Atomic Spectrometry Update. Atomic mass spectrometry
Jeffrey R.
Bacon
*a,
Jeffrey S.
Crain
b,
Luc
Van Vaeck
c and
John G.
Williams
d
aThe Macaulay Land Use Research Institute, Craigiebuckler, Aberdeen, UK AB15 8QH
bUnion Carbide Corporation, 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 26th July 2000
Abstract
This Update follows on from last year's,1 but only covers a period of approximately ten months as the production schedule has been brought forward by two months. The format remains the same with some minor changes in the section headings. 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. Routine 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, foods and beverages.4
Trends noted in last year's Update have generally been maintained. Although the scope of this review includes the rapidly increasing use of MS for speciation analysis, reported developments should involve or be intended for studies of speciation in natural systems. However, some degree of judgement has still been required to set the limits. Attention has continued to be paid mainly to sample preparation and introduction, in particular to meet the need for analysis of smaller samples. The sensitivity of many of the MS techniques is exemplified by the growing use of AMS in biomedical and pharmaceutical studies, made possible through the use of such low levels of radiocarbon tracers that the doses fall below natural background levels and regulatory limits. The overall trend in published material on ICP-MS continues to grow mainly in the applications and sample introduction areas, at the expense of fundamental studies. The exception to this is a considerable activity in the development and application of collision and reaction cells for reduction or removal of polyatomic species. Many applications exploit the isotope capability of MS, identified in last year's review as a growth area for ICP-MS.
1 Accelerator mass spectrometry (AMS)
1.1 Reviews
The substantial 52-page review of Fifield5 on state-of-the-art AMS was divided into two sections. The first covered the general methodologies, with specific examples for the commonly measured long-lived isotopes (for example, 26Al, 10Be, 14C, 36Cl and 129I). The second considered the wide range of typical applications grouped according to research area rather than individual isotopes. The review of Rucklidge et al.6 on the in-situ determination of trace elements in solid materials (Accelerator SIMS) included aspects of ion source design, neutral injection and applications. Hellborg et al.7 gave examples of 14C-labelled pharmaceuticals used in biomedical research to demonstrate the advantages (smaller, mg size, samples and shorter measuring times of less than an hour) of AMS over conventional radiometric methods.1.2 Instrumentation
The increasing demand for a high throughput of samples with high sensitivity has led to the commercial development of a high-intensity, Cs sputter, negative ion source, known as the MC-SNICS source.8 Cathode change times of 2 s and typical beam current rise times of 3 s could be achieved for the source, which held 134 samples under vacuum. A change of sample wheel was quick enough to allow data acquisition to be resumed within 45 min. Ion currents for 12C were typically 100 µA. The new source forms part of a system in use at the first commercial AMS service and is also to be installed at the Woods Hole facility in order to extend capabilities to the analysis of smaller samples than is currently possible.9 Carbon dating precisions of 0.1–0.2% have been demonstrated using a 40 sample MC-SNICS source installed at the AMS facilities in Japan.10Further details have been presented11 for the microbeam ion source in use at the AUSTRALIS facility in Australia for ultra-sensitive trace element and isotopic studies. Lead isotopes were measured as Pb4+ ions from either Pb−, PbO− or PbS− injected ions. Sulfur isotopes were measured as S2+ or S3+ ions from injected S− ions. A fast bouncing system, intended for high precision isotope ratio measurements, was demonstrated to be capable of precisions of 0.04% at the low energy side of the instrument.
The attractions of a compact microwave ion source over conventional sputter ion sources for radiocarbon analysis were considered by Wills and colleagues12 to be the ability to analyse samples containing as little as 100 nmole of C and, for samples containing <3 µmole of C, the simplified sample treatment and greater throughput. This would be an advantage in the analysis, for example, of atmospheric aerosols, which typically contain <1 µmole of C. Although the source was designed several years ago, further development commenced only recently and a fully operational system had not yet been demonstrated. Improvements in the current design included a reduced plasma chamber volume and inclusion of viewing ports to allow observation of the plasma discharge.
The importance of well designed injection systems was demonstrated by the report of experience with the system designed for the Shanghai minicyclotron AMS.13 Over five years' experience of operation of the system revealed a number of limitations which have required the design of an improved system. A feature of the new system was an analytical magnet to remove the high intensity ion beams (12C− and C2−) which had previously been injected into the cyclotron together with the14C− beam. Other problems included large image magnification, space charge effects and different trajectories for the three C ions. Litherland et al.14 have contemplated the revival of neutral injection, in particular for production of continuous positive ion beams with energies ranging from 2.5 MeV for light ions to 5 MeV for heavy ions. Such energies would be suitable for most elements and would allow optimization of the production of neutrals. Key features in the proposed design would minimize the drift distance, scattering and optical errors which cannot be controlled directly for neutral species. There have been few, if any, detailed studies on important aspects of this design, for example scattering during neutralization or re-ionization. The viability of the proposed system remained to be demonstrated.
Significant improvements in the determination of 26Al and 10Be have been observed by Matsuzaki et al.15 through the use of a ΔE-E type gas counter. The 10B interference on 10Be was completely removed but an unidentified interference remained. In the determination of 26Al, the 17O2+ interference was completely eliminated.
1.3 Developments in radiocarbon analysis
There have been limited developments in radiocarbon analysis in the period covered by this Update. The major thrust in analysis of small samples has resulted in the setting up in the UK of the first fully commercial company in the world to offer AMS analysis.16 The key areas of application were seen to be biomedical and environmental sciences. Achieved sensitivities of from 10−21 to 10−18 mole in mg samples allowed the use of 10 nCi or less of 14C-labelled drugs in biomedical studies. These doses are sufficiently small for the total radiation exposure to be only a fraction of the background radiation and for the generated waste to be generally classified as non-radioactive. Increased application of AMS was seen as a way to reduce the need for widespread use of animals in biomedical research. Other workers17,18 have also identified the advantages of AMS for the long term tracing of small doses of nutrients, toxins and drugs after administration to healthy human subjects.Sample preparation has been receiving attention in Japan. In order to reduce background levels, Haruhara et al.19 processed geological samples by reduction of CO2 (produced from the sample) to graphite, re-oxidation to pure CO2 and reduction to amorphous carbon used for the AMS analysis rather than the conventional graphite. Aramaki20 has been using radiocarbon as a tracer of sea-water. A new simplified method for stripping CO2 from sea-water reduced the processing time and increased the yield in comparison with other methods.
1.4 Developments in the analysis of elements other than carbon
Long-lived gas radionuclides could play an important role in environmental research as they are chemically inert with the consequence that their geochemical behaviour is simpler to understand than that of reactive elements. Kutschera and colleagues21 demonstrated that the long-lived radionuclide 81Kr was unaffected by anthropogenic contributions and was suitable for dating applications, in particular the dating of deep ice and old groundwater. Within experimental uncertainty (±30%), the 18Kr∶Kr ratios in pre-nuclear and modern Kr were found to be the same, of the order of 10−13. The main technical problem of separating 81Kr from the isobaric 81Br background was overcome by using a Superconducting Electron Cyclotron Resonance ion source coupled to a Superconducting Cyclotron which provided ions of sufficient energy (45 MeV) for efficient full stripping. The fully stripped 81Kr36+ ions were separated from the 81Br35+ ions in a TOFMS instrument. The lower Br concentration (and resultant background) of Ni stripper foils made their use preferable to that of Be foils, even though stripping efficiency was reduced.Data for atmospheric concentrations of 129I are scarce because of the analytical difficulties in determining 129I at very low concentrations. López-Gutiérrez et al.22 presented methods for the radiochemical extraction of 129I from atmospheric charcoal filters and determination of 129I concentration (LOD 104 atoms m−3) or the 129I∶I ratio (LOD 10−10) by AMS. Air was pumped for one week through an activated charcoal filter to sample a volume of 350 m3. The adsorbed iodine species were extracted following addition of NaI carrier to provide sufficient iodine for stable MS measurement of the I5+ ion.
Brauer et al.23 reported that the preparation of Al2O3 targets for the AMS determination of 26Al in digested brain samples resulted in a glass-like material thought to be aluminium oxyphosphate. Removal of P from acid-digested brain samples using a cation-exchange procedure produced an amorphous material after ashing that was easier to handle and gave higher Al beam currents in the AMS analysis.
Long-lived radionuclides produced by cosmic rays in extraterrestrial matter not only preserve information about the objects themselves but also provide information on the spectral distribution and constancy of the cosmic flux. Radiochemical separation of the nuclides is necessary in order to achieve AMS analysis at the low concentrations found. Merchel and Herpers24 have improved procedures for the separation of heavier nuclides (26Al, 10Be, 41Ca, 36Cl, 60Fe, 53Mn, 59Ni) from meteoritic matter with particular emphasis on the efficient removal of interfering isobars but less so on high chemical yields. Although some of the procedures are long and complex, the quality of separation was proved by pioneering measurements at several AMS facilities.
2 Glow discharge mass spectrometry (GDMS)
2.1 Reviews
Bogaerts and Gijbels25 published a general review with 86 references on applications and new developments in ion sources. Consideration of basic aspects of the method and fundamental research on the plasma was illustrated with examples of analysis of high purity metals and alloys, semiconductors and non-conducting samples. In addition to the use of GD as an analytical source, GDs are widely used in micro-electronic industry, lasers, various kinds of light sources and in the new technology of plasma displays. Another review26 dealt with the use and modelling of GD in these fields of application.Saprykin27 reviewed the present state-of-the-art of rf GD sources for high mass resolution MS analysis of solid materials. The separation in time and space of the desorption and ionization steps was seen as a major advantage for GDMS in comparison with, for example, SIMS. The rf technique has matured sufficiently for practical applications in survey analysis and the depth profiling of semiconductors and ceramics. Optimization of the source parameters and energy selection before mass analysis improved the S/N by two orders of magnitude. Standardless (semi-)quantitative analysis was considered feasible.
In their review of tuneable plasma sources, which formed an introduction to a special issue of this journal, Marcus et al.28 discussed the issue of 'chemical speciation' and the increasingly blurred edges of atomic and molecular spectroscopies. Over recent years, a growing amount of research has been focused on the introduction of separated analytes into existing spectrochemical methods, making the speciation experiment a matter of time-resolved elemental detection. The distinction between molecular species was made by the chromatographic separation. Recently, increasing attention has been given to spectrochemical sources which allow explicit information on molecular components to be obtained and thereby add an extra dimension to the speciation information.
2.2 Instrumentation
Guzowski et al.29 developed a compact, self-igniting and inexpensive dc GD ionization source for gaseous analytes. The introduction tube was mounted flush with the bottom of a cooled hollow cathode and ceramic sleeves confined the discharge region. Switching from the atomic to the molecular mode of operation was achieved by changing the gas composition from He to Ar and adjusting the operating pressure and current. Minimum sample flow rates for atomic ion detection were 20–90 pg s−1. A precision of 0.4% RSD for the 79Br∶81Br isotope ratio was attained over a period of 2.5 h.Saprykin30 reported on the incorporation of a magnet in the sample holder of rf GDMS for the analysis of non-conducting glass samples. Optimized conditions for the analysis of NIST SRMs 611 and 613 led to a 4-fold increase in atomic ion signals and a 10-fold reduction in cluster ion intensities. Elemental LODs were 10–100 ng g−1 and 100–1000 ng g−1 at mass resolutions of 300 and 3000, respectively. The repeatability was 5–15% RSD and the relative sensitivity factors (RSFs) fell within a range of only 0.2–3.
The removal of matrix ions in plasma source (GD or ICP-MS) TOFMS through the use of traditional two-plate deflectors involves the risk of backscattering and detection of deflected ions. Therefore, Hang et al.31 mounted two additional plates at an angle of 40°, forming a 'sleeve' to collect the scattered ions. Thereby, the ArH+ signal was eliminated without influencing the Ar+ or CO2+ peak intensities, whereas the Cu+ signal from NIST SRM 495 was reduced by 4 orders of magnitude without affecting the Fe+ signal.
2.3 Analytical methodology
Inoue and Saka32 presented a comparison of powder sample preparation methods for quantitative dc GDMS analysis of metals and alloys. Compressing the analyte into pin electrodes without binder caused significant variability, whereas mixing with a carbon binder led to contamination. Therefore, pressing the powders into an indium sheet was preferred. Preliminary sputtering for 60 min removed the surface contamination. The RSFs and the peak intensities of the analyte relative to the matrix ions were independent of the powder grain size (60 or 110 µm) or the loading density. The RSFs were within 20% of those for disc-shaped bulk samples for all elements, except P and V. Comparison of GDMS with wet chemical analysis showed an agreement within 25% for trace element concentrations <10 µg g−1 and 3% for concentrations above 0.1%.The secondary cathode technique for dc GDMS of non-conducting samples is based on the formation of an electrically conducting film on the sample by redeposition of metal atoms, sputtered from a tantalum mask. Wayne et al.33 investigated crater formation and the characteristics of films deposited on samples such as glass slides, Nd∶YAG, potassium titanyl phosphate (KTP) and glass NIST SRMs 610 and 612. Crater shape depended on sample composition. Whereas KTP gave near-cylindrical flat bottom craters, even after 8 h of sputtering, Nd:YAG samples yielded irregular craters. The film thickness varied between 1015 to 1016 atoms cm−2.
Lewis et al.34 demonstrated the usefulness of time-gated detection in pulsed GD for the determination of 40Ca+ in the presence of 40Ar+ by using a triple quadrupole instrument. Sampling at the correct time in the afterpeak regime avoided the interference from the Ar+ produced by electron ionization and enhanced the S/N of the Ca+ produced by Penning ionization. The linearity of the 40Ca+ signal intensity as a function of concentration showed complete elimination of the 40Ar+ interference. The optimal sampling time in the rf source came earlier than in the dc case, reflecting the faster ionization and recombination processes under rf conditions. In this demanding case of distinguishing between Ca+ and Ar+, LODs were 24 and 6 ppm for the rf and dc GD sources, respectively.
The use of TIMS for isotope ratio determinations in nuclear fuel samples requires laborious chemical separations. Chartier et al.35 evaluated high mass resolution GDMS as an alternative to ID TIMS for the determination of Er in nuclear fuels. The home-built Mattauch–Herzog instrument allowed a mass resolution of 800 to be achieved. The isotope ratios of Er measured by GDMS agreed within 0.5% with TIMS results, except for 162Er∶166Er. The 168Er∶238U ratio determined by GDMS in Er-doped molybdenum–uranium fuel samples was within 3% of the value from ID TIMS with a double spike solution. Although TIMS remained superior in accuracy, GDMS allowed rapid screening with acceptable accuracy to be achieved.
3 Inductively coupled plasma mass spectrometry (ICP-MS)
3.1 Reviews and fundamental studies
The ICP-MS technique has attracted many reviews in this reporting period, often on specific areas rather than the system as a whole. The book edited by Montaser, 'Inductively Coupled Plasma Mass Spectrometry', was notable as it covered many aspects of the technique, such as an introduction to ICP spectrometries for elemental analysis,36 plasma generation,37 instrumentation for low- (LR) and high-resolution (HR) analysis,38 analytical characteristics,39 fundamental considerations,40 novel applications,41 mixed gas and helium ICPs42 and MIPs.43 In 'Inductively Coupled Plasma Spectrometry: Its Applications', edited by Hill, there were a number of useful reviews, such as fundamental aspects,44 isotope ratio measurements,45 alternative plasmas and sample introduction46 and environmental applications.47 In addition to these, there was a 120-reference review of the advances in ICP-MS instrumentation and methodologies.48 Olesik et al.49 have reviewed sensitivity and matrix effects in ICP-MS, concentrating on aerosol processing, ion production and ion transport. In a broad review with 161 references, Todoli and Mermet50 considered acid interferences in atomic spectrometry, in particular analyte signal effects and their subsequent reduction. In a detailed review (60 references) of B isotope ratio analysis, Sah and Brown51 concluded that, with recent developments in HR technology, ICP-MS is becoming the technique of choice because of its convenience and high sample throughput. The authors highlighted the limitations to improved precision and accuracy as being analyte loss and isotope fractionation during sample preparation, low-concentration effects, difficult matrices and memory effects. Several solutions were offered, but the authors concluded by indicating that analytical limitations would continue to be due to sample-related issues rather than the instrumentation. Douthitt52 presented a comprehensive bibliography (192 references + talks and posters) of sector field ICP-MS, including both high resolution and multi-collector devices.Farnsworth and co-workers continue to report on the use of laser fluorescence for characterization of the ion beam behind the skimmer. Velocity profiles of Ar metastable atoms were determined from the Doppler shift of the exciting radiation and were found to be bimodal.53 This was considered by the authors to be indicative of a flow disturbance at, or slightly upstream of, the skimmer orifice. Additionally, the ion beam was characterized 8 mm downstream of the skimmer cone tip and at the apex of the sampling cone, using Ba and Sc ions, and Pb atoms.54 This revealed that Sc ion densities decreased the most rapidly. The addition of either Mg or Pb solutions suppressed the transmission efficiencies of both Ba and Sc. Both matrices were found to have approximately the same effect. A lower limit of 0.3% Ba+ transmission from the plasma into the 2nd vacuum stage was calculated. The group also reported the effect of a shield torch55 on Ba, Pb and Sc (at 200, 100 and 50 mg l−1, respectively) ion and atom densities as well as on those of Ar metastable species. This study showed that the shield had no effect on the density of test species upstream of the sampling cone, whereas inside the skimmer cone the densities dropped when the shield was floated. The velocities of the test elements did not change on floating or grounding the shield.
Pupyshev and co-workers56,57 proposed multicomponent quasi-equilibria thermodynamic models of thermochemical processes in the ICP and ion transport models of the ion beam for ICP-MS. The authors believed they demonstrated a good agreement between experimental data and theoretical results. The model could be used to determine the total ion current and average ion mass at the interface entrance. This was considered necessary for quantitatively analysing the processes of ion loss during their transport in the ion beam.
Stewart and Olesik58 reported on an investigation of space-charge effects of high concentrations of lead matrix ions on Li analyte ions in ICP-MS. This was achieved using time-resolved measurements with single droplet introduction. The ICP-MS instrument was vertically orientated, in contrast to a conventional (horizontal) arrangement, which was reported to achieve greater reproducibility and stability in droplet-to-droplet sample introduction using a monodisperse microparticulate injector. Typical variation in the droplet-to-droplet arrival times and MS peak analysis areas was better than 5% RSD. This precision allowed a more quantitative description of the space-charge effect on a single cloud of ions to be obtained. Both radial and axial space-charge effects were observed in the ion beam. However, these effects had different actions on the ion beam. It was reported that when the Pb concentration in the sample was sufficiently high the Li+ signal had a bimodal peak shape. This suggested to the authors that space-charge effects were localized in the region of highest charge density occurring just after charge separation.
Sartoros and Salin59 illustrated the automatic selection of internal standards in ICP-MS using a cluster analysis algorithm. Samples contained 25 analytes, spanning the atomic mass and ionization potential ranges, and a single sample–matrix element (e.g. Ba, K, Mg, Na, Pb or Zn). The cluster analysis algorithm used kinetic energy, ionization potential, oxide bond strength, hydride bond strength and electronegativity to group the analytes, although these variables were weighted differently in the various matrices. The authors described the performance of the clustering method and selection of internal standards as 'good' for most analytes in the various matrices studied.
Correct adjustment for detector dead time is important for accurate analysis. Held and Taylor60 described a calculation method based on isotope ratios for the determination of dead time and its uncertainty. Isotope ratios were used as they were more precise than ion current measurements and resulted in values with a measurable uncertainty. The authors reported typical relative uncertainties of the dead time for a continuous dynode detector of 10–20%.
Fairman and co-workers61 presented a practical method for the measurement of uncertainty associated with ICP-MS analysis. A cause and effect diagram was constructed to aid in the identification of uncertainty sources associated with the determination of 60Ni in aqueous samples. The uncertainty estimate was calculated from a combination of existing QC data and specially planned experiments. The dominant contributions to the uncertainty budget were method precision, instrument drift and bias measured as method recovery. The uncertainties associated with the concentration of the working standard and sample dilution were considered insignificant by the authors.
Barnes and co-workers62 characterized theoretically and verified experimentally memory effects of B. The authors reported that the memory effect originated from the tendency of B to volatilize as boric acid from the sample solution, forming a layer that coats the inside surface of the spray chamber. Interface components, ion lenses, quadrupole and other components were not considered responsible for memory effects at trace levels. Addition of a small amount of ammonia solution with the sample was found to be effective in reducing boron signals to background levels in less than 20 s in a conventional concentric nebulizer–double pass spray chamber system. This method converted B from the volatile boric acid form to non-volatile ammonium borate.
3.2 Instrumentation
There has been a flurry of publication activity on the development and use of devices inserted in the ion lens system at some point after the interface and before the mass analyser. These permit the reduction or removal of plasma-based polyatomic ion interferences. Tanner and co-workers produced a series of reports describing the dynamic reaction cell (DRC) for ICP-MS. The DRC63–67 is an enclosed quadrupole introduced between the ion lenses and mass analyser high vacuum chamber. The cell could be operated either in a pressurized mode with a reactive gas to eliminate interferences or unpressurized ('standard' mode). There was no degradation in analytical performance in standard mode. With the unit pressurized, LODs were greatly improved for elements such as Fe, normally subject to polyatomic interference. Applications of the new technology that have been demonstrated included: the removal of carbon- and chloride-based spectral interferences in the determination of As, Cr and Mn;68 the low ng l−1 level determination of Ca, Fe and K in 30% hydrogen peroxide;69 and the determination of 41 elements in high purity acids, with LODs in pg g−1. An interesting additional benefit of the device was an improvement in the precision of isotope ratio measurements.70 Collisions in the DRC broadened the residence time of ions, thereby damping short term fluctuations and effectively reducing plasma noise on the millisecond time scale. For example, for a 40 ppb Ag solution this led to a factor of two improvement in 109Ag : 107Ag ratio precision to 0.023%. This was close to the statistical limit of 0.012%.Jakubowski and co-workers reported on the instrumental aspects71 and analytical figures of merit72 of a hexapole collision and reaction cell in ICP-MS. Helium was used as buffer gas and H2 as a reaction gas. The latter was found to be an effective means of reducing typical Ar polyatomic species by up to 4 orders of magnitude. Molecular species formed in the cell could be suppressed through the use of a retardation electric field established by a dc hexapole bias potential of −2 V. The application of the cell gases also resulted in improved sensitivities, which were lowest for Be (7 × 107 counts s−1 per µg ml−1) and highest for Ba (6 × 108 counts s−1 per µg ml−1), with RSDs generally <0.1%. Elements such as As, Ca, Cr, Fe, K and Se could be determined in HNO3 and HCl or in methanol. The authors reported LODs of 6 pg ml−1 for Cr in 2% methanol, and 23 and 9 pg ml−1 for As and Se, respectively, in 0.28 M HCl. Shakra73 described the accurate determination of As in RM NASS-3 using the hexapole technology. The author reported high sensitivities due to the high ion transport efficiencies and removal of ArCl– polyatomic species. Intriguingly, it was reported that no sample preparation was necessary as the 'cation exchange modified the matrix by removing Ca and Na and so minimized the problems of CaCl− and Na suppression….'.
Koppenaal and co-workers74 described the analytical performance of a plasma rf quadrupole ion trap for both elemental and isotopic analysis. Overall performance was comparable to a conventional quadrupole system with LODs in the range 0.3–5 pg ml−1. Calibration graphs linear over almost six orders of magnitude were achieved by using variable injection times to control the ion density in the trap. Abundance sensitivity was from 105 to 106 and mass resolution was typically a factor of 10 better than a conventional quadrupole. Polyatomic species were decomposed by collisional dissociation techniques. Isotope ratio precision was reported to be around 0.7% but the authors believed that this could be improved by use of a pulse counting detection system.
Douglas and co-workers75 presented both a discussion of the principles of the quadrupole mass filter and a review of work carried out on quadrupoles operating in the second and third stability regions. The authors reported that operation of quadrupoles in these higher zones of stability permitted a resolution of up to 9000 (half height) at m/z 59 and suggested that a high resolution quadrupole ICP-MS system might be possible.
