Atomic spectrometry update: review of advances in atomic spectrometry and related techniques

Contents

• 1. Sample introduction
  • 1.1 Liquids
  • 1.2 Vapour generation
  • 1.3 Solids

• 2. Instrumentation, fundamentals and chemometrics
  • 2.1 Instrumentation
  • 2.2 Fundamentals
  • 2.3 Chemometrics

• 3. Laser-based atomic spectrometry
  • 3.1 Laser induced breakdown spectroscopy (LIBS)
  • 3.2 Laser atomic absorption spectroscopy (LAAS)
  • 3.3 Cavity ringdown spectroscopy (CRDS)
  • 3.4 Laser induced fluorescence (LIF)

• 4. Isotope analysis
  • 4.1 Reviews
  • 4.2 New developments
  • 4.3 Radiogenic isotope ratio analysis
  • 4.4 Stable isotope ratio analysis
  • 4.5 Nuclear forensics

• Conflicts of interest

• Glossary of abbreviations

• Acknowledgements

• References

---

Abstract

This review of 155 references covers developments in ‘Atomic Spectrometry’ published in the twelve months from November 2018 to November 2019 inclusive. It covers atomic emission, absorption, fluorescence and mass spectrometry, but excludes material on speciation and coupled techniques which is included in a separate review. It should be read in conjunction with the previous review¹ and the other related reviews in the series.^2–6 A critical approach to the selection of material has been adopted, with only novel developments in instrumentation, techniques and methodology being included. The common atomic spectrometric techniques such as FAAS, GFAAS, ICP-OES and ICP-MS have reached a level of maturity such that there are fewer novel developments. Exceptions to this are: the continued development of single particle and single cell analysis by ICP-MS; advances in miniaturised sources for vapour and liquid sample introduction; and coupling compact sources with ion trap and TOF mass spectrometers. New developments in LIBS were less pronounced this year, but there were some interesting couplings of the technique with other instrumentation such as QMS, RAMAN and acoustic spectroscopy. LA-MC-ICP-MS was the predominant technique to push the boundaries of isotope ratio analysis by lowering LODs, increasing precision and accuracy, and measuring smaller samples. Another advance was the progress of high-impedance resistors (10¹² Ω and 10¹³ Ω) from the research lab into useful applications, with solutions being found to address the slow response times of their amplifiers resulting in their useful incorporation into the multi-collector arrays for the detection of minor isotopes.

---

1. Sample introduction

1.1 Liquids

1.1.1 Sample pre-treatment.
1.1.1.1 Extraction and digestion. Several reviews of LPME have been published. The first,⁷ (91 references) addressed developments involving chemical reactions, including ion-pair extraction, complexation, chemical derivatization, phase transfer catalysis, and nanoparticle-assisted chemical reactions. Figures of merit for various methods were tabulated for ease of reference. The use of deep eutectic and supramolecular solvents was identified as a possible route to further automation. A second review⁸ of 48 references focussed on single drop, hollow fibre and dispersive LPME for metals. Further progress was signposted by the use of magnetic materials, new interfaces and automation. A third review⁹ (98 references) concentrated on the use of FI systems for automation of LPME, with particular focus on single drop, hollow fibre and dispersive liquid–liquid microextraction methods. Applications were tabulated by analyte for ease of reference. While there were fewer FI methods compared to batch procedures, the reviewers concluded that automation promises higher throughput and greater accuracy. Indeed, the development of microfluidic sample pre-treatment methods has long held out promise for on-line, miniaturised and automated systems. Addressing this directly, He et al.¹⁰ reviewed (20 references) the use of microfluidic chips coupled with ICP-MS for the analysis of trace elements in cells. Various extraction regimes, including SPME, LPME and magnetic microextraction were covered. The use of micro-arrays and single droplet extraction of cells were also addressed. The authors concluded that the development of packing materials with good selectivity, high adsorption capacity and fast reaction kinetics is necessary to improve sensitivity, selectivity and sample throughput.

A specific example of this type of development was published by Chen et al.,¹¹ who developed a method that utilised a chip-based array and magnetic SPE. Eight solid phase extraction columns packed with magnetic porous organic polymers (MOPs) with 1,3,5-tris(4-aminophenyl)benzene were integrated in parallel on a microfluidic chip for array microextraction. The microextraction channels were 500 μm wide and 50 μm high. A 4 cm length fused silica capillary (75 μm i.d.) was used as the connector to a Burgener nebuliser for ICP-MS analysis of Au, Bi and Pt, with LODs of 4.4, 3.4 and 8.6 ng L⁻¹ respectively. The method exhibited good tolerance to matrix elements typically found in biological samples and was validated by analysis of the CRM GSH-1A human hair (3.18 ± 0.15 μg L⁻¹ compared with the certified value of 3.40 ± 0.20 μg L⁻¹).

Ghorbani et al.¹² published a tutorial review (222 references) of dispersive SPME which included the basic theory of the technique with regard to partitioning between phases. The review was split into sections on traditional, vortex assisted and ultrasonic assisted DSPME, with comprehensive tables of applications organised by analyte, sorbent and sample. The authors acknowledged the difficulty of automating DSPME, presumably because of the necessity to separate the solid and liquid phases at some point.

1.1.1.2 Elemental tagging. Elemental tagging is mostly used for the indirect quantitation of proteins, peptides and nucleic acids. A number of methods have been developed: quantitation via an intrinsic hetero-atom in a protein, such as sulfur or selenium; chelation or covalent bonding of a metal tag to the target molecule and subsequent quantitation by direct calibration or ID-MS; labelling using an antigen–antibody immunoassay approach; or hybridisation of target strands of DNA or RNA with complementary labelled probes. There are also variations which combine elements of all of the foregoing. The number of publications in this area peaked several years ago, then declined in the face of the problems of routine implementation. Nevertheless, one of the main advantages of this approach is the ability to amplify the signal, which is possible in several ways. One of these, the label-free bioassay has been developed for the determination of DNA and proteins. A new variation on this has been published in two similar papers by the same group.^13,14 They used a DNA hairpin to trap Hg²⁺ ions (the T–Hg²⁺–T complex) which opened when hybridised to a complementary strand of the DNA/protein target molecule, thus releasing Hg²⁺ into solution. The free Hg²⁺ then took part in a cation exchange reaction with CdTe QDs releasing Cd²⁺, which was separated by a filtration membrane without separating the CdTe QDs. The target DNA was quantified by determining the Cd²⁺ before and after the selective recognition reaction. In the initial study¹⁴ LODs of 3 fM and 2 fM were obtained using CVG-AFS and ICP-MS respectively. The method was subsequently applied¹³ to the determination of DNA and carcinoembryonic antigen in serum samples, with LODs 0.2 nM and 0.2 ng mL⁻¹, respectively.

The use of NPs for amplification has also proved popular. Liu et al.¹⁵ developed an assay for Exo I nuclease. Poly-thymine DNA (polyT DNA) was immobilised onto magnetic beads, which acted as a substrate of Exo I and template for Cu NP formation. The method was poorly described in the paper, but it seems that Cu NPs were formed in situ on the DNA template by reduction of Cu²⁺ to Cu⁰ using ascorbic acid. However, in the presence of Exo I the polyT DNA was split into fragments, so formation of Cu NPs was not possible. Hence, the difference in the Cu measured using ICP-MS, with and without Exo I present, and after magnetic separation, was related to Exo I activity. An LDR of between 0.1 to 20 U μL⁻¹ and LOD of 0.029 U mL⁻¹ were obtained. Xiao et al.¹⁶ used Au NPs and primer-DNA in a dual amplification strategy. First, Au NPs were conjugated with hairpin-DNA (HP-DNA) which hybridised with the target DNA. A biotin-labelled DNA primer was also hybridised and an isothermal circular strand-displacement polymerization reaction undertaken to polymerise this with the AuNPs. These were then captured by streptavidin modified magnetic beads (SA-MBs), separated, and determined using ICP-MS. The method could detect DNA as low as 8.9 fmol L⁻¹ (45 zmol in 5 μL) and was used to determine target DNA in spiked human serum samples, with recoveries between 84 and 120%.

