floor area of 6.5 × 4.5 m. The argon gas density of 2 µg cm−3, necessary to achieve sufficient suppression of interfering molecular species, resulted in a relatively long (76 cm) terminal stripper, however. It was demonstrated that molecular interferences could be destroyed efficiently enough for radiocarbon analysis at natural levels to be feasible. From a study on the Zürich system of the ion beam interaction with the stripper gas, Jacob et al.22 concluded that AMS with terminal voltages of only 200 kV was possible. At that point, angular straggling of the ion beam became a limiting factor. Molecular species could be destroyed by collision with the stripper gas throughout the energy range studied (200–600 keV). A 1 MV system under construction at the Lawrence Livermore National Laboratory23 has been designed for high throughput (>300 samples per day) combined with attomole sensitivity at 3.5% precision. Schubank24 has proposed a new design of a low-energy table-top AMS instrument based on the mass analysis of high intensity +3 ion beams followed by charge exchange to form negative ions.The high throughput now achieved at many facilities and the production of large quantities of data have driven the development of fully automated instruments. Steier et al.25 have explored the possibility for computer-assisted and automated analyses using the VERA facility at Vienna. A multi-dimensional optimization algorithm, robust to noise, was developed as a new tool for computerised tuning of the ion optics. Operator involvement was still required at this stage but the reliability of the setup was greatly improved and the time needed for tuning was reduced by about half to 90 min. The isotope ratio measurements themselves were fully automated yet data could be evaluated on-line.
The pioneer of small, low energy AMS instruments, the minicyclotron facility at Shanghai, has been brought into routine operation after about 10 years of development. In a detailed report, Liu et al.26 described the solutions found to the unique problems encountered in this project, which included electric field penetration of the high voltage on the extraction deflector, unstable power supplies caused by rf interference, high electron loads destabilizing voltages, the influence of vacuum conditions on measurement precision, space charge effects and X-ray induced backgrounds. As the minicyclotron cannot simultaneously accelerate ions with different masses, ions must be accelerated and measured sequentially. Chen et al.27 investigated five different sequential acceleration schemes and adopted one which measured only the weak 13C− and 14C− ion beams and not the strong 12C− beam. Various other instrumental parameters were also investigated and optimized in order to achieve the first real 14C measurements on the instrument. Zhou et al.28 reported use of the system for analysis of archaeological samples (1 mg C) with a mass resolution of >3000 and a measurement precision of 1%.
Considerable effort continues to be put into improving the design and performance of the ion source. In a particularly informative paper, Southon and Roberts29 outlined the evolution of the Cs sputter ion source used at the Lawrence Livermore National Laboratory from a commercially available source. Modifications included a complete redesign of the sample changer to improve reliability and ease of maintenance, improved pumping and operation at higher sputter voltages to increase output and improved thermal isolation to reduce memory effects. Operating characteristics of the current source are a sample changing time of 3 s, a sample carousel change time from beam off to beam on of 15 min and a full source cleaning time of 3 h. The ion source will maintain performance for at least two weeks without intervention. Modelling of the interior of the sputter source by Brown et al.30 indicated that significant modification of the interior geometry of the source would double Cs+ ion production from the spherical ionizer and produce a significant increase in negative ion output. Redesign of the target shield and ionizer shroud increased the 12C− ion currents from 200–250 to ≥360 µA.
Von Reden et al.31 have reported the first direct comparison of the two leading sputter ion source designs. Both sources were mounted on the same system. Results from the two sources were in good agreement but the older, 59-sample, spherical ionizer source presented several operational problems including excessive maintenance requirements. On the other hand, the newer, 134-sample, high intensity design was considered to have several desirable features, in particular the smaller cathode geometry that allowed small samples to be analysed.
Gas ion sources are potentially useful for the radiocarbon analysis of small samples for which high precision is not feasible. Ramsey and Humm32 have described the testing of on-line combustion of samples in their further development of the hybrid (graphite or CO2) ion source at Oxford, which has been upgraded to incorporate a 40-target, sample changing mechanism. The combustion system was suitable for small samples (30–150 µg C) in radiocarbon dating, pre-screening of radiocarbon ages and biomedical applications. Cross-contamination was low at <0.25% and the combustion blank was 0.5 µg modern C. A microwave ion source, reported in last year's Update but not yet coupled to an AMS instrument, has been further developed by Schneider et al.33 to improve conversion efficiency of carbon atoms into ions. The plasma discharge chamber volume was reduced from 137 to 50 ml and the extraction aperture from 5 to 3 mm diameter. The configuration of the permanent magnet field remained unchanged as it has proved excellent at producing high currents at low charge states. The source is intended for samples for which a sputter source would be impractical, including very small samples (few µmol C), the monitoring of evolved gases and large numbers of small gaseous samples.
The use of a 10 GHz electron cyclotron resonance (ECR) ion source was found by Paul et al.34 to have two interesting features in the analysis of heavy elements. Efficient production of high-charge state ions in the ECR source ensured elimination of molecular species at the ion source stage. The linear acceleration based on velocity matching, together with the beam transport system, acted as a powerful mass filter for suppressing the background. System performance was improved by careful selection of the charge state of the accelerated ion (for example, 236U21+) and use of Xe pilot beam (124Xe11+) with a low memory effect. An abundance sensitivity of 1 × 10−14 has been demonstrated for lead isotopes.
Several elements of interest in analytical and environmental studies do not form sufficient quantities of negative ions to be useful for AMS analysis. Neutral injection would allow such elements to be analysed provided the elements could be transported to the terminal of the tandem accelerator and stripped to produce positive ions with sufficient efficiency. DeTurck et al.35 have studied the neutralization processes for Ar using an ion implantation accelerator. Singly charged positive ions were accelerated and passed through a differentially pumped charge-exchange cell, where a portion was neutralized. Maximum neutralization efficiency was achieved with low ion energy and increasing target thickness up to an equilibrium point. If the target thickness was below the equilibrium point, the neutral portion created from the incident Ar ion beam did not suffer significantly from gas scattering after passing the charge-exchange cell. Litherland et al.36 investigated charge neutralization for the design of a neutral injection system and found that multiple scattering during molecular dissociation was more important than single and double scattering due to neutralisation and re-ionization.
Brown and Gillespie37 have described an electrostatic deflector model, which allowed first-order simulation of different deflector geometries. The model was applied to the modelling of a spherical electrostatic analyser as a component of a low energy beamline recently installed at the Lawrence Livermore National Laboratory. Although initial tests showed that the new analyser was not behaving as predicted by the first-order model, modifications to correct for non-first-order electrostatic fields led to the beamline becoming operational.
Drifts in accelerator stability and beam transport components limit the precision that can be achieved in AMS. To minimize these effects, the isotopes of interest should be counted for as short a period as possible but fast switching at the high-energy end of the instrument has hitherto not proved feasible. Sie et al.38 have constructed a high-energy fast isotope switching system based on a pair of deflector plate sets, deflecting in the orbit plane, at the entrance and exit of the high-energy analysing magnet. Switching times of <10 ms per isotope led to precisions of better than 1‰ for lead isotope ratio measurements.
If heavy ions are stopped in a suitable target, characteristic projectile X-rays (PX) are produced, which can be used to identify the ion by its atomic number and thereby separate it from isobaric interferences. Two groups have continued to develop this method as a detector for AMS, the so-called PX-AMS technique. Persson et al.39 improved their original system for the detection of 22 MeV 59Ni8+ ions through use of a germanium target. In addition, the foil stripper in the accelerator was replaced with a gas stripper. Further improvement in the LOD to 4 × 10−959Ni∶Ni was achieved by chemically reducing the 59Co content of samples. He et al.40 improved the detection sensitivity of 79Se by more than two orders of magnitude through use of PX-AMS with an yttrium target. Targets of CdSe were used to produce high Se− ion currents and 78Se and 80Se, which interfered with 79Se in the PX detector, were removed by placing a slit after an electrostatic deflector in the high-energy section of the MS instrument. The detection limit was about 3.6 × 10−9 for 79Se∶Se. The detector has also been used41 for analysis of the radioisotope 64Cu, which must be determined immediately upon production because of its short half-life of 12.7 h.
Knie et al.42 have reported that use of a gas-filled magnet system combined with a multi ΔE (energy loss) ionization chamber, optimized for maximum background rejection, which allowed relatively high negative ion currents to be used and resulted in short measurement times in the determination of 60Fe, 53Mn and 63Ni. Best overall suppression of isobaric background was achieved with a N2 pressure of 7 mbar in the magnet chamber. The LOD of 2 × 10−16, 2 × 10−14 and 2 × 10−14 for 60Fe∶Fe, 53Mn∶Mn and 63Ni∶Ni, respectively, were superior to those achievable with other detectors.
Two other improved detector systems have been reported for the analysis of 36Cl. Jiang et al.43 used a gas-filled (10% methane in Ar, 35–50 Torr) TOF detector system in combination with a new particle identification system called “energy-loss and time-of-flight (ΔE-T) parabola”. Santos et al.44 designed and built a Bragg curve detector based on the principle that a particle passing through a gas (propane at 164 Torr) produces electrons which drift towards an anode and produce a signal. Both detectors successfully discriminated between 36Cl and 36S when operated with 72 or 64 MeV ions, respectively.
There have been few reports on the development of AMS for the trace element analysis of solid samples. Sie et al.45 have given full details of their system for the in-situ microanalysis of geological samples using a microbeam Cs source. The system featured a novel bouncing system on the high-energy side. Bouncing at a rate of at least 5 ms per isotope was essential to overcome source instabilities and to achieve the high precision required in isotope studies. A bouncing rate of <1 ms per isotope was used to achieve a precision of better than 1‰ for Pb isotope studies. Some long-term fractionation has been reported for the system and it was considered that computer control was required to stabilize the accelerator set up further.46 Datar et al.47 have used the related technique of accelerator SIMS to determine As in a GeSi matrix with a sensitivity significantly better than that possible with conventional SIMS using 10–20 keV primary ions. Litherland et al.36 investigated the use of cesium vapour between the target and local ionizer to overcome unstable negative ion beams, which resulted from surface charging upon the ion bombardment of insulators.
The coupling of chromatography with AMS is seen as a route to achieving cost-effective high-throughput systems. Hughey et al.48 have developed a prototype GC-MS designed for the detection of 14C and 3H in biomedical research. The effluent from the GC passed through a combustion furnace to produce CO2 for the gas-fed Cs sputter ion source. A Nafion™ drier was required between the furnace and ion source as the AMS source proved to be very sensitive to water vapour. Although the first on-line measurements of carbon in resolved GC peaks were reported, the complete AMS system was still under construction. Description of a system by Buchholz et al.49 as HPLC-AMS was somewhat misleading in that it was not an on-line system. Compounds eluted from the HPLC system were collected in fractions and graphitized separately for AMS analysis. A number of problems made matching of the two systems difficult. Accounting for the carbon inventory in the LC solvents, ion-pairing agents, samples and carriers complicated the analysis. The sensitivity of the AMS technique made background a problem, especially as it could vary during use of gradient elution. Limits of quantitation of about 10 amol 14C per HPLC fraction were, however, attained.
1.3 Developments in radiocarbon analysis
There has been a notable increase in the number of reports on radiocarbon analysis, in particular on sample preparation and the analysis of new types of small samples. Of particular note has been the use of compound-specific radiocarbon analysis (CSRA). Uchida et al.50 dated individual fatty acids in estuarine sediment samples and found that although the bulk organic component in the sediments had an age of 5000 years, some individual components were composed of modern carbon whereas others had an age of 17![[thin space (1/6-em)]](https://www.rsc.org/images/entities/char_2009.gif)
000 years. Samples were separated by preparative GC with cryogenic trapping prior to combustion and graphitization for AMS analysis. A similar study by McNichol et al.51 found that the dates of individual phenolic compounds isolated from lignin agreed to within 2% with that of the bulk. Although the Woods Hole instrument is currently dedicated to the analysis of sea-water,
this area of research is being reduced and Gagnon et al.52 considered CSRA to be a viable alternative. The instrument used a multi-dimensional preparative capillary GC system in which only selected peaks from a first column were injected onto a second column.Advanced research in atmospheric and marine science requires higher sensitivity than the tens of µg C commonly achieved. Currie et al.53 have reported an intrinsic modern-C quantification limit of 0.9 µg when using a small sample (25 µg) AMS target preparation facility and a microsample combustion–dilution facility. The overall processing blank for the analysis of Greenland ice was reduced from 7 µg total carbon to <1 µg. Klinedinst and Currie54 also described a correction methodology which can be applied when no blank data are available, for example when blanks in a sample collection system become contaminated through sampler malfunction.
Dee and Ramsey55 have described in detail their study of the range of possible reaction conditions during the preparation of graphite targets and the effect on the nature of the graphite produced. The optimum amount of iron catalyst was 2–2.5 mg. Less than this led to very fluffy graphite which was difficult to work with. More resulted in an increased background. A H2∶CO2 ratio of 2.1∶1 was chosen to prevent tubes bursting at higher ratios. The optimum temperature for the trapping of water vapour was 5°C. At higher temperatures, a high background was observed. At lower temperatures, the tubes would freeze. The optimum reaction temperature was in the range 550–650°C. At a temperature of <500°C, hydrocarbons were formed. At >750°C, the graphite would not compact. The studies have led to a preparation method as robust as possible with a high
success rate, a short reaction time and very simple apparatus requirements. A novel feature of the method of Yuan et al.,56 for the preparation of CO2 gas using an elemental analyser, was the collection system whereby the CO2 could be recovered if the graphitization failed. The apparatus of Pohlman et al.57 allowed continuous extraction, distillation and graphitization of a suite of organic and inorganic carbon pools within a single system.
The increased use of AMS in biomedical studies has highlighted a problem of radiocarbon contamination. Buchholz et al.58 identified the problem as being a legacy of earlier tracer work in most biological laboratories and outlined some of the measures that needed to be taken to minimize contamination. The three main sources of contamination were the laboratory and its furnishings, laboratory air and tools or vessels. The authors considered that there was a need to cultivate an appreciation of AMS sensitivity and the ease of cross-contamination.
Klinedinst and Currie59 have performed radiocarbon analyses of atmospheric particulate matter with an equivalent aerodynamic diameter of ≤2.5 µm (PM2.5) using techniques specially developed for small samples (<100 µg C). The preparation of carbonaceous material deposited on quartz fibre filters involved isolation of the carbon fraction of interest, combustion to form CO2 and reduction by manganese to graphite. Carbonate-C was removed from collected samples through exposure to HCl fumes. The Fe–C bead targets, used in place of the usual pressed graphite powder, were prepared in a closed system in order to minimize target preparation blank. It was concluded, perhaps unsurprisingly, that PM2.5 collected in both summer and winter seasons in an urban area was dominated by fossil-derived carbon but that summer episodes could be dominated by biomass-derived carbon.
A number of new sample preparation procedures have been presented for extracting C from a diverse range of matrices. Two groups have carried out similar studies on the extraction of C from fossil bones. Both Minami and Nakamura60 and Yuan et al.61 used XAD-2 chromatography to remove from collagen hydrolysates the humic and fulvic acids derived from exogenous organic matter. This was of particular benefit in the analysis of poorly preserved bones containing less than 1% extractable gelatin for which the traditional method of gelatin extraction proved inappropriate due to contamination with humic or fulvic acids. The gelatin extraction procedure was, however, simpler and gave good results for well-preserved samples. In studies of sediment and peat cores, Morgenroth et al.62 separated organic fractions (for example, plants and pollen) from the inorganic carbonate and silicate fractions by ultracentrifugation in ZnCl2 solution (2 g cm−3, 3000 rev min−1, 10 min). The dating of textiles requires the effective removal of modern contaminants such as fats and oils, food particles and humic substances. In order to investigate the effectiveness of traditional cleaning procedures, Turnbull et al.63 compared the traditional method of a series of acid and alkali washes with a new method in which the specific reaction between ninhydrin and silk or wool protein was used to extract the associated C. No significant differences were observed between the two methods so it was concluded that the traditional method is effective at removing non-protein sources of C and that the large number of modern dates being given for supposedly ancient textiles are in fact genuine.
The huge amount of data collected during routine 14C AMS measurements requires sophisticated processing tools to guarantee the quality and reliability of the resulting radiocarbon dates. The automatic evaluation system described by Puchegger et al.64 included a calibration program which was able to handle the peak associated with nuclear bomb testing. The on-line, real time evaluation of data, applied routinely before the measurement had been finished, allowed instrument instabilities and other malfunctions to be identified.
1.4 Developments in the analysis of elements other than carbon
Dedicated facilities for the measurement of tritium have been developed for two quite different applications. Friedrich et al.65,66 have built a system for depth profiling in carbon in order to understand and control the processes in nuclear fusion experiments. In order to achieve uniform sample sputtering, essential for depth profiling, and to avoid edge effects, the samples were scanned mechanically within a circle of 2.5 mm diameter and data only collected during sputtering at the centre of the crater. Ions sputtered from the side wall of the crater were not analysed. The LOD was about 1011 atoms cm−3 and the depth range >25 µm. The absorption of water vapour from residual gas, which limited the depth profile information, could be reduced through the use of an ultra-high vacuum ion source or a vacuum system with cryogenic pumps. Roberts et al.67 have developed a low-technology, compact system for the measurement of tritium tracers used in biological and environmental research. The system was based on a rf quadrupole linear accelerator (linac) which accelerated only hydrogen, thereby simplifying the input beam transport with no need for mass analysis. Molecular species (HD, H3) were dissociated by a foil placed after the linac and the 3H, mass separated from 1H and 2H, measured with a silicon charged-particle detector. Although the achieved LOD for 3H∶H was 1 × 10−13, a 100- to 1000-fold improvement in comparison with scintillation counting, it was expected to be improved further to 1 × 10−15. In addition, the mg sample requirement was 1000-fold lower than that of scintillation counting.The need to develop dating methods beyond the radiocarbon limit has seen increased use of the in-situ-produced cosmogenic radioisotope 10Be for dating. Child et al.68 have developed a method for the high and reproducible recovery of Be in the presence of excess Al and Ti. They modified an existing cation-exchange procedure to incorporate a washing step with 0.25 M H2SO4 containing 0.015% H2O2 which eluted titanium as a peroxide complex. The same group69 have reported the first 10Be concentrations measured in Antarctic ice cores and found that the effects of production, transport and deposition affect the interpretation of the 10Be record. In their study of prehistoric cave sediments and flints, Boaretto et al.70 encountered large errors in 10Be values due to B interferences, introduced by storage in glass containers, and high background 7Be. In addition, there were problems in removing atmospherically produced 10Be in the cleaning process.
The cosmogenic radioisotope 32Si (t½ = 140 yr) is an excellent candidate to provide time information in the range 100–1000 yr. Morgenstern et al.71 have reported the first AMS 32Si measurements in natural samples (rainwater, ice and snow). The principle AMS challenge was removal of isobaric 32S, which was achieved through use of a gas-filled magnet system. The LOD for 32Si∶Si was 2 × 10−15. Treacy et al.72 achieved isobaric suppression by using a new chemical separation procedure, a key feature of which was the conversion of SiO2 into Mg2Si, and the differential energy loss profile of 32Si5+ and 32S5+ ions after transmission through a thin carbon foil. Use of Mg2Si produced higher 28Si− currents (15–20 µA) than the use of SiO2 (5–10 nA), but further improvement of the method was required to reduce completely the SiO2.
A major objective in the measurement of 36Cl is discrimination from isobaric 36S. Nagashima et al.73 relied on two, independent chemical procedures (dissolution and precipitation) with a total processing time of 10 h to reduce the background 36Cl∶Cl to 1 × 10−14. This was considered to be unsatisfactory and the authors were seeking further improvements. They adopted a novel way for keeping the accelerator terminal voltage stable to within 0.1%. The AgCl target was mixed with pure graphite (AgCl∶C = 200∶1) so that 36Cl− and 12C3− beams were produced simultaneously in the ion source and the latter could be used as a pilot beam to produce feedback for controlling the acceleration voltage. Hatori et al.74 used a gas-filled magnet system (N2 at 2 Torr) to reduce the 36S background by a factor of 6 × 10−5. Their LOD was also 1 × 10−14 and their reproducibility 4%.
Nuclear processes and doses can be measured by determination of Ni isotopes produced in steel or copper. The detection of 63Ni, induced by fast neutrons in copper samples from Hiroshima and Nagasaki, requires a very sensitive technique because of the relatively small amounts of 63Ni expected (about 105–106 atoms g−1 copper). The group at Munich75,76 have optimized the ion source of their high-energy facility to reduce the background level of the 63Cu isobaric interference by two orders of magnitude. The basic features of the source were an open geometry with large pumping units and a water-cooled surface to reduce cross-talk and to improve the vacuum. All source components normally made of copper were replaced with copper-free materials. In particular, the target holders were made of high-purity graphite. After each run, the source was completely dismounted and cleaned by sand blasting. The combination of a large tandem accelerator (>10 MV) with a detection system featuring a gas-filled magnet system and multi-anode ionization chamber allowed for very sensitive measurements. The LOD for the 63Ni∶Ni ratio of 2 × 10−14 was limited by the 63Cu background. In contrast to this study, Persson et al.77 have presented the first investigation of low-energy (3 MV) PX-AMS for the measurement of 59Ni in irradiated steel samples. Although extensive sample preparation was used to remove the isobaric 59Co, the tail from the cobalt Kβ peak was still considered the most serious limitation to the LOD for 59Ni∶Ni of 1 × 10−7.
In a continuation of their previously reported work, Collon et al.78 have reported the first dating of groundwaters using 81Kr. Extensive analytical procedures were required to extract a reliable 81Kr signal from large (16 × 103 l) groundwater samples. The scale of the challenge is demonstrated by the fact that <1100 atoms of 81Kr l−1 were available for counting. This represents an activity of 1.1 × 10−10 Bq or one radioactive decay every 300 yr. Although the feasibility of the method has been established, it was considered that “routine” measurements for groundwater samples of convenient size (10–20 l) still rested on major technical improvements.
The AMS of 99Tc presents two particular challenges: the lack of a stable Tc isotope to form the matrix and isobaric interference from 99Ru, which is difficult to resolve from 99Tc. Two groups have developed different approaches to solving these problems. In the first reported measurement of low levels of 99Tc by AMS, Fifield et al.79 used high purity aluminium oxide as the bulk material with a known trace amount of 103Rh added as reference element. The contribution from 99Ru was quantified by measuring the 101Ru isotope. In contrast to conventional AMS, in which at least one isotope is measured as a beam current, in the procedure described all three isotopes (103Rh, 101Ru and 99Tc) were present in very low concentrations (<10−7). They were all measured in the +14 charge state using a propane-filled ionization chamber, which permitted partial, but not complete, separation of 99Tc from 99Ru ions. In a different procedure, Bergquist et al.80 prepared targets in a niobium matrix containing 95mTc as chemical yield monitor. The high-energy ion beam (99Tc13+ ions at 125 MeV) allowed rejection of 99Ru by up to two orders of magnitude. The sensitivity of 10–20 fg, similar to that achievable with TIMS, was limited by the need to measure time-dependent Tc∶Nb ratios.
The determination of 129I by AMS poses a number of challenges: in particular, problems of contamination; the host of molecular fragments that can appear at the detector together with I; and the difficulty of separating 129I and 127I. The importance of this analysis is demonstrated by scientists in a considerable number of facilities having reported their solutions to these challenges. Szidat et al.81,82 have given a good overview of some of the analytical problems, in particular the need for quality assurance and the control of contamination and blanks during sample preparation. The equilibration of intrinsic 129I with 127I carrier was considered not to be trivial and careful control of pH was required to prevent, for example, loss of I by evaporation. Buraglio et al.83 also demonstrated the importance of pH control, with increased 129I concentration observed upon acidification of samples, but considered storage not to be an issue. Szidat et al.84 and Lopez-Gutierrez et al.85 have given details of their own preparation procedures, developed to address some of these issues for soil, water and thyroid samples. The quality control procedures identified careful blank control as being important with evidence for contamination during long-term storage. These authors and others (Hatori et al.86 and Buraglio et al.87) used what can now be considered the standard approach to 129I detection, namely discrimination of high-energy 129I5+ ions from serious interferences (for example, 128TeH5+ and other molecular fragments) through use of a TOF spectrometer and ΔE−E ionization chamber. The choice of the +5 charge state is generally preferred over the +7 state because it produces higher transmission yet lower background count rates. Buraglio et al.87 found that an unidentified background peak could be removed either by careful selection of slit settings or through use of niobium as a target binder in place of the more commonly used silver binder. Yiou et al.,88 on the other hand, exploited the affinity of molecular iodine for silver to develop carrier-free extraction of µg quantities of I. Although more complicated than the traditional carrier method, the new procedure had the advantage of not requiring independent determination of the carrier and avoided potential problems in achieving complete equilibration. Biddulph et al.89 relied on a new high-energy beamline to resolve the 129I5+ signal so that their system remained compatible with radiocarbon analysis. Neither a low-energy electrostatic deflector nor a high-energy TOF detector was used. Extensive sample preparation procedures were needed. The need for careful control of sample preparation was demonstrated by a 129I interlaboratory comparison, which identified discrepancies of more than an order of magnitude for a suite of samples prepared in three laboratories but analysed in eight.90
A considerable effort has been directed at the determination of Pu isotopes by AMS. Wallner et al.91 reported that the large analysing magnet installed at the Munich facility was able to deflect even the heavy transuranic elements at the high energies achievable with the tandem accelerator. Final detection by TOF and multiple energy loss measurement resulted in the almost background-free measurement of individual actinides like 244Pu. An efficient chemical preparation method was used to extract Pu from 1 kg samples of terrestrial materials. In contrast to the high-energy approach, Priest et al.92 used a terminal voltage of only 3.5 MV to demonstrate the suitability of AMS for the detection of ultra-low concentrations of Pu in human urine samples. Levels of 239Pu as low as 2.5 × 10−16 g (550 nBq) were measured, albeit with large errors. This compares with the practical LOD for α-particle spectrometry, employing count times of several months, of 25 µBq. Oughton et al.93 have measured 240Pu∶239Pu ratios in samples contaminated by releases from a nuclear installation with a LOD of <1 fg. The measured ratios could be used to identify the presence of several different sources of the Pu.
A common theme throughout much of the reported work has been the need for strict analytical quality assurance and control and, in particular, for RMs. Wallner et al.94 identified the lack of a reliable 26Al standard when several 26Al standards were analysed in other laboratories and no accurate absolute values could be established. An independent standard was produced by irradiating Al-metal samples with a high neutron fluence. The isotopic ratio measured by AMS in the author's facility agreed within 1.4% with the ratios derived from activity measurements.
2 Glow discharge mass spectrometry (GDMS)
2.1 Reviews
Glow discharge sources and their analytical application in optical and mass spectrometry were covered by Caroli et al.95 in a comprehensive 57-page review with nearly 400 references. A tutorial on mechanisms in sputtering and ionization preceded the critical review of the innovative sources and their potential. The analysis of metals and non-conductive samples was surveyed with extensive referencing to the achieved LOD, reproducibility and accuracy.Bings et al.96 reviewed speciation analysis by chromatography coupled to GD- or ICP-TOFMS. The ability to record a full spectrum in 50 µs made TOFMS the method of choice for probing the rapidly changing analyte concentrations in eluents from GC, LC and capillary electrophoresis (CE). Furthermore, the actual sampling of the initial ion bunch within nanoseconds allowed high precision to be achieved for peak ratios. Particular attention was devoted to gas-sampling GD sources which allowed elemental and molecular information to be obtained in rapid succession by biasing the introduction electrode.
2.2 Instrumentation
Interface skimmer and nozzle design is critical for the coupling of high-pressure sources to low-pressure MS. Hang et al.97 investigated the different physics to be considered for sampling and ion extraction in ICP and GD. The gradual conversion of the transition flow into molecular flow aided the GD skimmer design whereas ICP devices had to deal with shock wave formation. Because few collisions between heavy and light atoms occurred and ions kept their thermal energy, the GD skimmer angle was not critical. Attention was needed, however, to avoid space charge and freezing between sampler and skimmer. The latter effect can cause fast polyatomic clustering even with few collisions.Particle beam (PB) technology was exploited by Gibeau and Marcus98 for the interfacing of LC to GD for the analysis of free metals and organic compounds in solution. The inorganic mass spectra contained only atomic ions and no oxide ions whereas organic analytes yielded results virtually identical to those obtained with 70 eV electron ionization sources. Concentrations of about 6 pg ml−1 for Ag, Cs and Fe and 13 pg ml−1 for caffeine were detected. The RSD was better than 5% over a concentration range of 10–500 µg ml−1. Marcus et al.99 extended the approach to the direct introduction of particulate matter, which flash evaporated upon impact on the heated inner surface of the GD source and was ionized by the 0.5–5 Torr He or Ar plasma. Sensitivity was determined by introduction of aliquots of the NIST SRM 1648 (Urban Particulate Matter). Elements such as Na and V yielded LOD in the range of 20–30 ng for samples of 50 µg whereas 7 ng of Fe could be detected in a 175 ng sample. The capability also to generate molecular ions from organic samples was considered to be of great interest for future applications of real-time environmental monitoring.
A simple and compact GD-MIP tandem ion source-TOFMS was developed in order to achieve improved ionisation.100 Neutral species sputtered from the GD source were ionized in the MIP. The MIP could be added without disassembling the original GD source. No boundaries between the GD and MIP could be observed. Microwave plasma boosting improved the LOD by a factor of 3–4 and reduced the background signals from Ar+ and ArH+. The systematic optimization of the discharge pressure and voltage and of the sampling distance by Su et al.101 increased signal enhancement factors up to 10-fold without a significant increase in the noise level.
Yang et al.102 designed a µs-pulsed Grimm-type source for TOFMS as a small additional chamber on an ultra-thin vacuum valve. The sample, simply pressed against an O-ring, stayed fully accessible for positioning. Adjustment of the extraction delay largely removed the background signals from argon, water, oxygen and nitrogen from the spectra. Isotope ratios were accurate to within 3% and the RSD was 0.12%. Analysis of the NIST SRM 1104 (Brass) over a period of 3 days yielded a RSD within 4.5% and the LOD were in the ppm range.
2.3 Fundamental studies
The modelling of GDs is proving to be an effective approach to understanding ion formation. Bogaerts and Gijbels103 modelled the behaviour of Ar2+ and Ar2+ ions in a dc GD cell. The calculations accounted for the various production and loss processes as well as transport by diffusion and migration. The Ar2+ ions were shown to orginate almost exclusively from two-electron ionization of Ar0 and to disappear by diffusion and recombination at the cell walls. The Ar2+ ions were mainly created by associative ionization. Sputtered Cu atoms and ions in dc and rf GD were modelled by combining a collisional radiative model with Monte Carlo simulations for their thermalisation after sputtering and for the transport of Cu+ in the sheath.104 The predicted differences in calculated erosion rates and ionic emission spectra in dc and rf sources were in agreement with experimental observations. Modelling the effect of H2 addition to an Ar GD plasma on the atom and electron density distributions yielded an estimate for H2 dissociation of the order of 5 and 67% for typical conditions used in GDMS and GDOES, respectively.105 The authors warned that the simplifications made in modelling might result in considerable uncertainties in these values.Itoh et al.106 studied the relationship between ionization potential and the relative intensities of the singly- and doubly-charged metal ions, dimers and argides of 39 high-purity materials in dc GDMS. The intensity ratio M2+∶M+ decreased linearly with the second ionization potential and was 10–1000 times lower for disc samples than for pins. In contrast, the intensity ratio MAr+∶M+ increased with the first ionization potential and was higher for discs than for pins. Reaction kinetics could be used to explain why the M2+ intensities did not correlate with the first ionization potential.