The analytical performance of an axial ICP-TOFMS has been reported by Adams and co-workers.76 Typical LOD values of 0.5–20 pg ml−1 were reported for 64 elements. Long term stability over 4 h for the raw signals ranged from 1% RSD for 7Li to 2.1% RSD for 40Ca, operating under cold plasma conditions. Under normal plasma conditions the corresponding values were 6.9% and 12.8% RSD for 208Pb and 59Co, respectively. Results from the analysis of 16 elements in SRM NIST 1643d (trace elements in water) were considered accurate and isotope ratio precision was in the range 0.07–0.7% RSD for a short data acquisition period.
Becker and Dietze77 investigated the capability of a double-focusing magnetic sector ICP-MS system with a shielded torch to determine long-lived radionuclides (241Am, 237Np, 239Pu,226Ra, 230Th and 238U) at the 1 ng l–1 concentration level. Four different nebulizer/spray chamber configurations were compared as part of the assessment. Overall, the microconcentric nebulizer with minicyclonic spray chamber offered the best performance, consuming the least sample mass (0.4 pg) and providing the highest sensitivities (2 × 109 and 0.4 counts s−1 ppm−1 with and without a shielded torch, respectively). The LODs were sub-pg l−1 and precision ranged from 1 to 2% RSD (n = 5) at the 1 ng l−1 concentration level. Ultrasonic nebulization provided a further 1 order of magnitude increase in sensitivity, but at the expense of a 25-fold increase in sample consumption. Unlike the other devices, a direct injection high-efficiency nebulizer exhibited a decrease in signal when used with a shielded torch. This was believed to be due to the high solvent load delivered by this type of nebulizer. Application of the methodology to the determination of Pu isotope ratios in aqueous solutions and radioactive wastes provided precisions of 0.2, 2 and 14% for concentrations of 1000, 100 and 10 pg l−1 (equivalent to absolute masses of 500, 50 and 5 fg), respectively.
Caruso and co-workers78 reported a practical application of a He-ICP low/reduced pressure-ionization source for MS detection of GC eluents from organobromine compounds and derivatized organotin compounds. The system allowed both total elemental and chemical structure information to be acquired. The LODs for bromobenzene, 1-bromoheptane and benzylbromide were 11, 6 and 4 pg, respectively. Molecular ions for all three organobromine compounds were obtained, as well as characteristic alkyl chain fragments for 1-bromoheptane which were reported to resemble electron impact mass spectra.
3.3 Sample introduction
An important parameter in LA is the focus of the beam. Wanner et al.81 described the use of an autofocus system, built to achieve reproducible ablation conditions. Automatic focusing on any sample with an accuracy of 10–50 µm in the focal plane of the laser, or at a height specified by the user, was reported to correct for surface roughness. The applicability of the system was demonstrated by analysis of archaeological samples with a lateral resolution of 50 µm. Absolute LODs of 0.1–1.4 pg were achieved.
The wavelength of the laser light is an important consideration. Some work continues with IR lasers, although there is general agreement that systems operating in the UV are more appropriate ablation sources. Wennrich et al.82 reported on the influence of pellet composition in IR LA-ICP-AES and -MS and found that the sensitivity for Ca, Fe, La, Mg, Mn, Sr, V and Zn in targets prepared from powders of geological materials depended on the binding material. Six binders were tested (LiBO2, KBr and powders of C, Cu, polyamide or PTFE). In some cases RhCl3 was added to the binder as an internal standard. The authors concluded that the use of suitable binding materials minimized differences in the ablation behaviour of analytes and internal standards associated with the variation of chemical and physical properties of the geological matrices. Günther and Heinrich compared the performance of two lasers operating in the UV83 and studied the influence of He–Ar carrier gas mixtures on sensitivity.84 They found that there were a few small advantages of a 193 nm wavelength excimer laser over a Nd∶YAG operating at a wavelength of 266 nm. A notable difference between the two lasers was that analyte signal intensity in a He–Ar carrier gas mixture (compared to Ar only) was enhanced by up to two-fold with the 266 nm wavelength laser and consistently 2–3 times with the 193 nm wavelength system. The use of He improved LODs for both systems and reduced visible (>1 µm size) particle deposition. However, memory effects were present with the 266 nm system yet absent with the 193 nm system. Reproducibility between multiple analyses was slightly better for the shorter wavelength laser than for the 266 nm system, yielding 2–5% RSDs for major and minor elements and 7–15% for concentrations below 10 µg g−1.
Mank and Mason85 reported on a study of elemental fractionation effects during ablation of deep craters (>200 µm depth and >1000 laser pulses) in depth-profile experiments on SRM glasses (NIST 610, NIST 126A). The effects of laser wavelength and power density, masking apertures and beam focus position were studied. Experiments were carried out with both Ar and He carrier gas on three commercially available LA systems. The authors advised the use of large ablation craters for depth-profile analysis and the addition of He to the Ar carrier gases to reduce elemental fractionation. For large diameter craters, the authors reported there was less interaction between ablated material and the surface of the crater wall in the intermediate and crater opening sections.
Inter-element fractionation and correction in LA-ICP-MS was discussed by Chen.86 In this study, using a 266 nm UV laser, the inter-element fractionation behaviour of Ga, V and Zn in the synthetic silicate glass SRM NIST 613 was found to be different from that in the quenched glass fused silicate rock RMs (BCR-2 and SY-4). The fractionation of Ga, V and Zn, relative to Ca, was less in SRM NIST 613 than in BCR-2 and SY-4. The element intensity, normalized to an internal standard, was not found to be linear with time for a LA period of 210 s. The author concluded that data acquisition using prolonged LA without a matrix match would not improve precision and accuracy for elements whose fractionation behaviour was different from that of the internal standard.
Calibration is an ongoing issue in LA-ICP-MS. Borisov et al.87 used three laser systems (20 ns KrF excimer, 6 ns and 35 ps Nd∶YAG) to study the LA behaviour of a suite of 10 Cu–Zn binary alloys. Non-linear calibration for both Cu and Zn was observed for all three lasers, despite significant differences in laser ablation mechanisms and good stoichiometry of ablated mass. Crater volume measurements were used to determine the mass of sample removed during repetitive LA from each sample. The authors believed that a change in mass ablation rate for samples with different compositions explained the observed phenomena. Linear calibration lines for Zn could be achieved by normalization to Cu signal intensity or crater volume. Sharp and co-workers88 described a novel approach to solving the problem of calibration that employed aqueous standards whose absorption coefficients were modified, by addition of a chromophore, to produce the desired ablation yield. Chromophores for the important laser wavelengths at 193, 248 and 266 nm were given. Investigation of the mechanism of ablation and parametric dependences for the modified aqueous standards led the authors to the conclusion that ablation proceeds by a three step process leading ultimately to ablation of the bulk liquid. This was considered important as such a process should not involve fractionation between elements. Calibration lines defined using this method were linear and reproducible, but internal correction was required to link calibration to real samples. The results obtained from the analysis of a low density polyethylene and SRM NIST 613 (trace elements in glass) using this method agreed with those obtained by other techniques or reference values.
Wang et al.89 described the direct LA-ICP-MS analysis of airborne particulates collected on PTFE-membrane filters. The effect of energy, pulse type and beam focus of the 1064 nm laser, as well as thermal properties and homogeneity of the sample and carrier gas flow rate, were all investigated. The authors revealed that the best ablation efficiency was obtained with a 160 mJ single laser shot operated in free-running mode and using a 6.5 mm defocus distance from the filter surface. Optimum transport efficiency was obtained at Ar flow rates of <0.8 l min−1. The same group also determined Cr in airborne particulates90 and found that the results correlated well with values obtained by high resolution solution ICP-MS. The LA approach provided the authors with a rapid direct analytical technique with a detection limit of 0.05 µg per filter.
Applications of LA-ICP-MS are increasing in number, in particular in geological sciences, as the technology becomes established. Booth et al.91 determined environmentally sensitive elements, such as As, Cd, Hg, Mo, Sb and Se in coals. Calibration was achieved using a series of uncertified coals and the method evaluated using South African coal RMs SARM 18, 19 and 20. Results were considered accurate and precise, even at low concentrations, with LODs in the ng g−1 range.
Campbell and Humayun92 determined platinum group elements PGEs in iron meteorites using a laser system coupled to a high resolution ICP-MS instrument. This offered high sensitivities, low background and flat topped peaks. Bulk composition was determined from both point and line-scan analyses using Co as an internal standard. The Pd, Rh and Ru distributions across a single taenite lamella were recorded by scanning a laser beam, of about 40 µm diameter, across the meteorite surface at 5 µm s−1. Rapid scanning allowed analyte and internal standard data to be obtained every 20 µm of travel by the laser beam. The authors described the technique as accurate with sub-µg g−1 LODs.
Shepherd et al.93 compared LA-ICP-MS and cryogenic-SEM-energy dispersive spectrometry (cryo-SEM-EDS) for the quantitative analysis of solutes in large (1–100 µm diameter), single primary-brine fluid inclusions in halite. Cryogenic SEM-EDS was used to determine major solutes (Ca, Cl, K, Mg, Na and SO4) and the data were used to convert the relative element concentration ratios determined by LA-ICP-MS (B, Br, Ca, Cl, K, Li, Mg and Sr) into true concentration units. Normalization was achieved by reference to Cl. The authors judged the two techniques to be perfectly complementary and yielded values for K and Mg that agreed to within 1σ. Analytical precision (RSD %) for the major solutes by cryo-SEM-EDS was 5–15% compared with 10–35% for LA-ICP-MS. The techniques were considered equally applicable to lower salinity fluid inclusions or inclusions containing daughter minerals. Ghazi and Shuttleworth94 used a novel approach to overcoming the perennial problems of calibration, by ablating artificial inclusions made in glass microcapillary tubes, in the determination of Ca, Sr and Rb in individual natural fluid inclusions in halite. The authors concluded that the method was quantitative and that uncertainties (RSD %) in the analysis of individual fluid inclusions ranged from 4% to 20%.
Butler and Nesbitt95 demonstrated the existence of non-random Ag, Au, Ba, In, Pb, Te, U and V distributions within the chalcopyrite wall of an immature black smoker chimney using high sensitivity ICP-MS and UV LA (with <30 µm resolution). A number of models were proposed for elemental deposition on or in the chimney walls. The authors concluded that this study illustrated the power of the LA-ICP-MS method and that such spatially resolved data had the potential to constrain models of elemental precipitation both in chimneys and in associated mounds.
Halliday and co-workers96 reported on preliminary in situU–Th isotope determinations, made at very high spatial resolution, using LA in combination with a multiple collector (MC) ICP-MS instrument. Initial studies were made on glass RMs (NIST 611 and 613) to determine precision and accuracy. Experiments on opal known to be in secular equilibrium with respect to 230Th, 234U and 238U gave reproducible δ234U determinations, in good agreement with TIMS data. The determination of 230Th∶238U ratios, although reproducible, was considered more problematic and was still under investigation.
Isotopic analysis of Zr in terrestrial zircon samples was carried out by Hirata and Yamaguchi97 using an enhanced sensitivity LA-MC-ICP-MS system. The precise determination of Zr ratios was required in order to determine the presence of 92Nb, which constrains the timescale of nucleosynthesis and formation of the early solar system. Precisions of ratio measurements were between 0.01 and 0.04% (2σ), depending on the ratio, a factor of 2–3 times poorer than those that could be achieved by solution analysis. The authors conceded that data obtained gave no indication of any radiogenic contribution from 92Nb, but added that they had clearly demonstrated that LA-MC-ICP-MS had the potential to become a strong tool for the detection of possible 92Zr excess in older zircon samples.
As a technique that causes minimal visible damage to samples, LA-ICP-MS is ideal for analysis of valuable antique artefacts. For example, Devos et al.98 used the technique to determine impurities in antique silver objects for authentication purposes. An alternative laser cell was designed as conventional devices were too small to accommodate the samples. The new cell was placed on the object to be analysed. A micro-amount was then ablated through an aperture in the bottom of the cell. The signals for Au, Bi, Cd, Pb, Sb, Sn and Zn, were normalized to the silver matrix. Standard silver materials were used for external calibration. Crater-to-crater repeatability of normalized signals was <10% RSD for most elements. Accuracy was validated with solution ICP-MS analysis of digested samples and XRF. Application of the method was illustrated by analysis of eight antique silver objects, including one forgery.
The use of LA preserves the temporal axis present in a sample that would be destroyed by a bulk dissolution approach. Appleton and co-workers99,100 demonstrated how this lends itself particularly well to the analysis of teeth, which themselves contain a unique record of the former owners' exposure to heavy metals. Trends in metal concentration on either side of the neonatal line (i.e., pre- and post-natal human enamel) were measured. Evidence for the consistent incorporation of elements such as Sb and Zn was presented, together with data for the occasional inclusion of Cr, Fe and Hg.
Several novel vaporization approaches were revealed during this review period. For example, Okamoto102 described the development of an alkylation scheme to promote the vaporization of Sb from a tungsten boat atomizer. A solution of methyllithium in diethyl ether was reacted with sample residue inside the atomizer so that Sb would be vaporized at 923 K. Under these conditions, Sb was thermally separated from the sample matrix and the alkylating reagent, thereby improving quantitation. The absolute LOD for Sb was 0.1 pg, and the RSD (n = 8) for replicate determinations of 20 pg Sb was 2.8%. In similar work, Lam et al.103 used an ETV device as athermochemical reaction system for the determination of Se in sediments. Sample digests were modified with citric acid so that, under appropriate conditions, Se vaporized as a molecular species at low temperature, thereby separating the analyte from the sample matrix. When the method of standard additions was used, an absolute LOD of 10 pg of Se was obtained. The authors were unable to apply ID techniques in this scheme, reportedly due to severe mass bias effects. In related work, Uggerud and Lund104 studied the effects of Ir and Pd matrix modifiers upon the determination of As and Sb by ETV-ICP-MS. The signal intensities of the analytes were found to increase as the mass of added Pd modifier was decreased (from 2 µg to 50 ng). Such effects were not observed for Ir modifier. Using either modifier, significant analyte losses were observed at temperatures in excess of 1173 K. The authors suggested Pd itself acts as an active carrier, whereas Ir interacts with the graphite, making carbon the main analyte carrier.
Lam and Sturgeon105 also developed a combined FI-HG/ETV system for the determination of As and Se in sea-water. The FI manifold was used to mix the sample with NaBH4. The resultant vapor was separated from the liquid stream and trapped by a reduced palladium bed inside a graphite furnace atomizer held at 573 K. The furnace temperature was stepped to 2773 K, at which point the analyte vapour was injected into the ICP torch. The LODs for As and Se were 14 and 7 ng l−1, respectively, and the RSD (n = 9) for As was 2.5% (at 1 µg l−1). Results for As and Se in RMs NASS-4 and NASS-5 agreed closely with the certified values.
Ben Younes and co-workers106 described a method for the vaporization and removal of silica during direct analysis of geological materials. The authors found that complete vaporization of silica was achieved at 2473 K. Hydrofluoric acid was found to produce two vaporization peaks, one below 753 K (attributed to SiF4) and a second at about 2773 K (attributed to unreacted silica and silicon carbide). Complete removal of silica was achieved when modifer volume and modifier–sample reaction time were optimized. For example, 20 µl of 50% HF effected complete removal of 125 ng of silica. The use of HF had no apparent deleterious effects during the lifetime of the atomizer tube, which exceeded 200 firings.
Mahoney and co-workers107 described their preliminary study of ETV sample introduction for ICP-TOF-MS, in which the duration of the ETV transient remains a problem when precise determinations are needed. In this study, 10 µl aliquots of a 100 µg l−1 multi-element solution were analysed using vaporization temperatures in the 2273–3273 K range. A high-speed digital-to-analogue converter (DAC) was used to average multiple 'scans' over the duration of a single ETV transient. Instrument LODs for 7 selected elements were less than 80 fg and, using a combination of ion counting and boxcar averaging, the dynamic range was greater than 6 decades. A pair of boxcar averagers were used to determine the Ag isotope ratio for which the RSD was less than 2%, despite cycle-to-cycle variations of 19% RSD.
Taking another approach, Langer and Holcombe108 developed an in-line transient extension (TEx) chamber to lengthen the duration of the ETV signal such that full mass scans were readily obtainable. The TEx chamber allowed exponential dilution of the ETV aerosol pulse. Diagnostic studies of the TEx chamber showed that the transient peak area decreased by 20 ± 4% compared with a conventional ETV-ICP-MS configuration. Comparison of peak areas obtained with and without the addition of 1 ng of NaCl to the sample, without the TEx chamber present, resulted in increases of between 8 and 40%, depending on the metal. The authors found that interruption of carrier gas flow into the TEx chamber by up to 2 min caused a 25–35% decrease in peak area, presumably due to settling or other loss mechanisms within the chamber. The addition of 1 ng NaCl led to a decrease in peak area when the sample was held for up to 2 min in the chamber.
Moor and co-workers110 described a new HG system developed for the determination of Se in biological materials. An acidified sample was mixed with NaBH4 at the tip of a cross-flow nebulizer. A modified Scott double-pass spray chamber served as the gas–liquid separator, providing 30 s wash-in and wash-out times. Data for RMs DORM-2 and DOLT-2 were calibrated by ID and external standards, and were in good agreement with certified values. The LODs for Se and some other hydride-forming elements (e.g., As, Sb and Sn) were less than 10 ng l−1, with typical precision of 2% RSD at analyte concentrations of 10 µg l−1. Volatile species of various other elements (e.g., noble metals) were also produced. In a series of more conventional studies, Hou and Narasaki111,112 described a HG system for the determination of Sb in natural waters. Total Sb was determined by first removing possible interferents (e.g., transition metals) with Chelex-100 resin. The resultant solution was treated with KI to promote reduction of SbV to SbIII, then the sample was treated with NaBH4 to produce SbH3 for analysis. Using a HR-ICP-MS system, the authors attained 92.5–96.2% Sb recovery (from 200 ng l−1 Sb) with a LOD of 0.7 ng l−1. By eliminating the KI treatment, the authors were also able to determine selectively SbIII. The SbV concentration was calculated from the difference between the total and SbIII concentration.
Use of desolvation systems has addressed a number of significant problems in ICP-MS. However, use of desolvation can lead to new problems, such as differential transport of analyte species which have different volatilities in samples and in calibration standards. In an interesting study, Al-Ammar and co-workers115 developed a technique for correction of these differences. A volatility correction factor was calculated from analyte intensities measured at two different spray chamber temperatures. This correction factor, applied without any prior knowledge of analyte speciation, was found to reduce errors in the determination of organosilicon and organochlorine compounds by 2–30-fold.
Nebulizer development continues to have a pronounced impact upon the determination of rare isotopes, and some investigators have consequently focused upon the problem of nebulizer selection for this application. For example, Becker and co-workers116 compared four different nebulizers for the determination of long-lived radionuclides by quadrupole ICP-MS. Sensitivity (signal per unit concentration) for ultrasonic nebulization (USN) was in the range of 420–850 million ions s−1, which was about an order of magnitude greater than that of a cross-flow nebulizer. However, an USN consumed about 26 times more sample than a microconcentric nebulizer (MCN). All four nebulizers gave LODs in the range 0.01–0.6 ng l−1. Based upon sample consumption, the authors concluded that MCN was the technique of choice for radionuclide determinations.
In related work, Jones117 studied the effect of total dissolved solids (TDS) upon the performance of several nebulizers used to determine radionuclides in high-level waste solution analogues. In the range of 0.25–1% TDS, the direct injection nebulizer was found to perform poorly, possibly due to matrix-related 'de-tuning' of the nebulizer efficiency. This study indicated that cross-flow nebulization gave the most stable signal and the lowest degree of matrix-related signal suppression.
In a related study, Wen et al.119 described a polyacrylonitrile (PAN) hollow fiber membrane derivatized with 8-hydroxyquinol-ine (8-HQ) for the concentration of REEs in sea-water. Recoveries were 91 to 107% at preconcentration factors as high as 300-fold. The authors noted that the cycle time of this FI system was significantly lower than that of comparable 8-HQ-based chelating systems. They also indicated that the derivatized PAN membrane was useable over a wide range of pH and may well be useful in other applications.
Lofthouse and co-workers120 examined the use of iminodi-acetate (IDA) and 8-HQ ligands in a miniaturized FI manifold. This study was novel in that the 8-HQ was immobilized upon a microporous silica frit rather than a microcolumn. Using a microconcentric nebulizer, preconcentration and matrix elimination was achieved within 3 min for both 'microcolumn' systems. Analyte recoveries were 91–102% for the IDA columns and 96–105% for the 8-HQ columns. Calibrations showed excellent linearity, with correlation coefficients greater than 0.999 for both column types.
Coedo and co-workers121 described a micro-scale mercury electrolysis cell for the determination of various elements in steel by FI-ICP-MS. The steel sample was prepared by microwave-assisted digestion in a mixture of HNO3, HCl, HF and H2SO4, and the digest was evaporated until fuming. The digest was then introduced into the electrolytic cell, where Hf, Th, U, Y, Zr and REEs were separated from the sample matrix. The resultant solution was loaded into the FI manifold and injected into the plasma. The LODs (in the solid sample) were a factor of 10 better than for a direct analysis procedure. The RSDs were 3.5% or less for analytes present at concentrations 10 times the quantitation limit and recovery was greater than 97%.
As noted above, other recent FI studies have focused upon the automation of existing sample preparation schemes. Wei and Jiang122 described an FI-HG system for ID determination of Tl in sea-water. Samples were prepared by adding 203Tl to the sample and oxidizing the TlI to TlIII with H2O2. A 1 ml aliquot of the treated sample solution was diluted 10-fold with approximately 0.5 M HCl, and 0.2 ml of the resultant solution was injected into the FI carrier stream (4% HCl, 1 mg l−1 Te). The carrier stream was merged with a second, alkaline, carrier stream containing NaBH4 and, after passing through a gas–liquid separator, the hydride vapors were injected into the ICP. At 1 µg l−1 of Tl, peak area RSD was 2.8% and the Tl isotope ratio was determined with a 2.4% RSD. Several elements, when present at a concentration of 1 mg l−1, were found to interfere with the analysis. Yet the results from the FI method agreed well with those of the method of standard additions when used for the analysis of the RM CASS-3 sea-water.
The list of separation methods interfaced with ICP-MS continues to grow. Ranville and co-workers123 described the combination of ICP-MS with sedimentation field-flow fractionation (FFF) for the analysis of soil colloids in the 0.05–1 µm size range. A UV detector was placed in-line to monitor particles eluting from the FFF device. The eluent flow was split at a T-connector with 50% passing to the mass spectrometer and the remainder going to a fraction collector (for subsequent off-line analysis). Mass spectra were recorded with peak jumping in the range m/z 23–238. Results showed enrichment of Fe in colloid fractions and depletion of kaolinite in the large colloid fractions.
Publications related to capillary electrophoresis (CE) ICP-MS have increased in the past year with particular emphasis on the CE-mass spectrometer interface. Schaumloffel and Prange124 described a new interface for CE-ICP-MS based upon a modified microconcentric nebulizer. The free aspiration uptake rate for this nebulizer of 6 µl min-1 allowed direct matching of the electro-osmotic flow to that of the plasma source uptake. Optimization of fluid mechanical properties prevented the development of laminar flow conditions within the CE system so that the system's high resolving power could be maintained. Optimized nebulizer conditions were independent of optimal CE conditions so that when the system was used in combination with a low dead-volume spray chamber, peak widths were as low as 3.5 s. This compared favourably with CE peak widths using UV detection.
Baker and Miller-Ihli125 carried out a comparison of cross-flow and microconcentric nebulizer (MCN) interfaces for the capillary zone electrophoresis (CZE) ICP-MS determination of Cd in metallothionein solutions. Although the MCN had the better sensitivity, it exhibited slightly poorer chromatographic resolution and had a tendency to clog. The LODs using the 114Cd isotope and 4 nl injections were 90 and 10 µg l−1 for the cross-flow and concentric nebulizers, respectively. Determinations of Cd in rabbit and equine metallothioneins agreed closely with values obtained by solution ICP-MS or ETAAS. The authors recommended use of the cross-flow nebulizer owing to its robustness and relative simplicity.
Non-nebulizer interfaces are also under development. In one noteworthy paper, Tian and co-workers126 described a moving-bed HG system for the determination of As species by CE-ICP-MS. In this system, the capillary effluent was mixed with KBH4 and the mixture placed dropwise upon a moving tape inside a small reaction vessel. Arsenic hydrides were subsequently swept into the plasma using a 1.1 l min−1 carrier gas flow. The LODs ranged from 68 fg for AsV to 120 fg for AsIII and recoveries were in the range 85–108%.