He et al.¹⁷ developed an approach using Au and Ag NPs, for the simultaneous determination of SMMC-7721 (human hepatocellular carcinoma) and A549 cells (human lung carcinoma) cells. In this method, DNA aptamers (Ap-SMMC and Ap-A549) which could recognize the target cells were immobilised on magnetic beads (MBs). Complementary DNA probes labelled with AuNp and AgNPs were hybridised with Ap-SMMC and Ap-A549 respectively. Two further AuNp and AgNP probes were then introduced which hybridised with the first, hence forming a chain to considerably amplify the signal. When solutions of the target cancer cells were introduced they hybridised with the MB-Ap-SMMC and MB-Ap-A549 sections of DNA, thus releasing the amplification chain into solution. After magnetic separation the AuNP-DNA and AgNP-DNA chains were determined using ICP-MS to yield an LOQ of approximately 50 cells for both SMMC and A549. The method was used for the simultaneous determination of these cells in spiked serum samples, with recoveries of between 93 and 108%. Another method of cell counting was reported by Liang et al.¹⁸ for the determination of Escherichia coli, Staphylococcus aureus, Listeria monocytogenes, Shigella dysenteriae, and Vibrio parahemolyticus. In this approach, noncanonical alkyne-D-alanine (aDA) was added to the bacterial strains and metabolically assembled into the peptidoglycan layer-supported bacterial cell walls. This was followed by subsequent conjugation of a 1,4,7,10-tetraazacyclododecane-1,4,7-tris-acetic acid-10-azidopropyl ethylacetamide-Eu complex (azide-DOTA-Eu). The tagged cells, which can be considered as Eu-engineered particles, were then determined using SP-ICP-MS, with 5-fold amplification relative to the single bacterial cells. Furthermore, the cells were labelled with lanthanide (Ln)-coded polyclonal antibody (Ln = ¹³⁹La, ¹⁴¹Pr, ¹⁴²Nd, ¹⁵²Sm, and ¹⁶⁰Gd) tags so the individual bacterial strains could be recognised simultaneously.

One popular way of quantifying proteins has been to adapt immunoassay methods to make them amenable to element specific detection by tagging with a metal or non-metal label. Alonso-Garcia et al.¹⁹ developed an assay using iodinated antibodies for the determination of transferrin in breast cancer cells. Cell lysate samples were mixed with biotinylated anti-transferrin and iodinated anti-transferrin antibodies to form a sandwich complex. Streptavidin-coated magnetic microparticles were then added to capture the biotinylated complexes, separated magnetically and determined using ICP-MS detection of ¹²⁷I⁺. The key to the procedure was the adaption of the well-established protocol used for the iodination of peptides and proteins (e.g. antibodies) using radioactive ¹²⁵I. In this work, non-porous, polystyrene beads coated with an oxidising reagent were used to iodinate with NaI. This resulted in efficient iodination of the antibody with minimum degradation, with a stoichiometry of 27 : 1 iodine moles per mole of antibody, thereby amplifying the signal. The LOD was estimated to be 140 ng mL⁻¹ of transferrin. Another novel immunoassay development was reported by Vaneckova et al.,²⁰ who combined CdQD-labelled antibodies with molecularly imprinted polymer technology. The CdQD–peptide–antibody–antigen complexes were immobilised on the MIP surface to allow removal of interferents, then determined using LA-ICP-MS detection of ¹¹¹Cd with an LOD of ∼3 μg.

Ren et al.²¹ developed a somewhat different approach for the quantitation of peptides by making use of his-tags (a string of histidine residues attached at the N or C terminus of a recombinant protein to facilitate cleanup). They managed to non-covalently label histidine residues with Ru using a THMA-coordinated Ru^III complex; the advantage of the approach being that selective labelling of the histidine segment did not disrupt the function of the target recombinant protein. For example, labelling of a recombinant form of serum transferrin (rhTf) containing an N-terminal His-tag resulted in incorporation of up to three Ru ions per protein. In this case there was no Ru loss from the species following its 24 h incubation with albumin, but this is always a concern. Hence, one way of maximising stability of the tagged protein is to label it using a covalent reaction. Ji et al.²² achieved this by synthesising an epoxysuccinyl-leucine-tyrosine-6-aminocaproic-lysine-amino-Boc-alkyne to irreversibly label cathepsin, which was in turn labelled with azido-DOTA-Eu. Quantitation of the activity of the cathepsin proteins was determined using HPLC-ID-ICP-MS, with an LOD of 25.5 fmol.

1.1.2 Nebulisation. The wide range of nebulisers available for ICP applications was addressed by Masone et al.²³ by considering choice of nebuliser for a particular ICP-OES analysis. Results of a performance comparison of concentric nebulisers and parallel path nebulisers provided a selection guide based on performance and design. Improvements in sensitivity and LODs in ICP-OES through IR heating of a cyclonic spray chamber were reported by Al Hejami and Beauchemin.²⁴ A ceramic beaded IR rope heater was used to heat the top surface, side-arm, torch connector and torch base of a quartz baffled cyclonic spray chamber. IR heating resulted in enhanced transport efficiency from 4.5% at room temperature to 15.7% at 150 °C. This resulted in improvements in sensitivity by 3.0- to 4.9-fold and LODs by 3.7- to 12-fold, depending on the element, compared to those with the same sample introduction system at room temperature. A micro-flow liquid sample introduction system (μ-dDIHEN), for direct nebulisation of small samples in ICP-MS, was described by Louvat et al.²⁵ The system was tested with sample loops of 10 and 50 μL at flow rates between 5 and 50 μL min⁻¹ using SF-ICP-MS and MC-ICP-MS. Three applications were examined: B isotopic ratio measurement of geological samples; multi-element analyses of natural water samples; and Au NP characterisation by SP-ICP-MS. Signal sensitivity increased with liquid uptake rate up to a maximum of 30 μL min⁻¹. B isotope ratios were determined successfully for reference material solutions with B concentrations between 10 and 200 μg L⁻¹. Transient versus continuous measurement modes for a series of natural samples were compared with the two modes providing comparable results. The transient mode was recommended for samples of less than 100 μL and B content of low ng. For multi-element analyses by SF-ICP-MS, the LOQs for most of the trace and ultra-trace elements were lower (up to 15 times) with the μ-dDIHEN than with a concentric nebuliser and spray chamber, due to the reduced sample volume. The promise of the system for NP characterisation was demonstrated with transport efficiencies of up to 85% recorded for 40 nm diameter Au NPs.

1.1.3 Thermal vaporisation. Improvements in sensitivity and LODs were obtained using a mixed-gas Ar–N₂ plasma in ETV-ICP-OES.²⁶ With 0.4 L min⁻¹ N₂ in the plasma gas flow and 20 mL min⁻¹ N₂ sheathing gas in the central channel of the ICP, sensitivity based on peak area increased by up to 32-fold (Ba) and LODs improved by up to 300-fold (Ba). Detection limits, based on 3 mg aliquots, ranged from 0.007 (Mo, Sb) to 100 (Mg) mg kg⁻¹ with the mixed-gas plasma versus 0.01 (Tl) – 3000 (Al) mg kg⁻¹ with an Ar plasma. Improvements were observed to be greater for ionic compared with atomic emission lines. The mixed-gas plasma was demonstrated to be more robust than an Ar plasma as evidenced by reduced suppression of the Ar 763.511 nm signal caused by the ETV effluent in the mixed-gas plasma. Accurate results were obtained for between 41 and 65 of the elements determined in three soil CRMs using another soil CRM for external calibration.