Saka and Inoue107 investigated the dependence of relative sensitivity factors (RSFs) on sputtering yields in dc GDMS using pin and disc samples. The correlation between sputter yield data from the literature for 40 elements and the RSFs yielded a correlation coefficient between 0.7 and 0.85. Although the RSFs for pin and disc samples were different, the degree of correlation with sputter yield data was similar.
2.4 Analytical methodology
Eanes and Marcus108 developed an integrated EXCEL-based program for multiple peak processing in, for example, rf GD ion trap MS. Up to 3 peaks in a maximum of 225 spectra, containing 2000 data points, could be processed to obtain background-corrected peak areas by the trapezoidal method or by successive addition of the channel intensities across the peak. The simultaneous fitting of several peaks in a given spectrum did not affect the precision of the baseline corrected peak area. The program was particularly useful for determining the optimum number of data points to be used in peak area calculations when only a limited number of points were available within each peak.The quantitative determination of C and N in steel by dc GDMS was optimized by Itoh et al.109 using RSFs measured from commercial RMs (IARM 152A and 157A) and in-house alloy steels. Analysis of over 20 steel samples with certified C and N contents yielded precisions of 1.6 and 2.9% for C and N, respectively. The accuracy was within 5–8%, except for N when nitride inclusions were present.
Guzowski and Hieftje110 interfaced a TOFMS with orthogonal extraction to a gas-sampling dc GD source for fast switching between atomic and molecular mass spectra of halogenated hydrocarbons introduced by GC. The operating mode was defined by the potential on the sample-introduction plate. Long-term temporal stability of the atomic and molecular signals from a brominated ethane yielded RSD of 13.3 and 6.1%, respectively. The RSD of peak ratios in isotopic signal clusters was 1.1 and 1.6% for atomic and molecular ions, respectively, and essentially limited by the counting statistics. The LOD were 3–110 pg s−1 in the atomic mode and 15–250 pg s−1 in the molecular mode using an exponential-dilution sample-introduction device and boxcar averagers for data collection. Use of single channel gated ion counting improved the LOD to 1–25 fg s−1 in the static atomic mode.111 The source was also combined with ETV for analysis of aqueous solutions.112 The analysis of copper and zinc nitrate, sulfate or acetate required addition of ammonium chloride for efficient transport through the interfacing capillary. This had not been the case in the analysis of halides.
Of all the matrices commonly encountered in the field of elemental analysis, polymer samples are the most challenging. Marcus113 demonstrated the capability of rf GDMS to generate elemental and molecular information from polymer multilayer samples. The plasma negative glow appeared to be a relatively mild energy regime, which facilitated the speciation of analytes. Fragments with m/z up to 200 were more useful than the element ratios for distinguishing between the polymers. A major advantage over, for example, SIMS was the erosion rate of 0.1–1 µm min−1, which allowed depths up to 100 µm to be readily investigated.
3 Inductively coupled plasma mass spectrometry (ICP-MS)
3.1 Fundamentals and instrumentation
It is interesting to note that studies of ICP-MS characteristics undertaken in the formative years of the technique remain relevant. Diegor and Longerich114 studied signal optimization of a commercial ICP-MS instrument using concentric and Babington-type nebulizers under various values of applied rf power, sampling depth and nebulizer gas flow. The authors noted that, while the findings were similar to those of earlier studies of first-generation instruments, there were differences, especially the interesting observation of two signal maximums. They emphasized the multidimensional characteristic of optimization since the three important parameters interacted to a very significant degree. As a consequence, it is very important when characterizing an instrument set-up to be aware of the interactions of all ICP parameters.Feldmann et al. reported on some studies of a hexapole collision and reaction cell in ICP-MS.115,116 Helium was used as a buffer gas and H2 as a reaction gas. Addition of the latter was found to be an effective means of reducing, through gas-phase reactions, typical argon-containing polyatomic ions (e.g., Ar+, Ar2+, ArO+) by up to four orders of magnitude. Elements such as As, Ca, Cr, Fe, K and Se could be determined in CH2O, HCl and HNO3. Molecular interferences generated in the cell could be suppressed by using a retarding electric field established with a dc hexapole bias potential of −2 V. For the elements investigated, application of the buffer and reaction gases resulted in improved sensitivities, which were lowest for Be with about 7 × 107 counts s−1 (cps) per µg ml−1 and highest for Ba with about 6×108 counts s−1 per µg ml−1 with RSD of better than 0.1%. The LOD were 6 pg ml−1 for Cr in 2% CH2O, and 23 and 9 pg ml−1 for As and Se, respectively, in 0.28 M HCl. For other elements, LOD of <1 pg ml−1 were realized in the medium and high mass range. Accuracy was assessed using NIST SRM 1643d (Trace Elements in Water).
Tanner et al.117 continued studies of the performance of the dynamic reaction cell (DRC) for ICP-MS. The bandpass of the DRC could be adjusted in order to suppress the appearance of new interferences produced through sequential reactions within the reaction cell. Alternatively, the DRC could be operated at low pressure (vented under collision-free conditions) to emulate conventional ICP-MS. A practical optimization procedure for both the vented and pressurized modes was described. Examples were given to demonstrate the suppression of both plasma-based and cell-based isobaric interferences. The analytical performance characteristics of the instrument, including efficiency of isobar rejection, LOD in clean water and in neat hydrogen peroxide, short- and long-term stability, determination of As in chloride solution and Se-isotope determination, were given.
Sloth and Larsen118 described the application of an ICP-MS instrument equipped with a DRC to the measurement of Se, isotope ratios and chromatographic detection of selenoamino acids. The potentially interfering argon dimers at Se m/z 74, 76, 78 and 80 were reduced in intensity by approximately five orders of magnitude through the use of CH4 as reactive cell gas in the DRC. When using 3% v/v CH3OH in water for C-enhanced ionization of Se, the sensitivity of 80Se was 104 counts s−1 per ng ml−1 of Se, and the estimated LOD was 6 pg ml−1. The almost interference-free detection of Se by ICP-DRC-MS made the detection of the 80Se isotope possible for detection of selenoamino acids after separation by cation-exchange HPLC. The LOD for Se with the HPLC-ICP-DRC-MS method was 3–5 pg .
The characteristics and capabilities of an ICP-MS with a DRC for dry aerosols and LA was described by Hattendorf and Günther.119 The dependence of the signals from species formed in the ICP or in the interface region was determined by the variation of the concentration of a reaction or buffer gas used. The differences between wet aerosols, generated with a standard cyclonic spray chamber and concentric nebulizer, and dry aerosols, generated by a desolvating nebulizer or laser ablation, were determined. The comparison of prominent background signals to ion signals from selected analyte ions was used to determine parameters that led to optimum signal-to-background ratios and analytical performance for LA analysis. Ammonia and H2 were used as reactive gases in the experiments. Additionally, He, Ne and Xe were used as a buffer gas to enhance thermalization in the DRC. The reaction rate with NH3 was found to be distinctly higher than with H2. However, side reactions with analyte ions, leading to additional interferences and analyte loss through the formation of clusters, were severe with NH3. Hydrogen, having a lower reactivity than NH3, reduced cluster formation and retained analyte sensitivity even at a high gas concentration. Therefore, the authors felt that H2 was better suited for methods that allow only short measurement times, like LA or ETV, to be used. The capabilities of the DRC for LA were demonstrated through the determination of Ca in a quartz sample and Nb in a chromium matrix, both of which suffered from either Ar ions or Ar-based interferences. Reduction of the background intensities and use of the most abundant isotope led to a reduction of the LOD for Ca in quartz by two orders of magnitude and an improvement of accuracy for the determination of Nb in a chromium-matrix.
Praphairaksit and Houk120 reported on the reduction of mass bias and matrix effects in ICP-MS with a supplemental electron source in a negative extraction lens. Electrons from a heated tungsten filament were created inside the extraction lens and driven out toward the skimmer. These electrons were believed to move through the ion path and reduce space charge effects between positive ions in the beam. The ion transmission was improved by factors of between two (Pb+) and 27 (Li+). The greater sensitivity improvement for low-mass ions led to a substantial reduction in mass bias. With the additional electrons, MO+∶M+ and M2+∶M+ abundance ratios increased, but could be minimized with a small reduction in aerosol gas flow rate. No new background ions were observed with this technique. The authors reported that, when the electron source was operated under high electron current mode, matrix effects could be significantly diminished and the mass dependence of matrix-induced suppression of analyte signals was essentially eliminated. Using FI to minimize solids deposition, the technique could tolerate a sodium concentration of up to 1% with a loss of analyte sensitivity of about 15%.
Sturgeon et al.121 presented analytical data illustrating the typical response characteristics of a commercial ICP with an orthogonal acceleration TOFMS. With optimum instrument response tuned at 103Rh, the LOD for a suite of elements representative of 9–238 m/z were estimated to be typically 1 ppt. Background counts across the mass range averaged 0.5 Hz; sensitivity for Rh was 7 MHz per µg ml−1; resolution (FWHM) ranged from 500 (7Li) to 2200 (238U); long-term drift over 700 min was 0.7% h−1; abundance sensitivity was 2.8 × 10−6 (low mass side) and 7.4 × 10−5 (high mass side); and mass bias ranged from 10% per mass unit at 24Mg to <1% per mass unit at m/z >80. Isotope ratio precision was demonstrated to be limited by counting statistics when the detector was operated in the pulse counting mode. Analyte signal suppression from 3500 ppm NaCl was between 60 and 80%.
Helium-based plasma devices continue to attract attention. Su et al.122 studied an atmospheric pressure He-MIP ion source coupled with an orthogonal acceleration TOFMS for elemental analysis. Studies of the relationship between ion signals and sampling distance revealed that background signals could be suppressed dramatically without sacrificing the signal intensities of analytes when the MIP plume was off the tip of the sampler orifice. This “off-cone” ion-sampling mode eliminated the spectral interference from air entrainment and the plasma-gas species. This made it possible to determine sensitively isotopes that suffered from spectral interference in ICP-MS and MIP-MS, such as 40Ca, 52Cr, 56Fe and 55Mn. In addition, the off-cone sampling placed little demand on the cooling device, as the high-temperature plasma was kept away from the sampling aperture. As a consequence, the lifetime of the sampler plate was extended. The authors reported that the system could provide a full width half maximum (FWHM) mass resolution of 1100 and LOD of tens of pg ml−1 for the elements studied. It was believed that these could easily be improved with an advanced detection system.
Iacone et al.123 described the acquisition of high-speed video images of He-ICP at a rate of 1000 frames s−1 under conditions suitable for MS. The authors believed that the experiments established, for the first time, that the He-ICP was a rotating discharge at the forward powers (500–1000 W) used for He-ICP-MS. The rotational frequency of the He-ICP ranged from 80–250 Hz, depending on the forward power and whether the analyte was introduced. When not coupled to the MS, the rotational frequency of the He-ICP decreased slightly as the forward power was increased. Coupling of the ICP to the MS resulted in enhanced plasma oscillations. The implications of these variations on the sensitivity and precision of analytical measurements were discussed and results were contrasted with data obtained for an Ar-ICP.
Okino et al.124 made measurements of plasma properties in an interface for a He-ICP-MS using a Langmuir probe. To overcome problems involving the Ar-ICP-MS, a He-ICP-MS device was developed using an enhanced vortex-flow torch. A secondary discharge in the interface region was the main problem for practical application, so a new vessel for measuring the discharge was constructed. The electron temperature, Te, and the electron number density, ne, values at an rf power of 700 W were 2.7 eV and 3.7 × 105 cm−3, respectively, for a position 6 mm from the sampler orifice. The Te values were higher, but the ne values much lower, than those of an Ar-ICP-MS. The probe measurements also indicated that the Mach disc position in this device was about 8 mm from the sampler orifice.
O'Connor et al.125 identified and determined tetraethyllead in fuel using low-pressure ICP-MS. The ICP, sustained at 6 W, with 7 ml min−1 of He and 1.8 ml min−1 of isobutane, was optimized for the determination of tetraethyllead, with a LOD of 7 pg. A chromatogram of SRM NIST 1637 II yielded a single peak, which was identified as tetraethyllead by comparison with an electron ionization spectra library. Quantitative analysis of the fuel yielded a tetraethyllead concentration of 13.1 ± 0.9 µg ml−1 as Pb, which was within the certified range of 12.9 ± 0.07 µg ml−1.
Milstein et al.126 reported quantitative and qualitative MS results for organotin compounds introduced into a mixed gas He/Ar low-power/reduced-pressure (LP/RP) ICP. The LOD for organotin compounds introduced at the pressure producing the optimum analyte signal were in the low pg range. The source demonstrated “tuneable” fragmentation with variation in the He/Ar mixed gas ratio. In addition, tuneable fragmentation with pressure variations similar to that previously reported with He-rf-GD was observed with the LP/RP-He-ICP.
Appelblad and Baxter127 described a model for simultaneous determination of detector dead time and mass discrimination factors in ICP-MS, as well as corresponding uncertainties. The model was tested on three representative isotope systems, In, Mg and Tl. This allowed the sampling time to be spent entirely on the isotopes of interest.
Appelblad et al.128 studied the performance characteristics of a magnetic sector ICP-MS instrument fitted with a platinum shield (Pt-shield) torch. Specifically, the importance of the optimization procedure and matrix effects caused by a sea-water matrix were assessed for 20 elements together with oxide and doubly charged ion formation. Use of a grounded Pt-shield allowed ion transmission to be increased, by a factor of up to 20, and resulted in improved instrumental LOD. The improvement in sensitivity was mass dependent, with the highest gain observed for elements at lower m/z. Severe spectral interferences from oxide species, observed when operating the Pt-shield grounded rather than in the floating mode, were often unresolved even in high-resolution mode. For example, the BaO+∶Ba+ ratio was ten to twelve times higher in the grounded mode. In addition, it was reported that non-spectral interferences from the sea-water matrix appeared to be more pronounced with the grounded Pt-shield, yielding a recovery of Ni of 55% compared with 93% in the floating Pt-shield mode. The authors concluded that all possible advantages and limitations of the use of the Pt-shield should be carefully considered prior to the analysis of real samples.
Stewart et al.129 studied the effect of gas flow entrainment on the gas sampling, ion sampling and ion detection processes in ICP-MS. Isolated, single droplets of sample from a monodisperse dried microparticulate injector (MDMI) were used in conjunction with time-resolved ICP-MS, photographs of ion cloud movement and time-gated imaging using a gateable, intensified charge-coupled device (ICCD) detector mounted on an imaging spectrometer. It was believed that gas flow entrainment into the sampling orifice could have a significant effect on the plasma gas velocities as far as 7 mm from the sampling orifice. However, the effects were found to be most pronounced within 3 mm of the sampling orifice. The trends in these results were consistent with theoretical calculations. Photographic images showed that plasma gas, initially as far as 3 mm off axis, adopted a curved path into the sampling orifice. Time-resolved emission images of Sr+ ion clouds approaching the sampling orifice demonstrated significant distortion of the ion cloud as it flowed into the sampling orifice. The widths of radially resolved La+ ICP-MS signal peaks were not significantly different whether sampled at 2 or 5 mm from the vaporization point. In contrast, ICP-MS signals measured on axis as a function of time clearly showed broadening due to diffusion. Probably some detected ions had originated from off-axis locations in the plasma.
A cool plasma ICP-MS system with a direct injection high efficiency nebulizer (DIHEN) was investigated by Minnich and Montaser.130 Taking a novel approach, a group of elements was investigated that was suitable for analysis in the cool plasma only by measurement of the metal oxide ion, because the sensitivity and precision of the atomic ion were inferior to those obtained for the metal oxide ion. The sensitivity and precision obtained for these molecular species were comparable to those obtained for the atomic ions of elements that were suitable for analysis using the cool plasma. Calibration curves for the metal oxide signal as a function of the metal concentration in the cool plasma were linear over six orders of magnitude, covering the concentration range from 1 ppt to 1 ppm.
Praphairaksit et al.131 developed and evaluated a low-flow torch specifically designed for ICP-MS. The outside of the torch was cooled by pressurized air flowing at about 70 l min−1 through a tube sealed on to the usual ‘outer’ tube of a standard mini-torch, so that cooling air could not enter the plasma. Although plasmas could be sustained at operating powers as low as 400 W with a 2 l min−1 outer Ar flow, stable and analytically useful plasmas could only be obtained at 850 W with an outer Ar flow rate of 4 l min−1. Under optimum operating conditions, the externally air-cooled plasma produced sensitivities, M2+∶M+ ratios, matrix effects and other analytical figures of merit comparable to those produced by a conventional torch whilst using much less argon, but MO+∶M+ ratios were slightly higher with the externally cooled torch.
The furnace atomization plasma ionization mass spectrometer (FAPIMS) of Stewart et al.132 was designed to overcome some of the limitations of ET-ICP-MS by performing the atomization and ionization within the same chamber. A small sample volume (2–10 µl) was injected through a dosing port into a transversely heated, integrated-contact-cuvette graphite tube furnace. The sample was vaporized and atomized directly into a capacitively coupled helium plasma formed between an axial graphite rod electrode and the grounded wall of the furnace. Although the distance between sampler and furnace should preferably be as small as possible, a minimum separation of 8 mm was required to avoid secondary discharges. In contrast to most grounded plasma MS sampling interfaces, a potential was applied to both sampler and skimmer cones to improve transmission. The sampling configuration used provided a precision of about 5%, a linear dynamic range of 3–4 orders of magnitude and absolute LOD of 20–500 pg for Cd, Co, Cs, Cu, Fe, Mg, Pb and Se. Although the FAPIMS technique gave sensitive analysis, further work was required to define the limitations on total sample and matrix load imposed by the low power rf plasma. Stewart and Sturgeon133 found that introduction of as little as 2–10% argon into the pure helium plasma gas increased by up to 10-fold the signal intensities of elements with first ionization potentials greater than 6 eV. Higher Ar contents led to plasma instability and accelerated erosion of graphite surfaces. A disadvantage of the mixed gas plasma was the appearance of Ar-based spectral interferences, which degraded the determination of species such as 52Cr+, 56Fe+ and 80Se+.
3.2 Sample introduction
Several workers have considered the issues of calibration in LA-ICP-MS. Bi et al.137 developed a method for the determination of trace element concentrationsin solidsusing solutions and an internal standard calibration. The method was evaluated using soil and glass RMs. For soil analysis, Mg was successfully used for internal correction (except for Cu and Ni) and Sr for glass samples. The heterogeneous distribution of Ni in the soil was suggested as a reason for its poor results.
Bellotto and Miekeley138 improved repeatability and accuracy in the quantitative determination of trace elements in mussel shells or carbonate-based materials by using a series of multi-element calibration standards. These were prepared by co-precipitation of twelve elements into a CaCO3 matrix, which improved the homogeneity of the resulting powder samples. Pressed powder discs of good mechanical stability were prepared at a pressure of 50 MPa, without the addition of a binder. A Nd:YAG (266 nm) laser was used in the Q-switched mode at a repetition rate of 10 Hz and an energy level of 3.5 mJ. Correlation coefficients for the linear calibration graphs over a concentration range of 1.5–400 µg g−1 for As, Ba, Cd, Co, Cr, Cu, Mn, Pb, Sn and Zn were generally better than 0.997. The LOD for all elements studied were in the sub-µg g−1 range.
Becker et al.139 addressed the problem of quantification in the determination of long-lived radionuclides by LA-ICP-MS, for which no suitable standard RMs were available. Their approach was to make synthetic standards by doping a graphite or concrete matrix with long-lived radionuclides, such as 237Np, 99Tc, 232Th, 233U, 235U or 238U. Two calibration procedures were considered: the correction of analytical results with experimentally determined relative sensitivity coefficients; and the use of calibration curves and solution calibration by coupling LA-ICP-MS to an ultrasonic nebulizer. The LOD of long-lived radionuclides investigated in a concrete matrix were determined in the pg g−1 range, for example 50 pg g−1 for 237Np by quadrupole LA-ICP-MS and 1.3 pg g−1 for 233U
by double-focusing magnetic sector LA-ICP-MS. Furthermore, LA-ICP-MS allowed precise and accurate isotope ratio measurements of Th and U in solid samples. For example, the 234U∶238U ratio of 0.000067 in radioactive reactor graphite was determined with a RSD of 1.1%.
Calibration procedures for the analysis of natural calcium carbonate-rich materials were evaluated by Craig et al.140 Calibration lines were constructed using: (a) commercially available glass reference materials; (b) high-purity calcium carbonate powder spiked with the analytes of interest; and (c) natural geological RMs. The analytes initially considered were Ba, Cd, Fe, Mg, Mn, Pb, Sr, U and V, although the low concentrations, or lack of data for some elements, particularly for Cd and U, precluded a complete assessment. Data were corrected using Ca as internal standard. The analysis of geological RMs using glass calibration lines showed a systematic offset in measured versus reference, or accepted, values. The authors concluded that the preferred calibration strategy would be to use selected natural geological RMs.
Borisov and co-workers141,142 investigated the effects of crater development on elemental fractionation and signal intensity by measuring Pb∶U ratios in NIST SRM 610 (Synthetic Glass). A comparison was made between two Nd:YAG lasers with wavelengths of 213 and 266 nm. Fractionation, dependent on the irradiance that should be >0.6 GW cm−2, was similar for both wavelengths. They also observed that if the laser beam was initially focused close to the sample surface fractionation increased and was influenced by the formation of a crater during repetitive pulsing at a single sample location. As the ratio of crater depth to radius increased, plasma sampling and/or an effective irradiance decrease could cause additional fractionation. A good correlation was found between the fractionation of 14 elements in NIST SRM 610 and the logarithms of their oxide melting temperatures. In a similar study,143 an investigation was made of fractionation during ablation of calcite and NIST glass SRMs using three UV wavelengths (157, 213 and 266 nm). Fractionation was observed for all wavelengths, depending in each case on the laser-beam irradiance and the number of laser pulses at each sample-surface location. The transparency of the sample was found to influence the amount of sample ablated at each wavelength and the extent of fractionation. The Pb∶Ca and Pb∶U ratios were again used as examples to demonstrate the degree of fractionation at the different wavelengths.
Jeong et al.144 investigated the relationship between laser-generated particles and ICP-MS instrument signal intensity using single-pulse LA sampling of solids. The particle size distribution of glass samples was measured using an optical particle counter for different LA conditions. Ablation of a new surface produced fewer particles and a lower signal intensity than a pre-ablated surface. The laser power density of 0.4–0.5 GW cm−2 was found to be a threshold value, across which particle size distribution changed. Also, laser beam diameter was a more influential parameter than power density in efficient particle generation. Particle loss during transport from the ablation chamber to the ICP-MS was found to be significant for a low carrier gas flow rate of 0.1 l min−1, while almost no loss was observed for a higher flow rate of 0.26 l min−1. Particle entrainment efficiency of the LA-ICP MS system was estimated and found to be a strong function of laser power density.
In a similar study but using TOF-MS, Leach and Hieftje145 investigated two data analysis techniques to improve single-shot measurement precision. Both techniques exploited the simultaneous, full-spectrum acquisition capability of a TOFMS instrument. In the first approach, a normalization factor was computed from the total ion current, which should be proportional to the total mass of sample ablated. This scheme resulted in an improvement in precision of greater than a factor of two, only moderately better than possible with a single internal standard. The enhancement in measurement precision was found to be concentration dependent, with the greatest improvement (10–50-fold) experienced by elements at high concentration. The second method correlated the attenuation of plasma matrix ions to analyte intensities. This technique demonstrated no statistically significant improvement in precision, limited by the relatively low S/N of the attenuated signals.
Bleiner et al.146 reported on the optimization of an excimer LA-ICP-TOFMS system for transient-signal acquisition. Various parameters were optimized for laser-generated aerosols. A reduced-volume ablation cell was designed and used in order to increase the sample density in the ICP. Results for 63 simultaneously measured isotopes in NIST SRM 610 (Trace Elements in Glass) led to LOD in the 1–100 µg g−1 range for a 80 µm diameter crater (10 Hz, 1.2 mJ pulse energy). The reproducibility of signal ratios was better than 2% RSD for transient signals using 102 ms integration time.
Bleiner et al.147 used the same LA-ICP-TOFMS system for depth profile analysis of various Ti-based coatings on steel and tungsten carbide. Laser parameters, such as repetition rate, pulse energy and spatial resolution were tested to allow optimum depth-related calibration curves. The best depth resolution obtained was 0.20 µm per laser shot. Variation of the volume of the ablation cell did not influence the depth resolution. The application to TiN- and TiC-based single layers showed the potential of LA-ICP-TOFMS as a complementary technique for fast depth-determinations on various coatings in the low to medium µm region.
Hoffmann et al.148 reported the analysis of green leaves of oak trees (Quercus robur)
by LA-ICP-MS in order to determine the spatial distribution of Al, Ba, Ca, Cu, Mg, Mn, Ni, Pb and Sr. Instrument operating parameters such as the laser wavelength and the pulse energy were optimized for sensitivity and reproducibility. Their method provided a lateral resolution down to 300 µm with the use of a 355 nm wavelength Nd:YAG laser (frequency trebled from 1064 nm) and a pulse energy of 50 mJ. Plant standards and cellulose, doped with multi-element solution standards, dried and pressed to pellets, were used as calibration samples. To compensate for signal fluctuations caused by the variation of the ablated sample mass, 13C was used as a “natural” internal standard. The accuracy of the calibration was verified with selected samples digested in HNO3
under high pressure (107 Pa at 170°C for 2 h) and then analysed by solution ICP-MS. Recoveries were between 93%
(Cu) and 108%
(Mn).
Ghazi et al.149 used LA with a high resolution ICP-MS system for in situdetermination of Ga concentration in treated human teeth. Gallium salts have an inhibitory effect on osteoclastic activity. The authors reported that Ga was proposed as a treatment for external root resorption. Their study examined the diffusion of Ga introduced as the nitrate, through root dentin from the prepared root canal toward the root surface. Root canals from extracted human teeth were bio-mechanically prepared, sealed apically, and closed off coronally with polyethylene tubing. A 1.0 M solution of gallium nitrate, chelated by 1.0 M citrate buffer (pH 7.4), was introduced into canals. Roots were then suspended in test-tubes of distilled water. The Ga-treated teeth were analyzed for 43Ca, 69Ga, and 71Ga. The NIST SRM 612 (50 ppm) was used for calibration and 43Ca used for internal correction. The precision for the tooth samples were between 15.4 and 19.1% RSD. The analysis revealed a significant uptake of Ga by the root dentin, ranging from about 2% for the upper portion of the root to about 0.8% for the areas in the lower sections. The analysis also revealed that a significant concentration gradient was established for gallium nitrate across the root dentin. The authors concluded that the results from this study suggest that, in an effort to inhibit root resorption, therapeutic concentrations of Ga could be achieved throughout the root dentin when the prepared root canal is treated with 1.0 M gallium nitrate.
Snell et al.152 investigated the thermal vaporization behaviour of various mercury compounds in organic solutions. A number of sulfur, palladium, and gold solutions were tested as modifiers for vaporization of mercury diethyldithiocarbamate. A propanolic solution of Pd(NO3)2 was found to give the best results when used in combination with thionyl chloride. Under these conditions, successive injections of 5 µg l−1 HgCl2 were reproducible to within 3% RSD and the LOD for Hg present as the diethyldithiocarbamate was 0.19 µg l−1. Whereas recoveries for HgCl2 and CH3HgCl were 100 and 115%, respectively, those for mercury metal and (CH3)2Hg were only in the range 48–65%. Further investigation showed that these losses, which could not be corrected by ID, took place during the drying portion of the ETV programme.
Wei et al.153 studied the effect of eight polyhydroxy compounds upon the thermal vaporization of water samples containing As, B, Cr, Mo, Sb, and Se. Additions of sorbitol and mannitol were found to enhance signals for As, B, Sb and Se by a factor of between 2.5 and 84. Under optimized furnace conditions, the LOD for these four elements were between 1 and 100 ng l−1, which was at least a factor of ten improvement compared to vaporization without modification. This enhancement was attributed to the formation of volatile polyhydroxy complexes of the analyte. On the other hand, elements that are typically present in natural water samples (Ca, K, Mg and Na,) were found to suppress sensitivity. This was attributed to significant matrix-induced changes in the volatility of the analyte complexes.
Moor et al.155 described a novel system for the determination of Se in biological media. Acidified sample digests were mixed with the NaBH4 reagent at the tip of a cross-flow nebulizer. The reaction time was 60 ms. A modified Scott spray chamber was used as the gas–liquid separator, yielding wash-in and wash-out times of 30 s. Analyses of the DORT-2 and DORM-2 RMs by external calibration and ID were found to agree well with the certified concentrations. The LOD for Se and other hydride-forming elements (As, Sb and Sn) were less than 10 ng l−1. In related work, Mester et al.156 used solid phase microextraction (SPME) and thermal desorption to determine hydrides of As, Sb, Se and Sn. A heated, glass-lined, splitless GC injector was placed at the base of the torch and used as the thermal desorption interface. Polydimethylsiloxane–carboxen showed best overall extraction capacity and enhanced selectivity for Sn hydrides. Instrument response was linear over 3.5 decades and LOD for As, Sb Se and Sn were 70, 310, 5300 and 8 ng l−1, respectively. Concentrations of As determined for the SLRS-4 and CASS-4 RMs agreed well with certified values. The technique was also successfully used for the determination of methylmercury.157
Li et al.158 used a mixture of oxalic acid and ammonium CeIV nitrate for the determination of Pb by HG. With continuous sample introduction, the sensitivity when using the hydride generator was about 7.5 times higher than when using ultrasonic nebulization. The LOD was 7 ng l−1 and precision was 0.21% RSD (n = 13) at a concentration of 1 µg l−1 Pb. No memory effects were observed when the generator was operated in combination with a FI manifold. Peak height and peak area from 13 successive injections were measured with precisions of 0.33 and 1.4% RSD, respectively. The presence of 10 mg l−1 FeIII or NiII did not interfere with the HG process, and determinations of Pb in several environmental RMs agreed closely with certified values.
Feng et al.159 used HG in combination with high resolution ICP-MS for the speciation of Sb. Addition of L-cysteine was used to reduce pentavalent Sb to the trivalent oxidation state. The LOD were 4.2 and 17 ng l−1 for SbIII and SbV, respectively. The precision was about 10% for both species. Recovery of the two species from river water samples was between 91 and 113%.