This year saw a surge in the number of papers related to GC-ICP-MS. Montes Bayon and co-workers127 developed an alternative interface for the determination of Hg, Pb, S, Se and Sn species by capillary GC-ICP-MS. The interface was constructed from a heated metallic tube into which the capillary exit was inserted. The tube was connected to a heated 'tee' into which the carrier gas was injected, thereby producing a sheath gas flow to prevent analyte condensation within the interface.
Derivatization techniques were used by Slaets et al.128 to develop a multicapillary GC-ICP-MS system for the determination of Hg species in biological media. Extracts of fish tissue were treated with sodium tetraethylborate and the resultant vapours collected in a purge and trap apparatus. The GC–plasma interface included a valve which allowed intermittent injection of a standard Hg vapour for optimization of analysis. The response was linear over two decades for methylmercury and one decade for HgII. The LODs were 0.2 µg kg−1 for methylmercury and 0.2 mg kg−1 for HgII. The measured concentrations of methylmercury and total Hg in the RMs DORM-1 and TORT-1 agreed well with the certified values.
The recent development of ICP-TOFMS has opened up new opportunities for the analysis of fast chemical transients. Leach and co-workers129 have evaluated a GC-ICP-TOFMS system for the determination of alkyltin and alkyllead compounds. The TOFMS system was limited to the acquisition of only 78 integrated spectra per second, yet this proved to be sufficient to measure all but the fastest transient signals (peak widths <50 ms). For tetramethyltin (1 s peak width), isotope ratio precision was about 3% RSD with an accuracy of ca. 0.3%. The LODs were in the low femtogram range for all compounds studied and the dynamic range was greater then six decades.
In an interesting paper related to GC-ICP-MS, Zapata et al.130coupleda capillary GC-MIP to a mass-selective detector for the determination of 16 organochlorine pesticides. The analytes were measured by selected ion monitoring at m/z 35. The authors reported that peak intensities registered by capillary GC-MIP-MS were equal to or greater than those registered by electron impact ionization GC-MS. Interference-free determinations at diesel oil concentrations as high as 70% were made possible through addition of N2 or O2 to prevent formation of carbon deposits. Recovery of 16 organochlorine pesticides from the test mixture fell within the 79–110% range.
Alternative plasma sources have also been investigated in combination with LC systems. Chatterjee and co-workers131 developed a high-power nitrogen MIP source for mass spectrometry of HPLC effluents. A system based on cation exchange was used to determine seven arsenic species. The authors found that there was no spectral interference of 40Ar35Cl+ upon 75As+ even at a chloride concentration of 1% (m/v). The LODs fell between 0.68 µg l−1, that of pentavalent As, and 22 µg l−1 for trimethylarsine oxide. Repeatability (within-day measurement precision) and reproducibility (between-day measurement precision) were 0.7–9.2% RSD and 6.5–11% RSD, respectively. The results for As species determined in the CRM-18 candidate RM (freeze-dried human urine) agreed with HPLC-ICP-MS values.
Despite the concerns of stagnation in the development of the technique noted in last year's review, more effort continues to be placed on the application of HPLC-ICP-MS rather than technical development. It is notable that there has been a marked reduction in the use of quadrupole mass analysers for these studies. Gonzales LaFuente and co-workers132compared quadrupole and double-focusing mass analysers for thedetermination of selenium species in urine. Both reversed-phase and vesicle-mediated HPLC were used with the columns coupled to a conventional nebulizer or an on-line microwave digestion–HG interface. Selenium sensitivity with the double-focusing mass analyser was 23–59 times higher than that with the quadrupole mass analyzer, yet the LODs were only improved 9-fold or less because of the higher background noise. Use of microwave digestion–HG provided greater sensitivity but higher background noise than the conventional nebulization interface. More compounds were detected by vesicle-mediated HPLC than by reversed-phase LC and detection limits were improved.
There appears to be a trend away from conventional nebulizer-based interfaces with HG-ICP-MS systems being developed for hydride-forming elements. Dagnac and co-workers133 used a LC-UV irradiation-HG ICP-MS system to determine As species in mussels with LODs of 25–130 ng l−1 and peak area reproducibilities (n = 5) of 0.67–3.1%. Goenaga Infante et al.134 used a vesicular HG interface in the determination of Cd species in human urine and rabbit liver metallothionein (MT). The HG efficiency was the same for both CdII and Cd bound to the main isoforms of rabbit liver MT. The HG system provided a 4–16-fold improvement in sensitivity compared with a conventional nebulization interface yet LODs were improved only 2–5-fold owing to high blank concentrations.
3.4 Interferences
The existence of spectral interferences in ICP-MS is now extremely well documented, although many authors cannot resist revisiting much of this area of study. Indeed, with the availability of cold plasmas, reaction/collision cells and high resolution MS instruments, the problems caused by spectral interferences may be finally in abeyance, although some interesting issues are still being addressed. Ramanujam et al.135 investigated interference effects on zinc isotope ratio analysis caused by polyatomic species of S and other major components of blood plasma (Ca, Cl, K, Na and P). A series of mineral solutions which simulated human plasma with respect to these elements was investigated and it was found that a mixture of all the elements interfered only with 64Zn and 70Zn. It was reported that interferences to 66Zn, 67Zn and 68Zn were minimal and that the presence of Na or S reduced the interference of 35Cl16O2 on 67Zn.Grotti et al.136 described a multivariate quantification of spectroscopic interferences caused by Ca, Cl, Na and S when presented together as matrix elements at concentration levels of 0.5–1000 g ml−1. The method was applied to modelling the spectral interferences, caused by the matrix, on the elements As, Cd, Co, Cr, Cu, Ni, Pb, V and Zn. In order to separate spectroscopic from non-spectroscopic interferences, analyte isotope ratios were measured. For mono-isotopic elements, signal ratios with and without the matrix were measured and correction for non-spectroscopic interferences was made by use of 103Rh and 45Sc internal standards.
Campbell and Törvényi137 reported suppression of Zn signal in the analysis of RM IAEA 392 algae, following microwave digestion with HNO3. Accurate analysis of the samples by FAAS indicated that all the Zn was in solution. The use of the method of standard additions led to an overestimate of the Zn concentration by about 17%, but a 1 + 9 aqueous dilution of the digests gave an 'acceptable' result. The suppression effect was not evident for eight other analytes studied.
In the determination of Se in urine, Gammelgaard and Joens138 made a five-fold dilution, together with inclusion of ethanoic acid (5% v/v), HNO3 (0.14 M) and Ga, In or Y as an internal standard, to overcome interferences. With instrument parameters of carrier gas flow (0.95 l min−1) and forward power (1.3 kW) optimized, the presence of the ethanoic acid was reported to enhance the sensitivity of Se by six-fold. Curiously, despite the presence of internal standards, the method of standard additions was used. The method produced LODs of 0.9 ng ml−1 and analysis of SRM NIST 2670 gave results that 'agreed' with the certified value.
3.5 Isotope ratio measurement
The use of sector-based ICP-MS instruments for isotope ratio measurements continues to gain momentum. These systems have simplified the determination of elements that have traditionally required complex sample preparation methodologies. Tomascak et al.,139–141 in a series of papers, reported on the application of MC-ICP-MS to the determination of Li isotope ratios. In the analysis of basalts, external precision of multiple replicate and duplicate measurements for a variety of sample types was in the order of ±0.11‰ (2σ). The technique allowed rapid (∼8 min sample−1) analysis of small samples (∼40 ng Li) relative to commonly used thermal ionization methods. The δ7Li of basalt RMs JB-2 and NIST 688 were measured at +5.1 and +2.8‰, respectively.Albarede and co-workers142 suggested that the lack of suitable analytical techniques was responsible for the paucity of Cu and Zn stable isotope data. Using a MC-ICP-MS instrument and an anion exchange sample clean-up procedure the authors were able to determine precise Cu and Zn isotopic compositions in a Cu RM NIST 976 and a Johnson–Matthey Zn solution, following correction for instrumental mass fractionation. This external normalization led to an internal precision of 20 ppm and an external reproducibility of 40 ppm. As little as 200 ng of each element was needed and isobaric interferences were considered small enough to be neglected. Isotopic fractionation was observed for Cu on the anion exchange resin.
Useful isotope ratio measurements are being made with single collector sector systems, despite their inherent limitations for this type of work. An obvious benefit of this instrumentation is its ability to resolve atomic and polyatomic species that occur at the same nominal mass. Hamester et al.143 demonstrated that external reproducibilities of <0.02% could be achieved for ratios without spectral interference and of <0.1% where an interference existed and a higher resolution mode was required to overcome the mass coincidence.
Latkoczy and co-workers144 described the determination of sulfur isotope ratios (34S∶32S) with a single collector sector system operating at resolution (m/Δm) of 4000. Instrument sensitivity was increased by a factor of two by use of a guard electrode on the torch together with a microconcentric nebulizer operating with a membrane desolvation unit. The desolvation unit also significantly decreased the signal from oxygen-based interfering polyatomic species. The LODs in solution were 0.01 ng ml−1. These were limited only by blank levels. Sulfur isotope ratios could be determined to a precision of better than 0.1% RSD at concentrations as low as 1 ng g−1. At higher concentrations a precision of 0.04% RSD could be achieved.
4 Laser ionization mass spectrometry (LIMS)
4.1 Fundamental studies
The 10−13 s pulses of modern lasers are shorter than the time scale for thermal diffusion (about 10−9 s) and ejection of species from the solid (about 10−11 s). Hence, formation of a cloud of hot electrons and subsequent processes must be considered. Kinetic energy distributions (KEDs) and angular number density data provide a key to the understanding of these processes. These fundamental data were measured by Pakhomov et al.145 using a simple TOFMS and target irradiation under a range of incidence angles. The mode-locked Ti:sapphire laser delivered about 1 mJ in 100 fs, giving power densities in the range of 1013 W cm−2 on the sample. The angular distribution of the ion energy and number densities obeyed a cosine function. About 2–3% of the 2 × 1013 atoms removed from the sample per pulse were ionized. The total kinetic energy of the ions was about 1.5% of the laser pulse energy. The rate of mass removal was correlated linearly with the conductivity of free electrons.Amoruso et al.146 investigated the KEDs of Al+ and Cu+ produced from metal targets under XeF laser irradiation at a wavelength of 351 nm and a fluence between 1 and 4 mJ cm−2. The KEDs were bimodal with a contribution below 1 eV from thermionic emission. The second component depended strongly on the laser fluence and was associated with emission of energetic ions from the laser produced plasma. In contrast to Cu+, the high energy contribution for Al+ was always present because the 3.5 eV photons were able to promote direct photoionization of the first excited state. Hence, inverse bremsstrahlung absorption at high electron density occurred in the hot plasma with high electron density, created at the beginning of the laser pulse, and steeply increased the degree of ionization. Therefore, the electronic structure and ionization potential of the evaporated atoms played a controlling role in the laser–vapour interaction and subsequent kinetics of the laser plasma production.
4.2 Instrumentation
Song et al.147 developed a laser ablation/ionization ion trap instrument for the analysis of metals and alloys. Solid samples were introduced through a hole in the ring electrode on the tip of a ceramic holder (2.5 mm in diameter for the sample). A XeCl excimer laser irradiated the sample at a wavelength of 308 nm and with an energy of 8–12 mJ per pulse. The trap was operated in the mass-selective instability mode. A He buffer gas pressure of 4 × 10−5 Torr gave unit mass resolution up to m/z 100. Laser ablation had to be adjusted to keep the number of ions under 104–105 in order to avoid space charging. The instrument's size and low vacuum requirements were believed to allow on-site analysis of single particles.A new dedicated laser TOFMS instrument for multi-element analysis of solids was based on a Q-switched Nd:YAG laser (spot diameter 30–70 µm; wavelength 1064 nm; repetition rate 0.1 Hz) operated in the reflection mode.148 The erosion depth was 100 nm per laser shot. Since no extraction voltage was applied, ions travelled with their own emission kinetic energy. The mass resolution of 250 was considered to be sufficient for isotopic analysis. The sensitivity of the instrument was demonstrated by the LOD of about 50 µg g−1 for the minor 114Sn+ isotope. Intense doubly charged ions were seen. Using RSFs determined by the analysis of 11 metal RMs, measured elemental concentrations agreed with certified values within a factor 2. This was considered as reasonable for a field instrument.
4.3 Analytical methodology
The excellent review (204 references) of Suess and Prather149 about on-line laser ionization of aerosols can be recommended to everyone working in the field of small particle analysis by MS. The unique benefits of dual TOF systems, allowing positive and negative ion spectra to be recorded from the same microvolume, made the analysis of even single particles in collected aerosols somewhat obsolete. Furthermore, the so-called aerosol TOFMS (aTOFMS) was sufficiently simple to allow field monitoring. Further research was needed on quantification and the processing of the huge aTOFMS data sets, which became a bottleneck. Therefore, Hinz et al.150 developed an automated classification procedure based on a fuzzy logic clustering algorithm. The particle classes could be defined 'manually', by selection from a database, or by fuzzy clustering of a given subset of mass spectra from the population under study.5 Resonance ionization mass spectrometry (RIMS)
Wendt et al.151 have presented a review on the basic principles and instrumentation of RIMS, in which the bandwidth of pulsed lasers limits the isotopic selectivity in the excitation step to typically less than 102. A major improvement could be achieved by laser irradiation along the axis of the fast moving atom beam with minimal kinetic energy spread. The Doppler shift, which strongly depended on the velocity of the atom beam, produced an artificial isotope shift and excitation selectivity could potentially be increased to 108. Multi-step diode laser systems might decrease the cost of instruments and yet provide high selectivity (108) and efficiency (>1%). Preliminary experiments on 41Ca showed an optical selectivity of 106, an overall efficiency of 10−6 and a LOD of 109 atoms. This was adequate for medical applications. An optical selectivity of 1015 and an efficiency of 10−3 should be feasible with a three-colour scheme and easy-to-use solid state lasers.Bushaw et al.152 calculated the line shapes in high-resolution double resonance ionization. The simulated line shapes predicted the experimental optical selectivities quite accurately. The main uncertainty came from some parameters which were experimentally difficult to evaluate with precision, such as the laser intensities and overlap in the interaction region. Double and triple resonance schemes were evaluated for 41Ca in the presence of 40Ca. For radiochemical dating, triple resonance experiments were required and the calculations could be used to estimate the feasible isotopic selectivity.
Wendt et al.153 investigated the use of multi-step-excitation narrow-bandwidth CW diode lasers and a quadrupole analyser for the determination of ultra-trace long-lived radioactive isotopes in the presence of a large surplus of stable isotopic neighbours. A compact, easy-to handle and reliable CW laser system for three-step resonance excitation was built specifically for the determination of 41Ca. Experiments confirmed that the specifications predicted by theoretical calculations could indeed be attained. The isotopic selectivity was 2 × 105 and the LOD of 109 atoms met the requirements for biomedical studies.
Isotope ratio determination in a challenging matrix such as chondritic meteorites was performed by Bisling et al.,154 using two-colour RIMS for the isotope selective excitation of Hg. Whereas the randomly polarized beams from older designs of lasers led to signal intensities which reflected directly the correct isotope ratios, this was not the case with modern equipment because the latter provided shorter pulses, narrower bandwidths, better polarization and sufficient power to induce power broadening of the line shapes. The data showed that anomalous odd-to-even isotope ratios could be explained by the specific population probabilities of excited magnetic sublevels, which were due to the linear polarization of the radiation.
Yorozu et al.155 investigated the detection of H isotopes using low resolution MS and direct laser irradiation of solids under resonant conditions. A thermal-cracking hydrogen-atom source, producing a gas phase beam of H and D, was used to calibrate the system. The possible interference of H2+ on the D+ signal was eliminated through the use of a wavelength of 243 nm, which allowed a 2 + 1 RIMS ionization scheme to be applied. Direct laser irradiation of D-implanted graphite at the resonant wavelength improved the ionization efficiency by a factor 103–104 in comparison with the non-resonant conditions. The optimal wavelength for D detection in direct ionization from the solid was slightly different from the wavelength required for D-atoms in the gas phase.
Whitacker et al.156 reviewed the application of resonant laser post-ionization in combination with primary ion bombardment for surface analysis and depth profiling of solids with low matrix effects and improved quantification. The selectivity of the resonance ionization process allowed the TOF equipment to be kept simple yet the sensitivity surpassed that of SIMS. Experimental depth profiles for Cu in silicon and silicon dioxide showed a dynamic range of over 6 orders of magnitude. High resolution imaging was shown to be feasible. The useful yield of Cu+ (3–8%) was substantially higher than that obtainable using SIMS. Unfortunately, use of resonant post-ionization for elements of major interest in semiconductor applications (C, N and P) would require a VUV wavelength of around 120 nm.
6 Secondary ion mass spectrometry (SIMS)
Distinction, based on the primary ion dose per unit of surface area, must be made between dynamic SIMS, commonly denoted by the acronym SIMS, and static SIMS (S-SIMS). The former method uses a primary ion beam at high current density for the layer-by-layer erosion of the sample and characterization of the chemical composition as a function of depth. The ultra-fast penetration of keV primary ions through the first monolayer at the surface limits the interaction time and energy deposition therein. Hence, the molecules at the outer surface layer stay intact. The gradual deposition of the keV energy into the solid along the penetration track of the primary ion destroys the molecular bonds in the subsurface. As a result, the fast erosion rate in SIMS means that the detected signals have their origin in the subsurface regions and only yield atomic information. In contrast, molecular information is available from mass spectra of S-SIMS because a reduced analysis time allows only the upper monolayer to be sampled. 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
Continuing analytical development has steadily pushed the limits in quantification, depth profiling and imaging. A critical review by Schumacher et al.157 demonstrated the need for dedicated instruments. Imaging at high lateral resolution required both primary ion bombardment and ion extraction along the normal to the sample surface. Hence, applications were limited to primary and secondary ions of opposite polarity, which was often not possible for quantitative work. In contrast, reduction of the matrix effect by the MCs+ method required tilting of the sample to optimize the Cs-surface concentration. Shallow depth profiling required low energy primary ions impinging on the sample under a grazing angle and fully decoupled ion optics for primary and secondary beams. Therefore, quadrupole instruments were preferred for depth profiling but magnetic mass analysers were necessary for isotope ratio determinations with multicollector detection.Aiming at an affordable instrument for applied research and process control, Eccles et al.158 built a benchtop SIMS system, fully controlled by user-friendly software. Nominal mass resolution was achieved up to m/z 300 and LODs for B and Li in silicon were 10 and 0.1 µg g−1, respectively. Sub-monolayers of silicones and fluorocarbons on the surface of electronic devices could be imaged with a lateral resolution of about 10 µm.
Gillen et al.159 developed a triplasmatron ion source for increased depth resolution by polyatomic primary ion bombardment. The gun allowed the generation of both SF5+ and F− primary ion beams in an ion microscope SIMS. A commercial filament-duoplasmatron was modified by the addition of a positively biased expansion cup in the extraction side of the anode to initiate a localized secondary arc discharge. The gun produced 200–300 nA of SF5+ with a current density of 0.5 mA cm−2, which resulted in a focused primary ion beam with a spot diameter in the µm range at currents of about 10 pA. Up to 3 µA of F− primary ions could be obtained by removing the cup. Depth profiling of a Ni–Cr multilayer (NIST SRM 2135a) showed a significantly increased depth resolution and reduced topography development when SF5+ was used instead of O2+.
Negative primary ions are particularly useful for analysis of insulators because sample charging is compensated by the electrons emitted initially. Hereć et al.160 reported the production of Cl− primary ions through the use of a modified Cs+ source in which Cs+ ions sputtered a CsCl target. The gun design with two co-axial chambers allowed positive or negative primary ion beams to be selected by simply biasing the electrodes.
Daolio et al.161 developed a two-dimensional quadrupolar electrostatic separator for simultaneous sampling of positive and negative ions by separate quadrupoles, each of which had its own detector. The high energy primary ions were projected through the separator without deflection, whereas the secondary ions with relatively low energy were deflected by the electrostatic field onto different trajectories. The quadrupoles transmitted ions with initial emission angles up to 10°. The energy resolution was 15–20%, expressed as full width at half maximum (FWHM), for ions with energies between 0 and 50 eV.
A new reflectron type TOF analyser was built by Smehtkowski et al.162 to perform SIMS and mass spectroscopy of recoiled ions (MSRI) on the same sample. The MSRI method measured the ions generated by binary interaction between the primary ions and the surface groups. This process generated only elemental ions, so that MSRI readily distinguished between H and D whereas H2+ and D+ had to be separated in SIMS analysis. The recoiled ions had different emission angles and higher energies than the secondary ions. Hence, choice of the primary beam incidence angle and reflector voltage allowed switching between MSRI and SIMS. Using the complementary surface information obtained from MSRI and SIMS, distinction could be made between different phases, for example, diamond, graphite or amorphous carbon.
6.2 Fundamental studies
In order to clarify the basic ion formation mechanisms, velocity distributions of Ag−, Au− and Cu− ions, generated from metal foils by 12.5 keV Ar+ and Cs+ ion beams, were studied.163 Plots of the normalized intensities as a function of 1/V revealed the existence of two different ionization mechanisms. The exponential dependence of the ion intensity on the emission velocity above 100 eV pointed to resonant charge transfer. The Gaussian function fitting the data at lower energy pointed to surface excitation. Ivanov et al.164 investigated the temperature dependence of the Ni+ and Ni2+ emission from polycrystalline nickel under Ar+ and Xe+ bombardment. The Ni+ emission peaked at the magnetic phase transition temperature as a result of the change in the work function. The decreased Ni+ emission, together with the increased Ni2+ signal at higher temperatures, resulted from ion–molecule reactions in the selvedge. The term 'selvedge' refers to the dense gas phase region created by an ionization event just above the sample surface. The exponential relationship of the signal intensity ratio Ni2+∶Ni+ with temperature pointed to thermodynamically controlled interactions. Okuyama et al.165 studied the formation of unexpected PdH− ions from polycrystalline palladium under Cs+ bombardment. Surface contamination was ruled out since the signal stayed almost constant with depth. Experiments with D2-flooding excluded the role of the residual gas. The data suggested that the hydrides were generated by direct emission from the solid, and not by ion–molecule reactions in the selvedge. The adsorbed Cs atoms would act as electron donors and stimulate the emission of hydride ions.The momentum transfer from primary ions to sample atoms in SIMS is generally explained by a linear collision cascade but there might be an important role to play for electronic processesinsputtering by low energy primary ion beams. Using a modified TOF SIMS, Sato et al.166 measured the secondary ion yield from MgO(001) under bombardment by Ar+ and Ar2+ with energies between 40 and 300 eV. The signal intensity ratio Mg+∶O+ increased strongly with the ion dose for Ar2+, whereas it was virtually dose-independent for Ar+. Since, in theory, sputtering by momentum transfer should be identical for singly and doubly charged ions, part of the O+ produced with Ar2+ had to be associated with electronic processes. The de-excitation of Ar2+ by capturing electrons from O2− could form O+, which would be emitted by Coulomb repulsion. The changing surface stoichiometry, caused by the preferential oxygen sputtering, would explain the dependence of the Mg+∶O+ intensity ratio on the ion dose.
Hereć et al.167 used subtraction of the secondary ion signal from the measured primary ion current as a simple experimental estimate of the implanted ion dose of 2.5 keV Cs+ and K+ primary ions in copper and silver. The dose derived from SIMS agreed within 10% with the results from simulation models currently used in ion beam implantation technology. The distribution range of the implanted ions was readily obtained by taking the second derivative of the 'collection' function, which related the implantation dose to the irradiation dose.
The resolution in depth profiling silicon under O2+ bombardment is limited by sputter damage processes, of which segregation is a major one, and by chemical effects. Deenapanray and Petravic168 investigated the segregation of several impurities in the surface oxide layers formed on silicon under 6–10 keV O2+ bombardment. The migration of Ca, Cr, Cu, Hf, Ta, Ti and Zr towards or away from the SiO2–Si interface could be rationalized by the formation enthalpy of the oxides, on the condition that it was normalized to the number of oxygen atoms. Specifically, the migration behaviour of Ca pointed to the formation of CaO2 instead of CaO in the ion implanted region. For migration towards the SiO2–Si interface, the normalized formation enthalpy must be higher than that of SiO2. Experimental data at increasing temperatures showed the importance of the impurity mobility in the segregation effects.