Use of carbon microparticles (CMs) as a physical carrier to enhance the analyte transport efficiency in ETV-ICP-MS was evaluated by Patocka et al.²⁷ The CMs were mixed with samples or calibration standard solutions and the resulting slurry introduced into the ETV. Optimised conditions included pyrolysis and vaporisation temperatures of 500 and 2700 °C for Au and 400 and 1900 °C for Tl, for concentrations of CMs of 1 g L⁻¹ for Au and 2.5 g L⁻¹ for Tl. LODs were determined as 0.016 ng L⁻¹ for Au and 0.026 ng L⁻¹ for Tl. The selected elements were determined accurately in several SRMs and recoveries and repeatabilities measured on calibration standards were found to be in the range: from 99 to 100% and 0.2 to 2.3% for Au; and from 100 to 111% and 2.9 to 6.7% for Tl.

1.1.4 Single particle analysis. The growing interest in the use of SP-ICP-MS continued in 2019. The technique is capable of simultaneously measuring NP size and number concentration of metal-containing NPs at environmental levels. Bustos et al.²⁸ reported on the validation of SP-ICP-MS capabilities for measuring the mean NP size and number size distribution of Au NPs. The validation data was obtained through calibration using NIST SRM 8013, comparison with HR-SEM reference analysis and an evaluation of the uncertainty associated with the measurement of the mean particle size to enable comparison of the SP-ICP-MS and HR-SEM methods. After optimising conditions for HR-SEM and SP-ICP-MS, both methods were used to characterise commercial Au NP suspensions of three different sizes (30, 60, and 100 nm) with four different coatings and surface charge at pH 7. The SP-ICP-MS measurements (corroborated by HR-SEM) demonstrated the existence of two distinct subpopulations of particles in the number size distributions for four of the 60 nm commercial suspensions. This was not recorded in the TEM data provided by the vendor, demonstrating the effectiveness of SP-ICP-MS for routine characterisation of commercial Au NP suspensions regardless of size or coating. Hence, SP-ICP-MS appears to be an effective tool for the detection and quantification of inorganic NPs.

While the sizing of NPs suspended in water is relatively straightforward by SP-ICP-MS, accurate mass quantification of NPs in complex media, such as consumer products and natural systems remains a challenge because the matrix has been shown to affect analyte sensitivity and lead to inaccurate NP sizing. An online microdroplet calibration system to size NPs in a single step, using ICP-TOF-MS, has been suggested as a solution to this matrix effect.²⁹ In this system, microdroplets used as the calibrant and nebulised NP-containing solutions were introduced concurrently into the ICP via a dual-inlet sample introduction system. Both the calibrant microdroplets and analyte NPs experienced the same plasma conditions, so they were subjected to the same matrix-related signal enhancement or suppression. Size determination of Ag and Au NPs was achieved, demonstrating successful matrix-independent mass quantification of analyte NPs in the presence of several matrix effects, including acid effects, space-charge effects, and ionisation suppression. The effect of non-spectral interferences in SP-ICP-MS was demonstrated by Loula et al.³⁰ In comparison to dispersions in water, non-spectral interferences caused by Na resulted in an under-evaluation of the diameter of As and Au NPs by 7% and 15% respectively at NaCl concentration of 450 mg L⁻¹, increasing to 28% and 41% at NaCl concentration of 4500 mg L⁻¹. In addition, as a result of lower transport efficiency, non-spectral interferences also led to a 9% reduction in the number of detected NPs for dispersions of both As and Ag NPs in 4500 mg L⁻¹ NaCl. In contrast, determination of NPs in matrices containing methanol gave results where Ag and As NP diameters were over-estimated by about 3% and 15% (at a methanol content of 1% v/v), and about 6% and 20% (at a methanol content of 2% v/v), respectively, in comparison to water dispersions. The organic carbon species behaved as surfactants and increased the transport efficiency; this led to an increase in the determined number concentration of NPs. In comparison to water dispersions, this was over-estimated by about 17% for Ag NPs and about 10% for As NPs at a methanol content of 5% v/v. Three sample introduction systems were compared with respect to transport efficiency (TE) and high throughout sample analysis using SP-ICP-MS.³¹ The systems evaluated were: a high-performance concentric nebuliser with a heated cyclonic spray chamber and a three-stage Peltier-cooled desolvation system (HPCN-APEX); a conventional sample introduction system (i.e. a commercially available nebuliser with a cyclonic spray chamber); and a total consumption (TC) system. The TEs observed for the HPCN-APEX and TC systems were almost 100% at sample uptake rates of 103 and 9.8 mL min⁻¹, respectively, whereas that of the conventional sample introduction system was only 10% at an uptake rate of 112 mL min⁻¹. The dried aerosol conditions in HPCN-APEX resulted in size and number detection limits 1.6-fold and 10-fold lower than those of the TC system. The HPCN-APEX showed better detection efficiency for smaller particles and lower particle numbers. The effects of using a collision/reaction cell (CRC) in ICP-QMS operated in single particle mode were assessed.³² The influence of: (i) various CRC gases; (ii) gas flow rate; (iii) NP size; and (iv) NP type was evaluated using Ag, Au and Pt NPs with both a traditional ICP-QMS instrument and a tandem ICP-MS. Using a CRC caused a significant increase in the NP signal peak width, and the effect was more prominent when heavier CRC gases, e.g., NH₃, were used. At a higher gas flow rate and/or for larger particle sizes (>100 nm), the NP signal duration was further prolonged. This effect of using a CRC was demonstrated by characterising custom-made 50 and 200 nm Fe₃O₄ NPs using different CRC approaches. The use of NH₃ to form the Fe(NH₃)₂⁺ product ion at 90 m/z induced a significant peak broadening compared to that observed when using H₂. This extension of transit time was attributed to the collisions/interactions of the ion cloud generated by a single NP event in the CRC. Based on these results, the influence of a longer peak width on the accuracy of SP-ICP-MS measurement data should be considered when using CRC technology to overcome spectral overlap. To mitigate the potential detrimental effect of using a CRC, the use of light gases and low gas flow rates was recommended.

Single-cell analysis using ICP-MS is of considerable interest in biological and medical analysis. However, differentiation of a single cell from a doublet remains a challenge. One suggestion is to encapsulate single cells into droplets on the platform of a microfluidic chip. However, the fabrication of these chips is sophisticated and has limited their use. Yu et al.³³ have described an off-the-shelf 3D microfluidic device, made by assembling commercially available parts and thus avoiding a specific manufacturing process. Uniform monodisperse microdroplets were generated from the 3D microfluidic device with a size variation of 1.5%, and the inner diameter of the 3D microfluidic device matched the nebuliser (150 μm). The 3D microfluidic device-time-resolved ICP-MS system was used to analyse Ag NPs (51 nm), and the result is in good agreement with a conventional acid digestion method, demonstrating the accuracy of the method. The uptake of Ag in HepG2 cells was studied by incubating with Ag⁺ or AgNPs under biocompatible conditions. The results revealed that the cell-to-cell variability in terms of the diversity of cells incubated with Ag NPs was wider than those cells incubated with Ag⁺ from the aspect of the content distribution of Ag at the single-cell level. Another approach to ensure single-cell sampling into an ICP involves a configured pressure-resistant interface for high-throughput and high-precision sampling.³⁴ An aqueous cell suspension was ejected and sheared into droplets by tangentially-flowing hexanol-continuous phases in the flow-focusing geometry of the interface. This configuration was found to trap a single cell into a droplet, with the a very low probability (<0.005%) of a single droplet containing two cells. The interface was coupled with time-resolved ICP-MS for quantifying NPs in single MCF-7 cells. To reduce the carbon deposition and plasma instability due to the use of hexanol, DMC was added to aid oxidation of the hexanol.