Several interesting developments in the field of aerosol processing were also noted. Olofsson et al.162 described a tandem spray-chamber assembly used in combination with a sector field ICP-MS instrument and a low-flow Micromist nebulizer. The assembly consisted of a cyclonic spray chamber, heated by an infrared lamp, placed upstream of a water-cooled double-pass spray chamber, that served both as a cooling stage and particle size filter. With this arrangement, sensitivity was increased 3-fold and MO+∶M+ ratios were decreased in comparison with those obtained with a conventional double-pass spray chamber only arrangement. However, matrix-induced signal suppression was greater with the tandem chamber arrangement, particularly for elements with high ionization potentials. Elsewhere, Sung and Lim163 developed a microconcentric nebulizer with double-membrane desolvation for the direct analysis of isopropyl alcohol. The desolvator consisted of two concentric membranes, each having a length half that of typical devices. Optimization of the sweep gas rate reduced the carbon loading by between 70 and 90%, about twice the accompanying reduction in analyte signal (Ca, Cr and Mn). The CeO+∶Ce+ ratio was found to be 0.16 at 210 µl min−1 uptake, but was reduced 3-fold when the uptake rate was halved.
Memory effects continue to be of concern in many laboratories. Hirata164 developed a flushing spray chamber of single-pass design with an internal volume of about 50 ml and an integrated continuous-flushing system. At a 1 mg l−1 metal concentration, 100- and 4-fold reductions in Hg and U washout times, respectively, were achieved relative to the standard double-pass spray chamber design. A similar reduction was observed using 1 µg l−1 Os solutions but at a 1 mg l−1 Os concentration washout times could not be defined. When chilled water (277 K) was used in the flushing system, instrumental figures of merit (sensitivity, stability and oxide ion ratios) were identical to those obtained from a conventional nebulizer–spray chamber arrangement.
Studies of trace element speciation by FI continued during this review period. Hirata et al.168 developed an FI system for the determination of trivalent and total Cr in seawater. Trivalent Cr was concentrated on Muromac A-1 (an iminodiacetate resin) at pH 3.0, and, after column washing with water, the analyte was eluted with 0.7 M HNO3. Total Cr was determined by the same procedure after hexavalent Cr had been reduced with a solution containing 2 mM hydroxylamine at pH 1.8. The LOD of this system was 20 ng l−1 Cr and the sample processing time 5 min. Yan and co-workers169,170 determined divalent and trivalent Fe in aqueous media using a PTFE knotted reactor FI system. Trivalent Fe was selectively chelated with pyrrolidinecarbodithioate (PDC) at HCl concentrations ranging from 0.07 to 0.4 M and the complex was absorbed by the knotted reactor. At lower acidity (1–4 mM HCl), di- and trivalent Fe were both chelated by PDC, allowing total Fe to be determined. At an optimized flow rate and retention time, Fe was concentrated 12-fold with a retention efficiency of 80%. The LOD was 80 ng l−1 Fe, throughput was 21 samples h−1 and the RSD (n = 11) for a solution containing 10 µg l−1 FeIII was 2.9%. Recoveries of di- and trivalent Fe spikes in natural water samples were between 95 and 103%.
A number of workers have investigated FI coupled to ICP-TOFMS for the determination of trace elements. Benkhedda et al.171 used a knotted reactor FI system with PDC chelation to determine 9 trace elements in biological and aqueous media. An ICP-axial-TOFMS instrument was used for analysis. A small volume of methanol stripped the analytes from the PDC and the eluate was introduced into the instrument using ultrasonic nebulization with membrane desolvation. At an optimized flow rate and retention time, the analytes were concentrated 5- to 70-fold. The LOD were between 0.5 ng l−1 (Sb) and 26 ng l−1 (Pb). The RSDs (n = 11) were <5% for analytes at a concentration of 0.2 µg l−1. Centineo et al.172 coupled FI to ICP-TOFMS for the simultaneous determination of seven elements in urine by HG. Compromise conditions for HG were established but carrier gas flow was optimized to give maximum vapour transport. For the analytes As, Bi, Ge, Hg, Sb, Se and Sn, LOD were in the low ng l−1 range and instrument response linear over 5 decades. Analyte recovery was better than 80% for all analytes. However, the accuracy of the Se determination could be compromised by the presence of Se species other than SeIV.
In last year's Update, we noted for the first time the widespread use of alternative plasmas and mass analyzers in combination with separation methods. Costa-Fernández et al.173 described improvements in a CE-ICP-TOFMS instrument for the determination of various elemental species. Three As species and two cobalt–cyanide complexes were fully separated in less than 70 s, despite the presence of Cr, Cu, Ni and V cyanide complexes. Typical CE peak widths were in the 1–3 s range. Each integrated mass spectrum represented the average of 5000 individual spectra, each of which was acquired at a rate of 20 kHz. Absolute LOD were between 1 and 20 pg and, for 10 successive sample injections, peak area and elution times were reproducible to within 4% and 1% RSD, respectively.
Chatterjee and co-workers continued their development of a high-power nitrogen MIP-MS system with HG174 and ultrasonic nebulization (USN)175 for the HPLC determination of Se and As species, respectively. For Se, the eluate from an anion-exchange column was mixed with an alkaline sodium borohydride solution in a continuous flow cell prior to analysis. Under optimized conditions, the LOD for selenous acid and selenomethionine were 0.73 and 8.7 µg l−1, respectively, and the reproducibility of three successive injections was better than 6% RSD. For As, the use of an USN enhanced signal strength 1.5- to 3-fold relative to use of a concentric nebulizer. The LOD using USN were between 0.2 (for arsenobetaine) and 4 µg l−1 (for trimethylarsine oxide). Within- and between-day precisions for replicate sample injections were less than 8% RSD, but some data were reproducible to within 1% RSD.
Several laboratories have focused their attention upon the interface between the mass spectrometer and the separation device. Goenaga Infante et al.176 described a HG interface for the determination of Cd species in fish cytosols by vesicle-mediated HPLC. The HG interface provided a 5-fold enhancement in sensitivity relative to traditional nebulization but the LOD were only improved 3-fold due to blank limitations and the poor signal-to-noise characteristics of the HG process. Tu and Qvarnstrom177 evaluated two different spray chamber–nebulizer combinations for the determination of Hg species by capillary zone electrophoresis (CZE). Under optimal CZE conditions, a cross-flow-nebulizer, double-pass spray chamber assembly gave best resolution and analysis times but the LOD could be improved 10- to 20-fold using a microconcentric nebulizer, cyclonic spray chamber assembly. Using the microconcentric nebulizer, absolute LOD were 2.3 pg for methylmercury and 1 pg for HgII.
Other investigations have focused upon refinement of the LC-ICP-MS interface. Gammelgaard and Jons178 compared USN and cross flow nebulization interfaces for the determination of four Se species by anion exchange LC. When the cross flow nebulizer was used, sensitivity for all 4 species was identical and all of the species responded similarly to changes in instrument operating conditions. Sensitivity was increased 14- to 21-fold when USN was used but all 4 species behaved differently with changes in operating conditions. The increase in sensitivity was greatest for selenate and trimethylselenonium ion. Chromatographic peak width was independent of the nebulizer used, but the LOD (approximately 0.2 µg l−1 Se) were somewhat better with USN. Ackley et al.179 compared the performance of concentric, microconcentric and Micro Mist nebulization interfaces for the analysis of methanolic eluates from a microbore LC system. The analyte signal was independent of sample flow rate when the eluate contained 20% methanol or more. At 70% methanol and 0.2 ml min−1 flow rate, the LOD for Sn were best using the Micro Mist and concentric nebulizers. Marchante-Gayón et al.180 compared three nebulizers and two separation methods for the determination of four Se species in commercial nutritional supplements. A hydraulic high-pressure nebulizer (HHPN) gave the best LOD (35 to 90 ng l−1). Although the LOD for concentric and microconcentric nebulization were about the same, they were somewhat poorer than that of the HHPN. Of the separation methods investigated, ion-pairing LC gave the best separation and LOD.
The effect of the interface upon the chemical separation is often overlooked. In an interesting series of publications, Day and co-workers181,182 made a comparison of indirect UV absorbance and ICP-MS with microconcentric nebulization for detection of species separated by CE. The migration times of low mass species (e.g., As compounds, REE cations) and the reproducibilities of migration time and peak area were relatively independent of the detection method. However, for metalloporphyrins (e.g., Vitamin B12), migration times using ICP-MS detection were several minutes less than those observed using UV detection. The difference was attributed to a “suction effect” characteristic of the self-aspirating nebulizer. In a subsequent study,183 it was found that a sol–gel frit placed at the downstream end of the CE device had a pronounced effect upon migration time and peak area. The authors noted that peak area was amplified by an order of magnitude with the frit in place.
Another frequently overlooked aspect of “hyphenated” ICP-MS is the effect of spectral interferences upon data integrity. Often, such interferences are ameliorated by the chemical separation on the “front end” of the mass spectrometer. However, Vanhaecke et al.184 described circumstances under which high spectral resolution and “cool plasma” conditions were needed to reduce spectral interferences encountered during the separation of Cr species from industrial process solutions. Whereas the interference of ArC+ upon Cr+ (due to co-elution of CrVI and bicarbonate during anion exchange) could be reduced using cool plasma conditions, that of 37ClO+ and 35ClOH+ upon Cr+ (due to co-elution of CrIII and Cl−) could not. Increased spectral resolution (R = 3000) reduced both types of spectral interference, but sensitivity was reduced 15-fold. Nonetheless, LOD of 50 and 120 ng l−1 for CrIII and CrVI, respectively, were attained using both interference reduction methods. It was concluded that high spectral resolution was the preferred approach for this specific application.
3.3 Interferences
Tromp et al.185 studied matrix interference diagnostics for the automation of ICP-MS in order to extend their total interference level (TIL) concept from ICP-AES. The TIL model was based on measurements with different interferents to determine a set of interference coefficients. Interference was assumed to be linear with interferent concentration, as that assumption allowed them to determine the model parameters with the fewest experiments. Although the TIL model was designed to indicate when a simple external standards calibration method was inadequate for a desired level of analytical accuracy, it was also tested on a simple form of internal standards. The TIL concept was used in both an initial calibration and a daily calibration mode for the interferents Al, Ba, Cs, K and Na and was found to work well in situations where both external and internal standards were used.A spreadsheet calculation procedure was set up by de Boer186 to calculate the correction factors that were part of elemental equations used to correct for polyatomic and doubly charged ion interferences in the ICP-MS analysis of complex samples. The factors were calculated after the analysis of an interference check solution. For typical environmental and biomedical applications, a set of 5–6 check solutions was needed in order to identify the origin of the interference. The calculation procedure compared the result for the quantitative isotope with the result for a control isotope (when feasible). The procedure proved to be robust when there was some analyte contamination present in the check solution.
Hu et al.187 studied the calibration of matrix effects and oxide ions for REEs in geochemical samples. Matrix effects were minimized by using synthetic calibration solutions that simulated the composition of natural rocks. Two separate internal standards of 115In and 103Rh were selected to compensate for drift of analytical signals. Interfering ions were calibrated by measuring the MO+∶M+ and MOH+∶M+ ratios with individual Ba, Ce, La and Pr solutions and calculating equivalent concentrations. The proposed method was applied to the analysis of the geological RMs AGV-1, BHVO-2, DNC-1, GSR-3 and RGM-1. The relative error between values obtained in this work and the recommended ones was <10% and the RSD <5%.
Gomez et al.188 used quadrupole ICP-MS with mathematical correction of interferences for the direct determination of the platinum group elements (PGE) at ultratrace levels in contaminated airborne samples. The interference contributions of HfO+ on 195Pt, of Pb2+, ArCu+, SrO+ and RbO+ on 103Rh and of ArCu+ ArGa+ and YO+ on 105Pd were studied in order to find the best analyte/interferent signal ratios. The influence of instrumental parameters such as nebulizer gas flow rate and plasma power was evaluated. The results showed that the Ar gas flow rate could not be used for interference alleviation. In general, a medium plasma power of 1350 W yielded the best ratio although Pb2+ formation was minimized at 1150 W. The LOD obtained were 3.3, 0.6 and 0.9 ng 1−1 for Pd, Pt and Rh, respectively, and the RSD (at 50 ng l−1, n = 5) was <3%. The determination of Pt could be made without major errors, with a HfO+ contribution of about 10%. The determination of Pd was hampered by the conjugate factors of very low Pd concentration in the samples and a matrix with a high content of Cu and Y whose interference contribution was enormous. The contribution of interferences to the Rh content in airborne particles was in the range of 50–75% and yet, after mathematical correction, the Rh content could be evaluated.
Becker and Dietze189 investigated the oxide ion formation of long-lived radionuclides (241Am, 237Np, 239Pu, 226Ra, 230Th and 239U) in a mixed aqueous solution using double-focusing sector field ICP-MS (DF-ICP-MS) with a shielded torch under hot plasma conditions. The measurements of the relative oxide ion intensities by DF-ICP-MS were performed using concentric, microconcentric and direct injection high-efficiency (DIHEN) nebulizers for sample introduction. The highest oxide ion formation rate was observed for Th with ThO+∶Th+ ratios of 0.13, 0.26, and 0.41 for concentric, microconcentric and DIHEN, respectively. In comparison with the configuration without plasma shielding, the application of the shielded torch in DF-ICP-MS yielded an increase in oxide ion formation for the concentric nebulizer by a factor of between 2.7 and 7, for the microconcentric nebulizer of between 1.6 and 13 and for the DIHEN of between 1.7 and 2.8. The direct injection of sample solution by DIHEN-DF-ICP-MS results in, for example, PuO+∶Pu+ ratios of about 0.16 and 0.35 without and with a shielded torch, respectively. Relative oxide ion intensities, using the concentric and the microconcentric nebulizers with a shielded torch DF-ICP-MS, decreased in the order, UO+∶U+ > NpO+∶Np+ > PuO+∶Pu+ > AmO+∶Am+ ≫ RaO+∶Ra+. The observed correlation of measured relative oxide ion intensities for ThO+∶Th+, PuO+∶Pu+ and UO+∶U+ with bond energies of these oxides allowed the estimation of unknown bond energies for AmO and NpO (730 and 670 kJ mol−1, respectively) to be made. The authors suggested that the presence of high oxide ion intensity of long-lived radionuclides in DF-ICP-MS could be used for analytical purposes by using the oxide ions as ‘analyte’ ions for interference-free isotope analysis.
Al-Ammar et al.190 investigated the mechanism by which carbon matrix species could cause non-spectroscopic matrix interference on B and Be during the determination of B in biological samples. The study indicated that C species manifest non-spectroscopic interference by two mechanisms. The authors believed that the major effect was by a charge transfer mechanism from C species to B and Be atoms in the central channel of the plasma and that the minor non-spectroscopic interference of C was through a space charge effect. The authors concluded that the large difference in the magnitude of the C charge transfer, non-spectroscopic matrix interference between Be and B, made Be unsuitable as an internal reference for B in solutions with more than 1500 µg ml−1 dissolved organic carbon (DOC). This DOC level was reported as approximately half that usually present in the final sample solution for B determination in biological samples. The authors pointed out that Be could still act as a suitable internal reference for B.
Tangen and Lund191 used partial least squares regression to study the effects of HNO3 concentrations and liquid flow rate on signals obtained with an ICP-MS system operated with a microconcentric nebulizer. Thirteen elements, covering the mass range 51–238 m/z were studied. The HNO3 concentration varied in the range 0.14–2.8 mol l−1 and its effect was found to vary significantly between the elements and to depend mainly on the first ionization potential of the element. The effect of the liquid flow rate was studied over the range 34.5–172.5 µl min−1. It was found to be significant (P = 0.05) only for Ag. It was possible to compensate completely for the acid matrix effect by the use of an appropriate internal standard.
Continuing the work reviewed in last year's Update, Al-Ammar et al.192 reported that the injection of 10–20 ml min−1 of NH3 gas into an ICP-MS instrument spray chamber eliminated the memory effect of B within a 2 min washing time for ppm levels of B in solution. The NH3 gas injection also reduced the B blank by a factor of four and enhanced the sensitivity by up to 90%. The LOD for B were improved from 12 and 14 to 3 and 4 ng ml−1, respectively, for the two ICP MS instruments studied. Using NH3 gas injection, trace B concentrations determined in RMs agreed well with certified values.
Du and Houk193 described the attenuation of metal oxide ions with a mixture of He and H2 gas in a hexapole collision cell. Metal oxide ions with different dissociation energies, such as CeO+, HoO+, LaO+, SmO+, WO+, YbO+ and ZrO+, were studied. By adjusting the collision conditions, especially the composition and flow rate of the collision gas, and the hexapole dc bias voltage, the MO+∶M+ ratio could be suppressed by a factor of up to 60 for CeO+ and LaO+ while maintaining about 20% of the original signal for analyte species. For the oxide ions studied, collisions with H2 improved the MO+∶M+ ratio more effectively for those ions with higher dissociation energies. The same collision conditions also served to remove the ArO+, ArN+ and Ar2+ from the background spectrum.
Koyanagi et al.194,195 described the construction of an ICP-selected-ion flow tube (ICP-SIFT) MS. This was used for the study of the kinetics and product distributions of reactions of ICP ions with neutral reagents. The operation of the combined ICP-SIFT instrument was illustrated for the determination of reaction rates and product distributions using the reactions of the isobaric pairs ArO+/Fe+, Ar2+/Se+ and ClO+/V+ with a range of neutral modifiers. These measurements led to the formulation of strategies for the use of ion–molecule reactions to remove isobaric interferences (termed chemical resolution). These strategies were considered useful in the selection of reagent gases used in ICP-MS reaction cells for improving LOD.
3.4 Isotope ratio measurement
Magnetic sector ICP-MS instruments continue to gain popularity, particularly for isotope ratio measurements. For example, Hassler et al.196 described a method for the rapid determination of both Os isotopic composition and PGE concentrations using magnetic sector ICP-MS. The method used the transfer of volatile OsO4 by an Ar gas stream into the instrument. Four sample dissolution methods were studied: microwave oven digestion, NiS fire assay, acid leaching and Carius tube digestion. The elimination of a nebulizer for Os isotope ratio measurements minimized the memory problems often associated with the determination of Os isotope ratios in liquid samples by ICP-MS. The external reproducibility of 187Os∶188Os measurements of an in-house standard solution was 0.78% (1 s, n = 13), for analyte masses of between 0.081 and 1.22 ng total Os.Townsend197 used magnetic sector ICP-MS with enhanced sensitivity to measure isotope ratios in low-concentration Os (≤1 ng g−l) solutions. A capacitive decoupling Pt-shield torch, developed in-house and using a plasma forward power of 1050 W, increased Os signals by about a factor of five compared with that with normal power of 1250 W. Without a Pt-shield, the ratio precisions (1σ, n = 10) of 186Os∶192Os, 187Os∶192Os and 187Os∶188Os were ≥1% at the 1 ng g−1 Os concentration level. With the Pt-shield in use, ratio precisions improved to between 0.2 and 0.8%. The total amount of Os used was about 2.5 ng per measurement per replicate. At concentrations of 10 ng g−1, precisions of 0.15–0.3% were measured, irrespective of whether the Pt-shield was used or not. Any improvement in precision offered by use of the Pt-shield was considered small compared with that expected from Poisson counting statistics. The author suggested that this was likely to be due to noise contributions from other sources, such as the sample introduction system and plasma flicker. No appreciable loss of ratio accuracy was observed at 100 pg g−1 Os concentrations, as expected from counting statistics, but poorer precisions of the order of 0.45–3%, (1 s, n = 5) were noted. The presence of Re was reported to have a detrimental effect on the precision of Os ratios involving 187Os, indicating that the separation of Re and Os samples was a necessary prerequisite for analysis.
Quetel et al.198,199 compared the performance of four commercial ICP-MS instruments using U isotopic measurements. Two quadrupole and two double-focusing magnetic-sector [one single detector and one multiple collector (MC)] instruments were compared. The same samples of the IRMM-072 series of RMs were used under routine conditions to measure 233U∶235U and 233U∶238U ratios, which varied by over almost three orders of magnitude from 1 to 2 × 103. Within expanded (k = 2) uncertainties, good agreement was observed between the certified values and the data internally corrected for mass-discrimination effects. For the MC instrument, expanded uncertainties varied from ±0.04% to ±0.24% for the 233U∶235U ratio, and from ±0.08% to ±0.27% for the 233U∶238U ratio. They were 1–5 times larger for the single detector instrument, and 10–25 times larger with both quadrupole instruments. For the MC instrument, repeatability of the measurements seemed to be limited by instrumental background, whereas counting statistics were the only limitation on the smallest ratios measured with the single-detector magnetic sector instrument. Mass discrimination was poorer, but more reproducible, in the case of magnetic sector instruments than for both quadrupole instruments.
Stirling et al.200 have developed a method for the rapid, in situmeasurement of U/Th isotopic compositions using LA sampling combined with MC-ICP-MS. The system used a Q-switched and frequency-quadrupled 266 nm Nd:YAG laser to ablate, at 150 µm scale resolution, samples containing 100 ppm levels of U. Synthetic glass standards and naturally occurring samples of zircon and opal were used to assess the precision and accuracy of the technique. The results of laser analyses on the glass zircon samples were indistinguishable from those obtained using solution nebulization. In addition, analysis of the opal gave values in excellent agreement with the expected secular equilibrium value of unity. The Nd:YAG laser, coupled with an all Ar gas system, produced large elemental fractionation effects between Th and U. The Th∶U fractionation was primarily controlled by ionization conditions in the plasma, transport efficiency of ablated particles and the composition of the sample matrix. The use of He instead of Ar in the ablation cell significantly improved the relative sensitivity of Th, and entirely eliminated the elemental fractionation between U and Th, while retaining accuracy and precision in U isotope measurement.
Belshaw et al.201 used MC-ICP-MS for the precise measurement of natural variations in the isotopic composition of Fe. The contributions of ArN and ArO species to the spectrum in the Fe mass region were minimized. The repeatability of the measurement of the 57Fe∶54Fe ratio of the IRMM-14 RM solution (Fe isotope standard) was better than 100 ppm at 95% confidence.
There is growing interest in the use of ICP-TOFMS systems for isotope measurements. Vanhaecke et al.202 evaluated the isotope ratio precision (n = 10), as a function of time, using multi-element standard solutions with analyte concentrations of 50–500 µg l−1. For an acquisition time of 30 s per replicate and an elemental concentration of 500 µg l−1, typical isotope ratio precisions of ≤0.05% RSD were obtained. The authors pointed out that an attractive feature of ICP-TOFMS was that this level of isotope ratio precision could be obtained for many ratios simultaneously. In contrast to what was expected, increasing the acquisition time per replicate to values of >30 s resulted in a slightly poorer isotope ratio precision. At short acquisition times (<10 s), isotope ratio precisions similar to, or better than, the best values ever reported for quadrupole-based instruments were obtained. Mass discrimination was observed to be comparable to that observed with other types of ICP-MS instrumentation (about 1% per mass unit at mid mass). The accuracy was evaluated by comparing Pb isotope ratios obtained by ICP-TOFMS with those obtained by TIMS.
Emteborg et al.203 used isotopic RMs to assess the performance of axial ICP-TOFMS for precision and mass bias for isotope ratio measurements and accuracy of ID. The strength of TOFMS is the inherent recording over the entire m/z range for the same ion bunch. Experimental RSD of <0.05% for isotope ratios were obtained at high signal levels measured in the analogue detection mode. Effects on isotope ratios derived from changes of instrumental parameters, such as detector voltage and transverse rejection ion pulse settings, were evaluated. Isotope ratios for different concentrations changed when operating at low detector voltages. An appropriate voltage setting was considered important but, as the authors pointed out, optimum voltages differed according to the age and history of the detector. Results obtained for ID measurements of Mg and Rb compared well with results from quadrupole ICP-MS and TIMS. Mass bias per mass unit was 13% for 6Li∶7Li but around only 0.2% in the high mass range. The long-term stability of ratios required measurements of an isotopic RM at regular intervals to correct for small variations in mass bias overtime.
Dombovari et al.204 developed a method for the measurement of Mg isotope ratios for the determination of Mg concentration in 26Mg-spiked plant nutrient solutions using the reverse ID technique. The ICP-MS instrument was optimized to achieve lowest RSDs in the 25Mg∶24Mg and 26Mg∶24Mg isotopic ratios. Long-term isotopic ratio measurements using 50 µg l−1 of NIST SRM 980 (26Mg-enriched Isotopic Standard) yielded RSDs of 0.14 and 0.23% for 25Mg∶24Mg and 26Mg∶24Mg, respectively. The accuracy of measured isotopic ratios was 0.1% to 0.3%. Synthetic isotopic mixtures of NIST SRM 980 and highly enriched 26Mg solutions were prepared. The RSD of the isotope ratios measurements ranged between 0.14 to 0.07% for the 25Mg∶24Mg ratio and 0.2 to 0.04% for the 26Mg∶24Mg ratio. The accuracy of this ID technique was confirmed with results obtained by ICP-AES and conventional ICP-MS.
Diemer and Heumann205 developed an ID method for the analysis of Cd, Cr, Hg and Pb in polyolefins. Results for Cd, Cr and Pb were compared with those obtained by the ID-TIMS reference method. The high first ionization potential and volatility of Hg meant that the element could be determined only by ID-ICP-MS. Detection limits of 5, 164, 9 and 16 ng g−1 were obtained for Cd, Cr, Hg and Pb, respectively, using sample weights of only 0.25 g. It was concluded that the ID-ICP-MS results agreed very well with those of ID-TIMS and demonstrated its suitability for routine analyses.
Bandura et al.206 described the effect of collisional damping and reactions in a DRC on the precision of isotope ratio measurements. The target gas pressure, entrance and exit lens potentials and DRC rod offset were optimized to broaden the 107Ag+ ion packets of 0.2 ms duration to 2–4 ms (FWHM). Operating the DRC under optimized conditions improved the precision of isotope ratio measurements obtained with a downstream quadrupole mass analyzer to the theoretical limit defined by the counting statistics. For Ag and Pb samples at 40 ng ml−l, measured with the cell pressurized by Ne, isotope ratio precisions of 0.02–0.03% RSD at a counting statistics error of 0.01–0.02% were obtained. All 4 isotopes of Pb were measured in the same acquisition. The 204Pb∶206Pb, 207Pb∶206Pb and 208Pb∶206Pb isotope ratios in NIST SRM 982 (Lead Isotopes) were determined at a concentration of 10 ng ml−1 with an external precision (n = 7) of 0.056–0.117%, 0.022–0.034% and 0.014–0.028% RSD and with an average inaccuracy of 340, 781 and 180 ppm, respectively. The NIST SRM 981 at 10 ng ml−l was used as external standard. Elimination of isobaric interferences in the DRC was also shown to remove effectively interferences that would occur in analysis by quadrupole ICP-MS. For example, the ArN+ and ArO+ interferences on Fe+ were removed by ion–molecule reactions with NH3. Collisional damping of the ion beam fluctuations was sustained. Internal precision was limited only by counting statistics. For Fe at 50 ng ml−l, the average internal precision for 19 samples (n = 9) of 54Fe∶56Fe and 57Fe∶56Fe ratios was 0.068% and 0.1% RSD, with a corresponding average counting statistics error of 0.053% and 0.082% and external precision of 0.055% and 0.056% RSD, respectively.
4 Laser ionization mass spectrometry (LIMS)
This section includes studies using laser irradiation of solid samples either under non-resonant or resonant conditions whereas the following section on RIMS covers studies using resonant ionization of atomic beams or gas phase neutrals.4.1 Review
The review of Winefordner et al.207 with over 200 references provided the reader with an excellent survey of laser applications in atomic spectroscopy and MS. The basic processes of the interaction between the laser beam and the analyte atoms were treated well and the analytical specifications (LOD, RSD, lateral and depth resolution, elemental selectivity and sensitivity to matrix effects) were summarized as a guide to the broad range of laser techniques.4.2 Fundamental studies
Lasers are proving to be versatile tools for the analytical atomization and desorption of molecular analytes as well as for various material processing steps. The latter application of direct technological interest has increased the number of fundamental studies on the physical characteristics of laser-generated species. These studies are also relevant to the development of new analytical approaches. Important issues are the kinetic energy (KE) distributions of the generated neutrals and ions, the formation of the detected ions by direct ejection from the solid or by recombination at some distance from the sample (selvedge ionization) and ion formation during (prompt ionization) or after the laser pulse. Selvedge ionization implies that the detected ions can contain elements from distinct, non-neighbouring places in the sample. Continuing ion formation over µs-periods makes the use of ion storage analysers instead of TOF systems beneficial. However, the narrow energy acceptance of ion traps limits the sampling of ions with broad KE distributions and makes the electrical potential at the point of formation (sample or selvedge) critical.Li et al.208 investigated the mass, velocity and angular distributions of laser-ablated species from SrBi2Ta2O9 at a wavelength of 355 nm. The KEs of Bi+ and SrO+ pointed to a supersonic expansion-type formation mechanism. The metal ions had higher KEs than the neutral atoms and their angular distribution fitted a cosn function, whereas that of the neutrals was described by a bicosine function (A cos α + B cosn α where α is the emission angle). It was concluded that the laser-generated ions were mainly produced by non-thermal processes such as photochemical and electron ionization in the plasma whereas the neutrals resulted from both thermal evaporation and non-thermal processes.
Willmott et al.209 investigated the KE distributions of ions generated from a doped garnet by laser irradiation at a wavelength of 248 nm and fluences between 0.8 and 4 J cm−2. The KEs of the M+ and MO+ (M = Gd, Nd, Sc) typically ranged from about 20 to 40 eV at low fluence, though some ions had KEs of up to 100 eV. The KE of ions in the high fluence regime was about twice that of ions generated at low fluences.
Reid et al.210 investigated the wavelength dependence of the relative abundance of the neutrals generated from tin and tin dioxide by irradiation at a power density of about 108 W cm−2, i.e., below the plasma threshold. The major species generated from the oxide target were SnxOx (x = 1–3) at a wavelength of 532 nm, whereas atomic Sn was dominant at a wavelength of 355 nm. Laser ablation of the metal yielded Sn and Sn2 with an increasing relative contribution of Sn at the shorter wavelength. The KE distribution of the neutrals from the oxide target fitted a Maxwell–Boltzmann distribution and showed little variation with wavelength except for that of Sn ejected from the oxide target. The bimodal distributions of the neutrals ablated from the metal target pointed to a hybrid electronic–thermal mechanism.
The formation mechanisms of M+ and MO+ by the irradiation of YBa2Cu3O7 − x at wavelengths of 266 and 1064 nm were studied in the presence or absence of oxygen in order to distinguish formation by the selvedge reaction of M+ from formation of MO+ by direct ejection.211,212 The bimodal KE distributions of M+ at both wavelengths pointed to their production by direct ejection, ionization of neutral atoms and recombination of M2+ with electrons. In contrast, the unimodal KE distributions of MO+ and their shift to higher energy when 266 nm was used instead of 1064 nm, reflected selvedge ionization of neutral oxides. When oxygen was admitted to the source, the metal ions readily formed MO+ ions.
Shibagaki et al.213 studied the occurrence of ‘delayed’ ion formation after the laser pulse for positive and negative carbon clusters from graphite. For positive ions, the prompt ionization bunch contained only Cn+ with n < 4 and disappeared when a delay of 5 µs was applied between the laser pulse and the ion extraction. Delayed extraction yielded a series of Cn+ of which the intensity distribution peaked at n = 15. The number of C atoms in the negative clusters stayed constant with increasing delay but the absolute intensity increased while that of the positive clusters decreased. The additional negative ion production was associated with the existence of a stable electron density in balance with the excess of positive ions immediately after the laser pulse.