Jiang and Alkemade169 studied the surface transients in Si and SiGe during O2+ bombardment with energies between 0.56 and 2 keV and incidence angles of between 45 and 77°. The transition widths, i.e., the depth at which equilibrium Si+ intensity was reached, were 3–4 nm for sub-keV energies and incidence angles above 60°. The transient width appeared to be about twice the mean penetration depth of the primary ion. This corroborated the hypothesis that the equilibrium between matrix ion emission and sputtering was reached simultaneously at oblique incidence angles.
Biersack170 used a simulation code which takes into account both sputtering and mixing processes to model the depth profile broadening of a gold marker layer in silicon. A gold layer was chosen so that chemical effects could be neglected. Because the matrix became completely amorphous at high doses, it was assumed that the recoil atom could stop in any position. The preferential motion forward of the recoil atoms, seen in the simulations, explained the deviation from the Gaussian profiles and the isotropic diffusion, predicted by the analytical approach. The simulations agreed very well with the experimental profiles for a Sb δ layer in silicon bombarded with 1 keV Ar+ at 45°, and therefore could be very useful to develop depth resolution functions (DRFs).
The analytical description of noise becomes a key factor in the deconvolution of depth profiling data because the maximum entropy method, for example, explicitly requires this parameter. Therefore, Makarov171 undertook a theoretical investigation of the noise in SIMS depth profiling and showed that noise deviations from Poisson's law pointed to instrument instability. The elaborated analytical expression for noise took into account the statistical variations as well as the primary ion current fluctuations.
Pursuing the concept of using a SIMS instrument for low energy ion scattering spectroscopy (LEIS), Pratt et al.172 investigated the KEDs of 16O+ ions, sputtered or elastically scattered from sulfide minerals under O2+ bombardment. The different contributions from binary elastic backscattering collisions could be identified with sufficient resolution to distinguish between 32S and 34S. For high mass elements, isotopic contributions could not be resolved. Differences between the measured and calculated mean kinetic energies were associated with instrumental factors and occurrence of inelastic processes at the surface. In contrast to the LEIS signal for Fe, that for S exhibited no clear correlation with the atomic concentrations in the different minerals, reflecting unique changes in the (sub)surface composition during 16O+ bombardment. This aspect needs further investigation.
The electronic sputtering by highly charged primary particles, such as Au69+ in the keV–MeV range, increases the secondary ion yield by more than two orders of magnitude in comparison with that produced by singly charged primary ions. Although this methodology is limited to a few laboratories, its analytical potential for semiconductors and biomolecular solids was demonstrated by Schenkel et al.173 Specifically, with a 255 keV Xe44+ ion beam, a LOD of 1010 atoms cm−3 was attained for uranium oxide. A thermally grown silicon dioxide layer on silicon, covered with a tungsten pattern, could be analysed with a lateral resolution superior to the spot diameter of the primary beam because the high ion yield allowed coincidence techniques to be used.
Kang et al.174 have reported modelling of depth profiles of ultrathin multilayers by dynamic Monte-Carlo simulations. They accounted for the generation of interstitial atoms and vacancies and for the subsequent annihilation of the latter. The shape of the depth profile for silicon dioxide δ layersbetween tantalum oxide layers with a thickness of 0.5 and 18 nm, respectively, proved to be determined entirely by the physical processes in the sample. Atomic mixing explained the shift in apparent depth of 1–3 nm towards the surface as well as the relationship between primary-ion energy and the decay length measured for a 1 nm tantalum oxide layer on silicon dioxide. The increasing primary ion energy caused the collision cascade to occur deeper in the subsurface and thereby minimized ion beam mixing in the thin surface layer. Simulations and experiments agreed to within 20%.
Higashi et al.175 studied the interface effect in depth profiling of In1 − xGaxAs1 − yPy/InP multilayers using two different methods. The first one was SIMS using 2 keV O2+ primary ions at an incidence angle of 81° and the second one was non-resonant laser ionization SNMS with 10 keV Ar+ primary ions impinging on the sample at 77°. The use of SNMS was needed to separate the effects due to the matrix from those due to the changing sputtering conditions close to the interface. The distortion of the experimental In SIMS profiles near the layer–InP interface could be qualitatively explained by simulations, accounting for the preferential sputtering and for variations in the secondary ion yield as a function of the surface oxygen levels.
6.3 Analytical methodology
Failure analysis of integrated circuits often requires characterization of particulate matter incorporated in the circuit matrix. Verkleij and Mulders176 demonstrated the use of a Ga+ liquid metal ion gun focused to less than 10 nm for the cross sectioning of small features. After milling the particles with a 30 keV beam, the surface was cleaned by low current ion bombardment and analysis performed using a quadrupole SIMS instrument attached to the same vacuum chamber. Images could be taken with a resolution of about 50 nm.Fried et al.177 elaborated a simple two-step approach to determine the depth of the interface between substrate and silicon oxide layers with thicknesses between 40 and 120 nm and different porosities. The interface position was first measured as a function of the sputter time from a 'normal' depth profile. Sputtering in a subsequent experiment was performed only over the time needed to expose the interface, so that the crater depth could be measured by profilometry. In spite of the effect of composition and porosity on the ion yield and the sputter rate of individual layers, SIMS results agreed within 10–20% with those calculated from optical methods such as reflection spectrometry and spectroscopic ellipsometry.
Sample rotation has been found to provide a means to improve the depth resolution for multilayers.178 When ion bombardment under 45° was used to reduce the beam-induced roughening, the decay length for a 30 nm tantalum oxide layer on tantalum was reduced from 1.88 nm in the stationary regime to 1.64 nm when the sample was rotated. Furthermore, the experimental interlayer distance in sandwich samples of ten alternating boron carbide and molybdenum layers, with thicknesses of 1.7 and 0.26 nm, respectively, stayed constant down to the bottom layers.
Rar et al.179 investigated the use of the dimer secondary ions Al2+ and Ga2+ to quantify depth profiles from GaAs–AlAs multilayers with Auger electron spectrometry used as reference method. The mixing, roughness and information (MRI) depth model was applied to simulate the signal intensities as a function of the layer structure. The MRI parameters such as mixing length and roughness, derived from Auger electron spectrometry, proved to be adequate for SIMS. A simple empirical formula was elaborated to relate concentration and signal intensity.
Matrix effects inherently hamper the quantification of depth profiles, but applications such as implantation studies only require the total area densities or doses of the implanted element which are obtained by integrating the density function over depth (or time). Wittmaack et al.180 elaborated a simple method to calibrate the impurity dose in layer systems of silicon and silicon dioxide. The rapid oxidation of silicon under O2+ bombardment impinging at 90° equalized the ionization probabilities of the impurity and the matrix as well as the sputter yields for silicon and silicon dioxide. Hence, the dose in a sample with any sequence of material composition could be determined by measuring the erosion rate in one standard with known dose. Application to B-implanted silicon and silicon dioxide samples yielded doses within 1% of those predicted from simulations.
Baldwin et al.181 investigated the sputter rates of plasma-nitrided surface layers on stainless steel. Using SEM to verify the layer thickness on cross sections, it was found that the sputter rates in SIMS were identical for the treated layer and the bulk metal, thereby facilitating the calibration of the depth scale.
Seki et al.182 elaborated a method for the quantitative determination of N in silicon without matrix effects. The duoplasmatron was used to implant additional 14N+ into the sample. The peak concentration was calculated from the fluence and the standard deviation of the projected range was measured from the SiN− signal. The peak concentration of the unknown was derived from the relative signal intensities at the maxima in the nitrogen distributions. Experimental and calculated concentrations agreed within 10%. Since the method was based on the implantation of low energy ions, flat crater bottoms were critical for the accuracy of the depth scale.
Farquhar et al.183 investigated the accurate determination of 13C∶12C to study graphite precipitation and carbon chemistry in crustal fluids. With thorough optimization of all mass fractionation effects, the reproducibility of the 13C∶12C ratio (0.1%, 2σ) was close to the counting precision of 0.08%. Considering the uncertainty from the conventional analysis of standards, the practical accuracy of the method was assessed to be 0.1% at the 95% confidence level, which was sufficient to distinguish between the C ratios in graphite grains from different sites and to study the process of graphite formation and fluid infiltration.
Compston184 reviewed the in situ U–Pb age determination for minerals by SIMS. Since determinations within individual grains allowed the µm scale variations of the Pb∶U ratios to be monitored, SIMS was more useful for geological dating than the more accurate IDMS. The sources of errors in a dedicated scanning high resolution ion microprobe (so-called 'SHRIMP') instrument were addressed systematically. The evaluation of the precision of the method was hampered by the fact that the sample volumes analysed were different for each analytical method investigated.
6.4 Quantification
Wittmaack185 critically analysed the applicability of the infinity velocity (IV) method to standardless SIMS quantification. The IV method is based on the assumption that all sputtered atoms are ionized and, hence, no matrix effects occur at IV. This implies that the sputtered atom flux and the instrument transmission are independent of ion energy. However, it was demonstrated that the experimental secondary ion beam intensities did indeed deviate from their exponential dependence on the IV. These deviations could be ignored by forcing a linear fit through data in the 80–250 eV energy range where the elements exhibited poor counting statistics. Furthermore, it was demonstrated that the conversion of IV data into concentrations should also take into account element-specific parameters such as the surface binding energy and the shape of the energy and angular distributions.The development of IR focal plane arrays requires accurate quantification of the dopant As and critical impurities such as Cu in the HgCdTe matrix. Wang et al.186 used the MCs+ technique to circumvent the need for high mass resolution and thereby improved the LOD of As to 5.2 × 1014 atoms cm−3 from that of 1 × 1016 atoms cm−3 with O2+ SIMS. The precision evaluated over a three month period was 15% (RSD). Hongo et al.187 measured MCs+ and MCs2+ to quantify electropositive and electronegative impurities in GaN and AlxGa1 − xN films (x = 0–0.17) with minimal matrix effects. The LODs for O and Si of about 1–2 × 1018 atoms cm−3 were significantly lower than those obtained using the atomic ions. The LODs for C, H and Mg were 2 × 1017 to 3 × 1015 atoms cm−3. The formation mechanism of MCs2+ was studied by changing the surface coverage and comparing the useful yields of ions from GaN and AlxGa1 − xN films. The OCs2+ ions resulted from recombination of OCs with Cs+ while HCs2+ could arise additionally from a reaction of H− with 2 Cs+.
Douglas and Chen188 derived a multivariate expression for the dependence on local oxygen content of RSFs obtained for 52Cr, 56Fe and 58Ni in silicon oxides using a 12 keV Ga+ beam. The RSFs in an oxide matrix under Ga+ bombardment were almost identical to those for the elements in silicon using 8 keV O2+ bombardment. Therefore, RSFs obtained using O2 bombardment could be used to predict LODs in native oxide layers when using Ga+ bombardment. The LODs of between 5 × 106 and 5 × 108 atoms cm−2 made it possible to quantify, by TOF SIMS, 'surface' concentrations (i.e., within a depth of 0.5 nm) two to four orders of magnitude lower than those detectable by TXRF. The latter is currently the method commonly used to study the deleterious influence of metallic surface contamination at trace levels in metal–oxide semiconductors.
Because light and volatile elements such as B, Be, F, H and Li exhibit heterogeneous distributions in natural phases, there is a continuing search for suitable RMs. Aurisicchio et al.189 evaluated the use of tourmaline megacrysts as potential standards. Homogeneity was checked by EPMA. The SIMS results for H and Li closely agreed with those obtained by AAS and GC thermogravimetric analysis. The accuracy was estimated to be better than 3, 10 and 20% for B, H and Li, respectively. Hoskin190 used SHRIMP for the determination of F in geological samples at µg g−1 levels, which are below the LOD achievable with EPMA. Use of energy filtering reduced isobaric interferences and matrix-related secondary ion yield differences. Calibration standards were made by adding CaF2 to glass and were quantified using EPMA. An accuracy within 5% was achieved by SIMS. The F concentration determined in the widely used NIST SRM 610 was 296 ± 16 µg g−1 (5.4% RSD). The ion yield for F was about 3 counts s−1 nA−1 per µg g−1 F.
Quantification in SIMS depends on the precise characterization of the matrix effect in sputtering and ionization. Suzuki et al.191 determined the respective yields for primary ion bombardment of Fe, FeO, Fe3O4 and Fe2O3 with Cs+ and O2+. Under Cs+ bombardment, the ion yield for oxides was 30% higher than that for the pure metal, whereas under O2+ bombardment the ion yields of iron and its oxides were the same.
Whenever preferential sputtering occurs in the depth profiling of thin coatings on sub-µm particles, the relative contributions to the signal from coating and core components change continuously during analysis. Verlinden et al.192 investigated the depth profiling of cubic silver halide microcrystals in the size range of 400–600 nm. To calibrate the depth scale, pure halide particles were sputtered completely after size determination by electron microscopy. The resulting sputter rates were used to estimate the sputter rate in mixed halides by linear combination, which should yield an overall uncertainty within 10%. The estimated sputter depth could localize interfaces located within 1/3 of the crystal length below the surface. Quantification by the intensity ratio method or by the MCs2+ technique yielded results which agreed within 10% with those obtained by SEM-EDXRF and XRF. The IV approach was inadequate to quantify these core-shell microcrystals.
6.5 Single and multidimensional analysis
Several methods are available for the evaluation of the depth resolution in multilayers: FWHM, steepness of the leading or trailing edge, decay length or depth over which the signal intensity ranges from 84 to 16%. Therefore, Moon and Kim194 studied SiO2 and GaAs δ-layers in tantalum oxide and silicon, respectively, to be used as possible RMs. The δ-layer thickness was 1 nm, whereas those of the tantalum oxide and silicon layers were 18.6 ± 0.3 nm and 83.8 nm, respectively. Experimental depth profiles could be fitted with a DRF, which was a convolution of two exponential functions with a Gaussian function and was based on 3 parameters, namely width, growth and decay constant.
Zhao et al.195 investigated the TOF SIMS depth profiling of multiple-quantum-well ZnTe–CdSe multilayers using the clusters CdSe2− and ZnTe2− normalized on the simultaneously detected Se– and Te– intensities, respectively. The sputtering rates were characteristically different for CdSe, ZnSe, ZnTe, ZnSe∶Cl and ZnxCd1 − xSe. The Cl−-doped ZnSe epilayer showed a lower sputtering rate and a reduced depth resolution for Cl− because of strong atomic mixing. The decay length stayed between 8 and 10 nm for up to 10 layers, but the growth length varied with depth from 6 to 16 nm. Hence, the knock-on effects did not depend on the number of sputtered layers whereas the mixing with the substrate did.
Kruger et al.196 demonstrated the use of line scans across bevelled structures as an alternative to conventional depth profiling. The depth resolution was limited by the surface roughness of the bevel instead of the escape depth. A surface roughness of less than 1–2 nm could be achieved for bevel angles of 0.01–0.1°. The bevel angles had to be chosen such that the lateral distance of individual layers became larger than the diameter of the primary beam. An ion dose of less than 1012 ions cm−2 minimized atomic mixing and surface perturbation. Depth profiles on SiGe quantum wells yielded a FWHM of 3.5 nm and a decay length of 0.8 nm, whereas 2.4 and 0.75 nm were found for layers of B and Sb, respectively, in silicon.
The problems encountered in the depth profiling of deep layers and interfaces in the analysis of InGaAsP laser diodes were circumvented by removal of the InP substrate through etching followed by 'back side' analysis.197 Upper layers were selectively removed by HCl and H3PO4. The surface roughness stayed within 1 nm. The steepness of the Zn depth profile was 20 nm per decade when starting from the unetched front side, and 13 nm per decade in the back side mode. The DRFs showed that atomic mixing and redeposition were similar for both the conventional and back side methods.
The current trend towards improvement in depth resolution has resulted in the interface roughness becoming a limiting factor in the analysis of, in particular, industrially prepared metallurgical samples. Therefore, Shimizu et al.198 developed a coating procedure by dipping samples in H2O2 to produce a thin oxide layer on zirconium with a smooth interface and sharp concentration gap between the two layers. The high slope of the depth profiles reflected the absence of substantial topography. The method can be adapted for application to other studies. Pushing the limits of depth resolution further will place greater demands on chemical preparation rather than on instrumental refinements.
Gillen et al.200 elaborated methods for imaging non-conducting non-planar samples without coating with a metal layer, the removal of which would have involved the risk of beam-induced damage of the surface under study. A 19.5 kV O− beam with 0.5 µm spot produced positive secondary ions from human hair, chosen as a test sample. The secondary electrons 'automatically' compensated the surface charge build-up. Increasing the acceleration voltage improved the image contrast by reducing, in proportion, the effects of residual sample charging. To produce negative secondary ions, a 14.5 keV Cs+ primary beam with 0.2 µm spot was used in combination with low energy electron bombardment. The distribution of F− in labelled cocaine could be imaged.
Semiconductor applications demand elemental information from increasingly smaller areas. Therefore, Stevie et al.201 used a focused Ga+ ion beam with a spot diameter of 15 nm in the imaging of cross sections or side surfaces of tilted samples. Summation of the secondary ion counts laterally and in depth allowed LODs of 0.1 atom%, which were sufficient to study the implantation profiles of, for example, Al, Cr, K, Li and Na without the aid of oxygen or caesium. To obtain 'survey' compositional information from small features, the quadrupole was scanned over small mass windows. In practice, a set of up to 10 selected peaks could be monitored in 0.16 s from an area as small as 30 × 30 nm. This performance is superior to that currently possible by other techniques.
Xu et al.202 exploited the combination of isotopic detection and imaging to study the diffusion of moisture at the SiO2–TiN interface, which caused decohesion of ceramic multilayers. The samples were immersed in H218O to allow discrimination between the moisture and the 16O from SiO2. The upper layers were sputtered away and images recorded of the interface plane, of which the intensity distributions could be quantitatively treated to calculate the diffusion coefficient.
The potential of the MCs+ method to reduce matrix effects in SIMS is being increasingly exploited for depth profiling studies. Metson and Prince203 studied the interface, located at 200 nm below the surface, between TiOxNy films of varying stoichiometry and silicon or copper substrates. The MCs+ method did not eliminate the signal intensity variations due to the different sputtering rates in deeper layers. As the Cs+ signal was assumed to reflect directly the change in the work function caused by the Cs+ implantation at the surface, MCs+ intensities were normalized on the Cs+ intensity. Whereas use of a low primary ion beam energy (for example, 3 keV) produced sharp interfaces and thereby reduced the knock-on effect, a 14.5 keV Cs+ beam smeared the film/metal interface over more than 20 nm. Yue et al.204 characterized the interface for a 200 nm Ti film, deposited on polished aluminium nitride, as a function of the annealing conditions in vacuum. The depth profiles at the interface readily revealed the formation of Ti–Al binary and Ti–Al–N ternary compounds, for which the kinetics of formation could be followed as a function of temperature and time. These results would have been difficult to obtain using normal SIMS because of the existence of matrix effects.
Parks et al.205 studied sub-µm features in a highly repetitive array sample by using a large primary ion beam. The region of interest consisted of deep trench capacitors at a distance of up to 8 µm inside a dynamic random access memory device. Chemical and mechanical polishing was applied to remove the upper layers, resulting in a final surface roughness of about 0.3 µm. Topography development in SIMS analysis was minimized by aligning the long axis of the trench perpendicular to the primary beam. A large spot diameter of 30–60 µm was used, because the highly repetitive array of cells and geometrical considerations allowed the signals to be linked with the distribution of elements on a sub-µm scale.
Hoshi et al.206 reported a comparison of quadrupole and TOF SIMS in the dual beam mode for the shallow depth profiling of B in silicon. The LODs for both methods were of the order of 12 × 1016 atoms cm−3 and the secondary ion signal intensities were stable to within 1.3%. The depth profiles agreed well except for the first 3 nm. The decay length was 1.5 nm in TOF- and 1.3 nm in quadrupole-SIMS. The disagreement for the transient regime was due to the differences in primary ion bombardment conditions for the two instruments.
Since SIMS has established itself as the method of choice for the analysis of semiconductor materials, ultra-shallow depth profiling has become of increasing interest. Kinoshita et al.207 evaluated use of a low energy Cs ion beam in quadrupole SIMS for ultra-shallow depth profiling of As in silicon. Reducing the primary ion impact energy from 5 to 2 keV at an incidence angle of 60° improved the decay length significantly. Normalization of the As signal to that of Si in the stable matrix region improved the agreement between experimental and calculated implantation doses to 10%.
Ormsby et al.208 demonstrated the advantages of ultra-low energy primary ions impacting perpendicularly on samples containing 10 δ-layers of B in silicon. The effect of primary ion energy, incidence angle and oxygen flooding was studied systematically. The combination of low energy and normal incidence allowed a constant high depth resolution throughout the different layers. The 10th δ-layer was profiled at an apparent depth of about 1280 nm. The FWHM and decay length of 2.6 and 0.9 nm, respectively, compared favourably with the values of 3.4 and 1 nm, respectively, obtained when using an incidence angle of 50°. Oxygen flooding decreased the depth resolution significantly.
The 3-D distribution of B and Li, derived from reactor water, in the oxide layer of corroded zircalloy fuel rod cladding was performed by a combination of depth profiling, line scanning and imaging with high lateral resolution.210 The exact thickness of the oxide layer was measured with SEM on cross sections. Charge compensation was not needed because zircalloy became conducting under Ga+ bombardment. The roughness of the oxide–metal interface limited the spatial resolution to about 1 µm, although 50–200 nm should in theory be feasible. In this study also, micro-analysis was ahead of the actual requirements for material analysis.
6.6 Static SIMS (S-SIMS)
There is a growing interest in methods which provide direct speciation information in addition to elemental analysis. Therefore, Cuynen et al.213 systematically investigated solid state speciation of oxides. The reproducibility attained 15–20% for both thin layers and bulk samples as long as elemental ions were normalized on elemental ions and cluster ions on clusters. The major part of the total ion current was carried by the elemental ions but the cluster ions at high m/z had high specificity, which compensated for their low intensity. The main signals at high m/z corresponded to simple adducts of the neutral analyte with stable ions, already detected as such in the low m/z range. Hence, deductive identification could be achieved in most cases. Reference spectra were only needed to distinguish between oxides containing the same metal in different oxidation states.
Gilmore and Seah214 dealt with the general problem of calibration of the detection efficiency of channel electron multipliers. Specifically, S-SIMS was chosen because of the possibility of bombarding with primary ions and imaging the secondary electrons. The secondary electron images showed local 'hot spots' with significantly different ion/electron conversion efficiency. The influence of the operating voltage on the electron images could be directly demonstrated. Line scans across these images allowed the detection efficiency over the different regions to be quantified.
During metal deposition on semiconductor surfaces, the metal can penetrate through the open lattice cells of the upper monolayer to adsorb or bind to a specific atom in the 2nd or 3rd layer. Gross et al.216 used angle-resolved S-SIMS to study the Al–GaAs {001} c {4 × 4} interface. Metallization, angle-resolved SIMS and reflection high-energy-electron diffraction measurements were carried out in the same vacuum vessel. The angular distributions of the desorbed Al+ and the Ga+ showed specific changes at Al overlayer coverages of 0.33 and 1 monolayer. Molecular dynamics simulations were used to elucidate the change in desorption mechanism brought about by the overlayer.
Hamoudi et al.217 investigated the speciation of C and O at the surface of spent MnO2–CeO2 wet oxidation catalysts. The simultaneous presence of oxides, adsorbed water, surface hydroxyl and organic oxygen, as well as of aliphatic and aromatic carbon, made X-ray photoelectron spectroscopy unsuitable for unravelling the distribution of the different forms of C and O. The molecular detection capability of S-SIMS allowed the contributions of, for instance, aliphatic chains and aromatics to be differentiated directly. Relative peak intensities such as C2H−∶C2− could be used to quantify the relative contributions of graphite, aromatic or aliphatic hydrocarbons.
Liang218 used S-SIMS for the speciation of sulfur species in high-sulfur coal. Analysis by the commonly used method (ASTM D 2492-84) gave a S content of 8.9%, mostly defined as organic sulfur. However, SIMS detection of the S2–S12− clusters showed that the sample contained mainly elemental sulfur. The mass accuracy was better than 15 millimass at m/z 300, except for the smallest peaks, and isotopic patterns were within 4% of those expected. The conclusion was backed up by analysis of industrial S8 samples. Hence, the fraction originally defined as 'organic sulfur' would be better defined as 'organic phase' sulfur.