1.2 Vapour generation

Chemical vapour generation is long established as a means of sample introduction for atomic spectroscopy, and generally involves either conversion to a volatile element, hydride or organometallic species. Wang et al.³⁵ reviewed (81 references) developments in CVG for Zn determination by AAS, AFS, AES and ICP-MS. There is little in the way of in-depth analysis of the methodology, but there are some useful tables of applications with figures of merit.

D’Ulivo³⁶ has comprehensively reviewed (190 references) the mechanism of CVG for the hydrides of Ge, Sn, Pb, As, Sb, Bi, Se, Te, Hg and Cd using aqueous boranes. The review started by stating that evidence suggests the derivatisation involves two competitive reactions. Derivatisation leads to the formation of the final hydride, EH_n (also involving reaction of the analytical substrate, EL_n) with species containing at least one B–H bond. These species can be BH₄⁻ or hydridoboron intermediates which are formed in a second competing reaction during reagent hydrolysis. The review then discussed the formation of side products and the effects of pH, hydrolysis reactions, additives and reactions involving amine-boranes. For anyone undertaking research into CVG this paper is highly recommended. In a contribution to this canon of knowledge, a separate paper by the same workers³⁷ reported on an investigation into the mechanisms of volatile Cd generation by tetrahydridoborate. They proposed that Cd^II forms hydroxyl complexes at high pH and BH₄⁻ undergoes acid hydrolysis to BH₃(H₂O) which is then deprotonated to form BH₃OH⁻. This then acts as the reductant, which is stronger than BH₄⁻. An important factor was the presence of dissolved oxygen which hindered the generation of volatile Cd species in the liquid phase. Oxygen removal (by degassing with Ar) and promotion of BH₃OH⁻ derivatisation (by increasing the pH) improved the generation of volatile Cd by 20-fold.

Ma et al.³⁸ developed a simple EVG cell made from two pyrolytic graphite tubes connected in series using a piece of polyethylene tube. The two graphite tubes, from a GFAAS instrument, served as cathode and anode respectively, an ac to dc wall plug adapter (12 V, 2 A) was used as the power supply, and there was no ion-exchange membrane, as is usually the case. A solution containing 40 ng mL⁻¹ As^III with thiourea and HCl was used as the electrolyte, and sensitivity was optimised at concentrations of 1% (m/v) thiourea and 10% (v/v) HCl respectively. The authors noted that the presence of thiourea was essential to generate a signal, which increased at higher concentrations, but was limited to 1% for practical reasons. The LOD for As was 0.3 ng mL⁻¹ using AFS detection and the method was validated by the analysis of CRMs GBW10045 rice flour and GBW10015 spinach.

Photochemical vapour generation was the subject of two reviews. Slachcinski³⁹ reviewed (113 references) PVG and CVG used with a variety of sources and with OES and MS detection, with particular reference to a mixed mode of sample introduction for volatile and non-volatile species. The review includes a comprehensive table of methods of sample introduction, reagent conditions, analytes and sample type. Zou et al.⁴⁰ reviewed (96 references) the use of nanomaterials (e.g. TiO₂) and metal–organic frameworks (MOFs) for PVG, used both a catalysts and preconcentration media. One of their main conclusions was that there is scope for refining PVG by trying other types of photocatalysts. This was supported by the work of Hu et al.,⁴¹ who developed a method for the PVG of Mo which was enhanced dramatically by Co²⁺ and Cu²⁺ ions in formic acid. Using electron paramagnetic resonance (EPR), and addition of 5,5-dimethyl-1-pyrroline N-oxide (DMPO) to the solution as a spin trap, they detected carboxyl and hydroxyl free radicals in the reaction mixture during PVG. They speculated that the Co²⁺ and Cu²⁺ induced generation of the ˙CO₂ free radical from the photolysis of formic acid, which then reacted to form volatile Mo(CO)₆. Addition of 20 mg L⁻¹ of Co²⁺ and 2.5 mg L⁻¹ of Cu²⁺ into 20% formic acid resulted in a 15-fold enhancement in the Mo signal, resulting in an LOD of 6 ng L⁻¹ using ICP-MS.

The use of a DBD to trap, atomise and/or generate volatile hydrides, prior to introduction into e.g. AAS for detection, has become popular over the last few years. Burhenn et al.⁴² Burhenn studied the atomisation mechanism of AsH₃ in a planar DBD using temporally and spatially resolved OES. They observed maximum emission from As (at 228.8 nm) 235 and 285 ns after plasma ignition for helium and argon DBDs respectively. The emission signal was concentrated in the region of the re-igniting coincident plasma (i.e. where the plasma was divided in a temporally sharp early plasma stage and a bright re-igniting coincident plasma) in both cases, and was also detectable on the glass surface. The authors postulated that this provided evidence that As was atomised and excited across the whole discharge channel, hence its applicability for introduction into AAS. In a related paper⁴³ by some of the same workers, results of a study into the feasibility of trapping SeH₂ in the same type of DBD were reported. Using a ⁷⁵Se tracer and trapping onto activated carbon, they were able to distinguish between the Se fractions trapped and volatilised at various times in the DBD process. Results indicated that the trapped Se fraction reached 92%, mostly in the central part of the DBD, but 28% of the Se remained after the volatilisation step. With an LOD of 12 pg mL⁻¹, this compared reasonably well with other trapping methods. In a third paper,⁴⁴ the same research group optimised the system for trapping Sn.

Liu et al.⁴⁵ took a slightly different approach by using the DBD for generation of the volatile species. They developed a liquid-spray DBD, which was made from a piece of copper wire surrounding the outside of the nozzle of a micro-nebuliser as one electrode, with the other inserted into the inside of a glass tube at a distance of 3 mm from the tip. The nebulised sample solution formed a discharge between the two electrodes on application of an ac voltage. This was contained within a spray chamber arrangement to filter out large droplets before introduction into ICP-MS. Se, Ag, Sb, Pb, and Bi ions in solution were simultaneously converted to volatile species, using only formic acid as a reagent. LODs were 10, 2, 5, 4 and 3 ng L⁻¹ respectively, for a 20 μL sample.

1.3 Solids

1.3.1 Direct methods.
1.3.1.1 Arc & spark. A compact spark-induced plasma spectroscopic device was developed by Doh et al.⁴⁶ The system consisted of a spark generator connected to tungsten electrodes, a custom-built delay generator, and two spectrometers to cover the UV-Vis range from between 214 and 631 nm. The system was evaluated by sampling Cu standards. Prominent spectral peaks were identified using the NIST database for atomic emission lines. The effectiveness of the proposed system was also tested with a lanthanide sample (Gd) and provided qualitative identification of the characteristic peaks. Semi-quantitative measurement for Si and Au was performed using variable amounts of each particulate. Silica microbeads in solution were applied to paper wafers, while Au NPs were sputter-coated onto Si wafers. Results showed a positive correlation between the intensity of the signal and the concentration of each type of particulate. The variation of signal intensity was investigated to determine the repeatability, and the coefficient of variation was lowered from 60% to 25% after averaging measurements of multiple ablations per observation.
1.3.1.2 Glow discharge. A Ne plasma was tested for the direct analysis of solid dielectric materials by pulsed GD-TOF-MS, with an emphasis on elements with high ionisation energy.⁴⁷ Samples of fluorine-doped potassium titanyl phosphate (KTP) single crystals were tested. A range of operating parameters were optimised, relative sensitivity factors were used for quantification and Ti was used for normalisation. As with conventional argon GD-TOF-MS, the most effective mechanism for the ionisation of F and O was attributed to high-energy electrons under short repelling pulse delays. The LODs for O (0.0005 mass%) and F (0.0002 mass%) were improved using the Ne GD compared with those obtained using Ar GD (0.001 and 0.01 mass%, respectively) due to more effective ionisation. The improvement was especially pronounced for F, due to the alleviation of plasma-based interferences. However, for other KTP constituents, K, P and Ti, the analytical performance was poorer, due to the lower sputtering rates in the Ne GD. The LODs for K and P were significantly increased, being 0.008 mass% and 0.01 mass%, respectively. The analytical performance for elements with lower ionisation energies was observed to be adequate when analysed in a single run along with O and F. The results of the study suggested that Ne is a suitable plasma gas for the determination of high ionisation energy elements in non-conducting matrices.