Liu et al.214 investigated the effect of the inherent ion stability on the composition of the binary clusters from solid mixtures of Ge, Pb, Si and Sn with P. The relative abundance of the clusters followed the intrinsic stability of a given element combination rather than the target composition. Photodissociation at 148 nm was used to determine the inherent cluster stability and a model to describe the cluster formation was proposed.
Ion formation from a frozen solution was investigated by Han et al.215 using irradiation of CeCl3 in D2O at a wavelength of 266 nm and a power density of 3 × 108 W cm−2. The velocity distributions of the Ce+, CeO+, (D2O)nCeO+ and (D2O)nD+ indicated that the hydrated metal absorbed the photon energy and initiated an explosive desorption. Ejected CeO+ became a nucleation core onto which water molecules were attached sequentially to form CeO(D2O)n. Hence, hydrated CeO+ species were formed through condensation processes in the selvedge rather than through a direct ejection from the solid.
Watanabe and Iguchi216 elaborated a theoretical model to predict the effect of the laser power density in one-step resonant ablation-ionization on the sensitivity and selectivity of the atomic ions. The particle flux from the solid to the vapour phase was calculated and the ionization probability estimated by considering both a non-resonant and a resonant component. The results agreed with the observations within experimental error for both resonant and non-resonant ion yields of metals at a threshold power density of about 108 W cm−2.
4.3 Instrumentation
Russo et al.217 developed an ion storage trap to accumulate the entire laser-generated ion population prior to ejection as a short pulse into a TOFMS. The sample was inserted through the ring electrode and ionized by a 6 ns laser pulse at a wavelength of 254 nm. The diameter of the ionizing beam on the sample was 10 µm. The ion trap storage device allowed ions to be accumulated from consecutive shots. Ions were trapped at different spatial positions, by varying the phase between the laser trigger and the rf-trapping field, to minimize space charge effects. The estimated LOD for Cu was 5 pg and the RSD of the measurement of Pb isotope ratios was <5%.Miniature ion traps have been used previously for the storage and isolation of ions in optical spectroscopy. Kornienko et al.218 investigated the use of simple cylindrical traps with an internal diameter of 1 mm for mass analysis of ions generated at a wavelength of 266 nm. Optimization of the trapping parameters allowed full mass spectra to be recorded with better than unit mass resolution from less than 100 ions in the trap. Sensitivity and resolution increased with the frequency of the ring electrode voltage but required a proportionally larger voltage amplitude.
The ion storage and mass analysis by FTMS features ultra-high mass resolution, but up to now the sensitivity of FT LMMS could not equal that of TOF LMMS. Optimized static electrical lenses incorporated in new ion transfer optics for FT LMMS with an external ion source improved the ion injection through the fringing field of the 4.7 T magnet.219 Calculations showed that angular acceptance of the ion trapping increased to initial emission angles of 35° from only 5° with the former commercial design. Experiments demonstrated that the transmission improved by a factor of 10 and that the LOD of 108 molecules in FT LMMS compared well to those of TOF LMMS, generally quoted for its ultimate sensitivity.
The so-called ‘aerosol’-TOFMS was modified for the atomization and ionization of free sub-µm particles by irradiating the particle at the focal point of a Nd:YAG laser beam.220 The plasma was much hotter than had been achieved previously and allowed all elements, including the electronegative elements, to be detected as positive singly- or doubly charged ions with relative abundances that closely reflected the composition of 10–70 nm sodium chloride, silica and polystyrene particles. The ion intensities could even be used for particle size determination and the mean diameters found were identical to those measured with electron microscopy and electrostatic classifiers.
4.4 Analytical methodology
Hong et al.221 modified a commercial ion trap MS for single particle analysis to use excimer laser ionization at a wavelength of 308 nm. The particles were injected through an orifice in the end cap by a solenoid-driven, particle-launching device. The ionizing laser was fired following particle detection by a He-Ne laser and a photomultiplier. An amount of 10−18 g of 137Cs could be detected with S/N of 4 from a single microparticle containing 3.1 pg g−1 of 137Cs.Aubriet et al.222 investigated the speciation of chromate compounds by TOF LMMS and FT LMMS, of which the mass spectra showed qualitative and quantitative differences. Specifically, the peak intensity ratios of the less stable to more stable cluster ions, e.g, CrO2−∶CrO3−, was much lower in FT LMMS than in TOF LMMS because of the longer time between the laser pulse and detection in the former. Both methods allowed speciation of the different reference compounds studied but separate databases were needed when fingerprinting was to be attempted.
Resonant laser ablation ionization uses the leading edge of the laser pulse to ablate neutrals from the solid, which are ionized by the trailing edge of the same laser pulse. This allows quantification to be achieved with fewer matrix effects than in, for example, SIMS. Dai et al.223 optimized the determination of Cu in aluminium by tuning the laser to a 2 + 1 transition scheme. The laser pulse energy proved to be very critical as irradiation above the threshold for signal saturation lowered the mass resolution and increased the flight times. This pointed to space charge effects and probably also to the onset of plasma processes in the ablated plume. The estimated LOD was <1 µg g−1 with a laser spot of 200 µm.
5 Resonance ionization mass spectrometry (RIMS)
Wendt et al.224 addressed the question of whether one can consider RIMS as an alternative to AMS for trace analysis of radioisotopes by reviewing the state-of-art and new developments in RIMS. The LOD of 106239Pu atoms achieved by modern pulsed laser TOF RIMS approached typical AMS figures. Use of an almost closed hot cavity for the interaction of the atoms with multiple laser pulses allowed about 5 × 10381Kr atoms to be detected from samples containing 1027 atoms. This implied a suppression of stable Kr isotopes as high as 1014. Quasi-collinear optical excitation of a fast atom beam gave LOD that were 10 times better than those of AMS for 90Sr, although the selectivity was worse by two orders of magnitude. Coherent multi-step RIMS, using single frequency CW lasers for isotope-selective excitation of a thermal atomic beam and a far-IR laser for ionization, was seen as a promising approach to reduce the cost of instrumentation. The authors concluded that RIMS could not be expected to compete with AMS but could find a use in those cases in which isobaric interferences or strong backgrounds hampered the use of AMS.Ionization efficiency in RIMS critically depends on the laser beam profile and its overlap with the analyte beam. Thattey et al.225 developed a direct laser beam imaging system featuring a resolution of about 20 µm per pixel to quantify the irradiation intensity distribution in planes cross-sectioning the atom beam. Multi-laser experiments were studied by looking at each laser separately and calculating the “interaction” integral, which was proved to correlate directly with the ion signal intensity detected. Hence, the calculated number of sample atoms in the “effective” ionization region could be used to determine, for example, excitation and ionization cross sections.
Calaway et al.226 systematically investigated the sources of noise in the TOFMS with resonant or resonantly enhanced multiphoton ionization (REMPI) of the neutrals generated by primary ion bombardment. Modified pulsing schemes for the primary ion beam and the lasers decreased the noise level by a factor of 100 to less than 5 × 10−5 counts per pulse. Thereby, the LOD for Ti in silicon or germanium improved from about 1 ppb to less than 100 ppt and 5 ppt for 106 and 108 shots, respectively.
Ultra-trace determination of the radioisotope 90Sr by single-, double- and triple-RI schemes was investigated by Bushaw et al.227 The optical selectivity for 90Sr was 2 × 104 with the double-resonance process, whereas unfavorable isotope shifts in the triple-resonance scheme reduced it by a factor of 100. Using double-resonance ionization, an overall detection efficiency and isotopic selctivity of over 10−5 and 1010, respectively, could be achieved for 90Sr spiked samples. A LOD for 90Sr of 0.7 fg was demonstrated in samples containing stable Sr isotopes in an excess of 1010.
RIMS provides an elegant method for probing the population of specific atomic states in neutral atoms in order to optimize, for example, the power broadening at high fluence. Philipsen et al.228 investigated the double-resonant photoionization of Sr atoms out of the ground state and the metastable triplet states generated from Sr I by 15 keV Ar+ ions. The saturation behaviour of both the excitation to intermediate states and the ionization into the auto-ionizing regime were studied and the photoionization cross-sections were determined.
Müller et al.229 applied single-mode ended-cavity diode lasers for the determination of 41Ca in concrete from a reactor, to overcome the high 40Ca content and the isobaric interferences of 41K and CaH. Free atoms were generated from a heated channel-shape graphite crucible, which kept the atom beam divergence within 10° to minimize Doppler broadening. The 43Ca isotope was used as a reference since it had the same hyperfine structure and excitation efficiency as 41Ca. The latter isotope could be determined down to isotopic abundances of 10−10. Accuracy and reproducibility were within 15% and mainly limited by the counting statistics.
The development of a GD source with RIMS post-ionization by Pibida et al.230 for trace analysis was motivated by the ability of the GD to produce a high flux of atoms from a near-infinite reservoir with a nearly constant ratio of neutrals-to-ions for different elements. Two lasers, one of which was tunable, were focused to a spot of 1 mm at a distance of 3 mm from the exit of the conventional cell for pin electrodes. The isotopic selectivity for Cs with Faraday cup detection was 105, but 108 should be feasible with an electron multiplier. The overall selectivity of the system was estimated to be up to 1012. Memory effects were found to be negligible.
6 Secondary ion mass spectrometry (SIMS)
The extensive relocation of atoms in the subsurface during bombardment of solid samples with primary ions of 1–25 keV prevents subsequent use of the subsurface for the generation of molecular information. Hence, dynamic SIMS, which uses a high primary ion current density for layer-by-layer erosion and depth profiling, is restricted to elemental analysis. In contrast, static SIMS (S-SIMS) allows the intact molecules in the upper monolayer to be detected by limiting the ion dose to avoid probing of the subsurface. Speciation and organic analysis are feasible but depth profiling is excluded. This review uses the acronym SIMS exclusively for dynamic SIMS. Hence, TOF SIMS and TOF S-SIMS refer to the same instrument in the dynamic and static mode, respectively. Because of their fundamentally different operational and analytical features, dynamic SIMS is treated in sections 6.1–6.5 and S-SIMS in section 6.6. The incidence and emission angles, mentioned throughout this section, are expressed as the angle between the ion beam and the normal to the sample surface.6.1 Instrumentation
Schuhmacher et al.231 developed new duoplasmatron ion optics for low energy primary ion beams. The design minimized space charge effects and improved the matching between source emittance and primary column acceptance. Variation of the extraction voltage allowed primary ion incidence angles to be varied from 44° to 66°. The spot diameter on the sample with 0.5 and 0.8 keV primary ions was about 50 and 30 µm, respectively. Experimental results demonstrated a LOD in the low 1016 atoms cm−3 range for B using 0.5 keV primary ions and an analysed spot diameter of only 33 µm.The beam-induced damage of the subsurface region hampers the conventional 3-D analysis by sputter etching and 2-D mapping. Therefore, Takanashi et al.232 used a high intensity, focused Ga+ beam at an incidence angle of 90° for exposing deeply recessed layers up to a depth of 100 µm. In this way, the layer of interest was subjected to significantly less beam damage than in the case of conventional sputter etching. The precision of the sectioning was sufficient to expose successive planes at a distance of 10 nm. Cross sectioning of thin layers at oblique angles could be performed for analysis of layers with a thickness less than the spot diameter.
Although SIMS has already seen decades of development, even basic devices such as sample holders still appear to need optimization. Yamazaki and Enda233 reduced the electrical field distortion above the specimen surface by optimizing the faceplate thickness and tapering off the edges of a commercial device. The otherwise significant variations of the ion intensities from spots near the edge of the tilted sample holders were eliminated and the exposed area in the upper mask could be reduced to facilitate charge compensation. Dérue et al.234 designed a dedicated cold stage for the analysis of frozen biological samples, allowing the original sample holder to be positioned accurately and reproducibly in front of the immersion lens. At the same time, the design was optimized to ensure vibration-free operation.
Ultra-high-speed position-computing electronics for resistive anode encoders were developed by using a fast self-resetting charge integrator and sub-range digital techniques to process pulse height distributions using a reasonably sized circuit.235 Dead times of less than 35 ns were demonstrated for uniformly spaced test pulses and ion imaging over 256 × 256 pixels was achieved with dead times of 60 ns, shorter than those of commercial systems by a factor of at least 5. Image resolutions of 1 µm could be obtained at ion intensities above 106 counts s−1 without significant image distortion.
Gilmore and Seah236 investigated the detection efficiency of microchannel plates as a function of the ion energy, mass and composition. A new model for the ion-induced electron emission provided an excellent description of the decreasing efficiency with increasing mass and decreasing impact energy. At impact energies of 5 keV, the detection efficiency for ions containing light elements (for example, hydrocarbon ions) was a factor of two lower than that of clusters of similar mass but containing only heavy atoms. At an impact energy of 20 keV, however, the detection efficiencies were identical. A simple method to determine the optimal microchannel plate voltage was given.
Responding to the increasing interest in the use of polyatomic primary ions, Dahl et al.237 developed single-voltage-gun optics for generation of a broad ReO4− beam from a solid-state perrhenate source. Ions emitted at a given angle from any position in the source were focused on the same spot and the focal lengths of the lenses stayed practically constant at different operating voltages. A uniform intensity distribution with a sharp cutoff was demonstrated over the beam spot of 2 mm diameter. The beam intensity and diameter varied by only 25 and 10%, respectively, for primary ions in the range of 2–5 keV.
The increasing demands for ultimate specifications in material characterization mean that no single technique can deliver all the required information. Therefore, Terzic et al.238 designed a combined thin-film analysis instrument including four of the most useful surface characterization methods, namely low energy ion scattering, direct recoil MS, recoil ion MS and SIMS. In order to decrease the crater effect in depth profiling without rastering or electrical gating, a broad alkali ion beam was generated by a thermionic source outside the ultra-high vacuum chamber. A feature of the magnetic quadrupoles used to transport and focus the primary ions was the ease of beam tuning.
6.2 Fundamental studies
Wittmaack239 investigated the beam-induced oxide layer formation by O2+ bombardment of silicon at various incidence angles. Specifically, the angular dependence of the stationary sputter yields was compared for 10 keV O2+ and 5 keV Ne+ ions, which provided the same energy per impinging projectile atom. The silicon sputter yield with O2+, normalized to that with Ne+, increased from 0.33 to 0.89 at incidence angles between 0° and 65°. Assuming that the experimental sputter yield was a linear combination of those for pure silicon dioxide and silicon, the silicon dioxide fraction at the surface could be calculated and was found to agree closely with the results from ion-induced electron emission studies. Hence, the ion yield could not be considered as a simple product of a mean sputter and mean ionization yield but local formation of silicon dioxide had to be taken into account.The growing interest in ultra-shallow depth profiling increases the importance of the transient effects, which occur before a steady state between ion implantation and sputtering of the sample is reached. In a series of papers, van der Heide investigated the transient regime when 0.9 and 6 nm silicon dioxide layers on silicon were bombarded by 0.75 and 1 keV Cs+ at an incidence angle of 60°.240−242 The beam-induced contribution to the transient effects was derived from X-ray photoelectron spectroscopy (XPS) measurements of Cs in the crater bottoms after different sputter times and agreed closely with the values from calculations by the transport of ions in matter (TRIM) code. The implanted Cs concentration increased to 5.5 atom% in the 6 nm silicon dioxide layer, beyond which it decreased to a steady 3% in the substrate. The exponential relationship of the silicon yield with the Cs concentration in the oxide layer was consistent with the work-function-controlled, resonance-charge-transfer process in the electron-tunnelling model. Localized dipole layer formation was proposed to explain the negative ion emission from the oxide. Evaporation of Cs on to the sample before analysis plus O2 flooding was suggested as a practical method to control the transient effects.
Pre-equilibrium sputter rates were studied by Ronsheim and Murphy243 for ultra-shallow depth profiling of B in silicon by sub-keV O2+ with oxygen flooding. The native oxide layer was etched off before analysis. Sputter rate equilibrium was reached after a fluence of 1017 ions cm−2 and the crater depths varied linearly with the fluence. Use of a 800 eV beam caused surface swelling by the implanted oxygen but, for primary ions of less than 500 eV, crater depth offsets were within the experimental uncertainty due to beam instability. Hence, low energy sputtering with oxygen flooding yielded a near-instantaneous ion yield equilibrium, at least in this example.
Licciardello et al.244 investigated the interfacial phenomena during depth profiling of silicon dioxide on silicon with the dual-beam technique, in which Ar+ is used for sputtering and Ga+ for analysis. The Ga+ yield showed a characteristic double-peaked structure at the silicon dioxide–silicon interface. The distance between the two peaks remained independent of the Ga+ energy between 15 and 25 keV, but decreased when the Ar+ energy was reduced from 5 to 1 keV. Correction of the sputter rate in the interfacial region was based on the SixOy cluster ion intensity distribution. The first near-surface maximum of the Ga+ yield was due to a change in the sputter rate, whereas the second maximum, located at the silicon dioxide–silicon interface, was associated with Ga-segregation at the interface. Deenapanray and Petravic245 studied the electrical-field-driven segregation of F, K and Li in p- and n-type silicon covered by an oxide layer. Specifically, K segregated at the oxide side of the interface whereas the high solid-solubility of Li in amorphous silicon caused its segregation at the silicon side.
The influence of bombardment conditions on depth resolution in the analysis of GaAsN/GaAs multiple quantum wells was studied for 1–3 keV O2+ and Cs+ primary ions at impact angles between 45 and 60°.246 At an incidence angle of 60°, the depth resolution was 3.6 and 5.8 nm for 1 keV O2+ and for 3 keV Cs+, respectively. The experimental decay length increased with decreasing primary ion energy for both O2+ and Cs+. The Zalm–Vriezema equation worked well for determining the sputter-induced decay length contribution as a function of the impact energy and angle, the surface binding energy and the averaged cohesive energies in the sample.
High-resolution Rutherford back scattering (RBS) was used to investigate the artifacts in SIMS depth profiling of N in 2.5–8 nm silicon oxynitride films when using a 750 eV Cs+ beam at an incidence angle of 55°.247 Although RBS and SIMS yielded similar shapes for the N-distribution, SIMS located the maximum of the distribution at just 1 nm below the surface whereas RBS located it at 2.3 nm below the surface. In addition, the N concentration was underestimated by SIMS by a factor of 3.6. Hence, it was proposed that a simple rescaling of depth and concentration could yield accurate data in SIMS.
Shaanan and Kalish248 elaborated a modification of the TRIM code to account for the changing composition during depth profiling of implanted doses above 1017 ions cm−2. Specifically, the simulations showed that beam-induced elastic recoils broadened the initial profile whereas modification of the sputter rate by the changing matrix composition made post-analysis measurement of crater depth irrelevant. Application to thin layer analysis showed that the exponential tail in the depth profiles could be fully explained by the beam-induced artifacts.
The use of depth resolution functions (DRF) is decisive for accuracy in ultra-high resolution experiments at depths of less than 10 nm. Hofmann249 investigated the fundamental limits of SIMS depth profiling when DRF, modelled by the commonly used mixing length–roughness–information depth (MRI) code, were used. The estimated theoretical limit for depth resolution was 0.7–1 nm on the conditions that: (1) complete mixing occurred; (2) that no additional diffusion, preferential sputtering or segregation occurred; (3) that the width of the atomic mixing zone was minimized by the use of low energy or polyatomic projectiles at a high incidence angle (80°); and (4) that topography development was minimized by sample rotation. Experiments with low energy SF5+ yielded a depth resolution of 1.4–1.6 nm, of which 0.4–0.6 nm could be assigned to the mixing length contribution. Hence, surface roughness and topography development were considered critical for ultimate depth resolution. Kasnavi et al.250 considered the use of a DRF derived from traditional models as inadequate for depth profiling of ultra-low energy As implants in Si using 0.4–0.75 keV O2+ with oxygen bleeding. ‘True’ implant profiles were recorded by the application of XPS in combination with monolayer chemical oxidation and etching. The varying ion yield in the near surface region and the ion mixing effects were difficult to account for completely and the physical models for the As implantation at low energy could be improved by Monte Carlo simulations. Therefore, the authors advocated the use of XPS to calibrate the physical ion implantation simulations, which then could be used to obtain better DRF.
A simple two-step erosion model was developed to describe the shift of δ-layer location in silicon with increasing depth when oblique low energy O2+ bombardment and oxygen flooding was used.251 It was assumed that the almost three-fold decrease in the erosion rate over the first 4–5 nm was a result of the oxygen incorporation. The subsequent relatively steady rate at greater depths was mainly affected by ripple formation. Calculation of the true sputtered depth as a function of time showed that the depth dependent erosion rate had to be determined using δ-doped samples with markers at least 50 nm and no more than 100 nm below the surface. Jiang et al.252 investigated the apparent depth shifts of B and Ge δ-layers bombarded with 1–4 keV O2+ primary ion beams at incidence angles of 40–80°. The distance between the layer centroids calculated from DRF proved to be 12.00 nm whereas TEM on cross sections yielded 11.97 nm. The apparent position of δ-layers was not affected by the primary ion energy as a result of the compensating effects from oxygen penetration depth and equilibrium oxygen concentration. However, the apparent depth of the B layer varied over 1 nm with the primary ion's impact angle whereas that of the Ge layer did not. The apparent outward shift of B (to the surface) with decreasing impact angle was explained by the outward diffusion of B during depth profiling with oblique oxygen beams. In the case of Ge, however, inward chemical diffusion compensated for the expected depth scale offset.
Lively discussions on the validity of the infinite velocity (IV) method to overcome the matrix effect in quantitative SIMS were published by van der Heide253 and Wittmaack.254 The first author surveyed the experimental evidence for the IV method and pointed out that the observed cases of non-linear dependence of the normalized intensity on the reciprocal velocity at energies under 200 eV were due to the existence of two ionization processes. Under these conditions the IV method could not be applied. Hence, the input data had to be restricted to the energy range stipulated by the IV method. In his reply, Wittmaack reviewed the experimental data contradicting the validity of the IV method and considered the use of correction factors in the evaluation program of van der Heide as an implicit recognition that the IV method failed for some elements.
Zalm255 critically evaluated the different approaches for the ultrashallow depth profiling of 2–3 nm silicon dioxide layers using O2+ primary ions. Use of high incidence angles resulted in increased matrix effects and the development of surface topography. Oxygen flooding could lead to segregation of elements such as Ge. The old, lattice valency model was resurrected to determine the local oxygen content of the eroded layer. It required, however, panoramic recording and, for quadrupole and magnetic analysers, the SiO+ signal intensity was proposed as an alternative means to monitor the local oxygen content.
Quantitative analysis of biological tissues and single particles in the 1–10 µm range remains a challenging application of SIMS. Grignon et al.256 determined the local concentrations of a 15N tracer in soybean leaf samples, using carbon and the natural isotope ratio of N in the embedding medium to standardize the instrumental conditions. Detection of CN− allowed RSD within 10% to be obtained and the measured concentrations agreed within 20% with conventional bulk MS analysis. Tamborini and Betti257 attained a RSD of 2–4% for the isotope ratios of U in µm particles. A mass resolution of 1000 was used as a compromise between sensitivity and separation of isobaric interferences. The method was validated with the soil RM, IAEA-375, and should be applicable to the analysis of 1 µm particles containing as few as 1010 atoms.
6.3 Analytical methodology
Fitzsimons et al.258 reviewed the analytical precision in stable isotope measurements. Assuming that isotope ratios were taken in such a way as to avoid time-fluctuations, the precision was determined by the Poisson counting statistics and SIMS could compete with conventional bulk analysis if counting times of several hours were used. The gradual shift in mass fractionation with time could be corrected for by the analysis of reference materials. The final precision of a corrected isotope ratio largely depended on the scatter in the sample and the reference material.The use of ion implantation to make standards for ultrashallow depth profiling has the disadvantage that the maximum of the implanted ion distribution is located relatively deep inside the target. Stevie et al.259 developed a method for ion implantation in silicon using a removable layer of silicon dioxide with a thickness of the order of the projected range. Etching the upper layer with HF yielded a standard in which the maximum of the implanted ion distribution was very close to the surface. Surface flatness after etching was within 0.5 nm and the measured B concentrations with and without oxide removal were identical. Surface analysis by TOF SIMS and TXRF confirmed that the oxide etching did not affect the (near) surface concentration.
A new vibrating orifice nebuliser was designed by Erdmann et al.260 for preparation of adequate reference materials for single particle analysis. Monodisperse uranium oxide particles of about 1 µm diameter could be generated from uranyl nitrate solutions. The isotopic composition determination was accurate within 0.4% for 235U∶238U and within a few % for the minor isotopes. The mass of 234U contained in a 1 µm particle was about 0.13 fg.
Sample charging during sputtering can cause redistribution of mobile alkali elements within the target. Although the application of a conductive coating can reduce this effect, it carries the risk of contamination. Therefore, precise matching of the electron penetration depth to the thickness of the oxide layer by adjusting the electron energy was proposed by McKinley261 for charge compensation during depth profiling of alkali elements in silicon dioxide layers on silicon. Results showed that silicon dioxide layers with a thickness between 0.3 and 1 µm could be repeatedly analysed for the dose calculation of the alkali elements.
The protection of air sensitive samples to be transported from the dedicated preparation chamber to an off-line analysis instrument is a critical issue. Lisowski262 found that a 30–50 nm layer of gold evaporated from a fine wire was an extremely convenient and efficient way to protect, for example, titanium hydride films for over 6 months.
Bose et al.263 used both SIMS and photoluminescence to measure quantum well widths down to 2.5 nm in an InGaAS/InP structure. Differences in the widths measured by the two methods were attributed to the different sputtering rates of InGaAs and InP and the presence of interfacial layers of InAsP and InGaAsP. A simple correction formula based on the geometry and thickness of the interfacial layers and on the relative sputtering rates for homogeneous InGaAs and InP layers allowed a good correlation to be achieved between quantum well widths determined by both methods.
The complementary use of S-SIMS and SIMS has been demonstrated to determine the carbon trapped in gate oxides of microelectronic devices as a result of clean room contamination.264 Specifically, S-SIMS was used to probe the surface contaminants before thermal oxidation and polysilicon deposition and SIMS was used afterwards to quantify the concentrations of trapped C of the order of 1012 atoms cm−2. The measured concentrations correlated well with the summed peak intensities of the organic contaminants detected by S-SIMS on the initial wafer surfaces.
6.4 Quantification
The MCs+-detection method is a form of in situ post-ionization, featuring a virtual absence of matrix effects but lower LOD than detection of atomic ions. Kolber et al.265 observed that the RSF determined for the quantification of BCN coatings on metals by the MCs+ method, depended on the surface charging when the potential changed by more than 50 V. Optimization of the MCs+ method resulted in an RSD of better than 2% for intensity ratios. The linear relationship between the intensity ratios and the concentrations demonstrated the virtual absence of matrix effects. Wang and Zhang266 showed that the RSF when using AsCs+ and InCs+ in Hg1 − xCdxTe materials stayed within the RSD of 5 and 15%, respectively, for values of x between 0.2 and 0.8. Since the sputter rate decreased linearly with increasing x, the individual layer thickness in Hg1 − xCdxTe multilayer structures could be determined directly from the MCs+ intensities. Willich and Wischmann267 reported an accuracy of better than 20% for the analysis of diamond-like carbon coatings containing 2–10 nm metal or metal carbide particles. Typical LOD of about 0.01 atom % and a depth resolution of about 20 nm were achieved. In addition to the absence of matrix effects, detection of electropositive and electronegative elements in the same polarity mode was considered a major advantage for the study of embedded nanoparticles.Using nuclear reaction analysis as a matrix-independent method to determine the dose of B implanted in the upper 1–2 nm of silicon, Magee et al.268 demonstrated that SIMS could attain an accuracy of better than 10% for implantation doses up to 1015 atoms cm−2. Depth profiling was achieved with 1.5 keV O2+ at an oblique incidence angle and with oxygen flooding. The effect of the transient regime in the first nm of the depth profile did not constitute a major bottleneck so that high surface concentrations up to 5% could be quantified even if diluted standards were used for calibration. Ottolini and Oberti269 quantified B, Ba, Be, F, H, Li, Th, U and Y in complex matrices using X-ray single-crystal structure refinement as a reference method for the quantitative mapping of elements in the absence of matrix effects. Discrimination of the low energy ions in SIMS made it possible to reduce both the interferences and the dependence of the relative ion yields on the matrix composition.
Because of its high transmission, TOF SIMS has the potential to meet the foreseen requirements for detection of metal contaminants on silicon at concentration levels of less than 108–109 atoms cm−2, which are below the LOD of the currently applied TXRF method. Zanderigo et al.270 evaluated the use of dipping and spinning methods to prepare adequate reference samples. The presence of the metals exclusively at the surface and the homogeneity of the metal distribution was verified by dual-beam depth profiling and imaging. Spin-coated samples yielded a RSD of 6% in comparison with that of 30% for the dip-coated substrates. Lazzeri et al.271 used spin-coated reference samples for quantitative analysis over 150 × 150 µm areas. The TOF SIMS signal intensities correlated very well with the TXRF data for average metal surface concentrations of 1011–1014 atoms cm−2. Estimated LOD for TOF SIMS ranged from 5 × 107 atoms cm−2 for K to 3–6 × 109 atoms cm−2 for Cr, Cu, Fe and Ni and 3 × 1010 atoms cm−2 for Zn.
Hsu et al.272 determined Au, Ir, Os and Pt in individual phases of iron meteorites using a Cs+ primary ion beam and negative ion detection with 40 eV energy filtering and a mass resolution of 1900. Under these conditions, LOD of less than 1 ppm for Os, 0.1 ppm for Ir and 10–20 ppb for Au and Pt could be achieved with a spatial resolution of 10–20 µm. The RSD for the determination of Au was better than 1% due to the low matrix effect of this element but increased to 5% for the other elements.
6.5 Single and multi-dimensional analysis
Ramanath et al.275 investigated the channelling-induced distortion of depth profiles through polycrystalline TiN/Ti/TiN(001) trilayers. The sputter rate of TiN(001) layers varied by more than 40% and the ion yield, which changed by a factor of 3, was dependent on the angle between the Cs+ beam and the crystal planes. This could be related to the channelling of primary ions and secondary recoils through the 0.106 nm wide channels between highly aligned (100) or (010) planes in TiN(001). The sharp decrease of the ion yield in the Ti-layer immediately before the Ti/TiN(001) interface was responsible for the loss in depth resolution observed. The distortions could be eliminated by the use of either sample rotation or O2+ primary ions.
The use of MSi2+-detection was explored for depth profiling ultrathin silicon oxynitride films of less than 3 nm.276 Use of the secondary Si+ matrix ions to cationize the elemental and cluster species parallelled the MCs+ method but eliminated the 1 nm pre-equilibrium regime of the latter method. At a mass resolution of 300, isobaric interferences caused a background to the Si2N+ signal corresponding to 1.5 and 0.25 atom % at depths of 0.1 and 0.5 nm, respectively. Use of a mass resolution of 5500 lowered the background to 0.1 and 0.015 atom % at apparent depths of 0.1 and 1 nm, respectively. The long-term reproducibility was characterised by a RSD below 2%.