Ingram et al.220 developed a statistical image classification to identify the mineral phases in basalt using a Kohonen network. The ion images were statistically grouped using a program which employed a generalized learning vector quantification technique. A set of selected ions including Al, Ca, Fe, K, Mg, Na, Si and Ti was used to typify mineral phases such as plagioclase, olivine, pyrene and magnetite/ilmenite. The basalt was pressed in indium foil and the Ga+ primary ion beam was rastered over a total area of 80 × 80 µm at an ion dose of less than 3 × 1012 ions cm−2. The lateral resolution within 1 µm was sufficient because the size of the mineral phases was typically hundreds of µm.
Gerlock et al.221 exploited the combination of imaging and isotope detection to study the UV and thermal degradation of automotive paints. The samples were treated under an 18O2 atmosphere, cross-sectioned and the 18O− distribution was imaged with a primary ion dose of less than 4 × 1012 ions cm−2. The lateral resolution was 1–2 µm.
7 Sputtered neutral mass spectrometry (SNMS)
7.1 Instrumentation
Sichi et al.222 developed a reflector TOFMS with new liquid metal ion gun and resonant laser post-ionization for analysis with a lateral resolution in the sub-µm range. The diameter of the primary beam spot on the sample was less than 10 nm. The neutral species were ionized with a combination of 2 YAG and 3 YAG-pumped dye lasers. Useful yields improved by at least 2 orders of magnitude in comparison with Ga+ SIMS, but some of the possible sensitivity gain was lost because TOFMS needed extremely short analysis pulses. To keep the erosion rate and analysis time within reasonable limits, additional sputtering without signal registration had to be applied. Nevertheless, a LOD for B in silicon of less than 1018 atoms cm−3 could be achieved.7.2 Analytical methodology
Plasma-based SNMS with quadrupole analysers is suitable for the elemental analysis of almost any kind of material as long as the isobaric interferences can be accounted for. Goschnick et al.223 investigated the use of KEDs to determine the relative contributions of CH2+, Li2+, N+ and Si2+ at m/z 14. The KEDs of atomic and cluster ions generated from several inorganic and organic materials exhibited characteristically different shapes and were used to deconvolute the KEDs measured in unknown samples. This approach reduced the quantification error in standard samples to 3% from the 66% deviations experienced with the common correction procedure using relative signal intensities. The approach has been applied to the depth profiling of a Si3N4–SiO2 bilayer on silicon.224 In the case of Si3N4, the signal at m/z 14 consisted of a linear combination of N+ and Si2+. To increase the depth resolution, data were acquired at only two discrete energies so that a depth resolution of 3 nm could be achieved.He and Becker225 investigated the feasibility of surface analysis with uniform elemental sensitivity by high power laser post-ionization of neutrals, sputtered by a 5 keV Ar+ beam. To drive elements into saturation, the flux density and/or the pulse duration could be increased. A comparison was made between a 'common' Nd∶YAG, yielding 520 mJ per pulse of 6 ns, and a mode-locked Nd∶YAG laser, producing 40 mJ per pulse of 35 ps at the same wavelength of 532 nm. Elemental RSFs were determined in alloys sputtered by Ar+. The use of 6 ns pulses gave uniform elemental ionization at significantly lower power in comparison with the 35 ps laser. Laser focusing would still allow further reduction to be made, on the condition that the timing matched the velocity distribution of the neutrals. Under saturation conditions, the RSFs were practically unaffected by shot-to-shot variations of the laser and differences in light absorption.
The emission of molecular instead of atomic species upon the sputtering of the sample may hamper quantification significantly. Therefore, Jenett et al.226 investigated the KEDs of BO+ and BO2+ ions in low-pressure hf SNMS of copper powder mixed with H3BO3. It was demonstrated that BO+ and BO2+ ions exhibit the same KED as the Ar+ plasma gas ions and that the emission of low-energy BO and BO2 molecules was of the same order of magnitude as that of fast atomic secondary B neutrals.
Goschnick et al.227 have demonstrated the benefits of using one instrument for both plasma SNMS and SIMS for elemental survey analysis and depth profiling, respectively, of size fractionated aerosols. Initial bombardment was achieved by Ar at 1.3 kV with 2 mA cm−2 for plasma SNMS and at 5 keV with 10 µA cm−2 in SIMS. The SNMS signal at m/z 14 was corrected for interferences using the intensity ratios to Si+ and C+. Although the matrix effect was ignored, atomic concentrations in SNMS could be determined within 30–40%. The SIMS erosion rate of 0.1 nm s−1 allowed the chemical compositions of the outer layers and the core of the aerosol particles to be distinguished.
7.3 Single and multidimensional analysis
In the hf mode, the ion bombardment is periodically interrupted to extract the electrons from the plasma towards the sample for charge compensation of insulating samples. Goschnick et al.228 investigated the use of plasma-based hf SNMS at 500 kHz for the depth profiling of non-conductive oxidic multilayers. A depth resolution of 5 nm was obtained for a silicon wafer coated with a 105 nm silicium dioxide layer. This corresponded to a relative depth resolution of 0.05.Higashi229 compared the depth profiling of AlAs–GaAs multilayers using either SNMS with laser post-ionization or SIMS. Since a quadrupole MS was used, the energy and incidence angles of the primary beam could readily be changed for optimal depth resolution. Although the ion yield of As from AlAs was nearly 1000 times that of As from GaAs when using SIMS, it was only twice as high when using laser post-ionization SNMS. The remaining 'matrix' effect was explained by the difference in sputter rates. Shallow depth profiling of B implants was compared using Ar+ SIMS, O2+ SIMS, laser SNMS and Auger electron spectroscopy. The Ar+ SIMS profiles deviated from the expected Gaussian curve for the first 5 nm because of B+ yield enhancement by the native oxide layer. In O2+ SIMS, the transient regime distortions occurred at depths up to 10 nm. In contrast, SNMS depth profiles agreed in detail with those obtained by Auger electron spectroscopy.
The small spot sizes feasible with liquid metal ion sources inherently limit the secondary fluxes. Willey et al.230,231 have used resonant post-ionization to acquire images with sub-µm lateral resolution. A low-energy defocused ion gun was used for sputtering in combination with a high energy primary beam with a spot diameter of 50 nm for TOF analysis. The implantation doses measured for Cu in pure silicon and silicon dioxide agreed within 3%, showing the absence of matrix effects. The dynamic range was 6 orders of magnitude. Imaging experiments allowed the distribution of Cu in and around a Cd inclusion in a CdZnTe film to be visualized with sub-µm resolution. The brightest and darkest pixels corresponded to a concentration of >25 µg g−1 and <5 µg g−1, respectively. Depth profiles of a 10 nm AlGaAs layer in GaAs, located 10 nm below the surface, showed projected range standard deviations of 2 and 2.1 nm for the rising and trailing edges, respectively.
Müller et al.232 has achieved one of the very first 3-D elemental analyses with SNMS using non-resonant laser post-ionization and high mass resolution TOFMS. In this study also, the dual beam approach was used with an unfocused beam for sputtering and a high intensity pulse for analysis of a sample containing a pattern of Schottky contacts on a GaAs substrate. The distributions of Al, As, Ga and Ti were measured with lateral and depth resolutions of 3 µm and 20 nm, respectively. Depth resolution, limited primarily by the penetration depth of the 22 keV Ga+ primary ions, would be improved significantly through use of low-energy Ar+ primary ions. Relative detection sensitivities, referenced to Al, were between 0.1 and 4. The lateral resolution was intentionally limited to avoid topography development, which would have reduced the depth resolution in the deeper layers.
8 Stable isotope ratio mass spectrometry (SIRMS)
8.1 Reviews
The review of Meier-Augenstein233 on compound-specific analysis by SIRMS covered the literature since the advent of commercial GC-combustion-SIRMS systems in 1990 to the beginning of 1998. The capability for measuring isotope distribution at the natural abundance level with great accuracy and high precision has resulted in the technique becoming widely used for the control of authenticity and the determination of origin of samples. The considerable demands of sample preparation and clean GC-separation were covered in the review. Two reviews have highlighted the increasing use of SIRMS in biochemical and biomedical applications. Metges and Petzke234 considered the usefulness and sensitivity of the technique, specifically for amino acid and protein metabolism studies, in a 50-reference review which outlined the instrumentation and principle methodologies for the measurement of 13C and 15N enrichments in amino acids. The use of continuous flow SIRMS in biomedical research, covered by the 41-reference review of Brazier and Elbast,235 was seen to have three main advantages: stable isotope labelling, the high performance of SIRMS and the potential of coupling sample preparation to the system in hyphenated techniques. A number of examples of application were presented to illustrate the discussion.8.2 Instrumentation
Jackson et al.236 have developed a highly sensitive static mass spectrometer for the analysis of small samples of atmospheric methane. The method, which uniquely used methane as the analyte, required <10 ml of ambient air (8 ng CH4) for analysis, thereby making replicated determinations of the isotopic composition of methane in small samples feasible for the first time. Initially a mass spectrometer was designed to measure methane as the analyte and subsequently an inlet apparatus, based on packed column GC, was built to separate methane quantitatively from air samples and to transfer the gas to the mass spectrometer. Full details were given for the protracted procedure required to remove all impurities for the isotopic analysis of air samples. Small air samples (10 ml) could be analysed for δ17M with a precision (1σ) of 0.26‰. The technique was considered ideal for studies of the stable isotopic composition of soil methane fluxes because the small sample requirement allowed sampling to be carried out with minimal disturbance of the ecosystem under investigation.Mason et al.237 used a vaporizer injector with programmable temperature to introduce underivatized anabolic steroids into a GC-SIRMS instrument. The injector allowed large volumes of samples to be injected while the injector liner remained cool. This resulted in more sample being presented for analysis and greater sensitivity. In addition, all the solvent could be removed prior to analysis, thereby eliminating solvent peak tailing, a major problem in GC-SIRMS. Higher initial GC oven temperatures could be used to reduce the GC run time and improve the chromatographic peak shape.
The determination of both δ13C and δ15N in the same organic sample, using continuous flow SIRMS, poses an analytical challenge in that the carbon peak is much larger than the nitrogen peak and overloads a system set up to determine the nitrogen isotopic composition. A commercial interface (Werner et al.238), placed between the elemental analyser and the mass spectrometer, has been designed to overcome this problem by diluting the oxygen peak with pure He, thereby attenuating the oxygen signal. Full details were given of the dilution mechanism which operated virtually free of isotopic fractionation. The precision for carbon isotope ratio measurements of atropine as a model compound was in the range 0.6–0.06‰ for sample sizes in the range 1 µg to 5 mg C, respectively. A correction for the blank signal was critical for samples containing <4 µg C.
Interfacing an elemental analyser with a quadrupole mass spectrometer was demonstrated by Russow and Goetz239 to be a low-cost alternative to conventional continuous flow SIRMS for analysis of enriched soil and plant samples. Up to 80 samples could be automatically processed in a batch, with each measurement taking about 15 min. The reproducibility (1σ) for >100 µg N was better than 0.2% and 0.12% for δ15N at natural abundance and >1 atom %, respectively. It was possible to determine δ15N and δ13C reliably at ≥8‰ and ≥2‰, respectively. Russow240 has also reported a procedure for the determination of 15N in 15N-enriched nitrite and nitrate in aqueous samples. The procedure was based on the reduction of the nitrate and nitrite to NO gas prior to analysis on the quadrupole instrument. Sich and Russow241 found that incorporation of cryotrapping for the determination of 15N abundance in NO and N2O, present at very low concentrations in atmospheric samples, improved the analytical performance considerably.
Hall et al.242 used a continuous flow SIRMS instrument coupled to a quadrupole mass selective detector in the study of contaminant degradation in order to acquire simultaneously mass spectral and stable carbon isotope data from a single chromatography run. Both the target contaminant and extracellular intermediates could be identified and quantified in the one run. Ion source non-linearities, observed when the mass spectrometer was tuned for high sensitivity, required peak height correction with an algorithm produced using isotopic standards over a range of concentrations.
Gleixner and Schmidt243 coupled a pyrolyser to a GC-combustion-SIRMS combination in order to determine group-specific C and N isotope ratios in small amounts of organic matter. They demonstrated that partial degradation of refractory organic acids from aquatic systems, using controlled pyrolysis, allowed intermolecular and intramolecular distributions of the isotopes in individual compounds to be assigned. It was possible to identify the origin of specific pyrolysis products as being of plant or secondary (bacterial or fungal) origin. The potential of the new system for providing novel insights into the dynamics of soil organic components was illustrated by Gleixner et al.244 in a study of the carbon turnover rates of individual soil components.
The presence of He carrier gas creates problems in hydrogen isotope analysis by continuous flow SIRMS because of the potential spectral overlap by He on the HD+ peak at m/z 3. Two major instrument manufacturers have taken very similar approaches to overcoming this problem. Merren245 patented an energy filter fitted to the ion detector assembly which removed scattered He+ ions of reduced energy. Hilkert et al.246 reported development of a new mass spectrometer for measurement of transient hydrogen signals with high precision and accuracy. A key feature was incorporation of a retardation lens integrated into the m/z 3 Faraday cup collector.
8.3 Analytical methodology
Two groups in the UK have made substantial advances in the use of laser-assisted fluorination in the high-precision determination of oxygen isotopes in small samples. Pillinger and colleagues247 reported optimization of the procedure for 17O isotope determination using infrared (10 µm) laser-assisted fluorination in conjunction with a dual inlet mass spectrometer of high (for a SIRMS instrument) resolving power of 250. Precisions (1σ) were typically 0.08 and 0.04‰ for δ17O and δ18O, respectively, for 0.5-2 mg of silicates and other oxide mineral grains. Samples were loaded into a nickel block containing 22 wells and were individually heated to fusion prior to controlled melting with the laser and fluorination in an atmosphere of BrF5. For extraterrestrial materials, the achieved precision permitted the offset between different (parallel) mass-dependent fractionation lines to be measured to <±0.02‰. This was substantially smaller than previously reported. The mass-dependent fractionation relationship between δ17O and δ18O in terrestrial rocks was also re-evaluated. Young and colleagues have used UV (255 nm wavelength) laser ablation in an atmosphere of fluorine for the microanalysis of Ca–Al-rich inclusions from the Allende meteorite248 and tooth enamel phosphate.249 A spatial resolution of 100 µm or less represented a 100–fold reduction in minimum sample size compared with other fluorination techniques. Analysis of minerals gave analytical precisions (1σ) of ±0.3‰ and ±0.4‰ for δ18O and δ17O, respectively. The data for tooth enamel phosphate demonstrated that the procedure offered a combination of spatial and analytical precision for in-situ sampling of biogenic apatites unobtainable by other methods.Laser extraction coupled with combustion has been developed by Wieser and Brand250 for the in-situ determination of δ13C in organic and inorganic materials. Carbonaceous compounds were volatilized by the laser and converted to CO2 by combustion. The combustion products were separated by GC prior to the SIRMS analysis. Plant samples could be analysed with a spatial resolution of 200 µm and δ13C values determined with a precision of ±0.3‰.
Methods for the elucidation of site-specific isotopic composition have been receiving considerable attention. In particular, two laboratories have independently developed similar methods for the determination of the intramolecular distribution of the nitrogen isotopes in N2O. The methods of Toyoda and Yoshida251 and Brenninkmeijer and Röckmann252 were both based on the analysis of the fragment 14NO+ and 15NO+ beams at m/z 30 and 31, respectively, in combination with standard measurement of unfragmented N2O+ at m/z 44, 45 and 46. The latter authors considered the analysis relatively easy although they expressed concern over some complications which included the occurrence of mass-independent fractionation of oxygen in atmospheric samples and the consequent invalidation of the algorithm used to derive 17O from measured 18O. In addition there was some evidence for the scrambling of isotopic information as a result of reactions in the ion source. They also identified the lack of a reference material for this type of analysis as a drawback. However, Toyoda and Yoshida prepared standard samples by the thermal decomposition of NH4NO3 which produced N2O with N at the centre position derived from nitrate and N at the end position derived from ammonium. The precision of their analysis was ≤0.1‰. This methodology has considerable potential in studies of the sources and sinks in the atmosphere of N2O which plays an important role in the greenhouse effect and regulation of the ozone layer. Although Zhang et al.253 were able to measure site-specific 13C information for glycerol, used as a test compound, with a standard deviation of ≤0.7‰, they admitted that the method was not very practical as each compound would need a dedicated reaction to produce the unique fragments needed for the analysis. Glycerol was transformed by periodic oxidation into formaldehyde, which contained carbons 1 and 3, and formic acid, which contained carbon 2. The former was converted into hexamethylenetetramine and the latter into barium formate for measurement of δ13C.
Mazeas and Budzinski254 found that compound-specific isotope analysis of polycyclic aromatic hydrocarbons (PAHs), isolated from marine sediments contaminated with petroleum, was not possible because of incomplete GC separation. It was demonstrated, however, that it was still possible to apportion the methylphenanthrenes to source using the isotopic composition of the sum of the methylphenanthrenes and a simple mass balance calculation. Kaneko et al.255 reported that in the separation of natural gas components a combustion furnace temperature of >900![[thin space (1/6-em)]](https://www.rsc.org/images/entities/char_2009.gif)
°C was required in order to achieve reproducible isotopic measurement of individual components. Analysis of pure methane gas gave a standard deviation of 0.08‰.
Abramson and colleagues256 reported that the use of on-line FI with chemical reaction interface-SIRMS gave performance superior to that of off-line combustion for detecting low levels of enriched 13C in mass balance studies. Samples flowed through a desolvation system prior to combustion in the microwave-powered chemical reaction interface to form 13CO2. It proved possible to quantify less than 100 ng ml−1 of excess 13C (1 µg ml−1 of 13C-labelled drug) in samples equivalent to 10 µl of urine. The authors concluded that the method was sensitive enough for mass balance studies to be carried out using doses as low as 1–2 µg g−1.
In continuous flow SIRMS, drift and non-linearity effects are usually compensated for by including in each analysis batch several reference samples with the mass of the determinand closely matched to that in the unknown samples. Ohlson and Wallmark257 devised a new calibration model for correction of analytical results for drift and elemental mass. Pairs of reference samples replaced the normal single reference with one of the pair having a constant mass and the other having a variable mass over a range up to seven times the minimum quantifiable mass. Concentration and isotope ratio data were corrected using two factors, one based on peak area and the other on analysis time. The correction factors were expressed as polynomials whose coefficients were obtained using an optimization algorithm which minimized the standard deviation for the reference data while making the mean equal to the reference value.
8.4 Sample preparation
The attractions of on-line sample preparation for oxygen isotopic analysis over off-line methods include faster throughput, less risk of fractionation and a smaller sample requirement. Wassenaar and Koehler258 described field sampling procedures for the collection of gaseous and dissolved O2. This was used in studies in which oxygen isotopic composition was used to follow transport or biogeochemical processes in terrestrial and aquatic systems. Samples collected in evacuated serum bottles were subsequently injected into a He carrier stream in a modified elemental analyser–SIRMS system. There was no need for cryogenic preconcentration of O2 or conversion into CO2 for analysis and the use of O2 allowed both δ17O and δ18O to be measured. Repeat injections (n = 35) over a two month period gave standard deviations of ±0.17 and 0.5‰ for δ18O and δ17O, respectively. Kornexl et al.259 were able to determine δ18O values in organic and inorganic samples containing 50–100 µg O with a standard deviation of ≤0.5‰ using on-line pyrolytic decomposition at 1400°C in the presence of nickelized graphite. It was possible to determine carbon isotope ratios of organic substances and nitrogen isotope ratios of inorganic nitrogenous compounds in the same sample run. The method of Loader and Buhay260 was based on high-temperature pyrolysis of organic samples over a carbon source followed by a rapid, non-contributive partial catalytic oxidation over nickel powder at 550-600°C. Initial results demonstrated precisions of ≤0.2‰ in the analysis of cellulose and silver nitrate.
Equilibration methods for the isotopic analysis of small water samples have come back into favour. Socki et al.261 adapted the traditional CO2–H2O equilibration method for use on the microscale (10 µl samples). A feature of the method was that the equilibration, shown to be complete within 28 h, was carried out in 6 mm tubes attached directly to the mass spectrometer inlet. In contrast to other methods, there was no need for dangerous and expensive chemicals and the water samples were preserved for further analysis. The reproducibility was ±0.1‰ and ±0.6–2.5‰ for oxygen and hydrogen isotopes, respectively. McConville and Karlin262 found that direct equilibration of CO2 with soil water was less time consuming and more reliable than methods based on separation of water from the soil matrix prior to analysis. The method could be used to analyse the δ18O tracer profiles in the unsaturated zone of field soils.
Martin and colleagues263 have demonstrated that the δ13C values of individual sugars (fructose, glucose and sucrose) separated from pineapple juice or concentrates allowed adulteration to be detected through use of specific natural isotope profiles. The sugar fraction was separated from organic acids by anion exchange and the individual sugars subsequently separated by HPLC. Although no significant isotope fractionation occurred when recovery was close to 100% and hydrolysis of sucrose was avoided, reliable interpretation of results required close control of the recovery and sucrose hydrolysis.
The determination of 13CO2 enrichments in the CO2 released spontaneously from acidified blood has been shown by Dangin et al.264 to be a feasible alternative to the traditional breath sampling method if the latter is not possible. Although there was an enrichment in blood samples relative to the corresponding breath samples, there was a strong statistical relationship between the results obtained by the two methods which gave identical experimental results. Analysis could be performed with good reproducibility for up to five days after taking the blood samples.
The preparation of samples for sulfur isotope analysis can be problematic in that analysis of SO2, although cheaper, safer and more convenient than analysis of SF6, must take into account the uncertainty associated with the oxygen isotope composition. The method of Halas and Szaran265 involved on-line low-temperature (650–700°C) thermal decomposition in copper reaction boats of BaSO4 (10 mg) mixed with NaPO3 (60 mg). Any SO3 formed was converted completely into SO2 and there was no memory effect for the S and O isotope ratios. The oxygen isotopic composition of the SO2 produced was controlled totally by the 18O content of the NaPO3 reagent as long as the reagent : sample ratio was ≥6. The method, suitable for small samples, had the advantages of being quick (10–11 min reaction time), economical in terms of materials and energy, and safe.
The availability and use of RMs is crucial for the production of valid isotope data. Verkouteren266 has reported the preparation of three CO2 isotopic RMs and their characterization for isotopic uniformity, stability and composition. An intercomparison exercise involving international expert laboratories allowed the new RMs to be standardized against the Vienna Pee Dee Bellemnite scale. A primary isotopic gas standard for sulfur in the form of SF6 (Valkiers et al.267) is now available commercially. The interesting study by Kornexl et al.268 highlighted the need for exact and reliable calibration of existing RMs as well as the production of new standards. Results for a set of laboratory standards, collected from nine European and North American laboratories and analysed using a new on-line pyrolysis system calibrated with Vienna Standard Mean Ocean Water and Standard Light Antarctic Precipitation, showed up discrepancies of up to 2‰ from the reported values.
9 Thermal ionization mass spectrometry
9.1 Instrumentation
Although the precision and accuracy of TIMS have steadily improved over the years, with advances in instrumentation and methodologies, commercial designs have remained based on instrumental platforms which date back to the single collector instruments of the 1970s. Schwieters et al.269 have briefly described a new commercial instrument in which virtually all components from the ion source to the detector system have been redesigned. It was expected that the new instrument would lead to significant improvements in precision and accuracy and would open up measurements not currently feasible.9.2 Ion formation
Yoneda and Heumann270 have investigated the formation of negative ions from mixed solutions of platinum group elements (Ir, Os, Pd, Pt, Rh and Ru), Au, Mo, Re and W, with the aim of determining these elements without prior separation. Single Pt filaments were loaded with Ba(NO3)2 solution (2 µl, 10 mg Ba ml−1) and sample solution (2 µl, approximately 80 µg ml−1 for each element) to give 20 ng Ba and about 170 ng of each element on the filament. The dominant negative thermal ions were, in order of ion beam intensities, ReO4−, OsO3−, IrO2−, RuO3− and MoO3−. In addition, Au−, Br−, Cl−, I−, Pt− and PtO− ions were observed but no negative species for Pd, Rh and W could be detected. The last three elements probably require a different filament arrangement so a combination of Pt and Re filament methods might prove more effective. The suppression effect of strong negative ions (for example, Br−, Cl− or ReO4−) on other negative ion species could be reduced but not removed by overnight baking at a low filament temperature.Kawai et al.271 have studied the process of calcium ionization on a rhenium filament loaded with CaI2 as the sample. The analysis revealed that three different processes were involved in ionization, CaI2 → Ca+, CaI → Ca+ and Ca → Ca+, at low, medium and high evaporation filament temperatures, respectively, indicating that thermal decomposition of the CaI2 was an important part of the process.