Fast flow GD-MS was evaluated for the quantification of metallic impurities and oxygen in solid samples using a commercial system.⁴⁸ Quantitation based on relative and absolute sensitivity factors was evaluated using three sample matrices (Al, Cu and Zn). Discharge conditions were optimised to favour matrix independent calibrations and improved standard relative sensitivity factors were achieved, based on multi-matrix calibrations. The sputtering rate corrected calibration was also presented as a multi-matrix calibration approach. The capabilities of GD-MS for oxygen determination were investigated using a set of new conductive samples containing oxygen with mass fractions in the percent range in three different matrices (Al, Mg and Cu), produced by a sintering process. The LODs obtained were poor (in the order of g kg⁻¹) due to reduced sensitivity of oxygen in GD-MS and high oxygen background signal. The absolute sensitivity procedure was demonstrated as a matrix-independent approach, providing quantitative values consistent with those obtained by carrier gas hot extraction.

In dc-GD-MS, the current, voltage and argon pressure can be selected with two degrees of freedom to define the analytical conditions. Paudel⁴⁹ used a commercially available system to study the quantification of impurities as a function of changes in these operating parameters. A Ta pin sample was analysed using conditions from 0.5 mA to 5 mA at 1000 V and from 600 V to 1500 V at 3 mA. The results indicated that some elements were more sensitive to changes in operating parameters than others. The variation in the quantification of impurities was associated with differences in the dependencies of the ion intensities of the trace and matrix elements under the discharge conditions. The difference in the relative sensitivity factors (RSFs) at different discharge settings were attributed to these findings.

A noticeable feature of pulsed GD-OES is the temporal response of the emission signal. This signal, generated by the pulsed discharge, comprises a sharp pre-peak and a subsequent plateau portion with smaller emission intensities. The intensity of the pre-peak can be between 10 and 20 times as large as the plateau-stage intensity, depending on the discharge parameters. Wagatsuma⁵⁰ reviewed the various suggested mechanisms, including: temporal variation in a degree of self-absorption, during a pulse duration on generation of gas pressure wave; or on a transient increase of the discharge current at the initial edge of a pulse which thus elevates the gas temperature. It was also mentioned that such emission characteristics could be utilised to obtain better analytical performance in OES. A detection method, using a pulsed bias current, was effective for controlling the emission response from the pulsed plasma, because it totally elevated the emission intensity of the plateau portion rather than the pre-peak, with little change in the background level.

While the liquid sampling-atmospheric pressure (LS-AP) GD has shown potential as an ionisation source for elemental/isotopic/molecular species analysis, up to this point it has been used almost exclusively in conjunction with trapping-type mass analysers. Williams and Marcus⁵¹ described the coupling of LS-AP-GD to a commercially available QMS. The instrument is capable of numerous MS/MS techniques as well as scanning for low mass elements which are not typically analysed in trapping-type MS instruments. The QMS differs appreciably in how it may be implemented in tandem mass spectrometry versus triple quadrupoles. For example, fragmentation methods such as collision-induced dissociation (CID) in Q2 and in-source CID can be affected to reduce background spectral contributions and remediate molecular interferences. With this coupling, a multi-parametric evaluation was performed by optimizing the analyte responses and SBR for Rb, Ag, Tl, and U as test elements. With the optimised conditions set, parent ion, product ion, and neutral loss scan methods were explored to observe the species formed in the LS-AP-GD. LODs for the test elements were determined in the range between 0.99 and 38 ng mL⁻¹ for 50 μL injections.

LS-AP-GD was paired with a monochromatic imaging spectrometer to interrogate the formation of plasma species.⁵² Spatial emission profiles were obtained and used to gain an understanding of species location in the plasma and the effects of plasma gas flow. As sheath gas and counter gas flow rates were increased visible differences in intensity and distribution of emission from the monitored species were observed. Silver was used as the test analyte due to its good sensitivity, wavelength positioning, and the ability to compare with previous work. The majority of analyte emission was found to occur at the tip of the solution electrode, where solutes are introduced to the plasma, whereas various background species emitted throughout the plasma. In addition to helping understand plasma operation, the studies also aided optimisation of the gas flow parameters for analyte emission intensity, SBR and SNR. The optimised gas flow parameters were found to be 0.7 and 0.1 L min⁻¹ for the sheath and counter gas respectively.

1.3.2 Indirect methods.
1.3.2.1 Laser ablation. The analytical potential of ns LA-ICP-SF-MS system was evaluated for fast and highly spatially resolved qualitative elemental distribution in single cells by Pisonero et al.⁵³ Initially, a low surface roughness (<10 nm) thin In–SnO₂ layer deposited on glass was used to investigate the size, morphology and overlapping of laser-induced craters obtained at different laser repetition rates, using AFM. Conical craters with a surface diameter of about 2 mm and depths of about 100 nm were measured after a single laser shot. The influence of the sampling distance on the LA-ICP-MS ion signal wash-out time was evaluated and a significant decrease of the transient ¹²⁰Sn⁺ ion signal was observed after slight variations (around 200 nm) around the optimum sampling position. Ultra-fast wash-outs of the ablation chamber of less that 10 μs reduced the aerosol mixing from consecutive laser shots even when operating the laser at high repetition rates. Fast and highly spatially resolved images of elemental distribution within mouse embryonic fibroblast cells and human cervical carcinoma cells, incubated with Au NPs and Cd QDs respectively, were obtained at the optimised operating conditions. Elemental distribution of Au and Cd in single cells was achieved using a high scanning speed and high repetition rate. The results obtained for the distribution of fluorescent Cd QDs within the cells were in good agreement with those obtained by confocal microscopy. The size, morphology and overlapping of laser-induced craters in the fixed cells were also investigated using AFM, observing conical craters with a surface diameter of about 2.5 mm and depths of about 800 nm after a single laser shot.

The ability to obtain good quality 2D LA-ICP-MS elemental maps has improved in recent years using “fast” LA cells and simultaneous ICP-TOF-MS detectors. Data are generally generated in line scanning mode by pulsed LA of the surface and ICP-MS sampling (signal integration) of one or more pulses. However, imaging artefacts such as blur, smear, aliasing and noise can be encountered that may deteriorate the image quality. van Elteren et al.⁵⁴ reported on the selection of optimal conditions for fast and high-quality 2D LA-ICP-MS elemental mapping to minimise these artefacts, using a modelling approach.

A new LA-IRMS method was proposed⁵⁵ for the highly resolved analysis of carbon isotope signatures in solid samples down to a spatial resolution of 10 μm. The setup included in-house-designed exchangeable ablation cells (3.8 and 0.4 mL, respectively) and an improved sample gas transfer manifold. This facilitated accurate δ¹³C measurements of an acryl plate standard down to 0.6 and 0.4 ng of ablated carbon, respectively. Initial testing on plant and soil samples confirmed that micro-heterogeneity of the natural ¹³C : ¹²C ratio could be mapped at a spatial resolution down to 10 μm.