Depth scale distortion in low energy SIMS was investigated by Schueler and Reich.277 Localization of 5 equally spaced boron δ-layers (interlayer distance of 5.4 nm) in silicon (covered with a 0.55 nm silicon dioxide layer) was achieved with O2+ primary ions (200–1200 eV) at incidence angles of 0° (without oxygen flooding) or 45° (with oxygen flooding). Both O2+ at 0° incidence and O2+ at 45° (with oxygen flooding) yielded the correct interlayer distances, but the first layer was detected at an apparent depth of 4.3 and 4.7 nm, respectively, instead of 5.4 nm. Bombardment with O2+ of less than 500 eV at an incidence angle of 45° and with oxygen flooding reduced the depth scale distortion near the surface in comparison to normal incidence at any energy between 150 eV and 1 keV. Oblique O2+ bombardment and oxygen flooding improved the decay length and the width of the fifth δ-layer in comparison with the use of a normal incident beam at the same energy.
The strong extraction field places a limit on the use of low energy ions in magnetic sector SIMS. Napolitani et al.278 found that a 1.5 keV O2+ beam without oxygen flooding provided sufficient depth resolution to characterize a 500 eV B implant in silicon beyond the equilibrium depth of 5 nm. The decay length of 2.3 nm agreed with the one calculated by implantation modelling. A LOD of 1015 atoms cm−3 and sputter rates as high as 25 nm min−1 could be achieved. McKinley et al.279 developed improved equations for the variation of the impact angle as a function of the primary ion energy, the voltage on the sample and the deflector electrodes. In addition, they developed a dedicated procedure for low energy beam alignment in magnetic sector instruments. The method was validated by depth profiling 0.25 keV and 1 keV B implants using raster sizes as small as 80 × 80 µm.
Another way to overcome the low energy limitation of magnetic SIMS is the use of polyatomic primary ions to reduce the impact energy per incident atom. Gillen et al.280 studied the ultrashallow depth profiling of B-implants, δ-layers and a Ni–Cr multilayer by using SF5+ primary ions from a modified hot filament duoplasmatron. Trailing edge decay lengths for the B implants and δ-layers were 1.3 and 2.3 nm when using 3 keV SF5+ at an impact angle of 52° (with oxygen flooding) and O2+ bombardment (with oxygen flooding), respectively. Oxygen flooding reduced the ripple formation under SF5+ bombardment by a factor of 5 but the roughness of the crater bottom was still greater than that of the one generated with O2+ primary ions. The apparent depth shift to the surface was 1 nm and 2–3 nm for SF5+ and O2+, respectively. Although the sputter yield with SF5+ was 4–6 times higher than that with O2+ or Ar+ at all energies and angles investigated, the useful yield of B was 2–3 times lower than with O2+ primary ions.
Since extreme depth resolution is usually not required over most of the depth profile, the use of two beam energies in the same depth profiling experiment was investigated by Cooke et al.281 Low energy primary ions were applied first to minimize the transient regime and subsequently high energy projectiles were used once the decay length became smaller than the theoretical limit at that energy. The procedure could be inverted for buried interfaces. Layer positions were identical when 250 eV ions were used for the entire depth and when the energy was switched from 250 to 1000 eV (after correction for the sputter rate). In the latter case, the analysis time was 10 times shorter. The B-implanted dose agreed with the expected value within 3%.
Brox et al.282 used dual-beam depth profiling for thickness determination of micro-electronic gate oxides. A 0.4–5 keV Cs+ beam at an incidence angle of 52.5° was used for sputter etching and a pulsed 10 keV Ar+ beam for analysis. The use of TOF SIMS allowed Si−, SiO− and SiO2− to be monitored simultaneously for interface localization. The transient width decreased from 8 to about 1 nm when the sputtering beam energy was reduced from 5 to 0.4 keV. The relationship between the layer thickness, determined by electron microscopy, and the sputter time needed to reach the interface was linear and independent of the Cs+ energy.
Stubbings et al.284 developed the use of a wavelet transform for the fusion of image information given by ions of different masses but originating from the same analyte compound. In comparison to conventional fusion techniques, the weighted wavelet averaging improved the perception of weak structures. A condition was that the wavelet used, the fusion rule and the decomposition level should be critically optimized. Further improvement by implementing colour schemes was anticipated. Wolkenstein et al.285 improved edge detection in images with low counting levels through use of a wavelet-transform-based algorithm. In contrast to the classical image processing, the blurring of the region boundary in low S/N regions of the image was reduced significantly. Features of both high and low intensity regions of steel, soldering-joint images could be distinguished.
Gavrilov et al.286 investigated the capabilities of the high resolution scanning ion microscope at the University of Chicago to study the grain boundary chemistry of alumina ceramics. The liquid-metal ion gun featured a lateral resolution of 20–50 nm and depth resolutions of 1–10 nm were common. Element mapping with a typical resolution as high as 35 nm was demonstrated. The sensitivity for all elements was in the ppm-range.
Gebhardt and Gavillet287 reported the accurate quantification of 10B images taken from irradiated, nuclear-reactor, carbide control rods. An 11 keV Cs+ beam with a spot diameter of about 30 µm was used at an incidence angle of 50° to image fields of 2.5 × 2.5 mm. The averaged 10B depletion, determined by TIMS of dissolved neighbouring material to be 51 ± 0.5%, was measured as 47 ± 2%.
The use of a novel moiré pattern test structure for improvement of the image resolution and simplification of the 2-D data deconvolution in dopant quantification was demonstrated by Ukraintsev et al..288 The lateral resolution of the order of 10 nm was mainly limited by the size of the photomask pixel. Dopant concentrations of the order of 1017 atoms cm−3 could be detected.
Qualitative and quantitative 3-D analysis was demonstrated by Kolber et al.289 for B incorporated in polycrystalline, chemical-vapour-deposited, diamond films. The direct imaging in the stigmatic mode limited the lateral resolution to about 1 µm but reduced the analysis time in comparison to raster scanning. However, the latter mode featured a lateral resolution of 0.2 µm and the electron multiplier detection improved sensitivity in comparison to the camera detection in the stigmatic mode. Raster scanning allowed the enrichment of B and C to be observed in the interface regions of less than 0.5 µm between the diamond facets. Pollak et al.290 studied the 3-D distribution of elements such as Al in hot, isostatic-pressed, steel samples. The enrichment of Ca, K, Mg, Na and S in the particle boundaries was shown as well as nitride precipitates with dimensions of only 50–100 nm. Although the lateral resolution of 2-D images was only 2 µm, the depth resolution was much better so that the linking of images from successive planes allowed the size of inclusions to be determined. Musser et al.291 investigated the distributions of 13 elements in rhenium coatings and granulates. Imaging allowed precipitates of 5–10 µm to be visualised. The C content was quantified using pellets of soot and 36–63 µm grains of rhenium powder as calibration standards. The LOD was about 0.2 ppm and the RSD of 18–24% was primarily determined by the local heterogeneity.
Rosner et al.292 studied two-component metal coatings of Al and Sn using a double channel plate detector for single ion detection in the imaging mode and software in-house for visualization of the 3-D information. The typical lateral resolution was about 1 µm. Because of the surface roughness, the depth scale was calibrated using sputter rate estimates from TRIM calculations. Although the software did not account for the topography and reconstructed the 3-D composition as if the surface was perfectly flat, the images were of sufficient quality to visualise the Sn inclusions. Using low current O2+ depth profiling, monolayer sensitivity could be achieved over frames of 8 × 8 µm.
6.6 Static SIMS (S-SIMS)
Hagenhoff295 reviewed the state-of-the-art applications of high resolution imaging by TOF-SIMS under static and transient conditions. For analytes with sufficient ion yield, a lateral resolution of better than 200 nm could be obtained routinely and in favourable cases even 50 nm could be attained. Low ion yields limited the lateral resolution of molecular imaging to the µm range and improvement was expected to come from new focused guns for polyatomic primary ions.
Speciation of individual compounds in complex matrices without a chromatographic separation imposes increasing demands on the detection technique. Therefore, Groenewold et al.299 demonstrated the use of an ion trap mass analyser to exploit the specificity of MS-MS experiments. Specifically, the quantification of a phosphonothioic acid-based nerve agent on mg quantities of soil particles was studied using ReO4− polyatomic primary ions. In such a matrix, detection of the atomic ions (C, H, N, P and S) was insufficient to quantify the analyte. Even the detection of protonated molecules could fail to characterize the analyte unambiguously. Therefore, ion traps offered the advantage of relatively easy MS-MS experiments, which can be analytically exploited to reduce the chemical background from other constituents. The LOD for the nerve agent improved from a 0.4 monolayer in the direct MS mode to a 0.002 monolayer (3 µg per gram of soil). The method was subsequently applied to the detection of the nerve agent present as a 10−4 monolayer on concrete samples.300
7 Sputtered neutral mass spectrometry (SNMS)
Oechsner et al.301 developed an advanced electron-gas SNMS prototype with XPS incorporated in the same main vacuum chamber. The sensitivity was increased by a factor of 10–100 at 200 eV by extracting the post-ionized species from the separate plasma chamber in an oblique direction in order to exploit the torus-like angular distribution. Analysis of the Ni-based NIST SRM 1245a showed a linear correlation between ion intensities and concentrations, except for light elements such as Be and C, because of their high ionization energy and velocity. The LOD of B was 200 ppb.The quantification of metals in organic matrices by non-resonant, multiphoton post-ionization was studied by Schnieders and Benninghoven302 using metal submonolayers sputter-deposited on different polymers as model systems. The RSF were similar to those obtained for alloys. The LOD of about 1 ppm of a monolayer were 10 times better than those obtained by SIMS. The analysis of metalloproteins with LOD of 1 ppm showed that the method could be applied equally well to metals bound to organic macromolecules as to metals adsorbed on surfaces.
The emission of secondary ions may re-introduce the problem of matrix effects in the SNMS methodology. Strong Al+ signals and minor Ti+ signals, without a proportional increase of the O+ signal, were observed in the hf plasma SNMS of an insulating, µm-thick, oxide layer on Ti–48Al–2Cr–2Nb.303 These were shown to be due to the secondary ions, which could form up to 10% of the detected ions in hf sputtering and plasma processes
Dreer et al.304 compared the performance of hf SNMS with SIMS, Auger electron spectrometry and hf GD-OES for the quantitative, sputter depth profiling of non-conducting silicon- and aluminium oxynitride films. The methods of choice for depth profiling proved to be hf SNMS and SIMS. Athough hf GD-OES had the highest sputter rate, it varied strongly with the oxygen content. Comparing hf SNMS with SIMS, the former featured the major advantage of a small and linear dependence of RSF on the oxygen content in the film. Even the use of the MCs+ method in SIMS yielded larger matrix effects than SNMS. Using proper calibration, SIMS and SNMS depth profiles and concentrations agreed within 10%. The sensitivity of SIMS was, however, superior.
8 Stable isotope ratio mass spectrometry (SIRMS)
8.1 Reviews
Compound-specific isotope analysis (CSIA) by GC-combustion-SIRMS is a unique tool that can provide answers to biochemical and physiological questions unobtainable by other analytical methods. It has continued to generate the majority of the SIRMS reviews in the period covered by this Update although most were focused on applications rather than instrumentation. The review of Meier-Augenstein305 was principally on the use of the technique in nutrition and metabolic research but also considered practical issues concerning GC-combustion-SIRMS. It included a useful table summarising typical features and specifications of MS systems used for stable isotope analysis. The 191-reference review of Lichtfouse306 can be recommended for its coverage of a wide range of applications in various scientific fields. Although there was scant coverage of instrumentation, the review of early developments set the context in which current novel applications became possible. The short review of Douthitt307 covered the present status of GC-SIRMS and considered future developments. The advantages that Brazier and Elbast308 considered continuous flow SIRMS to have for biomedical research included non-invasive methods in medical diagnosis and the use of very low levels of labels and naturally enriched tracers for metabolic studies.8.2 Instrumentation
A new inlet system, developed by Leuenberger et al.,309 was based on the well-known and simple open-split design used in GC-MS. The design worked both as a pressure control device and as a change over device for admission of sample and reference gas. Several advantages were identified over the conventional double inlet system with a metal bellows. These included improved reproducibility due mainly to highly controllable pressure and temperature adjustment, a markedly lowered memory effect due to uninterrupted gas flow through the ion source and removal of the asymmetric sample and reference inlet paths through use of a single inlet capillary. In particular, the constant gas admission to the ion source circumvented large surface adsorption and desorption effects which otherwise led to gas fractionation and required a long time for equilibrium conditions to be re-established.The introduction of a retardation lens into the Faraday cup of a 3 kV GC-combustion-SIRMS instrument has previously been shown to be the key to solving the challenge of measuring 2H∶1H ratios in helium carrier gas. Ruff et al.310 have used this method in the first reported measurement, by on-line GC-combustion-SIRMS, of 2H∶1H ratios in the important flavour compound, benzaldehyde. The measured 2H∶1H ratio was dependent on sample size but reproducible linearity was obtained for >0.6 µg of benzaldehyde injected onto the column.
Traditionally, SIRMS has been dominated by magnetic sector instruments, which provide optimum performance with multiple fixed cup detector arrays. The improved resolution and precision of modern quadrupole instruments has made them a potential alternative, however, and their use has been investigated by several researchers. In an innovative approach to developing cheaper methods for metabolic studies, Lecchi and Abramson311 coupled a benchtop quadrupole instrument to a chemical reaction interface in which amino acids reacted with a reactant gas (N2) to form molecular hydrogen. The use of 2H rather than 13C or 15N simplified synthesis of enriched compounds and reduced the cost of the study. In an on-line, continuous flow system, Russow and Schmidt312 used a reaction interface to convert ammonium (into N2 and nitrite) and nitrate (into NO) for the measurement of 15N tracers in soil extracts and aqueous samples. The achievable precisions (≤3%) did not compare well with traditional methods but were considered adequate for measurement of enriched 15N in nitrite and nitrate. Neither sensitivity nor precision was, however, adequate for the measurement of 15N in ammonium. Kupka and Winkler313 reported on the unique combination of an elemental analyser with a quadrupole instrument for simple, low cost and simultaneous determination of 12C∶13C and 14N∶15N ratios. Rosenorn et al.314 constructed a quadrupole instument for the isotopic analysis of CO2 and O3 and found the accuracy to compare well with that of other methods used (FTIR and MS-MS).
8.3 Analytical methodology
The development of laser-assisted fluorination systems has continued. Bao and Thiemens315 have described a novel method for the simultaneous determination of δ18O and δ17O in O2 generated directly from BaSO4. A reaction time of only 2–5 min was required for complete reaction of a barite sample heated in BrF5 (25–30 Torr) with a CO2 laser. Although the generation of O2 from barite was not quantitative, the δ18O value was independent of the O2 yield. Within-run precisions were <±0.8‰ and <±0.05‰ for δ18O and δ17O, respectively. The δ18O values were, however, consistently 9.4 ± 0.8‰ lower than values obtained by a graphite reduction method in which there was almost 100% conversion of sulfate into CO2. The need to apply a correction factor resulted in the optimum sample size being larger than 4 mg. In addition, the dependence of yield on sample particle size resulted in a requirement for samples to have a particle size of <5 µm. The analysis time of <30 min was in marked contrast to the days required by the graphite reduction method. The procedure was seen to have particularly strong application to the study of mass-independent isotopic anomalies in atmospheric molecules. Jones et al.316 found that UV LA and fluorination provided a combination of spatial (100 µm) and analytical (±0.4‰, 1σ, for δ18O) precision previously unobtainable for the in-situ sampling of apatite in tooth enamel.Monitoring the mixing ratio of CO2 using its isotopic composition allows estimates to be made of the spatial and temporal variations in, and relative magnitudes of, biospheric and oceanic carbon sources and sinks. Ferretti et al.317 have reported the first N2O-free, high precision (<0.05‰) measurement of δ13C and δ18O in atmospheric CO2 from small air samples (45 ml). This sample volume requirement was up to three orders of magnitude lower than that required using a dual-inlet instrument. On-line GC separation of CO2 and N2O from a whole air sample was combined with SIRMS using elevated ion source pressures. A specialized open-split interface was an integral part of the interface system and ensured a continuous flow of either sample gas or pure helium to the mass spectrometer. Precisions were 0.02 and 0.04‰ for δ13C and δ18O, respectively.
SIRMS has considerable potential for the on-line simultaneous elemental and isotopic analysis of samples. Tsunogai et al.318 have developed a continuous flow method for the determination of CO in small volumes of natural waters. Sample preparation was complex and led to an analysis time of 40 min per sample. The LOD for CO concentrations were 300 and 750 pmol using δ13C and δ18O, respectively. A wide variation in δ18O in the blank CO2 was considered the main reason for the poor reproducibility of δ18O and the use of oxidizers with lower blank values than those used by the authors or the direct introduction of CO into the mass spectrometer were considered necessary to improve precisions. Although Leuenberger et al.319 considered a new SIRMS technique for measuring CO2 concentrations in air samples to be feasible with performance as good as that achieved by the GC method, a number of operational problems were identified. The collector configuration of the instrument allowed only three ion beams to be focused simultaneously, so a peak jumping mode, less precise than simultaneous measurement, had to be employed. Some analyses with a new wide mass range instrument offering simultaneous measurement led to improved reproducibilities of <0.1‰. Superimposed on background problems, caused by preferential adsorption of CO2 on source components, were effects originating from the production of molecules in the ion source. There was evidence for the production of N2O and NO2 from excited N2 and O2, which yielded significant signals at m/z 44 and 46. Despite this, the authors considered these to be minor problems and identified the gas admission procedure, resulting in mixing of sample and standard gases and therefore a memory effect, as the main limitation on improved reproducibility.
Despite the importance of O2 in biogeochemical processes, there have been relatively few reported stable isotope studies for assessing the origins and cycling of the gas. A major limitation has been the prohibitive cost and time requirements associated with dual-inlet analysis. A new method employing continuous-flow SIRMS was reported in last year's Update1 and now Roberts et al.320 have presented a GC-SIRMS technique for the determination of δ18O in dissolved and gaseous O2. The new approach utilized an evacuated sample loop, which avoided the tendency for syringes and septa, used in the continuous flow method, to leak. The new procedure used a 0.5 nm molecular sieve column held at a constant 50°C to separate N2 and O2. A precision of ±0.3‰ or better for δ18O was achieved for gaseous and dissolved
samples spanning an environmentally relevant concentration range of 20–700 µM. The method also provided data for δ17O with a precision of ±0.5‰.
Headspace equilibrium analysis is an effective method for the sampling of H2. Gerst and Quay321 have reported a new technique for the determination of δ2H in atmospheric H2. Air samples, compressed into a high-pressure cylinder, were cooled to –192°C to condense the air and generate a headspace highly enriched in 2H. The 2H was separated from other gases by combustion and the water produced reduced back to 2H for MS analysis. The primary limitation of the method was the contamination experienced during sampling and storage in the high-pressure cylinders. In an extension of the method using equilibrium headspace analysis to determine δ13C in toluene, Ward et al.322 established that the determination of δ2H using continuous flow CSIA was also possible and was accurate and reproducible
to within ±5‰. A small absolute shift was observed in the δ2H value for a working standard obtained by headspace extraction in comparison with that obtained by direct injection. This was attributed to a small isotopic fractionation occurring during volatilization.
The paper of Salata et al.323 on the measurement of 13C in dissolved inorganic carbon
(DIC) gave a useful summary of the advantages and disadvantages of methods previously developed for the analysis. An alternative method was described for the rapid preparation and automated analysis of samples using a simple and inexpensive valve system. Carbon dioxide was liberated from samples by H3PO4 in an evacuated GC injection vial and, after a 15 h equilibration period, the vial headspace gas was injected into a GC-combustion-SIRMS instrument. Although no combustion was necessary, the reactor of the instrument was maintained at 940°C to ensure stable run conditions and to ease movement of the analyte through the reactor. The calibration graph was linear over a relatively narrow range corresponding to DIC concentrations of 0.2–4 mM.
Kracht and Gleixner324 used on-line thermal degradation of whole peat samples (0.2–1 mg) by flash pyrolysis to extract volatile products which were transferred to a GC-combustion-SIRMS instrument. The combination of structural information from pyrolysis GC-MS and isotope content enabled the authors to distinguish between five different biogeochemical processes of humification.
The use of SPMEfor the analysis of flavour compounds produced by lactic acid bacteria was evaluated by Goupry et al.325 using both liquid and headspace samples. It was shown that neither the analyte concentration nor the period of fibre exposure affected the measured δ13C values, but a small but consistent fractionation in the δ13C values was observed during extraction. Two of the five tested fibres were considered suitable. Long exposure times with a carboxene-poly(dimethylsiloxane) fibre were considered appropriate for dilute samples (≤0.1 mM) whereas short exposure times with a poly(dimethylsiloxane)–divinylbenzene fibre were used for concentrated samples (≤2 mM). An analytical precision of ±0.4‰ was achieved in the first reported analysis of volatile metabolites in a fermentation medium.
In a pair of detailed and well-presented papers, Coleman and colleagues described studies into the SIRMS of stable halogen isotopes and their potential applications. Firstly, Rosenbaum et al.326 modified a dual inlet method to measure the 37Cl∶35Cl ratio by monitoring CH337Cl+and CH335Cl+ at masses 52 and 50, respectively. Comparison with a TIMS method for the determination of the 37Cl∶35Cl ratio showed that results were in agreement, but that, in general, the performance of SIRMS was poorer with a 100-fold larger sample requirement and a 10-fold poorer reproducibility (≤0.1%, 2σ). The SIRMS technique was considered, however, to be better suited to the analysis of larger samples (>10 µmol Cl). No analytical methods have been published on the determination of Br ratios using SIRMS, probably due to expected analytical difficulties and the expectation that Br would behave in natural systems in a manner similar to Cl. In a method analogous to the Cl method reported above, Eggenkamp and Coleman327 measured, with a precision of <0.18‰, the Br ratio in pure CH3Br prepared from samples containing 2–5 mg Br. Samples were analysed on a dual-inlet, triple collector SIRMS instrument with the 81Br∶79Br ratio measured by monitoring m/z 96 and 94. The mass spectrometer was not designed for measuring isotope ratios close to unity so that, under normal operating conditions, the beam at m/z 96 saturated the collector. This was overcome by reducing the trap current to 100 mA to make the source less sensitive and to reduce the minor beam current to <10−10 A. Analysis of Br in natural samples revealed significant isotopic variations, which did not mirror the isotope variations found in the related element Cl.
8.4 Sample preparation
Conventionally, water and other liquid samples are prepared for 2H analysis by an off-line zinc reduction method. This method is laborious, expensive and time-consuming and better suited to samples at natural abundance rather than for enriched samples for which the required accuracy is poorer. Considerable attention has been given in recent years to alternative methods for the reduction of water. Ward et al.328 have described a method based on the reduction of small volumes (10 µl) of water to H2 by LiAlH4 in individual vials, from which gases were drawn directly into a dual-inlet mass spectrometer by an automated gas acquisition system. Results obtained by this method were more accurate and precise than those obtained by either zinc or uranium reduction. The method was considered most appropriate for the analysis of 20–50 µl of enriched samples. An equilibration method was preferred for large (0.25–1 ml) samples whereas pyrolysis was considered better for very small (<1 µl) samples at natural abundance. An automated platinum reduction and equilibration method for the analysis of enriched samples (200 µl) was found by Herd et al.329 to be less costly and less labour-intensive than the zinc reduction method yet to produce more reliable data (precisions of ±2.07 and ±4.36‰ for the platinum and zinc methods, respectively). A fully automated chromium reduction system, previously developed for the analysis of water samples and now commercially available,330 has been applied by Schoeller et al.331 to the determination of δ2H in physiological fluids (1.2 µl). Correction for H3+ interference was performed mathematically using a factor that was determined daily. The within-day and between-day precisions of 0.5–0.8 and 2.5–6.4 ‰, respectively, were similar to those obtained by other methods. However, the method was unable to cope with high levels of hydrogen-containing organic solutes for which an upper limit of 2–5 g l−1 was placed. In addition, memory effect was a disadvantage of the on-line system such that replicate injection of samples could be required to reduce the effect.Meijer and Li332 used a new system based on electrolysis with CuSO4 as electrolyte for the determination of δ17O and δ18O in water samples and found it to be an excellent alternative to the traditional water–CO2 equilibration method. The accuracy obtained was 0.07 and 0.10% for δ17O and δ18O, respectively.
A simple extraction technique, described by Stickrod II and Marshall333 for the determination of δ15N in nitrate from groundwaters, was based on the ion-exchange extraction of nitrate nitrogen and its conversion into solid AgNO3 which could be analysed rapidly using continuous flow SIRMS. The precision of analysis was <0.1‰ even with high concentrations of carbonate. Analysis of stored AgNO3 showed slight isotopic depletion over time so it was recommended that the δ15N analysis be performed within several days of preparation of the AgNO3.
The widespread adoption of GC-combustion-SIRMS has required sample preparation techniques for biochemical analysis to be developed. Metges and Daenzer334 found that the isotopic fractionation introduced by derivatization of amino acids as their N-pivaloyl isopropyl esters was reproducible and that empirical correction factors could be derived for each individual amino acid. The mean analytical error for δ13C was 0.26‰. A rapid screening assay, developed by Aguilera et al.335 for the measurement of δ13C in urinary androsterone and etiocholanolone, was based on deconjugation with β-glucuronidase, solid phase extraction and derivatization with acetic anhydride and pyridine. Bourgogne et al.336 have reported a detailed procedure for the analysis of exogenous hydrocortisone and cortisone in urine. As urinary levels of hydrocortisone were too low for direct analysis, the focus was on the main metabolites, tetrahydrocortisone (THE) and tetrahydrocortisol (THF). After purification by different solid phase extractions, THE and THF were oxidised to 5-β-androstanetrione for analysis.
Contamination in SIRMS is an important issue. Meijer et al.337 have presented an extremely detailed treatise on cross contamination (of CO2 samples with reference gas and vice versa) in dual inlet instruments and have discussed why and how it should be corrected. A method to measure actual contamination was presented as well as algorithms to correct it. The cross contamination, caused not by leaks but by surface adsorption and desorption effects in the inlet system and ion source, was considered to be the explanation for poor results in interlaboratory ring tests which were analysed in considerable detail. Cross contamination was also considered to be a valuable diagnostic tool, which could be used to optimize the instrument performance. Nelson338 found that Na glass and borosilicate glass vials, equipped with butyl rubber septa, could cause significant changes in the isotopic composition of CO2 gas. The magnitude of the changes varied from vial to vial. This was of particular significance for continuous flow SIRMS analysis because of the much smaller gas volumes used than in dual inlet systems. The source of the problem remained unidentified so caution was advised in the analysis of small gas samples.
9 Thermal ionization mass spectrometry (TIMS)
9.1 Instrumentation
Kostoyanov et al.339 have rebuilt a Russian commercial instrument in order to perform isotopic analysis of refractory metals using negative ionization. Basic details were given of the controlling hardware, software and recording system. Test analyses of in-house OsO2 standards gave a reproducibility for 187Os∶188Os and 189Os∶188Os of 0.3%, an order of magnitude poorer than the values reported by other laboratories.McCulloch and Esat340 used a charge capacitance system rather than high ohmic resistors in the electrometers for a multiple Faraday cup array for the measurement of 229Th, 230Th and 232Th. Low intensity ion beams of 10−13–10−16 A could be measured simultaneously. For uranium isotopic composition measurements, combined Faraday cup and electron multiplier modes were used with 235U being measured in both modes to correct for the multiplier gain.
9.2 Calibration
The paper of Thirlwall341 is a tour de force giving a very detailed discussion of the sources of analytical error in Pb isotope analysis, in particular those due to mass fractionation. Correction for mass fractionation based on the use of the SRM NIST 981 was considered unsound for accurate analysis because the Pb in the standard fractionated in a different manner to Pb in real samples. The error sources in an alternative method using a 207Pb–204Pb double spike were assessed in depth taking into consideration the results for 33 analyses of NIST 981 over a period of a year. The main sources of uncertainty, in order of magnitude, were contamination, mass fractionation, anomalous behaviour of 207Pb, uncertainty in the measurement of 204Pb, collector calibration in static data collection and the fractionation correction procedure. Precisions (2σ) for NIST 981 were <0.01% for 207Pb∶206Pb and <0.02% for ratios measured to 204Pb. Although these levels of analytical performance are impressive, the cost and time requirement of the double spike method can probably only be justified when such precisions are essential.As an alternative approach to the problems associated with the use of NIST 981, Woodhead and Hergt342 have proposed the practical use of rock standards with known lead isotope composition as ‘geologically realistic’ RMs. They used a double-spike procedure to establish the lead isotope composition in six USGS reference materials, which could be matrix-matched to samples of unknown composition. An interesting outcome of the study was that the first generation RMs (e.g., BCR-1) experienced some form of gross contamination during preparation and therefore had significantly different isotopic compositions from the second generation RMs (e.g., BCR-2) even though the latter were sampled from the same location as the former.
Mass fractionation effects in the negative ionization analysis of B were corrected by You et al.343 by normalizing the 11B∶10B ratio to an empirical 18O∶16O ratio. Four BO2− peaks at m/z 42–45 were measured. It was found that mass fractionation caused both the 11B16O18O∶10B16O16O and 11B16O16O∶10B16O16O ratios to vary linearly with the 11B16O18O∶11B16O16O ratio so the last ratio was normalized to a fixed 18O∶16O ratio. The measured precison (2σ) for a NIST SRM was ±0.11‰, similar to that achievable by positive ionization methods.
9.3 Analytical methodology
The 10B∶11B and 6Li∶7Li ratios can theoretically be measured to better than 1‰ precision by measurement of multiple peaks of the Li2BO2+ ion using ion counting detection. Bickle et al.344 found, however, that the relatively large Li+ beam emitted simultaneously with the smaller Li2BO2+ beam caused the Li remaining on the filament to fractionate. The precision (1σ) of repeat analyses of 100 ng Li loads was no better than 3‰. The authors concluded that it was not possible to correct for fractionation resulting from ionization as Li+. It was considered that improved analysis would be possible only if a molecular complex could be ionized with two or more masses without an accompanying Li+ beam. In the analysis of nine rock RMs, James and Palmer345 achieved a precision (1σ) of <0.3‰ for 6Li∶7Li by measuring the Li+ beam itself. Significant mass fractionation was observed at filament temperatures below 850°C but between 850 and 1200°C the 6Li∶7Li ratio gave excellent reproducibility. Refinements were also made to the sample preparation to achieve good separation of Li from Na using a small resin volume (2.7 ml) and sample size (100 ng Li). In a series of theoretical studies on errors in the Li2BO2+ method, Datta346−348 concentrated on the selection of pairs of isotopomers to be monitored in the analysis and the effect on propagated errors. A problem identified by Liu et al.349
in the measurement of boron isotopes was CNO− isobaric interference derived from nitrate in the loading reagent.The emission of Cs2BO2+ and Cs2Cl+ in the measurement of B and Cl isotopes, respectively, can be enhanced significantly by coating the tantalum filaments with graphite (75 µg). Previous studies showed that the properties of the graphite affected the thermal ion emission but the optimum graphite characteristics had not been established. Xiao et al.350 have studied six graphite samples from different sources and one high-purity carbon powder and concluded that graphite with small maximal distortion of the crystal lattice, corresponding to a perfect crystal lattice, was best for high-precision analysis. Such graphite resulted in a stable ion beam for several hours and did not cause fractionation during analysis.