9.3 Calibration
The paper of Johnson and Beard272 can be recommended for its detailed consideration of the fundamental aspects of mass fractionation and for its discussion of methods for the correction of fractionation, in particular in the analysis of Fe. A completely general derivation was presented which could be used for any appropriate isotope system and which was applicable to the mass fractionation laws known to occur in TIMS. A rigorous method for correction of fractionation was applied to the isotopic analysis of Fe using the double-spike method with error propagation assessed as a function of algorithm and spike isotope composition. The described procedure corrected for instrumentally produced mass fractionation yet measured natural, mass-dependent, isotopic variations in the samples. The precision of ±0.02–0.03% achieved for measurement of 54Fe∶56Fe was the best yet reported. Minimum propagated uncertainties corresponded to a spike consisting of 54Fe and58Fe in the ratio 9∶1.A number of calibration procedures for a multicollector instrument were used by Ramakumar and Fiedler273 in their assessment of accuracy of the total evaporation method in the measurement of 239Pu, 187Re and 238U. Procedures were presented for evaluation of magnet calibration, gain calibration, Faraday cup efficiency factors, amplifier linearity check and mass fractionation. The exercise demonstrated that the various correction parameters need to be checked frequently in order to maintain high accuracy and precision. No mass fractionation was observed in the total evaporation procedure.
9.4 Analytical methodology
The use of lead as a global environmental tracerrequireslead isotope data from the various source regions in order to interpret data for aerosol samples. Bollhöfer et al.274 have initiated a world-wide sampling campaign in which aerosol samples were collected using a standard sampling kit and all analyses carried out in a single laboratory under rigorous ultra-clean sample preparation and analysis conditions. The paper was notable for its very detailed account of the sampling and measurement procedures and for the estimation of blank levels of lead from various sources including filaments and loading reagents, beakers, ion exchange resin, filters and extraction reagents. The complete procedural blank, the major contributor to which was the ion exchange resin, was estimated to be between 14 and 120 pg. Those samples assessed by visual inspection to contain low levels of particulates were processed without any column chemistry in order to reduce the blank to about 15 pg. Approximately 400 pg of lead on a filter was required to minimize the influence of the filter blank to less than 4% (without column chemistry), corresponding to a collection time of one month for Antarctic air samples. Acceptable accuracy and precision (0.2% for 206Pb∶207Pb) was, however, possible with 120 pg size samples.Improvements to the traditional silica gel method for the analysis of lead were realised by fitting a timer to control the electrical power supply for the sample loading unit.275 The aim was to achieve a constant loading condition in order to obtain a very stable ion beam, leading to higher analytical precision than was previously possible. Under the conventional loading method the filament current is raised to approximately 2 A until the filament becomes dull red and fumes of phosphoric acid are produced. In the modified procedure the filament was subsequently heated rapidly (2.7 A for 1.8 s) to fuse the silica gel on to the filament. Stable ion beams were produced for sample sizes in the range 150–250 ng Pb. Emission became unstable for samples above 450 ng Pb. The optimum loading conditions were assessed to be the minimum possible amount of H3PO4 (1.2 µl of 1 N H3PO4) and 2 µl of silica gel suspension. A notable feature of the stable ion beam obtained was the lack of mass fractionation observed with the measured isotope ratios remaining constant over a period of 3 h.
Galer and Abouchami276 have reported that a triple-spiking procedure gave optimum conditions for the correction of instrumental mass fractionation in the analysis of common Pb. The spike used was a mixture of 204Pb, 206Pb and 207Pb.
A new technique has been reported by Richter et al.277 for the measurement of magnesium isotope ratios using MgF2 as the analyte. The main reasoning for using the MgF3− ion was that mass-dependent isotopic fractionation would be less for this ion than for the lighter Mg+ ion. Both evaporation and ionization filaments were covered to a thickness of 20 µm with BaF2 before application of Mg solution (1 µl, 1 µg l−1), H3BO3 (1 µl, 0.1 M) and a final layer consisting of AgF (30 µg). Magnesium, loaded as MgF2, evaporated from the filament as MgF2 molecules and the formation of MgF3− ions was enabled by the presence of BaF2 and AgF on the filaments. A typical ion current at m/z 81 was 5 × 10–14 A. The precision achieved was 3–4 times better than that possible with ICP-MS.
The need for RMs with calibrated isotope ratios has prompted Taylor and colleagues to establish the uranium isotopic composition of a set of six uranium ore samples278 and 12 commercially available uranium chemicals.279 The results obtained were traceable to SI units. The intriguing finding that all the chemical reagent samples contained the non-natural isotope 236U confirmed the origin of such reagents to be nuclear industry plants.
9.5 Sample preparation
Maxwell and Jones280 have worked up a series of column-based procedures for the rapid separation of actinide and other metallic impurities in plutonium scrap metals for TIMS analysis. Columns packed with the commercial resins UTEVA Resin®, TEVA Resin® and TRU Resin® were used, either singly or in combination, with vacuum-assisted elution to increase flow rates and throughput. A main goal was clean separation of 238Pu and 241Pu from 238U and 241Am, respectively, to eliminate isobaric interferences.The TIMS determination of 226Ra in fish otoliths requires the clean removal of Ba and Ca, which interfere in the analysis. Andrews et al.281 have developed an ion-exchange procedure which separates efficiently the Ra from Ba and Ca. The sample size requirement for this procedure was three times less than that of traditional techniques for the determination of Ra (radon emanation and α-spectrometry).
10 Other methods
10.1 Electrospray mass spectrometry (ESMS) and ion spray mass spectrometry (ISMS)
The widespread interest in the use of ESMS in elemental and speciation analysis continues, but there has been a degree of rationalization of the perceived potential of the technique. The trend towards using molecular information derived by ESMS to supplement the elemental information provided by ICP-MS has now become more apparent than previously noted but in several instances this has entailed the use of complex and expensive tandem instruments. The substantial review of Stewart282 provided the reader with a good overview of the current potential and limitations of ESMS for elemental speciation. Analytical considerations included quantitation, sensitivity, limitations, applications and future directions. Although current limitations (e.g., the high backgrounds due to chemical noise and signal suppression) were acknowledged and discussed, it was considered that continued development would improve the capability of the technique and widen its application.An understanding of the relationship between the ion species in solution and those observed in the ESMS spectrum remains fundamental in interpreting the results of speciation analyses. Any change in speciation induced by the electrospray process itself can clearly place severe limitations on the feasibility of the technique for this application. Schramel et al.,283 in a study of capillary electrophoresis (CE)-ESMS for trace element speciation, concluded that the only classes of compounds unrestrictedly suitable for ESMS detection were covalent organometallic compounds and strong metal complexes. Metal ions and weak metal complexes could undergo gas-phase ligand replacement, thereby losing the original structure information. The paper also presented solutions to some of the technical problems experienced in coupling CE with ESMS, for example, the optimum position of the capillary inside the ES tip, compatible buffer systems, concentration sensitivity and detection mode. Wang and Agnes284 studied the well-understood complexation reactions between EDTA and each of the alkaline earth metal ions to gain an insight into how idealized systems were perturbed by the ES process. Shifts in the equilibrium position, always towards increased complexation, were observed, with the degree of shift correlating generally with the rate of water exchange with free metal ions. These alterations in solution species were considered to be predictable so that the ESMS data still gave meaningful information. Striegel et al.285 found that polarizability, an inductive effect and solvent molecular size all played a role in determining the relative intensities of solvated cation peaks in the ESMS study of the binding of alkali metal cations to low molecular weight solvents.
The determination of Cu has dominated the use of ESMS for metal analysis. Following on from the conference presentation on electrochemical sample pre-treatment coupled on-line with ESMS for elemental analysis, as reported in last year's Update, Pretty and Van Berkel286 have presented full details of the method using Cu as the model determinand. Under optimum conditions, the test solution was passed through a thin layer electrochemical cell for 2 min at 60 µl min−1 and the CuII deposited on the vitreous carbon working electrode at an applied potential of −800 mV. Following washing of the cell, the deposited Cu was stripped and infused into the mass spectrometer. The LOD was 1.4 µg l−1 and the calibration graph was linear over two orders of magnitude. To gain the structural information not provided by such a procedure, tandem mass spectrometry was used to study copper (Stone and Vukomanovic287) and tin and arsenic (Pergantis288) species present in solution. Ions produced in the ES process were dissociated in a collision cell placed between the two mass spectrometers to produce daughter ions. Bantan et al.289 also used tandem MS in their study of low molecular weight Al complexes in human serum samples to identify the main species eluted under chromatographic peaks as aluminium citrate, aluminium phosphate and ternary aluminium–citrate–phosphate complexes. The LOD was 5 ng cm−3 with an RSD of 8%. Mester and Pawliszyn290 coupled in-tube solid phase microextraction with ESMS for the analysis of trimethyl- and triethyllead and found that high fragmentation voltages produced complete dissociation. It was considered that by programming the fragmentation energy it would be possible to obtain in parallel both molecular and atomic signals of lead compounds.
The study of selenium speciation has continued to receive considerable attention. The series of papers by Uden and co-workers291–293 using HPLC and the overview of Michalke et al.294 on the use of capillary electrophoresis, both coupled to either ICP-MS or ESMS, all came to the general conclusion that analysis times were several times less and the LOD at least 102 better when using ICP-MS. A major drawback of this technique, however, was the difficulty in identifying the eluted selenium compounds in particular as it would be unwise to rely entirely on the retention time of standards. Hence ESMS, and its pneumatically assisted version ISMS, has found a niche as a molecular MS technique for identification of chromatographic peaks quantified by ICP-MS. Uden and co-workers found that, although this worked well for pure compounds and some selenium-enriched natural samples, low Se concentrations and spectral interferences prevented species identification for real samples. Casiot et al.295 overcame the problems of poor sensitivity and peak suppression by decoupling the HPLC and MS and lyophilizing the eluted peaks prior to MS analysis, which was further improved through the use of salt-free mobile phases. The identities of selenium species were confirmed using the fragmentation pattern of the sulfur analogues of the selenium compounds.
McSheehy and Szpunar296 found that the potential of ISMS for identification of As species eluted by HPLC was hampered by the mono-isotopic nature of the element and the resultant difficulty in getting unambiguous attribution of a spectrum peak. They concluded that tandem MS was essential for the confirmation of arsenosugars present in extracts of algae. Schramel et al.297 found that, although high separation efficiency and low flow rates made CE well suited for coupling to ESMS, the geometrical dimensions of both systems required the use of a long CE capillary, up to 100 cm, and resulted in long analysis times. The application of pressure along the capillary during or after separation, which shortened the analysis time significantly, was investigated in terms of LOD, peak shape and resolution. The LOD for the simultaneous determination of six commonly encountered arsenic species were in the range 60–480 µg l−1 for all species except AsIII, for which it was 50 mg l–1. Shimizu et al.298 found that the post-column addition of methanol to the LC eluate improved the analysis of arsenic species both qualitatively and quantitatively. The repeatability of analysis improved from 3–5% to 0.5–2% when post-column addition was used at an optimized flow rate.
10.2 Gas chromatography-mass spectrometry (GC-MS)
Brede and colleagues have continued to develop the use of microplasma MS as detector for capillary GC. A key feature of this method is the very low plasma gas flow which allows the ion source to be located directly inside the high vacuum area of the mass spectrometer. A capacitively coupled rf He plasma, used for selective detection of organotin compounds, was sustained by a flow rate of only 1–3 ml min−1 of He.299 An additional flow of H2 at 0.15–1.5 ml min–1 prevented carbon deposition and minimized interactions in the ion source between Sn and fused silica. Both C and Sn were detected as positively charged atomic ions, which were expelled from the ion source and focused using electrostatic lenses. The LOD was 3.5 pg s−1 with an RSD (n = 4) of 6.3%. Detection of negative ions has been investigated as a means of avoiding problems associated with the presence of isobaric interferences, in particular H3O+ on F+, in the measurement of halogens using positive ions.300 The much cleaner background spectrum in the negative ion mode, compared to that in the positive ion mode, allowed F to be detected with both high sensitivity and selectivity to hydrocarbon compounds. In addition, the LODs for Br, Cl and I were improved ten-fold in the negative ion mode.An alternative plasma ion source for GC-MS, a capacitively coupled low-power atmospheric-pressure rf He plasma, was used by Guevremont and Sturgeon301 as a detector for organometallic speciation. Volatile analytes were admitted to the source through a coaxial rf-powered centre electrode. A key feature of the technique was that variation of the forward power and repositioning of the plasma relative to the differentially pumped extraction orifice of the mass spectrometer gave control over the relative intensities of parent and daughter ion products of the decomposition of the organometallic compounds. It was possible to produce any relative amount of atomic and molecular species in the mass spectrum. Ferrocene, methylmercury chloride, ethylmercury chloride and triethyllead chloride were used to illustrate the application of the technique.
The use of ion-trap instruments has found application in speciation and ID studies. Operation of an instrument in the tandem-MS mode for the determination of HgII and alkylated Hg, Pb and Sn species in human urine (Dunemann et al.302), gave S/N at least one order of magnitude better than that obtained with an instrument operated in the MS mode. Sample preparation was based on derivatization with tetraethylborate followed by headspace solid phase microextraction (SPME). The LOD were 7–22 ng l−1 for all the species investigated. A major advantage of the technique over, for example, GC coupled with ICP–MS, was the ability to identify species directly from the precursor and daughter ions. Barshick et al.303 demonstrated that the transient nature of the GC profile was not incompatible with the pulsed nature of the ion trap MS and that precision was in the range 0.2–2.5% (RSD) for the ID determination of inorganic mercury compounds. Inorganic salts were converted to methylmercury iodide which was sampled with a SPME fibre. Accuracy using the ID method at the 400 ng g−1 level was improved by factors of 30 and 14 over the calibration curve and standard addition methods, respectively.
The current trend towards speciation analysis has given GC-MS a new lease of life with a number of groups reporting the use of GC-MS to determine metal and metalloid speciation. Gomez-Ariza et al.304 used an anion exchanger (to retain inorganic species) in line with a hydrophobic phase (to retain organic species) for the preconcentration of Se compounds in aqueous environmental samples prior to GC-MS analysis. The LOD for six Se compounds were in the range 1.4–900 ng l−1 for 1 l water samples. Mester et al.305 determined dimethylarsinic and monomethylarsonic acids in 1 ml urine samples with LOD of 0.95 and 0.8 ng ml−1, respectively. Iwamura et al.306 converted organotin compounds in waters and sediment extracts to the ethylated compounds prior to GC-MS analysis. The LODs were in the ranges 0.16–3.5 ng l−1 and 0.48–11 µg kg−1 in waters and sediments, respectively. Toniazzo et al.307 developed a new methodolgy, based on extraction with appropriate solvents and GC-MS analysis, to identify and quantify the forms of elemental S present on the surface of pyrite. The comparison of hexane and methanol extractions allowed two types of S8 elemental S to be detected. One fraction was present on its own, whereas the second was associated with hydrophilic superficial compounds such as sulfate.
Investigations of the metabolism of NH4+, an important intermediate of protein metabolism, are difficult in vivo using isotopic techniques because of the low concentration of NH4+ in biological fluids and frequent artifactural dilution of the enriched NH4+. Yang et al.308 have developed a new GC-MS method, based on reaction of NH4+ with formaldehyde to form hexamethylenetetramine (HMT), for monitoring the 15N enrichment and NH4+ concentration in vivo. The 15N enrichment was amplified four times in the reaction as each molecule of HMT contained four atoms of N derived from NH4+.
10.3 Noble gas mass spectrometry
Rau and Putzka309 investigated the implantation and persistent thermal desorption effects of He and Ne in static quadrupole mass spectrometers paying particular regard to the impairment of accuracy of subsequent analyses. Measured desorption time constants were 90–477 min for temperatures between 330 and 470 K. Recommendations for reducing the observed problems included extended pumping, extrapolation of repeated readings back to the time of sample entry and enlargement of the spectrometer volume.10.4 Spark source mass spectrometry (SSMS)
The multi-ion counting (MIC) SSMS technique of Jochum and co–workers has continued to impress by achieving analytical figures of merit that would not generally be considered possible with SSMS. Stoll and Jochum310 analysed a set of eight natural glasses, prepared by direct fusion of rock chips, with the intention of creating geological RMs suitable for microanalytical purposes. A key feature of the paper was the determination of the overall analytical uncertainty (<2–7% depending on the element and its concentration) through consideration of 14 sources of error (for example, sampling, weighing, homogeneity, contamination and matrix effects). Results obtained by MIC-SSMS for 18 elements in the eight glasses were all within 10% of reference values obtained by other techniques with the exception of results for a highly depleted sample. Pfander et al.311 attained LOD of <0.005 µg g−1 in the determination of Nb, Y and Zr in rock samples, with precisions of 2–5% at concentrations ≥0.02 µg g−1. At lower concentrations, the precisions were 10%. A limitation on accuracy was the need to make corrections for the presence of complex isobaric interferences, in particular in Al-rich (>15% Al2O3) samples.Saprykin and colleagues312 have presented further details of gliding SSMS in which an rf spark discharge develops in vacuum across the surface of a dielectric. The special configuration of electrodes strengthened the electric field over the surface of the non-conducting sample and created optimum conditions for the sputtering and ionization of the sample material. Up to 50 trace elements were determined in the NIST glass SRMs 610 and 611 with LOD in the range 10–3–1 µg g−1 and reproducibilities in the range 10–25%. In comparison with conventional SSMS, the RSFs covered a wide range (0.4–26) and the spectrum consisted mainly of multiply charged ions with significantly lower cluster ion intensities. It was possible to get RSFs closer to unity by choosing an internal standard element with physical properties similar to those of the determinand rather than use a matrix element. Such an approach would be highly impractical in a multi-element analysis.
The analysis of pure materials continues to provide a role for SSMS. Wiedermann and co-workers used SSMS as a reference method to provide C concentrations (LOD of 1.4 × 1013 cm–3)313 and B and N concentrations (LOD for both elements of 4.4 × 1013 cm−3)314 in crystalline gallium arsenide. The combined use of three techniques (SSMS, GD-OES, and XPS) to analyse thin Ir–Si films allowed Kurt et al.315 to determine all elements in the periodic table at concentrations in the ppm range.
10.5 New methodologies
Atmospheric pressure chemical ionization (APCI) has been investigated as an interface for LC-MS in speciation and other studies. In preliminary communications, Rosenburg and Grasserbauer316,317 reported that it was possible, under optimized chromatographic conditions, to separate and detect organotin compounds down to the 100 pg (absolute) level. Reversed phase mode was necessary as the APCI interface could not tolerate the introduction of high concentrations of mobile phase additives or large salt loads. Although the method was attractive for the determination of trisubstituted organotin compounds, there were limitations in the determination of mono- and di-substituted compounds which resulted from the difficulties of chromatographic separation and significantly reduced sensitivity as compared with the trisubstituted compounds. Thomas318 determined the biocide zinc pyrithione in water samples by forming the copper chelate prior to HPLC-APCI-MS analysis. Selected ion monitoring of peaks at m/z 221 and 316 gave a LOD for zinc pyrithione of 20 ng l−1 and a linear calibration graph from 2.5 to 125 ng injected.Flame ionization has been revisited by Turk and colleagues,319 who replaced the ICP in a commercial ICP-MS instrument with an air–H2 flame. The mass spectral background was simple but the low temperature of the flame resulted in incomplete ionization. The technique was limited by spectral interferences caused by easily ionized elements and the formation of determinand–solvent ion clusters. The presence of polyatomic ions did, however, provide a wider choice of masses. Measurement of Ca isotope ratios, for example, could be made using CaOH+, which was ten times more sensitive than Ca+. The results suggested that flame ionization MS would have limited application yet might be suitable for the measurement of Ca and K isotope ratios in simple matrices with precisions of 0.5 and 0.2%, respectively.
References
- J. R. Bacon, J. S. Crain, L. Van Vaeck and J. G. Williams, J. Anal. At. Spectrom., 1999, 14(10), 1633 RSC.
- B. Fairman, M. W. Hinds, S. M. Nelms, D. M. Penny and P. Goodall, J. Anal. At. Spectrom., 1999, 14(12), 1937 RSC.
- M. R. Cave, O. Butler, J. M. Cook, M. S Cresser, L. M. Garden, A. J. Holden and D. L. Miles, J. Anal. At. Spectrom., 2000, 15(2), 181 RSC.
- A. Taylor, S. Branch, D. J. Halls, L. M. W. Owen and M. White, J. Anal. At. Spectrom., 2000, 15(4), 451 RSC.
- L. K. Fifield, Rep. Prog. Phys., 1999, 62(8), 1223 CrossRef CAS.
- J. C. Rucklidge, G. C. Wilson, A. E. Litherland, W. E. Kieser, J. A. Krestow and I. Tomski, AIP Conf. Proc., 1999, 475 (Pt. 2, Application of Accelerators in Research and Industry), 629. Search PubMed.
- R. Hellborg, B. Erlandsson, M. Kiisk, P. Persson, G. Skog, K. Stenstrom, S. Mattsson, S. Leide-Svegborn and M. Olofsson, AIP Conf. Proc., 1999, 475 (Pt. 2, Application of Accelerators in Research and Industry), 636. Search PubMed.
- R. L. Loger, J. E. Raatz, M. L. Sundquist, R. C. Garner and J. Barker, AIP Conf. Proc., 1999, 475 (Pt. 2, Application of Accelerators in Research and Industry), 640. Search PubMed.
- M. Bellino, K. F. Von Reden, R. J. Schneider, J. C. Peden, J. Donoghue, K. L. Elder, A. R. Gagnon, P. Long, A. P. McNichol, C. Odegaard, D. Stuart, S. Handwork and J. M. Hayes, AIP Conf. Proc., 1999, 475 (Pt. 2, Application of Accelerators in Research and Industry), 644. Search PubMed.
- M. L. Sundquist, R. L. Loger, R. E. Daniel, R. L. Kitchen, J. J. Kolonko and G. M. Klody, AIP Conf. Proc., 1999, 475 (Pt. 2, Application of Accelerators in Research and Industry), 661. Search PubMed.
- S. H. Sie, D. A. Sims, F. Bruhn, G. F. Suter and T. R. Niklaus, AIP Conf. Proc., 1999, 475 (Pt. 2, Application of Accelerators in Research and Industry), 648. Search PubMed.
- R. J. Schneider, K. F. Von Reden, J. M. Hayes and J. S. C. Wills, AIP Conf. Proc., 1999, 473 (Heavy Ion Accelerator Technology), 422. Search PubMed.
- Y. Liu, D. Li, M. Chen and X. Lu, Nucl. Sci. Tech., 1999, 10(2), 116 Search PubMed.
- A. E. Litherland, K. H. Purser and H. E. Gove, AIP Conf. Proc., 1999, 473 (Heavy Ion Accelerator Technology), 426. Search PubMed.
- H. Matsuzaki, K. Kobayashi, S. Hatori and K. Tanigawa, Tokyo Daigaku Genshiryoku Kenkyu Sogo Senta Shinpojumu, 1998, 7th, 173 Search PubMed.
- J. Barker and R. C. Garner, Rapid Commun. Mass Spectrom., 1999, 13(4), 285 CrossRef CAS.
- A. J. Clifford, A. Arjomand, S. R. Dueker, P. D. Schneider, B. A. Buchholz and J. S. Vogel, Adv. Exp. Med. Biol., 1998, 445 (Mathematical Modeling in Experimental Nutrition), 239. Search PubMed.
- B. A. Buchholz, A. Arjomand, S. R. Dueker, P. D. Schneider, A. J. Clifford and J. S. Vogel, Anal. Biochem., 1999, 269(2), 348 CrossRef CAS.
- Y. Haruhara, K. Kobayashi, K. Yoshida, S. Hatori and C. Nakano, Tokyo Daigaku Genshiryoku Kenkyu Sogo Senta Shinpojumu, 1998, 7th, 169 Search PubMed.