LA-ICP-MS is increasingly used to determine major, minor and trace element concentrations in Fe-rich alloys. In the absence of matrix-matched standards, standardisation is often based on silicate glass CRMs, an approach that can result in significant matrix effects. These matrix effects were quantified by Steenstra et al.⁵⁶ for a range of trace elements using ns-excimer LA-ICP-MS. They compared results obtained using LA-ICP-MS with those using EPMA The measurements obtained with LA-ICP-MS consistently overestimated the concentration of volatile elements in metals relative to concentrations measured by EPMA. In contrast, the concentrations of non-volatile and refractory elements were systematically underestimated. To quantitatively describe these discrepancies, the fractionation index (F_i) for element i, or the ratio between the EPMA and LA-ICP-MS determined elemental concentration was calculated. The F_i values were found to be independent of concentration and type of Fe-rich alloy, and ranged from >0.14 for the most volatile elements to 1.8 for the most refractory elements. In addition, a positive correlation was found between the F_i and the 50% condensation temperature of the elements studied, suggesting that matrix effects were predominantly the result of ablation-induced evaporative and/or melting processes. The results suggested that neglecting matrix effects can result in partitioning errors for many volatile and refractory elements.

Metal imaging using LA-ICP-MS is an established methodology. However, problems can arise in the determination of metal concentrations from erroneous calibration, standardization, and normalization. In addition, LA-ICP-MS users tend to process their measurements by commercial processing software. Such programs typically visualise the regional metal differences in colourful and vivid imaging maps, but these might not represent the actual signal densities correctly. An overview of the software packages available for data processing for LA-ICP-MS imaging was provided by Weiskirchen et al.⁵⁷ They considered software packages, data routines, macros, programming tools, scripts, algorithms, self-written patches and updates for existing programs presently in use for obtaining LA-ICP-MS imaging data.

1.3.2.2 Electrothermal vaporisation. The effect of sheathing the sample aerosol, with H₂, N₂ or water vapour, on the analytical performance of SS-ETV-ICP-OES was evaluated by Al Hejami and Beauchemin.⁵⁸ Significant improvements in sensitivity and LOD were observed for 11 elements compared with conventional SS-ETV-ICP-OES. The greatest improvement in sensitivity and LOD was obtained with a N₂ sheath: sensitivity improved by between 1.2- and 4-fold, and LODs improved between 2- and 13-fold, depending on the element. The improvements in sensitivity and LODs with sheathing gas or water vapour did not degrade accuracy, as verified through the multi-elemental analysis of a soil CRM using external calibration with another soil CRM and internal standardisation. The Ar 763.511 nm emission line was used to compensate for sample loading effects on the plasma.

2. Instrumentation, fundamentals and chemometrics

2.1 Instrumentation

A variety of novel sources, or variations of existing sources, have come on the scene in the last few years. Ambient ionisation is a form of soft ionisation performed directly on the sample at atmospheric pressure, and can be performed in a number of ways. Bierstedt et al.⁵⁹ used a laser induced microplasma, sustained between a pneumatic nebuliser and the inlet capillary of an ion-trap MS, as the ionisation source. They found that analytes such as caffeine, 3-aminoquinoline and chlorpyrifos were fully desolvated and produced primarily molecular ions. Both positive and negative ions were formed simultaneously for the same operating conditions, and polar analytes with pronounced gas phase basicities were most efficiently ionised.

Miniaturisation of instrumentation brings advantages of portability, lower cost and reduced reagent and power consumption. Hence, this has been a constant theme in this review over several years. Xia et al.⁶⁰ developed a capillary liquid electrode discharge (CLED) coupled with both OES and VG-AFS. The discharge capillary was formed from a copper-coated capillary held concentrically inside a quartz T-tube, extending 1 cm beyond each end. One end was inserted into the discharge chamber and a tungsten electrode was also inserted opposite to face the capillary. The discharge was facilitated by a sheath gas of Ar and use of a 8 kV, 30 kHz power supply. Samples (10 μL) were pipetted onto a PTFE platform and were drawn into the capillary naturally due to the force arising from the solution vaporisation induced by microplasma. Initially, OES detection was performed, but the vapour generated by the discharge was also swept out of the discharge cell so that AFS could also be performed. Hence, the LDR was 10⁶ for a sample consumption of only a few μL. For both Hg and Cd, LODs of 10 μg L⁻¹ and ∼0.04 μg L⁻¹ were obtained using OES and AFS respectively. Recoveries of between 91 and 114% were obtained for blood samples which had been subjected to a micro-extraction procedure.

Swiderski et al.⁶¹ reported a miniaturised GD system coupled with a renewable hanging drop electrode (HDE). The GD was generated from the continuously forming drop in a 3.5 mm gap between the HDE and a tungsten pin electrode. The HDE was formed from a 4 mm i.d. quartz tube inserted into (but set back 0.5 mm) a 6 mm i.d. graphite tube. A dc power supply of 2 kV was used. LODs for Ag, Cd and Zn were 2.1, 2.4 and 6.0 μg L⁻¹, respectively. The system did suffer from analyte signal suppression in the presence of acidic matrices or EIEs, so suitable sample pre-treatment was necessary for the analysis of a pig kidney CRM, yielding 95% and 104% recovery for Cd and Zn respectively.

Tao et al.⁶² developed a dispersive UV spectrometer utilising a modified digital micromirror device (DMD), in conjunction with a diffraction grating, for use in AFS. When non-dispersive HCL-AFS is used there is potential for possible interference (e.g. between Se and Pb lines), so a dispersive system that has high transmittance in the UV is sometimes desirable. DMDs can be used to control the paths of dispersed spectral lines toward a single detector and have been used for FAAS and FAES. In this work the DMD was modified by replacing the cover window with one made from fused silica to provide higher transmittance between 180 and 320 nm. The LODs for the simultaneous detection of Se (at 203.9 nm) and Pb at (283.3 nm) were 0.89 μg L⁻¹ and 0.42 μg L⁻¹ respectively, using CVG sample introduction and HCLs for excitation. The authors correctly concluded that there is scope for improvements in sensitivity, perhaps by optimisation of the optical systems and the use of EDLs.

A liquid sampling atmospheric pressure glow discharge (LS AP-GD) has been developed and coupled with a variety of MS systems to use as a field-deployable system. Hoegg et al.⁶³ have coupled this system with an ion trap MS with the aim of producing an HR system with fewer interferences. This seems to have been achieved, with m/Δm > 400 000 readily obtained across the mass range used for elemental analysis. Precision of the ²³⁵U : ²³⁸U IR measurement improved up to a mass resolution of 1 000 000 and a precision of 0.086% RSD was obtained for U concentrations of 100 ng mL⁻¹. The LOD for U was 1 pg mL⁻¹ for a 60 μL injection volume. This satisfied the International Target Values (ITVs) for Measurement Uncertainties in Safeguarding Nuclear Materials set by the International Atomic Energy Agency (IAEA). Hence, in a second paper,⁶⁴ the same group evaluated the LS-AP-GD-IT-MS system for IR measurements, comparing the results with the more commonly used methods of TIMS and ICP-SF-MS, which are used when high precision results are required. For example, a ²³⁵U : ²³⁸U ratio of 4.266922 was obtained for the analysis of CRM U-800 (certified value = 4.265622), with a total combined uncertainty of 0.04%. This uncertainty compared well with TIMS (0.014%) and ICP-SFMS (0.08%), and was better than ICP-QMS (0.467%).

2.2 Fundamentals

2.2.1 Fundamental constants. The usual trickle of publications of fundamental constants has continued over the review period. Stark broadening of Pb II 220.35 nm and 438.65 nm lines was determined⁶⁵ for LIBS ablation of a Pb target in vacuum and in an Ar atmosphere at different delay times and pressures. A T_e in the range from 7400 to 20 600 K was determined over a range of argon pressures from vacuum to 12 Torr. Likewise, the n_e varied between 0.01 × 10¹⁷ and 1.3 × 10¹⁷ cm⁻³. dos Santos et al.⁶⁶ determined the impact parameter for Dy I emission lines in an LA plasma in Ar, and determined an n_e of 3.1 × 10¹⁸ cm⁻³ and T_e ∼ 10 000 K. They highlighted the incomplete data for impact parameters of heavy elements in the literature, which their method aimed to address.