Magnesium isotopic fractionation in barred olivine chondrules from a meteorite has been measured by Misawa and Fujita351 using single collector peak jumping. A direct loading procedure for the determination of Mg isotope ratios, in which spinel samples were loaded directly on to a V-shaped rhenium filament without chemical separation, gave δ26Mg values about 1.6‰ lower than those obtained after chemical separation using a cation-exchange procedure. A mean fractionation-corrected 26Mg∶24Mg ratio of 0.139819 ± 0.000009 (2σ) was obtained for a set of terrestrial samples.
There have been few reported applications of TIMS to plant studies. Procedures have been developed by Midwood et al.352 to analyse plant tissues enriched in 26Mg and 41K, used as tracers in studies of uptake and internal cycling. Magnesium isotope analysis was insensitive (in the range 0.2–1.2 µg Mg) to the amount loaded on to the filament. The 26Mg∶24Mg ratio in a NIST SRM was measured as 0.13960 ± 0.00006. Filament loadings of 5 µg K were recommended to obtain a stable 39K∶41K ratio and minimize errors due to fractionation. Although measurement precision was 0.2%, the value obtained for the 39K∶41K ratio in a NIST SRM was 0.4% higher than the certified value. The issue of contamination and filament blanks for this easily ionized element
was not addressed.
Banks et al.353 have shown that the method for determination of δ37Cl by TIMS, using the Cs2Cl+ ion, is applicable to the analysis of Cl in fluid inclusions for which only small samples are available. Modifications to the method improved the precision and made it possible to determine small differences between samples of about 0.2‰. The sample preparation was modified to include four ion-exchange columns in order to improve the removal of F− and SO42− which interfere in the MS. Prepared samples (10 µg Cl) were loaded on to a tantalum filament and covered with graphite (50 µg). It was important to keep a constant Cl loading and Cl∶graphite ratio as these affected the Cs237Cl+∶Cs235Cl+ ratio. Rosenbaum et al.326 have used the same procedure for the analysis of waters and compared the analytical performance with a SIRMS procedure. Although both methods gave the same results, the TIMS method was superior by orders of magnitude in terms of sample requirement and reproducibility. Sample loadings of 25–35 µg Cl gave a reproducibility for δ37Cl of ≤0.2‰ (2σ). These authors also made the observation that measured ratios were strongly dependent on the amount of CsCl loaded on the filament. Mass fractionation, attributed to breakdown of the Cs2Cl+ ion into Cs+ and Cl− ions, was also observed.
Double-spike procedures are being widely adopted to overcome problems of isotopic fractionation during sample preparation and isotopic analysis. Nägler and Villa354 have presented a double spike (43Ca–48Ca) procedure for the processing and isotope analysis of very small (<100 ng) amounts of Ca derived from K-rich silicates. Uncrushed rock samples (≤90 mg) were used in order to avoid potential contamination problems. For analysis, samples were loaded (2 µl as chloride) on single rhenium filaments together with a Ta2O5 solution (1 µl). Measurement in a single collector peak-jumping mode gave an internal precision of typically 0.04% (2σ). A memory effect on the electrometer caused by the high intensity 40Ca+ beam was monitored by measuring a peak intermediate between m/z 40 and the following baseline. Herbel et al.355 have used a double spike (82Se–74Se) procedure and negative ionization to quantify the isotopic fractionation during bacterial respiratory reduction of selenium oxyanions. Estimated precision of the isotope ratios was ±0.2‰ at the 95% confidence level.
A simple method for the precise (<±0.5‰ at the 95% confidence interval) measurement of Mo isotope ratios was considered by Wieser and De Laeter356 to be an improvement over previous methods for several reasons. The sample deposition procedure was uncomplicated and involved placing an ascorbic acid solution containing Mo (20 ng to 2 µg), extracted from the sample by ion-exchange procedures, on a single rhenium filament and adding silica gel. The beam intensity and stability and fractionation proved to be consistent from sample to sample and a loading of 1 µg typically gave an ion current of 1 × 10−12 A, which was sustained for several hours. Although checked before each analysis, Ru and Zr isobaric interferences were found to be negligible. Fractionation corrections were based on measured data and not on previous, uncalibrated measurements.
Although TIMS has not previously been used for the ID determination of the very low concentrations of Cd in benthic forameniferal shells, Rickaby et al.357 found the method to be very reliable, precise and sensitive. Attention to the chemical separation of Cd from the calcium carbonate matrix and optimization of the loading procedure allowed a reproducibility of ±0.025 µmol mol−1 in the Cd∶Ca ratio to be obtained. Miniature ion exchange columns (3.2 mm id) were designed to enable Ca from typical benthic foramenifera (200 µg Ca after cleaning) to be loaded on a cation-exchange resin and Cd to be eluted quantitatively without risk of Ca breakthrough. Fresh resin was used for each sample to avoid the risk of incomplete resin cleaning and contamination by Ca. The blank of 1 pg permitted analysis of individual benthics and of planktonic foraminifera. Reproducibility of filament loading was improved by mixing the sample, dissolved in 1 µl of 0.2 M HNO3, with 1 µl silica gel in 0.25 M H3PO4 prior to loading. Using this procedure, a stable 116Cd ion beam of 3–5 × 10−14 A could be maintained for 2 h from a clean 40 pg loading of Cd.
The determination of the light REE by TIMS is usually made by measuring the oxide ion but this requires corrections to be made for isobaric interferences from 17O- and 18O-containing oxides. To remove the interferences, Liu et al.358 have used a method, previously developed for Re–Os analysis, in which the natural oxygen in the source chamber was replaced with 16O-enriched oxygen bled in at a pressure >1 × 10−6 mbar. At this pressure the oxygen in the oxide species was derived predominantly from the gas in the source chamber rather than from oxygen-containing compounds loaded on the filament. Precise (0.035–0.1‰ for Nd isotope ratios) and accurate determination of Ce, La and Nd isotope ratios could be made without correction for oxygen isobaric interference.
Although reported in the Chinese language, Deng and Wei359 have described a novel method for achieving enhanced sensitivity of TIMS analysis of Pu and U. The analytes were electroplated on to a single rhenium filament then overlaid with a thin layer of platinum. The sample atoms diffused thermally through the platinum layer before they could evaporate as atoms or ions. Because thermal diffusion is generally a slower process than direct evaporation and the platinum layer prevented the rapid escape of neutral species, the filament could be operated at a higher temperature than previously, giving higher ionization efficiency without loss of sample. Wallenius and Mayer360 used a total evaporation procedure to determine isotope ratios in very small amounts of Pu (6 ng) and U (40 ng) with a precision better than 0.2%. As all the sample was collected, no correction for mass fractionation was needed. This methodology allowed the age of plutonium material to be determined accurately using parent-daughter isotope relationships.
9.4 Sample preparation
Use of Na2BO2+ rather than Cs2BO2+ for the isotopic analysis of B has the advantage of keeping the measured ions within the mass range of commercial instruments, but suffers from problems caused by the presence of excess NaOH following conversion of B to Na2BO2. Ahmed et al.361 found that the simple step of neutralizing excess NaOH on the loaded filament improved measurement precision from 4 to 0.3‰. In addition, no mass fractionation was observed while maintaining the ion beam over a period of more than 6 h.The simple, two-step ion-exchange method for separating CrVI from natural waters, reported by Ball and Bassett,362 was based on the fact that CrVI exists as an anion but CrIII exists as a cation. The procedure, for which full details were given, avoided problems of incomplete retention of CrVI on the anion-exchange resin and of CrIII on the cation-exchange resin. Subsequent processing of the separated sample eliminated residual organic material prior to loading on the MS filament, which was warmed up slowly to allow metals with boiling points lower than Cr to be removed. Although analysis of a NIST SRM gave a reproducibility of 0.4‰, further improvements were required to reduce interferences that restricted the life of the Cr+ ion beam and to minimize mass fractionation.
Gioia and Pimentel363 have modified previously described methods to achieve clean extractions of Nd and Sm from geological materials. The method was based on the separation of the REE group using a cation-exchange resin followed by reversed-phase LC of Nd and Sm on the commercial LNSPEC resin [PTFE impregnated with di-(2-ethylhexyl)hydrogenphosphate]. Complete separation of Nd and Sm was achieved and Gd, the main isobaric interferent on Sm, was not detected. The external precision for the 143Nd∶144Nd ratio in three rock RMs was in the range 0.006–0.016‰.
A very sensitive method (LOD of 0.4 ng l−1 for 500 ml samples) developed by Schedlbauer and Heumann364 has allowed dimethylthallium to be detected in environmental samples for the first time. Thallium was extracted from sea-water using species-unspecific anion exchange, the dimethylthallium separated using species-specific extraction of inorganic Tl into MIBK and the Tl determined by a positive ionization ID TIMS procedure. Some 3–48% of Tl in sea-water was found to exist in the methylated form which was concluded to be of biogenic origin.
You and Bickle365 investigated the use of the TRUSPEC resin in acidic media for the separation of Th and U and subsequently applied it to the high-precision 230Th∶234U dating of sulfides and carbonates. In comparison with conventional ion-exchange methods, the new chromatographic procedure dramatically reduced the time for separation by more than half, and avoided several complicated steps, including Fe(OH)3 precipitation and ammonium titration. Column size could be reduced, thereby minimizing acid consumption. The ability to measure both abundance and isotopic composition with the same filament loading reduced sample preparation and instrument time. To avoid any possible memory effect, each TRUSPEC column (0.45 ml bed volume) was prepared freshly and was used only once. The analytical precision for the measurement of both Th and U concentrations was <1% and determination of the 235U∶238U ratio in a RM was precise (0.2%) and accurate (0.003%). Cheng et al.366 used a modified method, based on a series of successive scrubbing and oxidative leaches, to clean deep-sea corals for U–Th dating. The method was designed to remove exterior, modern contaminants from fossil samples and thereby improve accuracy and reproducibility.
10 Other methods
10.1 Electrospray mass spectrometry (ESMS) and ion spray mass spectrometry (ISMS)
There has been a degree of rationalisation in the use of ESMS for metal and metalloid analysis with the focus being strongly on the identification of molecular species. Use of ESMS for elemental analysis appears to have been all but abandoned as a viable technique. Four substantial reviews by Szpunar, Lobinski and colleagues,367−370 all essentially on the same subject of bio-inorganic speciation by hyphenated techniques, set the context within which ESMS has found a role. That of Szpunar,367 with 375 references, can in particular be recommended as it gives a detailed yet critical coverage of the whole subject area with summaries, in tabular form, of all reported applications. Although the review of Chassaigne et al.370 was the only one specifically on ESMS, all four reviews emphasised the importance of identifying unambiguously the identity of chromatographic peaks. This could not be done by retention times alone hence the role of ESMS. A number of shortcomings were identified. One was the incompatibility of the LOD of ESMS (10–100 ng ml−1) and ICP-MS (0.1 ng ml−1), which precludes parallel running of the two techniques. The sensitivity of ESMS to the presence of salts that suppress ionization require LC (SEC or reversed-phase) with salt-free buffers. A major challenge remains the development of sample preparation procedures that allow biochemical speciation in solid samples to be studied.The key question in ESMS in general is, what is the process by which gaseous ions are formed from charged droplets? There are two main schools of thought, which support either the charged residue model (evaporation of solvent from droplets to leave a charged molecule) or the ion evaporation model (release of solute ions from the droplet surface). There is, however, some evidence that both are applicable depending on the size of the ion. Although not specifically on the ionization of metals and metal-containing species, the substantial reviews of Kebarle and Peschke371 and Kebarle372 gave very good coverage of the mechanisms of ion formation in the electrospray and assessed critically the merits of the two models.
Two groups very active in the study of selenium speciation, in particular in Se-enriched garlic and yeast, both used ESMS identification of the Se species. Ip et al.373 used HPLC on-line with ESMS to identify the two major selenium compounds present in garlic and yeast as γ-glutamyl-Se-methylselenocysteine and selenomethionine. This was the first time that >90% of the chemical form of Se in a Se-enriched natural product could be identified. McSheehy et al.374 used SEC to isolate Se-containing amino acids from aqueous extracts of garlic samples followed by off-line reversed phase LC to remove salts. A single high-purity fraction identified by ESMS as γ-glutamyl-Se-methylselenocysteine was produced. A feature was the use of a triple quadrupole MS instrument to carry out collision-induced dissociation (CID) MS-MS analysis. Although the title of the paper proclaimed parallel ICP-MS and ESMS detection, this was not strictly so in that simultaneous analyses were not performed. Initial attempts by Shou et al.375 to measure an organoselenium compound (4-hydroxy-2-methyl-2-aminoethyl selenide) by direct ESMS were unsuccessful because of excessive fragmentation even under the mildest of source conditions (low extraction energies). This was overcome by post-column introduction of a crown ether which formed a complex with the Se compound and survived the ES ionization process with reduced fragmentation. The new procedure increased sensitivity 5-fold and gave a LOD of 5 pg µl−1 in urine.
Application of ESMS to arsenic speciation has seen development of a number of different approaches. Madsen et al.376 verified the identification of arsenosugars determined by HPLC-ICP-MS by using LC-ESMS with a variable fragmenter voltage which provided both elemental and molecular detection. In addition, quantification of the four arsenosugars gave values within 5% of the values given by ICP-MS. The LC eluent (20 mM NH4HCO3) contained 10% methanol in order to aid the electrospray ionization process. The use of a single quadrupole instrument meant that there were some problems in obtaining unambiguous identification of some of the spectral peaks. McSheehy et al.377 overcame this problem using CID of relevant protonated molecules to identify dimethylarsinoylriboside derivatives extracted from seaweed. The arsenosugar fraction was eluted from the size exclusion HPLC prior to the majority of other arsenic compounds. Wu et al.378 have reported the first use of a polypyrrole-coated capillary for the automated SPME of organoarsenic species in aqueous samples. By coupling to LC-ESMS, organoarsenic compounds in water samples and arsenobetaine in a CRM could be determined. Demesmay and colleagues379,380 have coupled capillary zone electrophoresis (CZE) with ion trap ESMS in studies of arsenic speciation. The end of the CZE capillary was concentric with the electronebulization needle of the MS and make-up liquid (methanol–H2O, 1∶1) was introduced between the two. The make-up liquid was added not only to enable electric contact to be made at the outlet of the separation capillary but also to assist the electronebulization process. Optimization of coupling conditions (geometry of the concentric interface, positioning of the separation capillary outlet, composition and flow rate of the sheath liquid, electronebulization and detection conditions) identified the first two as being critical for stability and sensitivity. The LOD ranged from 0.5 to 3.3 mg l−1 injected, corresponding to 15–60 pg, and the calibration curve was linear over the range 5–200 mg l−1. On-line preconcentration improved the LOD for arsenobetaine and arsenocholine to 110 and 160 µg l−1, respectively.
Mounicou et al.381 highlighted the attractive potential of CZE-ICP-MS and CZE-ESMS when used in parallel for the characterization of metal complexes with biomolecular ligands, in particular metallothioneins. They found that the mass spectra of CZE peaks were complex and indicated the co-existence of several metal complexes within each chromatographic peak. Although the complexes could in general be identified on the basis of their molecular mass, there remained ambiguity in identification. It proved particularly difficult to distinguish between Cu and Zn whose atomic masses overlap. Chassaigne et al.,382 of the same Pau group that reported the first study, found that the use of HPLC-ESMS to identify metallothioneins was more difficult than CZE-ESMS and that post-column acidification was necessary in order to increase sensitivity and precision of molecular mass determination. Polec et al.383 have reported the detection of two major Cd-containing metallothionein isoforms isolated from rat liver tissue using HPLC-ISMS. The Cd macromolecular fraction was isolated by SEC, de-salted, concentrated by lyophilization and analysed by microbore, reversed-phase HPLC with parallel ICP-MS and ESMS detection. The desalting and preconcentration (by lyophilization) were necessary for the successful application of ISMS because of its relatively poor sensitivity. The molecular mass of the ligand protein was determined following on-line, post-column acidification of the eluate to dissociate the complex.
The development of procedures for a wide range of speciation studies is an indication of the perceived value of ESMS for this type of work. Vacchina et al.384 used ISMS with CID to evaluate the identification of Cd-binding ligands in complexes extracted from plant cultures. A single HPLC-ICP-MS peak was demonstrated to contain a number of complexes and, although the particular ligands could be identified, the complexity of equilibria within the extracted fraction prevented elucidation of the stoichiometry of the complexes formed. Initial investigations by Jones-Lepp et al.385 on the determination of organotin compounds in water by combining extraction with solid-phase extraction discs with analysis by micro-capillary LC-ESMS showed the method to be sensitive with low-ppb detection. A main advantage of the procedure was the lack of derivatization and hydrolysis required in traditional GC and ICP-AES methods. The use of tropolone as an additive in the mobile phase improved the chromatography and helped to stabilize the ions formed in the electrospray process. Further developments of the method were, however, required to overcome the problems of electrospray instability. Speciation of actinides, lanthanides and fission products is a major challenge in the framework of nuclear fuel studies. Moulin et al.386 successfully applied, for the first time, ESMS to the detection of free uranyl UO22+, the monohydroxy UO2OH+ and the oligomeric species (UO2)3(OH)5+. Under appropriate experimental conditions, both chemical (solvent) and ESMS tuning (low cone voltage, high gas flow rate and temperature), it should be possible to perform semi-quantitative measurements of the free uranyl–monohydroxy complex. Using in-tube SPME and HPLC coupled to a quadrupole ES instrument, Mester et al.387 were able to achieve complete separation and detection of trimethyl- and triethyllead (TML and TEL, respectively) in aqueous samples. Precision was better than 5% with the LOD 11.3 and 12.6 ng ml−1 for TML and TEL, respectively.
The difficulties encountered when studying metal speciation in solutions have been highlighted by a number of studies. In a study of some anionic metal complexes, Moraes et al.388 found that, even under the mildest of ES ion-formation conditions, the tendency was to form singly charged atomic ions. This tendency was most marked for Cu and Fe for which the original oxidation state was not maintained and, under the conditions used, speciation was not possible. Quantitative analysis was problematic because of the ion suppression caused by competition between ions in the ES process. Although Mollah et al.389 considered negative ionization ISMS to have some potential for the identification of metal cations and anions, several problems were identified. Reduction of some metal ions, dependent on the metal concentration, occurred at the needle and made assignment of the oxidation state in solution difficult if not impossible. Multi-component systems led to complicated spectra and there were also problems with memory effects. In a study of alkali and alkaline earth species, Ross et al.390 concluded that analysis of electroactive species could be compromised by redox reactions associated with ES and that in-source fragmentation dramatically reduced the sensitivity towards labile species. The formation of spurious radical ions highlighted the importance of mild ES conditions for speciation analysis.
Elemental analysis by ESMS in its current state of development is not competitive with ICP-MS principally because of issues of cluster formation with the solvent and poor sensitivity. The spectra obtained are highly dependent on operating conditions. Shou and Browner391 found that, if suitable in-source CID conditions were chosen, complexation with a crown ether, added post-column, resulted in simplified spectra free from solvent cluster interferences. The improved hydrophobicity of the metal–ether complexes helped ion evaporation and the soft ionization conditions allowed the high-transmission ion cluster mode to be used. Oxidation state information of singly- and doubly charged metals was retained in the metal–crown ether complex ions. Alkali metals could be detected either as the metal–ether complex ions or as the free metal ions following dissociation of the complexes by in-source CID. Alkaline earth and transition metals were detected as the complex ions. The aim of the study was to demonstrate improved performance through use of complexation and the analytical figures of merit of the new system remained to be assessed.
10.2 Gas chromatography-mass spectrometry (GC-MS)
Speciation analysis has been one of the fastest progressing techniques of instrumental analysis in the last decade and this, together with improvements in analytical methodology, has led to a renaissance of GC-MS as a viable technique for determination of organometallic compounds in natural samples. The popularity of sodium tetraethylborate (NaBEt4) as a derivatization agent has grown over the last ten years and has been used for the determination of organomercury,392 organolead393 and organotin394–396 compounds. The review of Takeuchi et al.397 on the determination of organotin compounds in environmental samples includes a substantial section on the use of NaBEt4. Derivatization with NaBEt4 is carried out in aqueous solution, a big advantage over the traditional use of Grignard reagents, and can be easily incorporated into automated on-line systems. A disadvantage of NaBEt4 for the determination of organotin compounds is its sensitivity to matrix interferences and the low volatility of phenyltin compounds, which restricts analysis to the butyltin compounds.The use of SPME also has given impetus to the application of GC-MS to speciation studies. Beichert et al.392 have developed a method in which methylmercury, extracted from soil samples using subcritical water (held at high temperature and pressure to reduce polarity), was re-extracted during the derivatization with NaBEt4 using headspace SPME. By sampling the headspace, only the species of interest and no matrix components were extracted. The benefit of the new method was that both matrix interferences and the formation of artifactual methylmercury were avoided. Further development of the method, in particular for improvement of recovery, was, however, required for it to be suitable for the measurement of methylmercury at low concentrations in soil samples. Yu and Pawliszyn393 also used headspace SPME. They extracted alkyllead derivatized with D-labelled NaBEt4 in a method designed for the monitoring of tetraethyllead in waters and considered easily adaptable for field analysis. Mester and Pawliszyn398 used SPME to extract methylarsenic compounds from human urine following derivatization with thioglycol methylate. The LOD for di- and monomethylarsenic acids of 0.12 and 0.29 ng ml−1, respectively, compared favourably with those achievable by ICP-MS. The calibration was linear over the range 1–2000 ng ml−1. Guidotti399 found that headspace SPME was more sensitive than (LOD of 81 ng l−1) and as precise (10%) as direct SPME for the determination of selenite in drinking water. The selenite was selectively derivatized with NaBEt4 prior to extraction.
Ndungu and Mathiasson395 have reported a method for the determination of both ionic and neutral organotin compounds in which ionic compounds were derivatized off-line using NaBEt4 and the sample subsequently cleaned-up by microporous membrane liquid–liquid extraction prior to analysis. The sample solution was pumped through the donor channel of a membrane device and the organotin species extracted across the membrane into a stationary acceptor solution (isooctane). Although on-line derivatization was attempted, poor recoveries were experienced. The LOD in the ng l−1 range gave the method potential for the analysis of contaminated waters but were 10–20 times higher than required for the analysis of natural waters.
The determination of four gaseous arsenic species by Pantsar-Kallio and Korpela400 was notable for requiring less than 2 min analysis time. No sample pre-treatment was necessary thereby minimizing sample contamination and species conversion prior to analysis. The LOD for the four species were in the range 21–174 pg for 20 µl samples.
A new and efficient method developed by Looser et al.394 gave fast, reliable and simple determinations of six butyl- and phenyltin compounds in biota samples (40 mg). The method was based on cold methanolic digestion followed by aqueous ethylation and liquid–liquid extraction. The LOD were in the range 4–52 ng g−1.
In the screening of environmental water samples for inorganic lead and ionic alkyllead compounds, Baena et al.401 first confirmed the presence of Pb using a FIA procedure and then carried out Pb speciation analysis, on those samples containing Pb, using GC-MS. Although an FIA procedure was used for pretreatment of samples for the GC-MS analysis in order to minimize contamination, the derivatization with Grignard reagent (propylmagnesium chloride) was carried out off-line. This was seen as a major limitation on the method and alternative derivatization procedures were sought. Although a small blank signal, derived from the Grignard reagent, was observed for dimethyllead, no other blank correction was required. The calibration was linear over the range 0.02–5 ng ml−1 and precision (1σ) was 6%. The LOD in the range 1–4 ng ml−1 were considerably poorer than those obtainable by ICP-MS. In comparison with other pollutants (e.g., pesticides or polycyclic aromatic hyrdrocarbons), the organolead derivatives dramatically shortened (4-fold) the lifetime of the GC column.
Although GC-MS is not widely used for the determination of Se, reports are becoming more common. In a comparison of three derivatization agents for the determination of selenite in environmental samples, Gomez Ariza et al.402 found 4-chloro-o-phenylendiamine to give the highest sensitivity. No matrix interferences were observed in the analysis of tap, river and sea-waters. Ducros et al.403 used an ID procedure to show that 3% of total human plasma Se was bound to lipoproteins. All the oxidation states of Se were converted into selenite, which was subsequently derivatized with 4-nitro-o-phenylenediamine to form stable piazselenol.
10.3 Noble gas mass spectrometry
Beyerle et al.404 have described in considerable detail the design, setup and performance of a MS system for the analysis of noble gas isotopes and 3H from water samples. After extraction, purification and separation, the noble gases were measured on two double-collector magnetic sector instruments. One instrument was used for the determination of He isotopes, the second for the other gases. After analysis, the completely degassed water samples were resealed in their container and allowed to stand for a minimum of three months after which the 3He produced by decay of 3H was determined. A feature of the method was direct high-vacuum connection of the sample container (copper tube) to the extraction line thereby avoiding use of glass ampoules and O-ring connections. Typical precisions for the determination of noble gas and 3H concentrations were ±0.3–±1.0 and ±2.7%, respectively. Precisions for the measurement of 40Ar∶36Ar, 3He∶4He and 20Ne∶22Ne were ±0.2, ±0.7 and ±0.3%, respectively.Itaya et al.405 have built a new triple collector instrument for argon analysis consisting of a laser probe gas extraction system, purification lines, MS detection and ultra-high-vacuum pumping system. With the exception of glass viewing ports, the whole instrument was made of stainless steel. The extraction and purification lines had a very small dead volume in order to minimize the blanks. The new 20 cm radius, magnetic sector mass spectrometer was more stable and gave higher precision than the traditional 15 cm radius instrument with a single collector which was used previously.
It is only in the last six years that detailed noble gas measurements have been reported for petroleum. As part of a search for 129Xe excesses, Nuzzo et al.406 designed and built an off-line oxidation and purification system for the complete extraction and clean up of Xe from petroleum. After a three-stage clean-up procedure to remove the copious amounts of unwanted gases, the purified Xe was collected on activated charcoal held at –66°C at which temperature only Xe was completely adsorbed.
10.4 New methodologies
The complexity of FT ion cyclotron resonance has prompted a fresh approach to ion trapping. The orbitrap of Makarov407 employed orbital ion trapping in an electrostatic field and was a much simpler and more compact design. In the absence of any magnetic or rf fields, ion stability was achieved only by means of ions orbiting around an axial electrode. Orbiting ions also performed harmonic oscillations along the electrode with a frequency proportional to (m/z)−½. These oscillations were detected using image current detection and transformed into mass spectra using FT. High resolution up to 150000 for ions produced by laser ablation was demonstrated along with high-energy acceptance and wide mass range. Much more development is required before the instrument can have any potential for elemental or isotopic analysis. Although present at up to 6% isotopic abundance, the minor isotopes of lead
and tin could not be detected in the analysis of a lead–tin solder.Spanel and Smith408 have developed a new method for the determination of isotopic abundance ratios of 2H and 18O in water samples using selected ion flow tube MS. The principle of the method is that the precursor ion species (H3O+) are injected into a fast-flowing helium carrier gas in which they react with trace gases in an air or headspace sample (in this case from H2O) introduced into the carrier gas via a downstream entry port. The processes of ion formation were discussed in considerable detail. The precursor and product ions were detected and counted by a downstream quadrupole MS system. The 2H and 18O contents of the cluster ions H82HO4+ and H918OO3+ at m/z 74 and 75, respectively, were determined by reference to the majority cluster ion, H9O4+, at m/z 73. A feature of the method was that only accurate quantitative mass analysis was required and not the use of RMs and repeated calibration.
Although Rosenberg et al.409 considered LC-atmospheric pressure chemical ionization (APCI)-MS to be a viable alternative to GC methods for the determination of organotin compounds, the LOD of 20–65 pg (on column) were several orders of magnitude poorer than those possible by GC or LC-ICP-MS methods. Monobutyl- and monophenyltin were undetectable at environmentally relevant concentrations due to their even poorer sensitivities and unfavourable chromatographic behaviour. An advantage of the new procedure over GC methods was the avoidance of derivatization and its associated systematic errors. In addition, the APCI interface parameters could be chosen for either specific detection of the adduct ion or quasi-element specific detection of the Sn+ ion. The initial extraction of the organotin compounds was identified as being the most critical step for successful quantitation. Extraction was incomplete leading to underestimation of the true concentrations in the sample unless the method of standard additions was used.
Appendix: Glossary of terms
| AES | atomic emission spectrometry |
| AMS | accelerator mass spectrometry |
| APCI | atmospheric pressure chemical ionization |
| CE | capillary electrophoresis |
| CID | collision induced dissociation |
| CRM | certified reference material |
| CSIA | compound-specific isotope analysis |
| CSRA | compound-specific radiocarbon analysis |
| CZE | capillary zone electrophoresis |
| DF | double-focusing |
| DIC | dissolved inorganic carbon |
| DIHEN | direct injection high efficiency nebulizer |
| DOC | dissolved organic carbon |
| DRC | dynamic reaction cell |
| DRF | depth resolution function |
| ECR | electron cyclotron resonance |
| ESMS | electrospray mass spectrometry |
| ETV | electrothermal vaporization |
| FAPIMS | furnace atomization plasma ionization mass spectrometry |
| FI | flow injection |
| FIA | flow injection analysis |
| FT | Fourier transform |
| FTIR | Fourier transform infrared |
| FWHM | full width half maximum |
| GC | gas chromatography |
| GC-MS | gas chromatography mass spectrometry |
| GDMS | glow discharge mass spectrometry |
| GDOES | glow discharge optical emission spectrometry |
| HG | hydride generation |
| HHPN | hydraulic high pressure nebulizer |
| HPLC | high-performance liquid chromatography |
| ICCD | intensified charge-coupled device |
| ICP-AES | inductively coupled plasma atomic emission spectrometry |
| ICP-MS | inductively coupled plasma mass spectrometry |
| ID | isotope dilution |
| ISMS | ion spray mass spectrometry |
| IV | infinite velocity |
| KE | kinetic energy |
| LA | laser ablation |
| LC | liquid chromatography |
| LIMS | laser ionization mass spectrometry |
| LMMS | laser-microprobe mass spectrometry |
| LOD | limit of detection |
| LP/RP | low-power reduced-pressure |
| MC | multi-collector |
| MDMI | monodisperse dried microparticulate injector |
| MIBK | methyl iso-butyl ketone |
| MIP | microwave-induced plasma |
| MRI | mixing length–roughness–information depth |
| MS | mass spectrometry |
| PB | particle beam |
| PDC | pyrrolidinecarbodithioate |
| PGE | platinum group element |
| PX | projectile X-rays |
| RBS | Rutherford back scattering |
| REE | rare earth element |
| REMPI | resonantly enhanced multiphoton ionization |
| RIMS | resonance ionization mass spectrometry |
| RM | reference material |
| RSD | relative standard deviation |
| RSF | relative sensitivity factor |
| SEC | size exclusion chromatography |
| SIFT | selected-ion flow tube |
| SIMS | secondary ion mass spectrometry |
| SIRMS | stable isotope ratio mass spectrometry |
| SNMS | sputtered neutral mass spectrometry |
| SPME | solid phase microextraction |
| SRM | standard reference material |
| S-SIMS | static secondary ion mass spectrometry |
| TEM | transmission electron microscopy |
| TIL | total interference level |
| TIMS | thermal ionization mass spectrometry |
| TOF | time-of-flight |
| TOFMS | time-of-flight mass spectrometry |
| TRIM | transport of ions in matter |
| TXRF | total reflection X-ray fluorescence |
| USN | ultrasonic nebulization |
| UV | ultraviolet |
| XPS | X-ray photoelectron spectroscopy |
References
- J. R. Bacon, J. S. Crain, L. Van Vaeck and J. G. Williams, J. Anal. At. Spectrom., 2000, 15(8), 1025 RSC.