- T. Aramaki, JAERI-Res., 1999(99-007), 1. Search PubMed.
- P. Collon, D. Cole, B. Davids, M. Fauerbach, R. Harkewicz, W. Kutschera, D. J. Morrissey, R. C. Pardo, M. Paul, B. M. Sherrill and M. Steiner, Radiochim. Acta, 1999, 85(1–2), 13 Search PubMed.
- J. M. Lopez-Gutierrez, M. Garcia-Leon, C. Schnabel, A. Schmidt, R. Michel, H.-A. Synal and M. Suter, Appl. Radiat. Isot., 1999, 51(3), 315 Search PubMed.
- R. D. Brauer, J. D. Robertson, P. Sharma and R. A. Yokel, Nucl. Instrum. Methods Phys. Res., Sect. B, 1999, 152(1), 129 Search PubMed.
- S. Merchel and U. Herpers, Radiochim. Acta, 1999, 84(4), 215 Search PubMed.
- A. Bogaerts and R. Gijbels, Fresenius' J. Anal. Chem., 1999, 364(5), 367 CrossRef CAS.
- A. Bogaerts, J. Anal. At. Spectrom., 1999, 14(9), 1375 RSC.
- A. I. Saprykin, Ind. Lab. (Diagn. Mater.), 1999, 65(2), 83 Search PubMed.
- R. K. Marcus, E. H. Evans and J. A. Caruso, J. Anal. At. Spectrom., 2000, 15(1), 1 RSC.
- J. P. Guzowski Jr., J. A. C. Broekaert, S. J. Ray and G. M. Hieftje, J. Anal. At. Spectrom., 1999, 14(8), 1121 RSC.
- A. I. Saprykin, J. Anal. Chem. (Transl. of Zh. Anal. Khim.), 1999, 54(7), 671 Search PubMed.
- W. Hang, X. Yan, D. M. Wayne, J. A. Olivares, W. W. Harrison and V. Majidi, J. Anal. At. Spectrom., 1999, 14(9), 1523 RSC.
- M. Inoue and T. Saka, Anal. Chim. Acta, 1999, 395(1–2), 165 CrossRef CAS.
- D. M. Wayne, R. K. Schulze, C. Maggiore, D. W. Cooke and G. Havrilla, Appl. Spectrosc., 1999, 53(3), 266 Search PubMed.
- C. L. Lewis, E. S. Oxley, C. K. Pan, R. E. Steiner and F. L. King, Anal. Chem., 1998, 71(1), 230 CrossRef CAS.
- F. Chartier, M. Aubert, M. Salmon, M. Tabarant and Bich Hang Tran, J. Anal. At. Spectrom., 1999, 14(9), 1461 RSC.
- A. Montaser, J. A. McLean, H. Liu and J.-M. Mermet, A. Montaser (Ed.), An introduction to ICP spectrometries for elemental analysis. Inductively Coupled Plasma Mass Spectrometry, Wiley-VCH, New York, NY, USA, 1998, 1–31. Search PubMed.
- I. L. Turner and A. Montaser, A. Montaser (Ed.), Plasma generation in ICPMS. Inductively Coupled Plasma Mass Spectrometry, Wiley-VCH, New York, NY, USA, 1998, 265–334. Search PubMed.
- P. J. Turner, D. J. Mills, E. Schroder, G. Lapitajs, G. Jung, L. A. Iacone and D. A. Haydar, A. Montaser (Ed.), Instrumentation for low- and high-resolution ICPMS. Inductively Coupled Plasma Mass Spectrometry, Wiley-VCH, New York, NY, USA, 1998, 421–501. Search PubMed.
- G. Horlick and A. Montaser, A. Montaser (Ed.), Analytical characteristics of ICPMS. Inductively Coupled Plasma Mass Spectrometry, Wiley-VCH, New York, NY, USA, 1998, 503–613. Search PubMed.
- D. J. Douglas and S. D. Tanner, A. Montaser (Ed.), Fundamental considerations in ICPMS. Inductively Coupled Plasma Mass Spectrometry, Wiley-VCH, New York, NY, USA, 1998, 615–679. Search PubMed.
- H. E. Taylor, R. A. Huff and A. Montaser, A. Montaser (Ed.), Novel applications of ICPMS. Inductively Coupled Plasma Mass Spectrometry, Wiley-VCH, New York, NY, USA, 1998, 681–807. Search PubMed.
- A. Montaser and H. Zhang, A. Montaser (Ed.), Mass spectrometry with mixed-gas and helium ICPS. Inductively Coupled Plasma Mass Spectrometry, Wiley-VCH, New York, NY, USA, 1998, 809–890. Search PubMed.
- H. Zhang, A. Montaser and B. S. Zimmer, A. Montaser (Ed.), Mass spectrometry with microwave-induced plasmas. Inductively Coupled Plasma Mass Spectrometry, Wiley-VCH, New York, NY, USA, 1998, 891–939. Search PubMed.
- G. O'Connor, E. H. Evans and S. J. Hill, Fundamental aspects of ICP-MS. Inductively Coupled Plasma Mass Spectrometry. Its Application, Sheffield Academic Press, Sheffield, UK, 1999, 119–144. Search PubMed.
- F. Vanhaecke, L. Moens, P. Taylor and S. J. Hill, Use of ICP-MS for isotope ratio measurements. Inductively Coupled Plasma Spectrometry. Its Application, Sheffield Academic Press, Sheffield, UK, 1999, 145–207. Search PubMed.
- S. J. Hill, A. Fisher, G. O'Connor and E. H. Evans, (Ed.), Alternative plasmas and sample introduction systems. Inductively Coupled Plasma Spectrometry. Its Application, Sheffield Academic Press, Sheffield, UK, 1999, 208–244. Search PubMed.
- K. L. Sutton, J. A. Caruso and S. J. Hill, Environmental applications of plasma spectrometry. Inductively Coupled Plasma Spectrometry. Its Application, Sheffield Academic Press, Sheffield, UK, 1999, 245–272. Search PubMed.
- F. Vanhaecke and L. Moens, Fresenius' J. Anal. Chem., 1999, 364(5), 440 CrossRef CAS.
- J. W. Olesik, I. I. Stewart, J. A. Hartshome and C. E. Hensman, Spec. Publ. - R. Soc. Chem., 1999, 241 (Plasma Source Mass Spectrometry), 3. Search PubMed.
- J.-L. Todoli and J.-M. Mermet, Spectrochim. Acta, Part B, 1999, 54, 895 CrossRef.
- R. N. Sah and P. H. Brown, Biol. Trace Elem. Res., 1998, 66(1–3), 39 Search PubMed.
- C. B. Douthitt, ICP Inf. Newsl., 1999, 25(2), 90 Search PubMed.
- J. E. Patterson, B. S. Duersch and P. B. Farnsworth, Spectrochim. Acta, Part B, 1999, 54(3–4), 537 CrossRef.
- B. S. Duersch and P. B. Farnsworth, Spectrochim. Acta, Part B, 1999, 54(3–4), 545 CrossRef.
- B. S. Duersch, J. E. Patterson and P. B. Farnsworth, J. Anal. At. Spectrom., 1999, 14(4), 615 RSC.
- A. Pupyshev, A. Lutsak, E. Krakovska and S. Ruzickova, Formation of ions and their transport for ICP-MS: a theoretical description and an experimental confirmation. Proc. - Semin. At. Spectrochem., 14th, Stroffek Publishing, Kosice, Slovakia, 1998, 330–333. Search PubMed.
- A. A. Pupyshev, V. N. Muzgin and A. K. Lutsak, J. Anal. At. Spectrom., 1999, 14(9), 1485 RSC.
- I. I. Stewart and J. W. Olesik, J. Am. Soc. Mass Spectrom., 1999, 10(2), 159 CrossRef CAS.
- C. Sartoros and E. D. Salin, Spectrochim. Acta, Part B, 1999, 54(11), 1557 CrossRef.
- A. Held and P. D. P. Taylor, J. Anal. At. Spectrom., 1999, 14(7), 1075 RSC.
- V. J. Barwick, S. L. R. Ellison and B. Fairman, Anal. Chim. Acta, 1999, 394(2–3), 281 CrossRef CAS.
- A. Al-Ammar, R. K. Gupta and R. M. Barnes, Spectrochim. Acta, Part B, 1999, 54B(7), 1077 CrossRef.
- V. I. Baranov and S. D. Tanner, J. Anal. At. Spectrom., 1999, 14(8), 1133 RSC.
- S. D. Tanner and V. I. Baranov, At. Spectrosc., 1999, 20(2), 45 Search PubMed.
- U. Vollkopf, V. Baranov and S. Tanner, Spec. Publ. - R. Soc. Chem., 1999, 241 (Plasma Source Mass Spectrometry), 63. Search PubMed.
- S. D. Tanner and V. I. Baranov, J. Am. Soc. Mass Spectrom., 1999, 10(11), 1083 CrossRef.
- S. D. Tanner and V. I. Baranov, Spec. Publ. - R. Soc. Chem., 1999, 241 (Plasma Source Mass Spectrometry), 46. Search PubMed.
- K. Neubauer and U. Vollkopf, At. Spectrosc., 1999, 20(2), 64 Search PubMed.
- U. Vollkopf, K. Klemm and M. Pfluger, At. Spectrosc., 1999, 20(2), 53 Search PubMed.
- D. R. Bandura and S. D. Tanner, At. Spectrosc., 1999, 20(2), 69 Search PubMed.
- I. Feldmann, N. Jakubowski and D. Stuewer, Fresenius' J. Anal. Chem., 1999, 365(5), 415 CrossRef CAS.
- I. Feldmann, N. Jakubowski, C. Thomas and D. Stuewer, Fresenius' J. Anal. Chem., 1999, 365(5), 422 CrossRef CAS.
- F. A. Shakra, Spec. Publ. - R. Soc. Chem., 1999, 241 (Plasma Source Mass Spectrometry), 120. Search PubMed.
- G. C. Eiden, C. J. Barinaga and D. W. Koppenaal, J. Anal. At. Spectrom., 1999, 14(8), 1129 RSC.
- Z. Du, D. J. Douglas and N. Konenkov, J. Anal. At. Spectrom., 1999, 14(8), 1111 RSC.
- X. Tian, H. Emteborg and F. C. Adams, J. Anal. At. Spectrom., 1999, 14(12), 1807 RSC.
- J. S. Becker and H.-J. Dietze, J. Anal. At. Spectrom., 1999, 14(9), 1493 RSC.
- J. W. Waggoner, L. S. Milstein, M. Belkin, K. L. Sutton, J. A. Caruso and H. B. Fannin, J. Anal. At. Spectrom., 2000, 15(1), 13 RSC.
- S. F. Durrant, J. Anal. At. Spectrom., 1999, 14(9), 1385 RSC.
- D. Guenther, S. E. Jackson and H. P. Longerich, Spectrochim. Acta, Part B, 1999, 54(3–4), 381 CrossRef.
- B. Wanner, C. Moor, P. Richner, R. Broennimann and B. Magyar, Spectrochim. Acta, Part B, 1999, 54, 289 CrossRef.
- R. Wennrich, P. Morgenstern, G. Liebergeld and G. Werner, Chem. Anal. (Warsaw), 1999, 44(3B), 523 Search PubMed.
- D. Gunther and C. A. Heinrich, J. Anal. At. Spectrom., 1999, 14(9), 1369 RSC.
- D. Gunther and C. A. Heinrich, J. Anal. At. Spectrom., 1999, 14(9), 1363 RSC.
- A. J. G. Mank and P. R. D. Mason, J. Anal. At. Spectrom., 1999, 14(8), 1143 RSC.
- Z. Chen, J. Anal. At. Spectrom., 1999, 14(12), 1823 RSC.
- O. V. Borisov, X. L. Mao, A. Fernandez, M. Caetano and R. E. Russo, Spectrochim. Acta, Part B, 1999, 54, 1351 CrossRef.
- F. Boue-Bigne, B. J. Masters, J. S. Crighton and B. L. Sharp, J. Anal. At. Spectrom., 1999, 14(11), 1665 RSC.
- C.-J. Chin, C.-F. Wang and S.-L. Jeng, J. Anal. At. Spectrom., 1999, 14(4), 663 RSC.
- C.-F. Wang, C.-J. Chin, S.-K. Luo and L.-C. Men, Anal. Chim. Acta, 1999, 389(1–3), 257 CrossRef CAS.
- C. A. Booth, D. A. Spears, P. Krause and A. G. Cox, Fuel, 1999, 78(14), 1665 CrossRef CAS.
- A. J. Campbell and M. Humayun, Anal. Chem., 1999, 71(5), 939 CrossRef CAS.
- T. J. Shepherd, C. Ayora, D. I. Cendon, S. R. Chenery and A. Moissette, Eur. J. Mineral., 1998, 10(6), 1097 CAS.
- A. M. Ghazi and S. Shuttleworth, Analyst (Cambridge, U.K.), 2000, 125(1), 205 Search PubMed.
- I. B. Butler and R. W. Nesbitt, Earth Planet. Sci. Lett., 1999, 167(3–4), 335 CrossRef CAS.
- C. H. Stirling, J. N. Christensen and A. N. Halliday, Mineral. Mag., 1998, 62A(Pt. 3), 1462 Search PubMed.
- T. Hirata and T. Yamaguchi, J. Anal. At. Spectrom., 1999, 14(9), 1455 RSC.
- W. Devos, C. Moor and P. Lienemann, J. Anal. At. Spectrom., 1999, 14(4), 621 RSC.
- K. M. Lee, J. Appleton, M. Cooke, F. Keenan and K. Sawicka-Kapusta, Anal. Chim. Acta, 1999, 395(1–2), 179 CrossRef CAS.
- F. Lochner, J. Appleton, F. Keenan and M. Cooke, Anal. Chim. Acta, 1999, 401(1–2), 299 CrossRef CAS.
- D. C. Gregoire and R. E. Sturgeon, Spectrochim. Acta, Part B, 1999, 54(5), 773 CrossRef.
- Y. Okamoto, J. Anal. At. Spectrom., 1999, 14(10), 1631 RSC.
- J. W. H. Lam, R. E. Sturgeon and J. W. McLaren, Spectrochim. Acta, Part B, 1999, 54(3–4), 443 CrossRef.
- H. T. Uggerud and W. Lund, Spectrochim. Acta, Part B, 1999, 54(11), 1625 CrossRef.
- J. W. Lam and R. E. Sturgeon, At. Spectrosc., 1999, 20(3), 79 Search PubMed.
- M. E. Ben Younes, D. C. Gregoire and C. L. Chakrabarti, J. Anal. At. Spectrom., 1999, 14(11), 1703 RSC.
- P. P. Mahoney, S. J. Ray, G. Li and G. M. Hieftje, Anal. Chem., 1999, 71(7), 1378 CrossRef CAS.
- D. Langer and J. A. Holcombe, Appl. Spectrosc., 1999, 53(10), 1244 Search PubMed.
- C. J. Park and S. A. Yim, J. Anal. At. Spectrom., 1999, 14(7), 1061 RSC.
- C. Moor, J. W. H. Lam and R. E. Sturgeon, J. Anal. At. Spectrom., 2000, 15(2), 143 RSC.
- H.-B. Hou and H. Narasaki, At. Spectrosc., 1999, 20(1), 20 Search PubMed.
- H.-B. Hou and H. Narasaki, Anal. Sci., 1999, 15(9), 911 Search PubMed.
- E. Debrah and G. Legere, At. Spectrosc., 1999, 20(2), 73 Search PubMed.
- E. Debrah and G. Legere, Spec. Publ. - R. Soc. Chem., 1999, 241 (Plasma Source Mass Spectrometry), 20. Search PubMed.
- A. S. Al-Ammar, R. K. Gupta and R. M. Barnes, J. Anal. At. Spectrom., 1999, 14(5), 793 RSC.
- J. Sabine Becker, R. S. Soman, K. L. Sutton, J. A. Caruso and H.-J. Dietze, J. Anal. At. Spectrom., 1999, 14(6), 933 RSC.
- V. D. Jones, ASTM Spec. Tech. Publ., 1998, 1344 (Applications of Inductively Coupled Plasma-Mass Spectrometry to Radionuclide Determinations: Second Volume), 41. Search PubMed.
- X. P. Yan, R. Kerrich and M. J. Hendry, Anal. Chem., 1998, 70(22), 4736 CrossRef CAS.
- B. Wen, X.-q. Shan and S.-g. Xu, Analyst (Cambridge, U.K.), 1999, 124(4), 621 Search PubMed.
- S. D. Lofthouse, G. M. Greenway and S. C. Stephen, J. Anal. At. Spectrom., 1999, 14(12), 1839 RSC.
- A. G. Coedo, I. Padilla, T. Dorado and F. J. Alguacil, Anal. Chim. Acta, 1999, 389(1–3), 247 CrossRef CAS.
- M.-T. Wei and S.-J. Jiang, J. Anal. At. Spectrom., 1999, 14(8), 1177 RSC.
- J. F. Ranville, D. J. Chittleborough, F. Shanks, R. J. S. Morrison, T. Harris, F. Doss and R. Beckett, Anal. Chim. Acta, 1999, 381(2–3), 315 CrossRef CAS.
- D. Schaumloffel and A. Prange, Fresenius' J. Anal. Chem., 1999, 364(5), 452 CrossRef CAS.
- S. A. Baker and N. J. Miller-Ihli, Appl. Spectrosc., 1999, 53(4), 471 Search PubMed.
- X.-D. Tian, Z.-X. Zhuang, B. Chen and X.-R. Wang, At. Spectrosc., 1999, 20(4), 127 Search PubMed.
- M. Montes Bayon, M. Gutierrez Camblor, J. I. Garcia Alonso and A. Sanz-Medel, J. Anal. At. Spectrom., 1999, 14(9), 1317 RSC.
- S. Slaets, F. Adams, I. Rodriguez Pereiro and R. Lobinski, J. Anal. At. Spectrom., 1999, 14(5), 851 RSC.
- A. M. Leach, M. Heisterkamp, F. C. Adams and G. M. Hieftje, J. Anal. At. Spectrom., 2000, 15(2), 151 RSC.
- A. M. Zapata, C. L. Bock and A. Robbat Jr., J. Anal. At. Spectrom., 1999, 14(8), 1187 RSC.
- A. Chatterjee, Y. Shibata, J. Yoshinaga and M. Morita, J. Anal. At. Spectrom., 1999, 14(12), 1853 RSC.
- J. M. Gonzalez LaFuente, J. M. Marchante-Gayo, M. L. Fernandez Sanchez and A. Sanz Medel, Talanta, 1999, 50(1), 207 CrossRef CAS.
- T. Dagnac, A. Padro, R. Rubio and G. Rauret, Talanta, 1999, 48(4), 763 CrossRef CAS.
- H. Goenaga Infante, M. L. Fernandez Sanchez and A. Sanz-Medel, J. Anal. At. Spectrom., 1999, 14(9), 1343 RSC.
- V. M. S. Ramanujam, K. Yokoi, N. G. Egger, H. H. Dayal, N. W. Alcock and H. H. Sandstead, Biol. Trace Elem. Res., 1999, 68(2), 143 Search PubMed.
- M. Grotti, C. Gnecco and F. Bonfiglioli, J. Anal. At. Spectrom., 1999, 14(8), 1171 RSC.
- M. J. Campbell and A. Torvenyi, J. Anal. At. Spectrom., 1999, 14(9), 1313 RSC.
- B. Gammelgaard and O. Joens, J. Anal. At. Spectrom., 1999, 14(5), 867 RSC.
- P. B. Tomascak and R. T. Helz, Mineral. Mag., 1998, 62A/Pt., , 3, 1523 Search PubMed.
- P. B. Tomascak, R. W. Carlson and S. B. Shirey, Chem. Geol., 1999, 158(1–2), 145 CrossRef CAS.
- P. B. Tomascak, F. Tera, R. T. Helz and R. J. Walker, Geochim. Cosmochim. Acta, 1999, 63(6), 907 CrossRef CAS.
- C. N. Marechal, P. Telouk and F. Albarede, Chem. Geol., 1999, 156(1–4), 251 CrossRef CAS.
- M. Hamester, D. Wiederin, J. Wills, W. Kerl and C. B. Douthitt, Fresenius' J. Anal. Chem., 1999, 364(5), 495 CrossRef CAS.
- T. Prohaska, C. Latkoczy and G. Stingeder, J. Anal. At. Spectrom., 1999, 14(9), 1501 RSC.
- A. V. Pakhomov, A. J. Roybal and M. S. Duran, Appl. Spectrosc., 1999, 53(8), 979 Search PubMed.
- S. Amoruso, V. Berardi, R. Bruzzese, R. Velotta, N. Spinelli and X. Wang, Appl. Surf. Sci., 1999, 138, 250 CrossRef.
- K. Song, K.-H. Hong, M. Yang, H. Cha, J. Lee and G.-H. Lee, J. Korean Phys. Soc., 1999, 35(3), 217 Search PubMed.
- D. M. Woll, M. Wahl and H. Oechsner, Fresenius' J. Anal. Chem., 1999, 365(1–3), 70 CrossRef CAS.
- D. T. Suess and K. A. Prather, Chem. Rev. (Washington, D. C.), 1999, 99(10), 3007 Search PubMed.
- K.-P. Hinz, M. Greweling, F. Drews and B. Spengler, J. Am. Soc. Mass Spectrom., 1999, 10(7), 648 CrossRef CAS.
- K. Wendt, K. Blaum, B. A. Bushaw, C. Gruning, R. Horn, G. Huber, J. V. Kratz, P. Kunz, P. Muller, W. Nortershauser, M. Nunnemann, G. Passler, A. Schmitt, N. Trautmann and A. Waldek, Fresenius' J. Anal. Chem., 1999, 364(5), 471 CrossRef CAS.
- B. A. Bushaw, W. Noertershaeuser and K. L. Wendt, Spectrochim. Acta, Part B, 1999, 54(2), 321 CrossRef.
- K. Wendt, K. Blaum, P. Muller, W. Nortershauser, A. Schmitt, N. Trautmann and B. A. Bushaw, J. Korean Phys. Soc., 1999, 35(3), 143 Search PubMed.
- P. Bisling, J. Dederichs, B. Neidhart and C. Weitkamp, Fresenius' J. Anal. Chem., 1999, 364(1–2), 79..
- M. Yorozu, Y. Okada and A. Endo, Appl. Phys. B: Lasers Opt., 1999, 69(2), 129 CrossRef CAS.
- T. J. Whitaker, K. F. Willey, W. R. Garrett and H. F. Arlinghaus, Adv. Sci. Technol. (Faenza, Italy), 1999, 19 (Surface and Near-Surface Analysis of Materials), 97. Search PubMed.
- M. Schuhmacher, B. Rasser, E. De Chambost, F. Hillion, T. Mootz and H.-N. Migeon, Fresenius' J. Anal. Chem., 1999, 365(1–3), 12 CrossRef CAS.
- A. J. Eccles, T. A. Steele and A. W. Robinson, Appl. Surf. Sci., 1999, 144, 106 CrossRef.
- G. Gillen, R. L. King and F. Chmara, J. Vac. Sci. Technol., A, 1999, 17(3), 845 CrossRef CAS.
- J. Herec, J. Filiks, M. Sowa, J. Sielanko and D. Maczka, Nukleonika, 1999, 44(2), 119 Search PubMed.
- S. Daolio, B. Facchin, C. Pagura, A. Tolstogouzov and N. Konenkov, Rapid Commun. Mass Spectrom., 1999, 13(9), 782 CrossRef CAS.
- V. S. Smehtkowski, A. R. Krauss, D. M. Gruen, J. C. Holecek and J. A. Schultz, J. Vac. Sci. Technol., A, 1999, 17(5), 2634 CrossRef.
- P. A. W. Van der Heide, Nucl. Instrum. Methods Phys. Res., Sect. B, 1999, 157(1–4), 126 Search PubMed.
- V. P. Ivanov, S. N. Trukhan and A. I. Borodin, Int. J. Mass Spectrom., 1999, 188(3), 183 CrossRef CAS.
- F. Okuyama, M. Kaneko, S. Senda, Y. Katada and M. Tanemura, Int. J. Mass Spectrom., 1999, 189(2–3), 173 CrossRef CAS.
- T. Sato, S. Takagi and T. Gotoh, Surf. Sci., 1999, 433, 711 CrossRef.
- J. Herec, J. Filiks and J. Sielanko, Nukleonika, 1999, 44(2), 103 Search PubMed.