The ionisation energy of ²⁰⁶Po was determined by Raeder et al.⁶⁷ using two-step laser resonance ionisation spectroscopy (RIS) to yield a value of the first IP = 8.418072(3) eV. In the same issue of the journal, Fink et al.⁶⁸ reported their result for the IP of ²⁰⁸Po to be 8.4180700(18) eV, obtained using RIMS. Both of these values were obtained with high precision compared to the existing literature. Vernon et al.⁶⁹ performed simulations of the relative atomic population distributions of ion beams of elements between atomic number 1 and 89, by calculating neutralisation cross sections using a semi-classical impact-parameter approach and simulation of the subsequent population decay. They then used collinear RIS to determine the actual relative populations of the 5s²5p ²P_1/2 and 5s²5p ²P_3/2 states of In to test their methodology, finding that the results were consistent with the calculations.

Urbina et al.⁷⁰ used a 3D Boltzmann plot method to determine atomic transition probabilities, using LIBS, for seven W I lines. After correcting the equations for self-absorption, values between 10 and 35% of the recent literature were determined.

Isotope shifts of 53 saturated absorption lines of Ti, in the wavelength range from 695 to 1005 nm, were measured by Kobayashi et al.⁷¹ Lines in b³F_J → y³D_J−1 (J = 2, 3, 4) were found to be approximately 15 MHz FWHM for ⁴⁶Ti, ⁴⁷Ti, ⁴⁸Ti, ⁴⁹Ti, and ⁵⁰Ti. These lines were split into Zeeman components of ΔJ = ±1 and, because the Zeeman components of the even-mass isotopes were only the components of M_J = 0, ⋯, ± J, the centres of those could be easily determined. The odd-mass isotopes had more complex hyperfine structure, so the Zeeman effect was calculated for the condition of an intermediate field approximation to determine the unperturbed line positions for ⁴⁷Ti and ⁴⁹Ti.

2.2.2 Diagnostics.
2.2.2.1 Plasmas. Modelling the spectroscopic transitions of atoms and ions in plasmas is a useful endeavour because it helps the researcher understand and predict the effects of operating conditions and concomitant species. To this end, Gupta et al.⁷² developed a collisional-radiative model for Ar/N₂ plasmas using fully relativistic electron impact excitation cross sections of Ar. In order to validate the model they compared their calculated intensities for three emissions lines at 750.4, 763.5 and 811.5 nm with calculated and experimental data from the literature. They also calculated the effect of varying the N₂ fraction from 0% to 10%, finding that T_e decreased from 1.0 to 0.43 eV and n_e varied between 2.3 × 10¹¹ cm⁻³ and 2.8 × 10¹¹ cm⁻³.

Wiess⁷³ performed a systematic analysis of Fe I and Fe II line emission spectra from an Ar and a Ne Grimm-GD. Excitation of Fe I occurred largely by electron collisions under non-LTE conditions, hence Boltzmann plots could not be used to estimate the temperature. A major ionisation mechanism in the Ar GD was found to be asymmetric charge transfer between Ar⁺ and Fe⁰ to form excited Fe II states. This was followed by radiative decay to low-lying Fe II then excitation by electron collisions to intermediate Fe II of ca. 13 eV (∼5 eV relative to the ground state), then emission from this state. In the Ne GD the formation of intermediate short-lived states followed the initial decay. A semi-quantitative empirical collisional-radiative model was also developed.

Matsuta⁷⁴ compared the CFS and AA spectra of O I 844.6 nm and Ar I 842.5 nm lines in pulsed and continuous rf-GDs. Different splitting behaviours were observed for the O I and Ar I lines in the CFS and AA spectra depending on whether the GD was operated in pulsed or continuous mode. It was not entirely clear from the paper what these differences were but e.g. in the pulsed mode the self-absorption profile of the CFS radiation revealed a Zeeman splitting of ∼500 MHz but the AA spectrum showed no splitting. This was attributed to the left- and right-handed circularly polarised CFS radiation for higher and lower off-resonance frequency.

Ariga et al.⁷⁵ studied the radial and axial distribution of M⁺ in the ICP with and without 5% v/v 2-propanol present in order to study the ‘carbon enhancement’ effect observed in ICP-MS. Their main finding was that the addition of 2-propanol shifted the maximum M⁺ intensity towards the sampling cone by ∼5 mm.

Guo et al.,⁷⁶ used time-resolved particle image velocimetry and 3D computer simulation to map particle trajectory, velocity, and residence time in a conical ICP torch. Key to their study was the use of spherical mesophase graphite particles of ∼9.41 μm diameter. The advantages of using these particles were: a Stokes number <0.1, which meant that they followed gas streamlines; high sublimation point and ionisation potential so particles remained intact during transit through the plasma; and intense orange C I emission around 589 nm which could be easily distinguished using a regular camera. The paper contains some useful photographic images demonstrating that particles travelling along the central channel were closer to the induction region in the conical ICP torch compared to a conventional design. The authors concluded that, for this reason, particles would be exposed to higher ionisation temperatures, and hence superior analytical performance for the conical torch. Maximum central gas velocities were 35 m s⁻¹ (at 7 L min⁻¹ flow rate and 900 W power) and 30 m s⁻¹ (at 15 L min⁻¹ and power of 1500 W) for the conical and conventional torches respectively. The conical torch also had a 500 K higher central temperature and a shorter sampling depth; the upshot of this being that it took between 277 and 500 μs for the carrier gas to reach the ionisation point in the conical torch, compared with between 606 and 2530 μs in the conventional torch, and 3× faster ionisation in the former compared to the latter.

In a separate paper,⁷⁷ the same workers evaluated the analytical performance of the conical torch for axial ICP-OES. They found figures of merit comparable to a conventional torch, but gas consumption was 56% less and power ∼50% lower using the conical torch. The effect of up to 5% NaCl concentration on Mg II (280.2704 nm) signal was also studied, and 15% less signal suppression was observed using the conical torch (at 900 W) compared with a conventional torch (∼1400 W).

2.2.2.2 Graphite furnaces. GFAAS has been suggested for the analysis of NPs by several authors. Brandt and Leopold⁷⁸ studied and compared the atomisation mechanism of Au ions and NPs as a step towards validation of the approach. The authors observed the parameters of atomisation delay (t_ad) and atomisation rate (k_at) and other peak characteristics (FWHM, peak asymmetry and appearance time (t_app)) and their concentration dependent trends to compare the mechanisms. Ionic Au^III displayed a rising trend in t_ad with increasing Au concentration, whereas t_ad was nearly constant for Au NPs over this concentration range. On the other hand, AAS peaks of ionic Au^III revealed constant t_app, while AuNPs exhibited a shift in appearance time. Moreover, peak asymmetry differed for ionic Au^III in comparison to Au NPs. These differences suggest different atomisation mechanisms involved in the evaporation of Au atoms introduced into the GF as either ionic Au^III solution or Au NP hydrosol.

2.2.3 Interferences. Axial ICP-OES offers lower LODs, but often at the expense of greater matrix effects, compared with side-on viewing. Carter et al.⁷⁹ tried to address this problem using signals from plasma background species of Ar, H and O with PCA and affinity propagation (AP) clustering, to evaluate the effects of complex matrices on ionic emission lines of Cd, Co, Cr and Pb. Matrix effects were quantified based on Euclidean distance in Principal Component space from an average calibration curve point. The statistical model was developed by running solutions with different levels of concomitants and monitoring background signals; then external standard calibration solutions and samples were run while monitoring both analyte and background signals. In operation, the principal advantage of the technique was that it could be performed on-the-fly, because the background species used to monitor changes in the plasma were simultaneously recorded with the analytical signals. Accurate results for seawater analysis were obtained using external standard calibration (EC) when the Euclidean distance was <1 for a given sample, but more work is required to make the method generally applicable.