- B. Fairman, M. W. Hinds, S. M. Nelms, D. M. Penny and P. Goodall, J. Anal. At. Spectrom., 2000, 15(12), 1606 RSC.
- M. R. Cave, O. Butler, S. R. N. Chenery, J. M. Cook, M. S. Cresser and D. L. Miles, J. Anal. At. Spectrom., 2001, 16(2), 194 RSC.
- A. Taylor, S. Branch, D. J. Halls, L. M. W. Owen and M. White, J. Anal. At. Spectrom., 2001, 16(4), 421 RSC.
- J. S. Becker and H. J. Dietze, Int. J. Mass Spectrom., 2000, 197, 1 CrossRef CAS.
- C. M. Barshick, D. C. Duckworth and D. H. Smith, Inorganic Mass Spectrometry: Fundamentals and Applications, Dekker, New York, 2000. Search PubMed.
- T. P. Rao and V. M. Biju, Crit. Rev. Anal. Chem., 2000, 30(2–3), 179 Search PubMed.
- W. Kutschera, R. Golser, A. Priller and B. Strohmaier, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, VII Search PubMed.
- L. K. Fifield, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 134 Search PubMed.
- W. Kutschera, AIP Conf. Proc., 1999, 495(Experimental Nuclear Physics in Europe), 407 Search PubMed.
- H. E. Gove, Radiocarbon, 2000, 42(1), 127 Search PubMed.
- M. Hotchkis, D. Fink, C. Tuniz and S. Vogt, Appl. Radiat. Isot., 2000, 53(1–2), 31 Search PubMed.
- M. Suter, S. W. A. Jacob and H. A. Synal, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 144 Search PubMed.
- R. C. Garner and D. Leong, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 892 Search PubMed.
- K. W. Turleltaub and J. S. Vogel, Curr. Pharm. Des., 2000, 991 Search PubMed.
- J. S. Vogel, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 884 Search PubMed.
- R. Braucher, D. L. Bourlès, E. T. Brown, F. Colin, J.-P. Muller, J.-J. Braun, M. Delaune, A. Edou Minko, C. Lescouet, G. M. Raisbeck and F. Yiou, Chem. Geol., 2000, 170, 95 CrossRef CAS.
- L. A. Currie, Radiocarbon, 2000, 42(1), 115 Search PubMed.
- W. Kretschmer, Acta Phys. Pol., 2000, 31(1), 123 Search PubMed.
- H. A. Synal, S. Jacob and M. Suter, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 1 Search PubMed.
- H. A. Synal, S. Jacob and M. Suter, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 161, 29 Search PubMed.
- S. A. W. Jacob, M. Suter and H. A. Synal, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 235 Search PubMed.
- T. J. Ognibene, T. A. Brown, J. P. Knezovich, M. L. Roberts, J. R. Southon and J. S. Vogel, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 47 Search PubMed.
- R. B. Schubank, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 288 Search PubMed.
- P. Steier, S. Puchegger, R. Golser, W. Kutschera, A. Priller, W. Rom, A. Wallner and E. Wild, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 161, 250 Search PubMed.
- Y. Liu, M. Chen, D. Li, X. Lu, L. Shen, S. Xu and G. Chen, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 160(3), 424 Search PubMed.
- M. B. Chen, X. S. Lu, D. M. Li, Y. H. Liu, W. J. Zhou, G. S. Chen, L. G. Shen, S. L. Xu and Y. X. Zhang, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 193 Search PubMed.
- W. J. Zhou, M. B. Chen, Y. H. Liu, D. Donahue, J. Head, X. F. Lu, A. J. T. Jull and L. Deng, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 201 Search PubMed.
- J. Southon and M. Roberts, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 257 Search PubMed.
- T. A. Brown, M. L. Roberts and J. R. Southon, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 344 Search PubMed.
- K. F. von Reden, M. Bellino, P. Long, R. J. Schneider and R. Loger, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 247 Search PubMed.
- C. B. Ramsey and M. J. Humm, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 242 Search PubMed.
- R. J. Schneider, K. F. von Reden, J. M. Hayes, J. S. C. Wills, W. G. E. Kern and S. W. Kim, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 252 Search PubMed.
- M. Paul, D. Berkovits, I. Ahmad, F. Borasi, J. Caggiano, C. N. Davids, J. P. Greene, B. Harss, A. Heinz, D. J. Henderson, W. Henning, C. L. Jiang, R. C. Pardo, K. E. Rehm, R. Rejoub, D. Seweryniak, A. Sonzogni, J. Uusitalo and R. Vondrasek, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 688 Search PubMed.
- T. M. DeTurck, D. L. Knies, K. S. Grabowski, G. K. Hubler, D. J. Treacy and A. C. Mignerey, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 293 Search PubMed.
- A. E. Litherland, R. P. Beukens, J. P. Doupe, W. E. Kieser, J. S. Krestow, J. C. Rucklidge, I. Tomski, G. C. Wilson and X. L. Zhao, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 206 Search PubMed.
- T. A. Brown and G. H. Gillespie, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 338 Search PubMed.
- S. H. Sie, D. A. Sims, T. R. Niklaus and G. F. Suter, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 268 Search PubMed.
- P. Persson, M. Kiisk, B. Erlandsson, M. Faarinen, R. Hellborg, G. Skog and K. Stenstrom, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 188 Search PubMed.
- M. He, S. Jiang, S. S. Jiang, S. Y. Wu, Z. Zhan-Xia and D. Yue-An, Nucl. Sci. Tech., 2000, 11(1), 1 Search PubMed.
- M. He, S. Jiang, S. S. Jiang, Q. Chen, J. C. Qin, S. Y. Wu, Y. E. Dong and Z. X. Zhao, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 177 Search PubMed.
- K. Knie, T. Faestermann, G. Korschinek, G. Rugel, W. Ruhm and C. Wallner, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 717 Search PubMed.
- S. Jiang, M. He, S. Y. Wu, Z. X. Zhao and Y. Dong, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 305 Search PubMed.
- G. M. Santos, J. C. Acquadro, P. R. S. Gomes, R. M. Anjos, R. L. Neto, N. Added, N. H. Medina, N. Carlin, M. M. Coimbra, M. L. di Tada, R. G. Cresswell and L. K. Fifield, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 310 Search PubMed.
- S. H. Sie, D. A. Sims, T. R. Niklaus, F. Bruhn, G. Suter and G. Cripps, Nucl. Instrum. Methods Phys. Res. Sect. B, 1999, 158(1-4), 201 Search PubMed.
- O. V. Makarova, T. Rajh, M. C. Thurnauer, A. Martin, P. A. Kemme and D. Cropek, Environ. Sci. Technol., 2000, 34(22), 4797 CrossRef CAS.
- S. A. Datar, L. Y. Wu, B. N. Guo, M. Nigam, D. Necsoiu, Y. J. Zhai, D. E. Smith, C. Yang, M. E. Bouanani, J. J. Lee and F. D. McDaniel, Appl. Phys. Lett., 2000, 77(24), 3974 CrossRef.
- B. J. Hughey, P. L. Skipper, R. E. Klinkowstein, R. E. Shefer, J. S. Wishnok and S. R. Tannenbaum, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 40 Search PubMed.
- B. A. Buchholz, S. R. Dueker, Y. M. Lin, A. J. Clifford and J. S. Vogel, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 910 Search PubMed.
- M. Uchida, Y. Shibata, K. Kawamura, M. Yoneda, H. Mukai, A. Tanaka, T. Uehiro and M. Morita, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 583 Search PubMed.
- A. P. McNichol, J. R. Ertel and T. I. Eglinton, Radiocarbon, 2000, 42(2), 219 Search PubMed.
- A. R. Gagnon, A. P. McNichol, J. C. Donoghue, D. R. Stuart and K. von Reden, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 409 Search PubMed.
- L. A. Currie, J. D. Kessler, J. V. Marolf, A. P. McNichol, D. R. Stuart, J. C. Donohue, D. J. Donahue, G. S. Burr and D. Biddulph, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 440 Search PubMed.
- D. B. Klinedinst and L. A. Currie, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 545 Search PubMed.
- M. Dee and C. B. Ramsey, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 449 Search PubMed.
- S. X. Yuan, X. H. Wu, S. J. Gao, J. X. Wang, L. Z. Cai, Z. Y. Guo and X. Y. Lu, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 458 Search PubMed.
- J. W. Pohlman, D. L. Knies, K. S. Grabowski, T. M. DeTurck, D. J. Treacy and R. B. Coffin, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 428 Search PubMed.
- B. A. Buchholz, S. Freeman, K. W. Haack and J. S. Vogel, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 404 Search PubMed.
- D. B. Klinedinst and L. A. Currie, Environ. Sci. Technol., 1999, 33(23), 4146 CrossRef CAS.
- M. Minami and T. Nakamura, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 462 Search PubMed.
- S. X. Yuan, X. H. Wu, S. J. Gao, J. X. Wang, L. Z. Cai, K. X. Liu, K. Li and H. J. Ma, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 424 Search PubMed.
- G. Morgenroth, H. Kerscher, W. Kretschmer, M. Klein, M. Reichel, T. Tully and I. Wrzosok, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 416 Search PubMed.
- J. Turnbull, R. Sparks and C. Prior, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 469 Search PubMed.
- S. Puchegger, W. Rom and P. Steier, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 274 Search PubMed.
- M. Friedrich, W. Pilz, G. Sun, R. Behrisch, C. Garcia-Rosales, N. Bekris and R. D. Penzhorn, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 161–163, 216 Search PubMed.
- M. Friedrich, W. Pilz, G. Sun, R. Behrisch, C. Garcia-Rosales, N. Bekris and R. D. Penzhorn, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 655 Search PubMed.
- M. L. Roberts, R. W. Hamm, K. H. Dingley, M. L. Chiarappa-Zucca and A. H. Love, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 262 Search PubMed.
- D. Child, G. Elliott, C. Mifsud, A. M. Smith and D. Fink, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 856 Search PubMed.
- A. M. Smith, D. Fink, D. Child, V. A. Levchenko, V. I. Morgan, M. Curran, D. M. Etheridge and G. Elliott, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 847 Search PubMed.
- E. Boaretto, D. Berkovits, M. Hass, S. K. Hui, A. Kaufman, M. Paul and S. Weiner, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 767 Search PubMed.
- U. Morgenstern, L. K. Fifield and A. Zondervan, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 605 Search PubMed.
- D. J. Treacy, D. L. Knies, K. S. Grabowski, G. K. Hubler, T. M. DeTurck and A. C. Mignerey, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 321 Search PubMed.
- Y. Nagashima, R. Seki, T. Takahashi and D. Arai, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 129 Search PubMed.
- S. Hatori, H. Nawata, M. Ohseki, H. Matsuzaki, T. Misawa and K. Kobayashi, Nucl. Instrum. Methods Phys. Res., Sect. B., 2000, 172, 211 Search PubMed.
- W. Ruhm, K. Knie, G. Rugel, A. A. Marchetti, T. Faestermann, C. Wallner, J. E. McAninch, T. Straume and G. Korschinek, Health Phys., 2000, 79(4), 358 Search PubMed.
- G. Rugel, T. Faestermann, K. Knie, G. Korschinek, A. A. Marchetti, J. E. McAninch, W. Ruhm, T. Straume and C. Wallner, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 934 Search PubMed.
- P. Persson, B. Erlandsson, K. Freimann, R. Hellborg, R. Larsson, J. Persson, G. Skog and K. Stenstrom, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 160(4), 510 Search PubMed.
- P. Collon, W. Kutschera, H. H. Loosli, B. E. Lehmann, R. Purtschert, A. Love, L. Sampson, D. Anthony, D. Cole, B. Davids, D. J. Morrissey, B. M. Sherrill, M. Steiner, R. C. Pardo and M. Paul, Earth Planet. Sci. Lett., 2000, 182(1), 103 CrossRef.
- L. K. Fifield, R. S. Carling, R. G. Cresswell, P. A. Hausladen, M. L. di Tada and J. P. Day, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 168(3), 427 Search PubMed.
- B. A. Bergquist, A. A. Marchetti, R. E. Martinelli, J. E. McAninch, G. J. Nimz, I. D. Proctor, J. R. Southon and J. S. Vogel, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 328 Search PubMed.
- S. Szidat, A. Schmidt, J. Handl, R. Michel, H.-A. Synal, C. Schnabel, M. Suter and J. M. Lopez-Gutierrez, Kerntechnik, 2000, 65(4), 160 Search PubMed.
- S. Szidat, A. Schmidt, J. Handl, D. Jakob, R. Michel, H. A. Synal and M. Suter, J. Radioanal. Nucl. Chem., 2000, 244(1), 45 CrossRef CAS.
- N. Buraglio, A. Aldahan and G. Possnert, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 518 Search PubMed.
- S. Szidat, A. Schmidt, J. Handl, D. Jakob, W. Botsch, R. Michel, H. A. Synal, C. Schnabel, M. Suter, J. M. Lopez-Gutierrez and W. Stade, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 699 Search PubMed.
- J. M. Lopez-Gutierrez, H. A. Synal, M. Suter, C. Schnabel and M. Garcia-Leon, Appl. Radiat. Isot., 2000, 53(1–2), 81 Search PubMed.
- S. Hatori, M. Ohseki, H. Nawata, H. Matsuzaki, T. Misawa and K. Kobayashi, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 299 Search PubMed.
- N. Buraglio, A. Aldahan and G. Possnert, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 161, 240 Search PubMed.
- F. Yiou, K. Diel, F. X. Martin and G. Raisbeck, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 395 Search PubMed.
- D. L. Biddulph, J. W. Beck, G. S. Burr, D. J. Donahue, A. L. Hatheway and A. J. T. Jull, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 693 Search PubMed.
- M. L. Roberts and M. W. Caffee, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 388 Search PubMed.
- C. Wallner, T. Faestermann, U. Gerstmann, W. Hillebrandt, K. Knie, G. Korschinek, C. Lierse, C. Pomar and G. Rugel, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 333 Search PubMed.
- N. D. Priest, G. M. Pich, L. K. Fifield and R. G. Cresswell, Radiat. Res., 1999, 152(6, Suppl.), Search PubMed.
- D. H. Oughton, L. K. Fifield, J. P. Day, R. C. Cresswell, L. Skipperud, M. L. di Tada, B. Salbu, P. Strand, E. Drozcho and Y. Mokrov, Environ. Sci. Technol., 2000, 34(10), 1938 CrossRef CAS.
- A. Wallner, Y. Ikeda, W. Kutschera, A. Priller, P. Steier, H. Vonach and E. Wild, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 382 Search PubMed.
- S. Caroli, O. Senofonte and G. Modesti, Adv. At. Spectrosc., 1999, 5, 173 Search PubMed.
- N. H. Bings, J. M. Costa-Fernandez, J. P. Guzowski, A. M. Leach and G. M. Hieftje, Spectrochim. Acta, Part B, 2000, 55(7), 765 CrossRef.
- W. Hang, X. M. Yan, D. M. Wayne, J. A. Olivares, W. W. Harrison and V. Majidi, Anal. Chem., 1999, 71(15), 3231 CrossRef CAS.
- T. E. Gibeau and R. K. Marcus, Anal. Chem., 2000, 72(16), 3833 CrossRef.
- R. K. Marcus, M. A. Dempster, T. E. Gibeau and E. M. Reynolds, Anal. Chem., 1999, 71(15), 3061 CrossRef CAS.
- Y. X. Duan, Y. X. Su and Z. Jin, J. Anal. At. Spectrom., 2000, 15(10), 1289 RSC.
- Y. X. Su, Y. X. Duan and Z. Jin, Anal. Chem., 2000, 72(22), 5600 CrossRef.
- C. Yang, M. Mohill and W. W. Harrison, J. Anal. At. Spectrom., 2000, 15(9), 1255 RSC.
- A. Bogaerts and R. Gijbels, J. Appl. Phys., 1999, 86(8), 4124 CrossRef CAS.
- A. Bogaerts and R. Gijbels, Spectrochim. Acta, Part B, 2000, 55B(3), 279 CrossRef CAS.
- A. Bogaerts and R. Gijbels, J. Anal. At. Spectrom., 2000, 15(4), 441 RSC.
- S. Itoh, R. Hasegawa and M. Amano, Mater. Trans., 2000, 41(2), 346 Search PubMed.
- T. Saka and M. Inoue, Anal. Sci., 2000, 16(6), 653 Search PubMed.
- R. C. Eanes and R. K. Marcus, Spectrochim. Acta, Part. B, 2000, 55(4), 405.
- S. Itoh, H. Yamaguchi, T. Yoshioka, T. Kimura and T. Kobayashi, Mater. Trans., 2000, 41(4), 516 Search PubMed.
- J. P. Guzowski and G. M. Hieftje, J. Anal. At. Spectrom., 1999, 15(1), 27 RSC.
- J. P. Guzowski and G. M. Hieftje, Anal. Chem., 2000, 72(16), 3812 CrossRef.
- J. P. Guzowski, J. A. C. Broekaert and G. M. Hieftje, Spectrochim. Acta, Part B, 2000, 55(8), 1295 CrossRef CAS.
- R. K. Marcus, J. Anal. At. Spectrom., 2000, 15(9), 1271 RSC.
- W. G. Diegor and H. P. Longerich, At. Spectrosc., 2000, 21(3), 111 Search PubMed.
- I. Feldmann, N. Jakubowski and D. Stuewer, Fresenius' J. Anal. Chem., 1999, 365(5), 415 CrossRef.
- I. Feldmann, N. Jakubowski, C. Thomas and D. Stuewer, Fresenius' J. Anal. Chem., 1999, 365(5), 422 CrossRef.
- S. D. Tanner, V. I. Baranov and U. Vollkopf, J. Anal. At. Spectrom., 2000, 15(9), 1261 RSC.
- J. J. Sloth and E. H. Larsen, J. Anal. At. Spectrom., 2000, 15(6), 669 RSC.
- B. Hattendorf and D. Guenther, J. Anal. At. Spectrom., 2000, 15(9), 1125 RSC.
- N. Praphairaksit and R. S. Houk, Anal. Chem., 2000, 72(18), 4435 CrossRef.
- R. E. Sturgeon, J. W. H. Lam and A. Saint, J. Anal. At. Spectrom., 2000, 15(6), 607 RSC.
- Y. X. Su, Y. X. Duan and Z. Jin, Anal. Chem., 2000, 72(11), 2455 CrossRef CAS.
- L. A. Iacone, M. G. Minnich, W. R. L. Masamba, S. H. Nam, A. Montaser and R. G. Michel, Spectrochim. Acta, Part B, 1999, 54(13), 1919 CrossRef CAS.
- A. Okino, D. Kaneko, H. Yabuta and E. Hotta, Bunseki Kagaku, 2000, 49(11), 869 Search PubMed.
- G. O'Connor, L. Ebdon and E. H. Evans, J. Anal. At. Spectrom., 1999, 14(9), 1303 RSC.
- L. S. Milstein, J. W. Waggoner, K. L. Sutton and J. A. Caruso, Appl. Spectrosc., 2000, 54(9), 1286 Search PubMed.
- P. K. Appelblad and D. C. Baxter, J. Anal. At. Spectrom., 2000, 15(5), 557 RSC.
- P. K. Appelblad, I. Rodushkin and D. C. Baxter, J. Anal. At. Spectrom., 2000, 15(4), 359 RSC.
- I. I. Stewart, C. E. Hensman and J. W. Olesik, Appl. Spectrosc., 2000, 54(2), 164 Search PubMed.
- M. G. Minnich and A. Montaser, Appl. Spectrosc., 2000, 54(9), 1261 Search PubMed.
- N. Praphairaksit, D. R. Wiederin and R. S. Houk, Spectrochim. Acta, Part B, 2000, 55(8), 1279 CrossRef.
- I. I. Stewart, R. Guevremont and R. E. Sturgeon, Anal. Chem., 1999, 71(22), 5146 CrossRef CAS.
- I. I. Stewart and R. E. Sturgeon, J. Anal. At. Spectrom., 2000, 15(9), 1223 RSC.
- R. Kafemann, S. Adlerstein and R. Neukamm, Fisheries Res., , 2000, 46, 313 Search PubMed.
- S. R. Thorrold, C. M. Jones, S. E. Campana, J. W. McLaren and J. W. H. Lam, Limnol. Oceanogr., 1998, 43(8), 1826 Search PubMed.
- G. Veinott, T. Northcote, M. Rosenau and R. D. Evans, Can. J. Fish. Aquat. Sci., 1999, 56(11), 1981 CrossRef.
- M. Bi, A. M. Ruiz, B. W. Smith and J. D. Winefordner, Appl. Spectrosc., 2000, 54(5), 639 Search PubMed.
- V. R. Bellotto and N. Miekeley, Fresenius' J. Anal. Chem., 2000, 367(7), 635 CrossRef CAS.
- J. S. Becker, C. Pickhardt and H. J. Dietze, Int. J. Mass Spectrom., 2000, 202(1–3), 283 CrossRef CAS.
- C. A. Craig, K. E. Jarvis and L. J. Clarke, J. Anal. At. Spectrom., 2000, 15(8), 1001 RSC.
- H. C. Liu, O. V. Borisov, X. L. Mao, S. Shuttleworth and R. E. Russo, Appl. Spectrosc., 2000, 54(10), 1435 Search PubMed.
- O. V. Borisov, X. L. Mao and R. E. Russo, Spectrochim. Acta, Part B, 2000, 55(11), 1693 CrossRef.
- R. E. Russo, X. L. Mao, O. V. Borisov and H. Liu, J. Anal. At. Spectrom., 2000, 15(9), 1115 RSC.
- S. H. Jeong, O. V. Borisov, J. H. Yoo, X. L. Mao and R. E. Russo, Anal. Chem., 1999, 71(22), 5123 CrossRef CAS.
- A. M. Leach and G. M. Hieftje, J. Anal. At. Spectrom., 2000, 15(9), 1121 RSC.
- D. Bleiner, K. Hametner and D. Gunther, Fresenius' J. Anal. Chem., 2000, 368(1), 37 CrossRef CAS.
- D. Bleiner, A. Plotnikov, C. Vogt, K. Wetzig and D. Gunther, Fresenius' J. Anal. Chem., 2000, 368(2–3), 221 CrossRef CAS.
- E. Hoffmann, C. Ludke, J. Skole, H. Stephanowitz, E. Ullrich and D. Colditz, Fresenius' J. Anal. Chem., 2000, 367(6), 579 CrossRef CAS.
- A. M. Ghazi, S. Shuttleworth, S. J. Angulo and D. H. Pashley, J. Anal. At. Spectrom., 2000, 15(10), 1335 RSC.
- Y. Okamoto, Fresenius' J. Anal. Chem., 2000, 367(3), 300 CrossRef CAS.
- Y. Okamoto, Fresenius' J. Anal. Chem., 2000, 366(3), 309 CrossRef CAS.
- J. P. Snell, E. Bjorn and W. Frech, J. Anal. At. Spectrom., 2000, 15(4), 397 RSC.
- W. C. Wei, P. H. Chi and M. H. Yang, J. Anal. At. Spectrom., 2000, 15(11), 1466 RSC.
- A. A. Menegario and M. F. Gine, Spectrochim. Acta, Part B, 2000, 55(4), 355 CrossRef.
- C. Moor, J. W. H. Lam and R. E. Sturgeon, J. Anal. At. Spectrom., 2000, 15(2), 143 RSC.
- Z. Mester, R. E. Sturgeon and J. W. Lam, J. Anal. At. Spectrom., 2000, 15(11), 1461 RSC.
- Z. Mester, J. Lam, R. Sturgeon and J. Pawliszyn, J. Anal. At. Spectrom., 2000, 15(7), 837 RSC.
- J. Li, F. Lu, T. Umemura and K. I. Tsunoda, Anal. Chim. Acta, 2000, 419(1), 65 CrossRef CAS.
- Y. L. Feng, H. Narasaki, L. C. Tian and H. Y. Chen, At. Spectrosc., 2000, 21(1), 30 Search PubMed.
- B. W. Acon, J. A. McLean and A. Montaser, Anal. Chem., 2000, 72(8), 1885 CrossRef.
- M. Huang, A. Hirabayashi, T. Shirasaki and H. Koizumi, Anal. Chem., 2000, 72(12), 2463 CrossRef.
- R. S. Olofsson, I. Rodushkin and M. D. Axelsson, J. Anal. At. Spectrom., 2000, 15(6), 727 RSC.
- Y. Sung and H. B. Lim, Microchem. J., 2000, 64(1), 51 CrossRef CAS.
- T. Hirata, J. Anal. At. Spectrom., 2000, 15(11), 1447 RSC.
- K. H. Lee, M. Oshima, T. Takayanagi and S. Motomizu, J. Flow Injection Anal., 1999, 16(2), 255 Search PubMed.
- K. H. Lee, M. Oshima, T. Takayanagi and S. Motomizu, Anal. Sci., 2000, 16(7), 731 Search PubMed.
- M. Oshima, K. H. Lee, Y. H. Gao and S. Motomizu, Chem. Lett., 2000, (11), 1338.
- S. Hirata, K. Honda, O. Shikino, N. Maekawa and M. Aihara, Spectrochim. Acta, Part B, 2000, 55(7), 1087 CrossRef.
- X. P. Yan, M. J. Hendry and R. Kerrich, Anal. Chem., 2000, 72(8), 1879 CrossRef CAS.
- X. P. Yan, Presented at the 7th Analytical Chemistry Division of Chinese Chemical Society and the 7th National Analytical Atomic Spectrometry Conference, Chongqing, China, October 16–19, 2000..
- K. Benkhedda, H. Goenaga Infante, E. Ivanova and F. C. Adams, J. Anal. At. Spectrom., 2000, 15(10), 1349 RSC.
- G. Centineo, M. M. Bayon and A. Sanz-Medel, J. Anal. At. Spectrom., 2000, 15(10), 1357 RSC.
- J. M. Costa-Fernandez, N. H. Bings, A. M. Leach and G. M. Hieftje, J. Anal. At. Spectrom., 2000, 15(9), 1063 RSC.
- A. Chatterjee, Y. Shibata and M. Morita, J. Anal. At. Spectrom., 2000, 15(8), 913 RSC.
- A. Chatterjee, Y. Shibata, J. Yoshinaga and M. Morita, Anal. Chem., 2000, 72(18), 4402 CrossRef CAS.
- H. Goenaga Infante, M. L. Fernandez Sanchez and A. Sanz-Medel, J. Anal. At. Spectrom., 2000, 15(5), 519 RSC.
- Q. Tu and J. F. W. Qvarnstrom, Analyst, 2000, 125(4), 705 RSC.
- B. Gammelgaard and O. Jons, J. Anal. At. Spectrom., 2000, 15(5), 499 RSC.
- K. L. Ackley, K. L. Sutton and J. A. Caruso, J. Anal. At. Spectrom., 2000, 15(9), 1069 RSC.
- J. M. Marchante-Gayon, C. Thomas, I. Feldmann and N. Jakubowski, J. Anal. At. Spectrom., 2000, 15(9), 1093 RSC.
- J. A. Day, K. L. Sutton, R. S. Soman and J. A. Caruso, Analyst, 2000, 125(5), 819 RSC.
- K. L. Ackley, J. A. Day and J. A. Caruso, J. Chromatogr., 2000, 888(1–2), 293 CrossRef CAS.
- C. B'Hymer, J. A. Day and J. A. Caruso, Appl. Spectrosc., 2000, 54(7), 1040 Search PubMed.
- F. Vanhaecke, S. Saverwyns, G. D. Wannemacker, L. Moens and R. Dams, Anal. Chim. Acta, 2000, 419(1), 55 CrossRef.
- J. W. Tromp, R. T. Tremblay, J. M. Mermet and E. D. Salin, J. Anal. At. Spectrom., 2000, 15(6), 617 RSC.
- J. L. M. de Boer, J. Anal. At. Spectrom., 2000, 15(9), 1157 RSC.
- S. H. Hu, S. L. Lin, Y. S. Liu and S. Gao, Chem. J. Chin. Univ.-Chin., 2000, 21(3), 368 Search PubMed.
- M. B. Gomez, M. M. Gomez and M. A. Palacios, Anal. Chim. Acta, 2000, 404(2), 285 CrossRef CAS.
- J. S. Becker and H. J. Dietze, Int. J. Mass Spectrom., 2000, 202(1-3), 69 CrossRef CAS.
- A. S. Al-Ammar, E. Reitznerova and R. M. Barnes, Spectrochim. Acta, Part B, 1999, 54(13), 1813 CrossRef CAS.
- A. Tangen and W. Lund, Spectrochim. Acta, Part B, 1999, 54(13), 1831 CrossRef.
- A. S. Al-Ammar, R. K. Gupta and R. M. Barnes, Spectrochim. Acta, Part B, 2000, 55(6), 629 CrossRef CAS.
- Z. Du and R. S. Houk, J. Anal. At. Spectrom., 2000, 15(4), 383 RSC.
- G. K. Koyanagi, V. I. Baranov, S. D. Tanner and D. K. Bohme, J. Anal. At. Spectrom., 2000, 15(9), 1207 RSC.
- G. K. Koyanagi, V. V. Lavrov, V. Baranov, D. Bandura, S. Tanner, J. W. McLaren and D. K. Bohme, Int. J. Mass Spectrom., 2000, 194(1), L1 CrossRef CAS.
- D. R. Hassler, B. Peucker-Ehrenbrink and G. E. Ravizza, Chem. Geol., 2000, 166(1–2), 1 CrossRef CAS.
- A. T. Townsend, Fresenius' J. Anal. Chem., 2000, 367(7), 614 CrossRef CAS.
- C. R. Quetel, J. Vogl, T. Prohaska, S. Nelms, P. D. P. Taylor and P. De Bievre, Fresenius' J. Anal. Chem., 2000, 368(2–3), 148 CrossRef CAS.
- C. R. Quetel, T. Prohaska, M. Hamester, W. Kerl and P. D. P. Taylor, J. Anal. At. Spectrom., 2000, 15(4), 353 RSC.
- C. H. Stirling, D. C. Lee, J. N. Christensen and A. N. Halliday, Geochim. Cosmochim. Acta, 2000, 64(21), 3737 CrossRef CAS.
- N. S. Belshaw, X. K. Zhu, Y. Guo and R. K. O'Nions, Int. J. Mass Spectrom., 2000, 197, 191 CrossRef CAS.
- F. Vanhaecke, L. Moens, R. Dams, L. Allen and S. Georgitis, Anal. Chem., 1999, 71(15), 3297 CrossRef CAS.
- H. Emteborg, X. Tian, M. Ostermann, M. Berglund and F. C. Adams, J. Anal. At. Spectrom., 2000, 15(3), 239 RSC.
- J. Dombovari, J. S. Becker and H. J. Dietze, Int. J. Mass Spectrom., 2000, 202(1–3), 231 CrossRef.
- J. Diemer and K. G. Heumann, Fresenius' J. Anal. Chem., 2000, 368(1), 103 CrossRef CAS.