- P. N. K. Deenapanray and M. Petravic, J. Appl. Phys., 1999, 85/8, Pt., , 1, 3993 Search PubMed.
- Z. X. Jiang and P. F. A. Alkemade, Surf. Interface Anal., 1999, 27(3), 125 CrossRef CAS.
- J. Biersack, Nucl. Instrum. Methods Phys. Res., Sect. B, 1999, 153(1–4), 398 Search PubMed.
- V. V. Makarov, Surf. Interface Anal., 1999, 27(9), 801 CrossRef CAS.
- A. R. Pratt, K. Franzreb and N. S. McIntyre, C. R. Acad. Sci. Ser. II: Sci. Terre Planetes, 1999, 328(5), 309 Search PubMed.
- T. Schenkel, A. V. Hamza, A. V. Barnes, M. W. Newman, G. Machicoane, T. Niedermayer, M. Hattass, J. W. McDonald, D. H. Schneider, K. J. Wu and R. W. Odom, Phys. Scr., T, 1999, T80A (IX International Conference on the Physics of Highly Charged Ions, 1998), 73. Search PubMed.
- H. J. Kang, W. S. Kim, D. W. Moon, H. Y. Lee, S. T. Kang and R. Shimizu, Nucl. Instrum. Methods Phys. Res., Sect. B, 1999, 153(1–4), 429 Search PubMed.
- Y. Higashi, A. Masamoto and Y. Homma, J. Surf. Anal., 1999, 5(1), 136 Search PubMed.
- D. Verkleij and C. Mulders, Micron, 1999, 30(3), 227 CrossRef CAS.
- M. Fried, O. Polgar, T. Lohner, S. Strehlke and C. Levy-Clement, J. Lumin., 1998, 80(1–4), 147 CrossRef CAS.
- P. Konarski, Nukleonika, 1999, 44(2), 143 Search PubMed.
- A. Rar, D. W. Moon and S. Hofmann, J. Surf. Anal., 1999, 6(1), 29 Search PubMed.
- K. Wittmaack, J. J. Lee and S. B. Patel, Appl. Phys. Lett., 1999, 74(26), 3969 CrossRef CAS.
- M. J. Baldwin, S. Kumar, J. M. Priest, M. P. Fewell, K. E. Prince and K. T. Short, Thin Solid Films, 1999, 345(1), 108 CrossRef CAS.
- S. Seki, H. Tamura and H. Sumiya, Appl. Surf. Sci., 1999, 147(1–4), 14 CrossRef CAS.
- J. Farquhar, E. Hauri and J. Wang, Earth Planet. Sci. Lett., 1999, 171(4), 607 CrossRef CAS.
- W. Compston, Mineral. Mag., 1999, 63(3), 297 Search PubMed.
- K. Wittmaack, Surf. Sci., 1999, 429(1–3), 84 CrossRef CAS.
- L. Wang, L. H. Zhang and J. Li, J. Electron. Mater., 1999, 28(6), 793 Search PubMed.
- C. Hongo, M. Tomita and M. Suzuki, Appl. Surf. Sci., 1999, 144, 306 CrossRef.
- M. A. Douglas and P. J. Chen, Surf. Interface Anal., 1998, 26(13), 984 CrossRef CAS.
- C. Aurisicchio, F. Demartin, L. Ottolini and F. Pezzotta, Eur. J. Mineral., 1999, 11(2), 237 CAS.
- P. W. O. Hoskin, Geostand. Newsl., 1999, 23(1), 69 Search PubMed.
- S. Suzuki, K. Yanagihara, S. Hayashi and T. Mori, J. Surf. Anal., 1999, 5(2), 278 Search PubMed.
- G. Verlinden, R. Gijbels and I. Geuens, J. Am. Soc. Mass Spectrom., 1999, 10(10), 1016 CrossRef CAS.
- B. Gautier, G. Prudon and J. C. Dupuy, Surf. Interface Anal., 1998, 26(13), 974 CrossRef CAS.
- D. W. Moon and K. J. Kim, J. Surf. Anal., 1999, 6(1), 34 Search PubMed.
- J. Zhao, M. Na, P. J. McKeown, H. Chang, E. Lee, H. Luo, J. Chen, T. D. Wood and J. A. Gardella Jr., J. Vac. Sci. Technol., A, 1999, 17(1), 224 CrossRef CAS.
- D. Krueger, K. Iltgen and R. Kurps, Diffus. Defect Data, Pt. B, 1998, 63 (Beam Injection Assessment of Defects in Semiconductors), 465. Search PubMed.
- Y. Kawashima, T. Ide and K. Shimizu, J. Surf. Anal., 1999, 5(1), 116 Search PubMed.
- K. Shimizu, B. J. Flinn, P. C. Wong and K. A. R. Mitchell, Can. J. Chem., 1998, 76(12), 1796 CrossRef CAS.
- K. L. Gavrilov, S. J. Bennison, K. R. Mikeska, J. M. Chabala and R. Levi-Setti, J. Am. Ceram. Soc., 1999, 82(4), 1001 Search PubMed.
- G. Gillen, S. Robertson, C. Ng and M. Stranick, Scanning, 1999, 21(3), 173 Search PubMed.
- F. A. Stevie, S. W. Downey, S. R. Brown, T. L. Shofner, M. A. Decker, T. Dingle and L. Christman, J. Vac. Sci. Technol., B, 1999, 17(6), 2476 CrossRef CAS.
- G. Xu, T. E. Mates, D. R. Clarke, Q. Ma and H. Fujimoto, Mater. Res. Soc. Symp. Proc., 1999, 563 (Materials Reliability in Microelectronics IX), 257. Search PubMed.
- J. B. Metson and K. E. Prince, Surf. Interface Anal., 1999, 28(1), 159 CrossRef CAS.
- R. Yue, Y. Wang, Y. Wang and C. Chen, Surf. Interface Anal., 1999, 27(2), 98 CrossRef CAS.
- C. C. Parks, H. Glawischnig, M. Levy, R. Stengl and C. Dieseldorff, J. Vac. Sci. Technol., A, 1999, 17(4), 1130 CrossRef CAS.
- T. Hoshi, L. Zhanping, M. Tozu and R. Oiwa, J. Surf. Anal., 1999, 5(1), 132 Search PubMed.
- J. Kinoshita, K. Shiozawa and T. Yamasaki, J. Surf. Anal., 1999, 5(1), 185 Search PubMed.
- T. J. Ormsby, D. P. Chu, M. G. Dowsett, G. A. Cooke and S. B. Patel, Appl. Surf. Sci., 1999, 144, 292 CrossRef.
- M. Wolkenstein, T. Stubbings and H. Hutter, Fresenius' J. Anal. Chem., 1999, 365(1–3), 63 CrossRef CAS.
- O. Gebhardt, Fresenius' J. Anal. Chem., 1999, 365(1–3), 117 CrossRef CAS.
- L. Van Vaeck, A. Adriaens and R. Gijbels, Mass Spectrom. Rev., 1999, 18(1), 1 CrossRef CAS.
- A. Adriaens, L. Van Vaeck and F. Adams, Mass Spectrom. Rev., 1999, 18(1), 48 CrossRef.
- E. Cuynen, L. Van Vaeck and P. Van Espen, Rapid Commun. Mass Spectrom., 1999, 13(23), 2287 CrossRef CAS.
- I. S. Gilmore and M. P. Seah, Appl. Surf. Sci., 1999, 144, 113 CrossRef.
- P. E. Vickers, J. E. Castle and J. F. Watts, Appl. Surf. Sci., 1999, 150(1–4), 244 CrossRef CAS.
- S. H. Goss, G. L. Fisher, P. B. S. Kodali, B. J. Garrison and N. Winograd, Phys. Rev. B: Condens. Matter Mater. Phys., 1999, 59(16), 10662 Search PubMed.
- S. Hamoudi, F. Larachi, A. Adnot and A. Sayari, J. Catal., 1999, 185(2), 333 CrossRef CAS.
- H. Liang, Chin. Sci. Bull., 1999, 44(13), 1242 Search PubMed.
- R. Avci, A. M. Hagenston, N. L. Equall, G. S. Groenewold, G. L. Gresham and D. A. Dahl, Surf. Interface Anal., 1999, 27(8), 789 CrossRef CAS.
- J. C. Ingram, G. S. Groenewold, J. E. Olson, A. K. Gianotto and M. O. McCurry, Anal. Chem., 1999, 71(9), 1712 CrossRef CAS.
- J. L. Gerlock, T. J. Prater, S. L. Kaberline, J. L. Dupuie, E. J. Blais and D. E. Rardon, Polym. Degrad. Stab., 1999, 65(1), 37 CrossRef CAS.
- H. Shichi, S. Osabe, M. Sugaya, K. Kanchori and Y. Mitsui, J. Surf. Anal., 1999, 5(2), 212 Search PubMed.
- J. Goschnick, C. Natzeck and M. Sommer, Appl. Surf. Sci., 1999, 144, 31 CrossRef.
- J. Goschnick, C. Natzeck and M. Sommer, Surf. Interface Anal., 1999, 18(1), 56 CrossRef.
- C. He and C. H. Becker, J. Appl. Phys., 1999, 86(6), 3463 CrossRef CAS.
- H. Jennett, X. Ai, V.-D. Hodoroaba, I. Iga and L. M. Tao, Nucl. Instrum. Methods Phys. Res., Sect. B, 1999, 155(1–2), 13 Search PubMed.
- J. Goschnick, C. Natzeck and M. Sommer, Appl. Surf. Sci., 1999, 144, 201 CrossRef.
- J. Goschnick, C. Natzeck, M. Sommer and F. Zudock, Thin Solid Films, 1998, 332(1–2), 215 CrossRef CAS.
- Y. Higashi, Spectrochim. Acta, Part B, 1999, 54(1), 109 CrossRef.
- K. F. Willey, H. F. Arlinghaus and T. J. Whitaker, Appl. Surf. Sci., 1999, 144, 36 CrossRef.
- K. F. Willey, H. F. Arlinghaus and T. J. Whitaker, J. Vac. Sci. Technol., A, 1999, 17/4, Pt., , 1, 1127 Search PubMed.
- U. Muller, M. Schittenhelm, R. Schmittgens and H. Helm, Surf. Interface Anal., 1999, 27(10), 904 CrossRef CAS.
- W. Meier-Augenstein, J. Chromatogr., A, 1999, 842(1–2), 351 CrossRef CAS.
- C. C. Metges and K. J. Petzke, A. E. El-Khoury (Ed.), The use of GC-C-IRMS for the analysis of stable isotope enrichment in nitrogenous compounds. Methods Invest. Amino Acid Protein Metab., CRC, Boca Raton, FL, USA, 1999, 121–134. Search PubMed.
- L. Brazier and W. Elbast, Analusis, 1999, 27(3), 218 Search PubMed.
- S. M. Jackson, G. H. Morgan, A. D. Morse, A. L. Butterworth and C. T. Pillinger, Rapid Commun. Mass Spectrom., 1999, 13(13), 1329 CrossRef CAS.
- P. M. Mason, I. Gilmour, C. Pillinger, S. E. Hall, E. Houghton and M. A. Seymour, Analyst (Cambridge, U.K.), 1998, 123(12), 2405 Search PubMed.
- R. A. Werner, B. A. Bruch and W. A. Brand, Rapid Commun. Mass Spectrom., 1999, 13(13), 1237 CrossRef CAS.
- R. Russow and A. Goetz, Arch. Agron. Soil Sci., 1998, 43(5), 349 Search PubMed.
- R. Russow, Rapid Commun. Mass Spectrom., 1999, 13(13), 1334 CrossRef CAS.
- I. Sich and R. Russow, Rapid Commun. Mass Spectrom., 1999, 13(13), 1325 CrossRef CAS.
- J. A. Hall, J. A. C. Barth and R. M. Kalin, Rapid Commun. Mass Spectrom., 1999, 13(13), 1231 CrossRef CAS.
- G. Gleixner and H. L. Schmidt, ACS Symp. Ser., 1998, 707 (Nitrogen-Containing Macromolecules in the Bio- and Geosphere), 34. Search PubMed.
- G. Gleixner, R. Bol and J. Balesdent, Rapid Commun. Mass Spectrom., 1999, 13(13), 1278 CrossRef CAS.
- T. O. Merren, Eur. Pat. Appl. EP 952607 A2 27 Oct 1999, 11 pp. (English). CLASS: ICM: H01J049-30. APPLICATION: EP 1999-302963 16 Apr 1999. PRIORITY: GB 1998-8319 20 Apr 1998..
- A. W. Hilkert, C. B. Douthitt, H. J. Schluter and W. A. Brand, Rapid Commun. Mass Spectrom., 1999, 13(13), 1226 CrossRef CAS.
- M. F. Miller, I. A. Franchi, A. S. Sexton and C. T. Pillinger, Rapid Commun. Mass Spectrom., 1999, 13(13), 1211 CrossRef CAS.
- E. D. Young, D. W. Coutts and D. Kapitan, Geochim. Cosmochim. Acta, 1998, 62(18), 3161 CrossRef CAS.
- A. M. Jones, P. Iacumin and E. D. Young, Chem. Geol., 1999, 153(1–4), 241 CrossRef CAS.
- M. E. Wieser and W. A. Brand, Rapid Commun. Mass Spectrom., 1999, 13(13), 1218 CrossRef CAS.
- S. Toyoda and N. Yoshida, Anal. Chem., 1999, 71(20), 4711 CrossRef CAS.
- C. A. M. Brenninkmeijer and T. Rockmann, Rapid Commun. Mass Spectrom., 1999, 13(20), 2028 CrossRef CAS.
- B.-L. Zhang, M. Trierweiler, C. Jouitteau and G. J. Martin, Anal. Chem., 1999, 71(13), 2301 CrossRef CAS.
- L. Mazeas and H. Budzinski, Analusis, 1999, 27(3), 200 Search PubMed.
- N. Kaneko, T. Maekawa, S. Igari and S. Sakata, Chishitsu Chosasho Geppo, 1999, 50(5–6), 383 Search PubMed.
- P. Chen, Y. Teffera, G. E. Black and F. P. Abramson, J. Am. Soc. Mass Spectrom., 1999, 10(2), 153 CrossRef CAS.
- K. E. A. Ohlsson and P. H. Wallmark, Analyst (Cambridge, U.K.), 1998, 124(H4), 571 Search PubMed.
- L. I. Wassenaar and G. Koehler, Anal. Chem., 1999, 71(21), 4965 CrossRef CAS.
- B. E. Kornexl, M. Gehre, R. Hofling and R. A. Werner, Rapid Commun. Mass Spectrom., 1999, 13(16), 1685 CrossRef CAS.
- N. J. Loader and W. M. Buhay, Rapid Commun. Mass Spectrom., 1999, 13(18), 1828 CrossRef CAS.
- R. A. Socki, C. S. Romanek and E. K. Gibson Jr., Anal. Chem., 1999, 71(11), 2250 CrossRef CAS.
- C. McConville, R. M. Kalin and D. Flood, Rapid Commun. Mass Spectrom., 1999, 13(13), 1339 CrossRef CAS.
- J. Gonzalez, G. Remaud, E. Jamin, N. Naulet and G. G. Martin, J. Agric. Food Chem., 1999, 47(6), 2316 CrossRef CAS.
- M. Dangin, C. Desport, P. Gachon and B. Beaufrere, Am. J. Physiol., 1999, 276/1. Pt., , 1, E212 Search PubMed.
- S. Halas and J. Szaran, Anal. Chem., 1999, 71(15), 3254 CrossRef CAS.
- R. M. Verkouteren, Anal. Chem., 1999, 71(20), 4740 CrossRef CAS.
- S. Valkiers, H. Kipphardt, T. Ding, R. Damen, P. De Bievre and P. D. P. Taylor, Int. J. Mass Spectrom., 1999, 193(1), 1 CrossRef CAS.
- B. E. Kornexl, R. A. Werner and M. Gehre, Rapid Commun. Mass Spectrom., 1999, 13(13), 1248 CrossRef CAS.
- J. Schwieters, D. Tuttas and B. Windel, Mineral. Mag., 1998, 62A(Pt. 3), 1360 Search PubMed.
- S. Yoneda and K. G. Heumann, Bull. Natl. Sci. Mus., Ser. E: (Tokyo), 1998, 21, 1 Search PubMed.
- Y. Kawai, M. Nomura, Y. Fujii and T. Suzuki, Int. J. Mass Spectrom., 1999, 193(1), 29 CrossRef CAS.
- C. M. Johnson and B. L. Beard, Int. J. Mass Spectrom., 1999, 193(1), 87 CrossRef CAS.
- K. L. Ramakumar and R. Fiedler, Int. J. Mass Spectrom., 1999, 184(2–3), 109 CrossRef CAS.
- A. Bollhofer, W. Chisholm and K. J. R. Rosman, Anal. Chim. Acta, 1999, 390(1–3), 227 CrossRef CAS.
- S. Nohda, Geochem. J., 1999, 33(2), 133 Search PubMed.
- S. J. G. Galer and W. Abouchami, Mineral. Mag., 1998, 62A/Pt., , 1, 491 Search PubMed.
- S. Richter, M. Berglund and C. Hennessy, Fresenius' J. Anal. Chem., 1999, 364(5), 478 CrossRef CAS.
- S. Richter, A. Alonso, W. De Bolle, R. Wellum and P. D. P. Taylor, Int. J. Mass Spectrom., 1999, 193(1), 9 CrossRef CAS.
- S. Richter, A. Alonso, R. Wellum and P. D. P. Taylor, J. Anal. At. Spectrom., 1999, 14(5), 889 RSC.
- S. L. Maxwell III and V. D. Jones, Nucl. Mater. Manage., 1998, 27(2), 1190 Search PubMed.
- A. H. Andrews, K. H. Coale, J. L. Nowicki, C. Lundstrom, Z. Palacz, E. J. Burton and G. M. Cailliet, Can. J. Fish. Aquat. Sci., 1999, 56(8), 1329 CrossRef CAS.
- I. I. Stewart, Spectrochim. Acta, Part B, 1999, 54B(12), 1649 CrossRef CAS.
- O. Schramel, B. Michalke and A. Kettrup, Fresenius' J. Anal. Chem., 1999, 363(5–6), 452 CrossRef CAS.
- H. Wang and G. R. Agnes, Anal. Chem., 1999, 71(19), 4166 CrossRef CAS.
- A. M. Striegel, P. Piotrowiak, S. M. Boue and R. B. Cole, J. Am. Soc. Mass Spectrom., 1999, 10(3), 254 CrossRef CAS.
- J. R. Pretty and G. J. Van Berkel, Rapid Commun. Mass Spectrom., 1998, 12(21), 1644 CrossRef CAS.
- J. A. Stone and D. Vukomanovic, Int. J. Mass Spectrom., , 1999(185–187), 227 Search PubMed.
- S. A. Pergantis, (Dept. Chem., Birkbeck College, Univ. London, London, UK WC1H 0PP). Presented at European Winter Conference on Plasma Spectrochemistry, Pau, France, January 10–15, 1999..
- T. Bantan, R. Milacic, B. Mitrovic and B. Pihlar, J. Anal. At. Spectrom., 1999, 14(11), 1743 RSC.
- Z. Mester and J. Pawliszyn, Rapid Commun. Mass Spectrom., 1999, 13(20), 1999 CrossRef CAS.
- M. Kotrebai, M. Birringer, J. F. Tyson, E. Block and P. C. Uden, Anal. Commun., 1999, 36(6), 249 RSC.
- M. Kotrebai, M. Birringer, J. F. Tyson, E. Block and P. C. Uden, Analyst (Cambridge, U.K.), 2000, 125(1), 71 Search PubMed.
- M. Kotrebai, S. M. Bird, J. F. Tyson, E. Block and P. C. Uden, Spectrochim. Acta, Part B, 1999, 54(11), 1573 CrossRef.
- B. Michalke, O. Schramel and A. Kettrup, Fresenius' J. Anal. Chem., 1999, 363(5–6), 456 CrossRef CAS.
- C. Casiot, V. Vacchina, H. Chassaigne, J. Szpunar, M. Potin-Gautier and R. Lobinski, Anal. Commun., 1999, 36(3), 77 RSC.
- S. McSheehy and J. Szpunar, J. Anal. At. Spectrom., 2000, 15(1), 79 RSC.
- O. Schramel, B. Michalke and A. Kettrup, J. Anal. At. Spectrom., 1999, 14(9), 1339 RSC.
- N. Shimizu, Y. Inoue, S. Daishima and K. Yamaguchi, Anal. Sci., 1999, 15(7), 685 Search PubMed.
- C. Brede, S. Pedersen-Bjergaard, E. Lundanes and T. Greibrokk, J. Chromatogr., A, 1999, 849(2), 553 CrossRef CAS.
- C. Brede, S. Pedersen-Bjergaard, E. Lundanes and T. Greibrokk, J. Anal. At. Spectrom., 2000, 15(1), 55 RSC.
- R. Guevremont and R. E. Sturgeon, J. Anal. At. Spectrom., 2000, 15(1), 37 RSC.
- L. Dunemann, H. Hajimiragha and J. Begerow, Fresenius' J. Anal. Chem., 1999, 363(5–6), 466 CrossRef CAS.
- C. M. Barshick, S. A. Barshick, E. B. Walsh, M. A. Vance and P. F. Britt, Anal. Chem., 1999, 71(2), 483 CrossRef CAS.
- J. L. Gomez-Ariza, I. Giraldez, E. Morales and J. A. Pozas, Analyst (Cambridge, U.K.), 1998, 124(1), 75 Search PubMed.
- Z. Mester, G. Vitanyi, R. Morabito and P. Fodor, J. Chromatogr., A, 1999, 832(1–2), 183 CrossRef CAS.
- T. Iwamura, K. Kadokami, D. Jin-ya, Y. Hanada and M. Suzuki, Bunseki Kagaku, 1999, 48(6), 555 Search PubMed.
- V. Toniazzo, C. Mustin, J. M Poprtal, B. Humbert, R. Benoit and R. Erre, Appl. Surf. Sci., 1999, 143(1–4), 229 CrossRef CAS.
- D. Yang, M. A. Puchowicz, F. David, L. Powers, M. L. Halperin and H. Brunengraber, J. Mass Spectrom., 1999, 34(11), 1130 CrossRef CAS.
- I. Rau and A. Putzka, Nucl. Instrum. Methods Phys. Res., Sect. B, 1999, B155(1–2), 53 Search PubMed.
- B. Stoll and K. P. Jochum, Fresenius' J. Anal. Chem., 1999, 364(5), 380 CrossRef CAS.
- J. A. Pfander, K. P. Jochum, A. Sassen, B. Stoll, P. Maissenbacher and M. Murmann, Fresenius' J. Anal. Chem., 1999, 364(5), 376 CrossRef CAS.
- A. I. Saprykin, J. S. Becker and H.-J. Dietze, Fresenius' J. Anal. Chem., 1999, 364(8), 763 CrossRef CAS.
- B. Wiedermann, H. C. Alt, J. D. Meyer, R. W. Michelmann and K. Bethge, Fresenius' J. Anal. Chem., 1999, 364(8), 768 CrossRef.
- B. Wiedemann, G. Radlinger, H. C. Alt, K. G. Heumann and K. Bethge, Fresenius' J. Anal. Chem., 1999, 364(8), 772 CrossRef CAS.
- R. Kurt, V. Hoffmann, R. Reiche, W. Pitschke and K. Wetzig, Fresenius' J. Anal. Chem., 1999, 363(2), 179 CrossRef CAS.
- E. Rosenberg and M. Grasserbauer, (Inst. Anal. Chem., Vienna Univ. Technol., 1060 Vienna, Austria). Presented at European Winter Conference on Plasma Spectrochemistry, Pau, France, January 10–15, 1999..
- E. Rosenberg and M. Grasserbauer, (Inst. Anal. Chem., Univ. Technol., 1060 Vienna, Austria). Presented at Instrumental Methods of Analysis. Modern Trends and Applications (IMA '99), 19–22 September, 1999, Chalkidiki, Greece..
- K. V. Thomas, J. Chromatogr., A, 1999, 833(1), 105 CrossRef CAS.
- L. L. Yu, G. C. Turk and S. R. Koirtyohann, J. Anal. At. Spectrom., 1999, 14(4), 669 RSC.
| This journal is © The Royal Society of Chemistry 2000 |