- D. R. Bandura, V. I. Baranov and S. D. Tanner, J. Anal. At. Spectrom., 2000, 15(8), 921 RSC.
- J. D. Winefordner, I. B. Gornushkin, D. Pappas, O. I. Matveev and B. W. Smith, J. Anal. At. Spectrom., 2000, 15(9), 1161 RSC.
- Y. Li, X. J. Yang, Y. X. Tang and Q. Z. Qin, Appl. Spectrosc., 2000, 54(5), 645 Search PubMed.
- P. R. Willmott, P. Manoravi and K. Holliday, Appl. Phys. A: Mater. Sci. Process., 2000, 70(4), 425 Search PubMed.
- S. A. Reid, W. Ho and F. J. Lamelas, J. Phys. Chem. B, 2000, 104(22), 5324 CrossRef CAS.
- S. M. Park and J. Y. Moon, Appl. Phys. A: Mater. Sci. Process., 1999, 69(Suppl), S695 Search PubMed.
- S. M. Park and J. Y. Moon, J. Appl. Phys., 1999, 86(12), 7139 CrossRef CAS.
- K. Shibagaki, T. Kawashima, K. Sasaki and K. Kadota, Jpn. J. Appl. Phys. Part 1, 2000, 39(8), 4959 Search PubMed.
- J. B. Liu, X. P. Xing, P. Liu and Z. Gao, Int. J. Mass Spectrom., 2000, 202(1–3), 261 CrossRef.
- C. C. Han, Y. L. W. Han and Y. C. Chen, Int. J. Mass Spectrom., 1999, 189(2–3), 157 CrossRef.
- K. Watanabe and T. Iguchi, Appl. Phys. A: Mater. Sci. Process., 1999, 69(Suppl.), S845 Search PubMed.
- R. E. Russo, G. L. Klunder, P. Grant and B. D. Andresen, Appl. Phys. A: Mater. Sci. Process., 1999, 69(Suppl.), S895 Search PubMed.
- O. Kornienko, P. T. A. Reilly, W. B. Whitten and J. M. Ramsey, Rapid Commun. Mass Spectrom., 1999, 13(1), 50 CrossRef.
- L. Van Vaeck, P. Van Espen, R. Gijbels, G. Baykut and F. H. Laukien, Eur. J. Mass Spectrom., 2000, 6(3), 277 Search PubMed.
- W. D. Reents and Z. Z. Ge, Aerosol Sci. Technol., 2000, 33(1–2), 122.
- K. H. Hong, K. S. Song, H. K. Cha, M. Yang, J. M. Lee, C. W. Lee and G. H. Lee, J. Radioanal. Nucl. Chem., 1999, 241(3), 533.
- F. Aubriet, B. Maunit and J. F. Muller, Int. J. Mass Spectrom., 2000, 198(3), 189 CrossRef.
- S. T. Dai, L. Jin, W. Lu, R. T. An, L. L. Tai and D. Y. Chen, Appl. Phys. A: Mater. Sci. Process., 1999, 69(Suppl.), S167 Search PubMed.
- K. Wendt, N. Trautmann and B. A. Bushaw, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 172, 162 Search PubMed.
- S. S. Thattey, B. M. Suri, P. N. Bajaj and V. Dev, Rev. Sci. Instrum., 2000, 71(1), 32 CrossRef.
- W. F. Calaway, M. P. McCann and M. J. Pellin, Mater. Res. Soc. Symp. Proc., 1999, 551, 83 Search PubMed.
- B. A. Bushaw and W. Nortershauser, Spectrochim. Acta, Part B., 2000, 55(11), 1679 CrossRef.
- V. Philipsen, J. Bastiaansen, G. Verschoren, P. Lievens, E. Vandeweert, R. E. Silverans and H. H. Telle, Spectrochim. Acta, Part B, 2000, 55(10), 1539 CrossRef.
- P. Müller, K. Blaum, B. A. Bushaw, S. Diel, C. Geppert, A. Nahler, W. Nortershauser, N. Trautmann and K. Wendt, Radiochim. Acta, 2000, 88(8), 487 Search PubMed.
- L. Pibida, J. M. R. Hutchinson, J. Wen and L. Karam, Rev. Sci. Instrum., 2000, 71(2), 509 CrossRef.
- M. Schuhmacher, B. Rasser and F. Desse, J. Vac. Sci. Technol., 2000, 18(1), 529 Search PubMed.
- K. Takanashi, H. Wu, Y. Kuramoto, Z. H. Cheng, T. Sakamoto, M. Owari and Y. Nihei, Surf. Interface Anal., 2000, 30(1), 493 CrossRef.
- H. Yamazaki and T. Enda, Anal. Sci., 2000, 16(5), 483 Search PubMed.
- C. Derue, D. Gibouin, F. Lefebvre, B. Rasser, A. Robin, L. Le Sceller, M. C. Verdus, M. Demarty, M. Thellier and C. Ripoll, J. Trace Microprobe Tech., 1999, 17(4), 451 Search PubMed.
- A. Liu, B. Woo and R. W. Odom, Rev. Sci. Instrum., 2000, 71(11), 4144 CrossRef.
- I. S. Gilmore and M. P. Seah, Int. J. Mass Spectrom., 2000, 202(1-3), 217 CrossRef.
- D. A. Dahl, A. D. Appelhans and M. B. Ward, Int. J. Mass Spectrom., 1999, 189(1), 39 CrossRef.
- I. Terzic, N. Bundaleski, Z. Rakocevic, N. Oklobdzija and J. Elazar, Rev. Sci. Instrum., 2000, 71(11), 4195 CrossRef.
- K. Wittmaack, Surf. Interface Anal., 2000, 29, 721 CrossRef CAS.
- P. A. W. van der Heide, Appl. Surf. Sci., 2000, 157(3), 191 CrossRef.
- P. A. W. van der Heide, Surf. Sci., 2000, 447(1–3), 62 CrossRef.
- P. A. W. van der Heide, M. S. Lim, S. S. Perry and J. W. Rabalais, J. Chem. Phys., 2000, 113(22), 10344 CrossRef.
- P. A. Ronsheim and R. J. Murphy, J. Vac. Sci. Technol., 2000, 18(1), 501 Search PubMed.
- A. Licciardello, L. Renna and S. Pignataro, Surf. Interface Anal., 2000, 30(1), 243 CrossRef.
- P. N. K. Deenapanray and M. Petravic, J. Appl. Phys., 2000, 87(5), 2178 CrossRef.
- Y. Kudriavtsev, A. Villegas, A. Godines, P. Ecker, R. Asomoza, S. Nikishin, C. Jin, N. Faleev and H. Temkin, Surf. Interface Anal., 2000, 29(6), 399 CrossRef.
- K. Kimura, K. Nakajima, Y. Okazaki, H. Kobayashi, S. Miwa and K. Satori, Jpn. J. Appl. Phys. Part 1, 2000, 39(7B), 4663 Search PubMed.
- M. Shaanan and R. Kalish, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 171(3), 332 Search PubMed.
- S. Hofmann, Surf. Interface Anal., 2000, 30(1), 228 CrossRef CAS.
- R. Kasnavi, Y. Sun, G. Mount, P. Pianetta, P. B. Griffin and J. D. Plummer, Electrochem. Solid State Lett., 2001, 4(1), G1 CrossRef.
- K. Wittmaack, J. Vac. Sci. Technol., 2000, 18(1), 1 Search PubMed.
- Z. X. Jiang, P. F. A. Alkemade, C. H. Tung and J. L. F. Wang, J. Vac. Sci. Technol., 2000, 18(2), 706 Search PubMed.
- P. A. W. van der Heide, Surf. Sci., 2000, 453(1–3), L328 CrossRef.
- K. Wittmaack, Surf. Sci., 2000, 453(1–3), L332 CrossRef.
- P. C. Zalm, Mikrochim. Acta, 2000, 132(2–4), 243 CrossRef.
- N. Grignon, J. Jeusset, E. Lebeau, C. Moro, A. Gojon and P. Fragu, J. Trace Microprobe Tech., 1999, 17(4), 477 Search PubMed.
- G. Tamborini and M. Betti, Mikrochim. Acta, 2000, 132(2–4), 411 CrossRef.
- I. C. W. Fitzsimons, B. Harte and R. M. Clark, Mineral. Mag., 2000, 64(1), 59 Search PubMed.
- F. A. Stevie, R. F. Roberts, J. M. McKinley, M. A. Decker, C. N. Granger, R. Santiesteban and C. J. Hitzman, J. Vac. Sci. Technol. B, 2000, 18(1), 483 CrossRef.
- N. Erdmann, M. Betti, O. Stetzer, G. Tamborini, J. V. Kratz, N. Trautmann and J. van Geel, Spectrochim. Acta Part B, 2000, 55(10), 1565 CrossRef.
- J. M. McKinley, F. A. Stevie, C. N. Granger and D. Renard, J. Vac. Sci. Technol., 2000, 18(1), 279 Search PubMed.
- W. Lisowski, A. H. J. Van den Berg, D. Leonard and H. J. Mathieu, Surf. Interface Anal., 2000, 29(4), 292 CrossRef.
- D. N. Bose, P. Banerji, S. Bhunia, Y. Aparna, M. B. Chhetri and B. R. Chakraborty, Appl. Surf. Sci., 2000, 158(1–2), 16 CrossRef.
- J. J. Guan, G. W. Gale and J. Bennett, Jpn. J. Appl. Phys. Part 1, 2000, 39(7A), 3947 Search PubMed.
- T. Kolber, K. Piplits, S. Dreer, E. Mersdorf, R. Haubner and H. Hutter, Appl. Surf. Sci., 2000, 167(1–2), 79 CrossRef.
- L. Wang and L. H. Zhang, J. Electron. Mater., 2000, 29(6), 873 Search PubMed.
- P. Willich and U. Wischmann, Mikrochim. Acta, 2000, 132(2–4), 419 CrossRef.
- C. W. Magee, D. Jacobson and H. J. Grossmann, J. Vac. Sci. Technol., 2000, 18(1), 489 Search PubMed.
- L. Ottolini and R. Oberti, Anal. Chem., 2000, 72(16), 3731 CrossRef CAS.
- F. Zanderigo, S. Ferrari, G. Queirolo, C. Pello and M. Borgini, Mater. Sci. Eng., 2000, B73(1–3), 173 Search PubMed.
- P. Lazzeri, A. Lui, L. Moro and L. Vanzetti, Surf. Interface Anal., 2000, 29(11), 798 CrossRef.
- W. Hsu, G. R. Huss and G. J. Wasserburg, Geochim. Cosmochim. Acta, 2000, 64(6), 1133 CrossRef CAS.
- D. W. Moon, J. Y. Won, K. J. Kim, H. J. Kim, H. J. Kang and M. Petravic, Surf. Interface Anal., 2000, 29(6), 362 CrossRef.
- M. P. Seah, S. J. Spencer, I. S. Gilmore and J. E. Johnstone, Surf. Interface Anal., 2000, 29(1), 73 CrossRef.
- G. Ramanath, J. E. Greene, I. Petrov, J. E. Baker, L. H. Allen and G. Gillen, J. Vac. Sci. Technol., 2000, 18(3), 1369 Search PubMed.
- M. A. Douglas, S. Hattangady and K. Eason, J. Electrochem. Soc., 2000, 147(5), 1893 CrossRef.
- B. W. Schueler and D. F. Reich, J. Vac. Sci. Technol., 2000, 18(1), 496 Search PubMed.
- E. Napolitani, A. Carnera, R. Storti, V. Privitera, F. Priolo, G. Mannino and S. Moffatt, J. Vac. Sci. Technol., 2000, 18(1), 519 Search PubMed.
- J. M. McKinley, F. A. Stevie, T. Neil, J. J. Lee, L. Wu, D. Sieloff and C. Granger, J. Vac. Sci. Technol., 2000, 18(1), 514 Search PubMed.
- G. Gillen, M. Walker, P. Thompson and J. Bennet, J. Vac. Sci. Technol., 2000, 18(1), 503 Search PubMed.
- G. A. Cooke, T. J. Ormsby, M. G. Dowsett, C. Parry, A. Murrell and E. J. H. Collart, J. Vac. Sci. Technol., 2000, 18(1), 493 Search PubMed.
- O. Brox, K. Iltgen, S. Hellweg and A. Benninghoven, J. Vac. Sci. Technol., 1999, 17(5), 2191 Search PubMed.
- S. Chandra, D. R. Smith and G. H. Morrison, Anal. Chem., 2000, 72(3), 104A.
- T. C. Stubbings, S. G. Nikolov and H. Hutter, Mikrochim. Acta, 2000, 133(1–4), 273 CrossRef CAS.
- M. Wolkenstein, T. Kolber, S. Nikolov and H. Hutter, J. Trace Microprobe Tech., 2000, 18(1), 1 Search PubMed.
- K. L. Gavrilov, S. J. Bennison, K. R. Mikeska and R. Levi-Setti, Acta Mater., 1999, 47(15–16), 4031 CrossRef.
- O. Gebhardt and D. Gavillet, J. Nucl. Mater., 2000, 279(2–3), 368 CrossRef.
- V. A. Ukraintsev, P. J. Chen, J. T. Gray, C. F. Machala, L. K. Magel and M. C. Chang, J. Vac. Sci. Technol., 2000, 18(1), 580 Search PubMed.
- T. Kolber, K. Piplits, R. Haubner and H. Hutter, Fresenius' J. Anal. Chem., 2000, 365(8), 636.
- C. Pollak, B. Kriszt and H. Hutter, Mikrochim. Acta, 2000, 133(1–4), 261 CrossRef CAS.
- S. Musser, H. Wildner and H. Hutter, Mikrochim. Acta, 2000, 133(1–4), 253 CrossRef CAS.
- M. Rosner, C. Eisenmenger-Sittner and H. Hutter, Mikrochim. Acta, 2000, 133(1–4), 267 CrossRef CAS.
- L. VanVaeck, A. Adriaens and R. Gijbels, Mass Spectrom. Rev., 1999, 18(1), 1 CrossRef CAS.
- A. Adriaens, L. VanVaeck and F. Adams, Mass Spectrom. Rev., 1999, 18(1), 48 CrossRef.
- B. Hagenhoff, Mikrochim. Acta, 2000, 132(2–4), 259 CrossRef.
- R. Van Ham, A. Adriaens, L. Van Vaeck, R. Gijbels and F. Adams, Nucl. Instrum. Methods Phys. Res. Sect. B, 2000, 161, 245 Search PubMed.
- H. B. Lu, C. T. Campbell, D. J. Graham and B. D. Ratner, Anal. Chem., 2000, 72(13), 2886 CrossRef.
- J. C. Vickerman, A. Oakes and H. Gamble, Surf. Interface Anal., 2000, 29(6), 349 CrossRef CAS.
- G. S. Groenewold, A. D. Appelhans, G. L. Gresham, J. E. Olson, M. Jeffery and J. B. Wright, Anal. Chem., 1999, 71(13), 2318 CrossRef.
- G. S. Groenewold, A. D. Appelhans, G. L. Gresham, J. E. J. M. Olson and M. Weibel, J. Am. Soc. Mass Spectrom., 1999, 11(1), 69.
- H. Oechsner, W. Bock, M. Kopnarski and M. Muller, Mikrochim. Acta, 2000, 133(1–4), 69 CrossRef CAS.
- A. Schnieders and A. Benninghoven, Anal. Chem., 2000, 72(18), 4289 CrossRef.
- X. T. Ai, H. Hutter, M. Gritsch, H. Borner, J. D. Sunderkotter, H. Bubert and H. Jenett, Fresenius' J. Anal. Chem., 2000, 366(1), 41 CrossRef CAS.
- S. Dreer, P. Wilhartitz, K. Piplits, H. Hutter, M. Kopnarski and G. Friedbacher, Mikrochim. Acta, 2000, 133(1–4), 75 CrossRef CAS.
- W. Meier-Augenstein, Curr. Opin. Clin. Nutr. Metab. Care, 1999, 2(6), 465 CrossRef.
- E. Lichtfouse, Rapid Commun. Mass Spectrom., 2000, 14(15), 1337 CrossRef CAS.
- C. B. Douthitt, Analusis, 1999, 27(3), 197 Search PubMed.
- J. L. Brazier and W. Elbast, Analusis, 1999, 27(3), 218 Search PubMed.
- M. Leuenberger, P. Nyfeler, H. P. Moret, P. Sturm and C. Huber, Rapid Commun. Mass Spectrom., 2000, 14(16), 1543 CrossRef CAS.
- C. Ruff, K. Hor, B. Weckerle, P. Schreier and T. Konig, J. High Resolut. Chromatogr., 2000, 23(5), 357 CrossRef.
- P. Lecchi and F. P. Abramson, J. Am. Soc. Mass Spectrom., 2000, 11(5), 400 CrossRef CAS.
- R. Russow and G. Schmidt, Isot. Environ. Health Stud., 1999, 35(4), 274 Search PubMed.
- H. J. Kupka and R. Winkler, Isot. Environ. Health Stud., 1998, 34(1–2), 217 Search PubMed.
- T. Rosenorn, J. Monster and M. S. Johnson, Asian Chem. Lett., 2000, 4(1–2), 101 Search PubMed.
- H. Bao and M. H. Thiemens, Anal. Chem., 2000, 72(17), 4029 CrossRef.
- A. M. Jones, P. Iacumin and E. D. Young, Chem. Geol., 1999, 153(1–4), 241 CrossRef CAS.
- D. F. Ferretti, D. C. Lowe, R. J. Martin and G. W. Brailsford, J. Geophys. Res., 2000, 105(D), 6709 CrossRef CAS.
- U. Tsunogai, F. Nakagawa, Y. Hachisu and N. Yoshida, Rapid Commun. Mass Spectrom., 2000, 14(16), 1507 CrossRef.
- M. Leuenberger, P. Nyfeler, H. P. Moret, P. Sturm, A. Indermühle and J. Schwander, Rapid Commun. Mass Spectrom., 2000, 14, 1552 CrossRef CAS.
- B. J. Roberts, M. E. Russ and N. E. Ostrom, Environ. Sci. Technol., 2000, 34(11), 2337 CrossRef CAS.
- S. Gerst and P. Quay, J. Geophys. Res., 2000, 105(D), 26433 CrossRef.
- J. A. M. Ward, J. M. E. Ahad, G. Lacrampe-Couloume, G. F. Slater, E. A. Edwards and B. Sherwood Lollar, Environ. Sci. Technol., 2000, 34(21), 4577 CrossRef CAS.
- G. G. Salata, L. A. Roelke and L. A. Cifuentes, Mar. Chem., 2000, 69(1–2), 153 CrossRef.
- O. Kracht and G. Gleixner, Org. Geochem., 2000, 31(7–8), 645 CrossRef.
- S. Goupry, N. Rochut, R. J. Robins and E. Gentil, J. Agric. Food Chem., 2000, 48(6), 2222 CrossRef CAS.
- J. M. Rosenbaum, R. A. Cliff and M. L. Coleman, Anal. Chem., 2000, 72(10), 2261 CrossRef.
- H. G. M. Eggenkamp and M. L. Coleman, Chem. Geol., 2000, 167(3–4), 393 CrossRef.
- S. Ward, M. Scantlebury, E. Krol, P. J. Thomson, C. Sparling and J. R. Speakman, Rapid Commun. Mass Spectrom., 2000, 14(6), 450 CrossRef CAS.
- S. L. Herd, W. H. Vaughn and M. I. Goran, Obes. Res., 2000, 8(4), 302 Search PubMed.
- M. Gehre, R. Hoefling, P. Kowski and G. Strauch, Anal. Chem., 1996, 68, 4414 CrossRef CAS.
- D. A. Schoeller, A. S. Colligan, T. Shriver, H. Avak and C. Bartok-Olson, J. Mass Spectrom., 2000, 35(9), 1128 CrossRef CAS.
- H. A. J. Meijer and W. J. Li, Isot. Environ. Health Stud., 1998, 34(4), 349 Search PubMed.
- R. D. Stickrod II and J. D. Marshall, Rapid Commun. Mass Spectrom., 2000, 14(14), 1266 CrossRef CAS.
- C. C. Metges and M. Daenzer, Anal. Biochem., 2000, 278(2), 156 CrossRef CAS.
- R. Aguilera, T. E. Chapman and D. H. Catlin, Rapid Commun. Mass Spectrom., 2000, 14(23), 2294 CrossRef CAS.
- E. Bourgogne, V. Herrou, J. C. Mathurin, M. Becchi and J. de Ceaurriz, Rapid Commun. Mass Spectrom., 2000, 14(24), 2343 CrossRef CAS.
- H. A. J. Meijer, R. E. M. Neubert and G. H. Visser, Int. J. Mass Spectrom., 2000, 198(1–2), 45 CrossRef.
- S. T. Nelson, Rapid Commun. Mass Spectrom., 2000, 14(4), 293 CrossRef.
- A. I. Kostoyanov, V. V. Manoilov, Y. M. Efis and M. V. Rodionov, Instrum. Exp. Tech., 2000, 43(1), 91 CrossRef.
- M. T. McCulloch and T. Esat, Chem. Geol., 2000, 169, 107 CrossRef CAS.
- M. F. Thirlwall, Chem. Geol., 2000, 163, 299 CrossRef CAS.
- J. D. Woodhead and J. M. Hergt, Geostand. Newsl., 2000, 24(1), 33 Search PubMed.
- C. F. You, J. S. Shen and T. Lee, Chenggong Daxue Xuebao, Keji, Yixuebian, 1999, 34, 1 Search PubMed.
- M. J. Bickle, H. J. Chapman and C. F. You, Int. J. Mass Spectrom., 2000, 202(1–3), 273 CrossRef.
- R. H. James and M. R. Palmer, Chem. Geol., 2000, 166(3–4), 319 CrossRef CAS.
- B. P. Datta, Rapid Commun. Mass Spectrom., 2000, 14(8), 696 CrossRef CAS.
- B. P. Datta, Rapid Commun. Mass Spectrom., 2000, 14(8), 706 CrossRef CAS.
- B. P. Datta, Rapid Commun. Mass Spectrom., 2000, 14(17), 1571 CrossRef CAS.
- W. G. Liu, Z. C. Peng, Y. K. Xiao, Z. F. Zhang and Z. R. Wang, Nucl. Sci. Tech., 2000, 11(2), 79 Search PubMed.
- Y. K. Xiao, H. Z. Wei, Y. M. Zhou, Y. H. Wang and W. G. Liu, Anal. Chim. Acta, 2000, 420(1), 95 CrossRef CAS.
- K. Misawa and T. Fujita, Meteorit. Planet. Sci., 2000, 35(1), 85 Search PubMed.
- A. J. Midwood, M. F. Proe and J. J. Harthill, Analyst, 2000, 125(3), 487 RSC.
- D. A. Banks, R. Green, R. A. Cliff and B. W. D. Yardley, Geochim. Cosmochim. Acta, 2000, 64(10), 1785 CrossRef CAS.
- T. F. Nägler and I. M. Villa, Chem. Geol., 2000, 169, 5 CrossRef CAS.
- M. J. Herbel, T. M. Johnson, R. S. Oremland and T. D. Bullen, Geochim. Cosmochim. Acta, 2000, 64(21), 3701 CrossRef CAS.
- M. E. Wieser and J. R. De Laeter, Int. J. Mass Spectrom., 2000, 197, 253 CrossRef CAS.
- R. E. M. Rickaby, M. J. Greaves and H. Elderfield, Geochim. Cosmochim. Acta, 2000, 64(7), 1229 CrossRef.
- Y. Z. Liu, A. Masuda and M. Inoue, Anal. Chem., 2000, 72(13), 3001 CrossRef CAS.
- Z. G. Deng and Z. Z. Wei, Yuanzineng Kexue Jishu, 1999, 33(6), 553 Search PubMed.
- M. Wallenius and K. Mayer, Fresenius' J. Anal. Chem., 2000, 366(3), 234 CrossRef CAS.
- S. Ahmed, M. A. Awan and U. u. A. M. Rehman, J. Radioanal. Nucl. Chem., 2000, 243(3), 723 CrossRef CAS.
- J. W. Ball and R. L. Bassett, Chem. Geol., 2000, 168(1–2), 123 CrossRef CAS.
- S. Gioia and M. M. Pimentel, An. Acad. Bras. Cienc., 2000, 72(2), 219 Search PubMed.
- O. F. Schedlbauer and K. G. Heumann, Anal. Chem., 1999, 71(24), 5459 CrossRef CAS.
- C. F. You and M. J. Bickle, J. Geol. Soc. China, 1999, 42(3), 319 Search PubMed.
- H. Cheng, J. Adkins, R. L. Edwards and E. A. Boyle, Geochim. Cosmochim. Acta, 2000, 64(14), 2401 CrossRef.
- J. Szpunar, Analyst, 2000, 125(5), 963 RSC.
- J. Szpunar and R. Lobinski, Pure Appl. Chem., 1999, 71(5), 899 CrossRef.
- R. Lobinski and J. Szpunar, Anal. Chim. Acta, 1999, 400, 321 CrossRef.
- H. Chassaigne, V. Vacchina and R. Lobinski, Trends Anal. Chem., 2000, 19(5), 300 CrossRef CAS.
- P. Kebarle and M. Peschke, Anal. Chim. Acta, 2000, 406(1), 11 CrossRef CAS.
- P. Kebarle, J. Mass Spectrom., 2000, 35(7), 804 CrossRef CAS.
- C. Ip, M. Birringer, E. Block, M. Kotrebai, J. F. Tyson, P. C. Uden and D. J. Lisk, J. Agric. Food Chem., 2000, 48(6), 2062 CrossRef CAS.
- S. McSheehy, W. Yang, F. Pannier, J. Szpunar, R. Lobinski, J. Auger and M. Potin-Gautier, Anal. Chim. Acta, 2000, 421(2), 147 CrossRef.
- W. Z. Shou, M. M. Woznichak, S. W. May and R. F. Browner, Anal. Chem., 2000, 72(14), 3266 CrossRef.
- A. D. Madsen, W. Goessler, S. N. Pedersen and K. A. Francesconi, J. Anal. At. Spectrom., 2000, 15(6), 657 RSC.
- S. McSheehy, M. Marcinek, H. Chassaigne and J. Szpunar, Anal. Chim. Acta, 2000, 410(1–2), 71 CrossRef CAS.
- J. C. Wu, Z. Mester and J. Pawliszyn, Anal. Chim. Acta, 2000, 424(2), 211 CrossRef CAS.
- C. Demesmay and J. L. Rocca, Spectra Anal., 1999, 28(211), 19 Search PubMed.
- L. Debusschere, C. Demesmay and J. L. Rocca, Chromatographia, 2000, 51(5–6), 262 Search PubMed.
- S. Mounicou, K. Polec, H. Chassaigne, M. Potin-Gautier and R. Lobinski, J. Anal. At. Spectrom., 2000, 15(6), 635 RSC.
- H. Chassaigne, S. Mounicou, C. Casiot, R. Lobinski and M. Potin-Gautier, Analusis, 2000, 28(5), 357 Search PubMed.
- K. Polec, O. Garcia-Arribas, M. Perez-Calvo, J. Szpunar, B. Ribas-Ozonas and R. Lobinski, J. Anal. At. Spectrom., 2000, 15(10), 1363 RSC.
- V. Vacchina, R. Lobinski, M. Oven and M. H. Zenk, J. Anal. At. Spectrom., 2000, 15(5), 529 RSC.
- T. L. Jones-Lepp, K. E. Varner, M. McDaniel and L. Riddick, Appl. Organomet. Chem., 1999, 13(12), 881 CrossRef.
- C. Moulin, N. Charron, G. Plancque and H. Virelizier, Appl. Spectrosc., 2000, 54(6), 843 Search PubMed.
- Z. Mester, H. Lord and J. Pawliszyn, J. Anal. At. Spectrom., 2000, 15(6), 595 RSC.
- M. C. B. Moraes, J. G. A. B. Neto and C. L. Do Lago, Int. J. Mass Spectrom., 2000, 198(1-2), 121 CrossRef CAS.
- S. Mollah, A. D. Pris, S. K. Johnson, A. B. Gwizdala III and R. S. Houk, Anal. Chem., 2000, 72(5), 985 CrossRef.
- A. R. S. Ross, M. G. Ikonomou and K. J. Orians, J. Mass Spectrom., 2000, 35(8), 981 CrossRef.
- W. Z. Shou and R. F. Browner, Anal. Chem., 1999, 71(16), 3365 CrossRef CAS.
- A. Beichert, S. Padberg and B. W. Wenclawiak, Appl. Organomet. Chem., 2000, 14(9), 493 CrossRef CAS.
- X. M. Yu and J. Pawliszyn, Anal. Chem., 2000, 72(8), 1788 CrossRef CAS.
- P. W. Looser, M. Berg, K. Fent, J. Muhlemann and R. P. Schwarzenbach, Anal. Chem., 2000, 72(20), 5136 CrossRef.
- K. Ndungu and L. Mathiasson, Anal. Chim. Acta, 1999, 404(2), 319.
- A. Gallino, F. Magno, L. Tallandini, T. Passaler, G. U. Caravello and P. Pastore, Rapid Commun. Mass Spectrom., 2000, 14(6), 373 CrossRef CAS.
- M. Takeuchi, K. Mizuishi and T. Hobo, Anal. Sci., 2000, 16(4), 349 Search PubMed.
- Z. Mester and J. Pawliszyn, J. Chromatogr. A, 2000, 873(1), 129 CrossRef CAS.
- M. Guidotti, J. AOAC Int., 2000, 83(5), 1082 Search PubMed.
- M. Pantsar-Kallio and A. Korpela, Anal. Chim. Acta, 2000, 410(1–2), 65 CrossRef CAS.
- J. R. Baena, S. Cardenas, M. Gallego and M. Valcarcel, Anal. Chem., 2000, 72(7), 1510 CrossRef CAS.
- J. L. Gomez Ariza, J. A. Pozas, I. Giraldez and E. Morales, Talanta, 1999, 49(2), 285 CrossRef CAS.
- V. Ducros, F. Laporte, N. Belin, A. David and A. Favier, J. Inorg. Biochem., 2000, 81(1–2), 105 CrossRef.
- U. Beyerle, W. Aeschbach-Hertig, D. M. Imboden, H. Baur, T. Graf and R. Kipfer, Environ. Sci. Technol., 2000, 34(10), 2042 CrossRef CAS.
- T. Itaya, T. Okada, Y. Honjou and H. Hyodo, Shizen Kagaku Kenkyusho Kenkyu Hokoku, Okayama Rika Daigaku, 1999, 25, 41 Search PubMed.
- J. Nuzzo, M. Hyman, M. W. Rowe, M. N. Rao, R. L. Palma and J. Westrich, Proceedings of the Indian Academy of Sciences—Earth and Planetary Sciences, 2000, 109(1), 89 Search PubMed.
- A. Makarov, Anal. Chem., 2000, 72(6), 1156 CrossRef.
- P. Spanel and D. Smith, J. Am. Soc. Mass Spectrom., 2000, 11(10), 866 CrossRef CAS.
- E. Rosenberg, V. Kmetov and M. Grasserbauer, Fresenius' J. Anal. Chem., 2000, 366(4), 400 CrossRef CAS.
| This journal is © The Royal Society of Chemistry 2001 |