Collision/reactions cells (CRCs) have now become standard additions to Q-ICP-MS instruments. The technology is less easily adapted for TOF-MS because of transmission losses. However, Burger et al.⁸⁰ developed such a system and used it to reduce background interferences due to Ar-based ions. They introduced H₂ as a reaction gas to suppress Ar⁺ and Ar²⁺ signals, thereby lowering the LODs for ⁴⁰Ca and ⁸⁰Se by 10³× and 10¹× respectively, but the formation of hydride, oxide and hydroxide molecular ions degraded the ability to quantify P, K and Sc. They also introduced microdroplets and used the high TOF-MS acquisition frequency to investigate the effect of the CRC on signal broadening, which occurred on the tens to hundreds of μs time scale. However, these changes in signal structure were deemed inconsequential for low-dispersion LA sample introduction where the ablation transient signal is comparatively long.

2.3 Chemometrics

Metrology is a subject which does not necessarily set the pulse racing, but it is fundamental to making accurate and traceable measurements. In the chemical sciences, analytical atomic spectrometry has led the way, as exemplified in a review (193 references) by Sargent et al.⁸¹ which covered the role of ICP-MS in inorganic chemical metrology. The review largely covered the contributions to the Consultative Committee for Amount of Substance: Metrology in Chemistry and Biology (CCQM) since 1993, during which time there were 56 international comparisons involving ICP-MS measurements undertaken by 16 different national institutes. After briefly reviewing key developments in the technique, such as interferences, sample introduction and IDA, there was a more in-depth analysis of the subjects of traceability, calibration and uncertainty estimation, and the particular problems associated with speciation analysis. The importance of IR measurements and ID-MS was also considered. The review speculated on the future metrological prospects for SP-ICP-MS, LA-ICP-MS and protein quantitation.

An important part of the metrological approach is the estimation of the uncertainty of a measurement. Fujiwara et al.⁸² evaluated the measurement uncertainty associated with a relatively new sample preparation technique, laser ablation in liquid (LAL), used to pretreat SiC prior to analysis using ID-ICP-MS. LAL was used to produce NPs by irradiating the surface of SiC, 3 mm below the surface of water, with a Nd:YAG laser. This had the effect of changing the bulk sample into many 200 nm size particles, increased the surface area and converted surface SiC into Si and amorphous C, making it possible to decompose SiC. Expanded uncertainty (k = 2) for the determination of Ti was <10% and 3.4% using external calibration and ID-MS respectively. The smaller uncertainty with ID-MS was speculated to be due to the rapid sampling interval (10 μs) which was faster than the frequency of signal variability.

Another novel calibration method, called multi-flow calibration (MFC), was developed by Williams et al.⁸³ The approach utilised a single calibration standard and multiple nebulisation gas flow rates. Analytical signals for the calibrant and samples were recorded at different flow rates and plotted on x and y axes. The analyte concentration in the sample was calculated by multiplying the standard concentration by the calibration plot slope. Accuracy was evaluated by determining Cr, Cu, Fe and Mn in drinking water, river sediment and tomato leaves CRMs, with recoveries between 91 and 112% and 84 and 134% for MFC and external standard calibration respectively. LODs were 20, 5, 7 and 2 mg L⁻¹ respectively.

3. Laser-based atomic spectrometry

Atomic spectrometry techniques where the laser is used as either an intense energy source or a source of precise wavelength (e.g. LIBS, LAAS and LIF) are considered. However, studies related to LA-ICP-MS/AES, and to the use of lasers for fundamental studies of the properties of atoms or for thin film deposition are not reviewed.

3.1 Laser induced breakdown spectroscopy (LIBS)

This section describes the latest instrumental developments and fundamental studies related to LIBS, but it does not cover detailed applications. Other recent reviews can be found on LIBS applications. For instance, a concise overview, by Jolivet et al.,⁸⁴ focused on elemental imaging and recent advances and future trends in LIBS applications to agricultural materials and their food derivatives were overviewed by Senesi et al.⁸⁵ LIBS of organic compounds was evaluated and reviewed by Moros et al.,⁸⁶ who highlighted that a broad variety of species may be formed depending on the irradiation conditions. In particular, the strong influence of the surrounding atmosphere on the spectral signature was discussed. The manuscript also reviewed the common emitting species populating laser plasmas of organic compounds, the routes to their formation, the dynamics of such species, the physical parameters that they confer to the plasma, and some details on the modelling of organic plasmas. A broad overview of LIBS was reported by Costa et al.,⁸⁷ including origin of LIBS, fundamentals, different sample preparation strategies, use of chemometric tools for data processing, advantages and limitations with regard to qualitative and quantitative analysis, most relevant applications and future trends.

In the review year several international conferences were dedicated to recent progress in LIBS. In particular, the 10^th Euro-Mediterranean Symposium on LIBS held in Brno, Czech Republic on September 8–13 (http://www.emslibs.com/), and the 3^rd Asian Symposium on Laser Induced Breakdown Spectroscopy held in Jeju, Republic of Korea, August 27–30 (http://www.aslibs2019.org). A panel of several experts, with a variety of backgrounds, commented on recent developments and on the potential LIBS transition from an experimental and lab-based method to one that is increasingly found in commercial and industrial use.⁸⁸ The usefulness of LIBS to rapidly create an HR elemental map on large 2D surfaces was highlighted. However, a practical device where data can be interpreted with chemometric algorithms for autonomous classification of the surfaces is still missing.

3.1.1 Fundamental studies. Spatio-temporal distributions of atomic and ionic emission line signals were investigated⁸⁹ in the early stages of a laser-induced plasma, produced after the ablation of a zeolite-like potassium gallosilicate, at low pressure conditions and using an IR CO₂ pulsed laser. A lower continuum background was observed at LIBS operating conditions, which permitted detection of species in higher ionisation states or elements with short lifetimes, such as oxygen. Stark broadening was used to measure plasma electron density, which was found to decrease by one order of magnitude (from 10¹⁷ to 10¹⁶ cm⁻³) at 2 cm from the sample surface. The plasma was further characterised in terms of electron temperature and the average expansion velocity of various plasma species. Time scales for the onset of atomic and molecular emission in a fs laser-induced plasma were measured by Carrasco-Garcia et al.⁹⁰ The fs time-resolved images were acquired using a fs pump(@800 nm) – probe(@400 nm) microscope in combination with OES. A time-bracketing procedure was employed to extract information from very narrow time windows. Ejection of material from silicon was observed several ns before any photon emission was detected. This time gap was noticed to be material dependent; in particular, the timing difference in the emission of the CN band at 388.1 nm in organic samples was larger. Different samples (e.g. nylon and Teflon), where formation of the CN species was considered via direct release from the target or by gas-phase recombination in the plasma, respectively, were investigated.

Two algorithms, based on the exploitation of self-absorbed lines for the characterisation of LIBS plasma, were critically evaluated by Safi et al.⁹¹ These methods were the columnar density Saha–Boltzmann plot method of Cristoforetti and Tognoni (G. Cristoforetti, and E. Tognoni, Spectrochim. Acta, Part B, 2013, 79–80, 63–71) and the C-sigma model of Aragon and Aguilera (C. Aragon, and J. A. Aguilera, J. Quant. Spectrosc. Radiat. Trans. 2014, 149, 90–102). After evaluation of their advantages and drawbacks, the authors developed an extended C-sigma approach that allowed a rapid and more accurate fit of the experimental data, as the approach was also able to deal with lines with different Stark broadening.

Calibration free (CF)-LIBS is a desirable aim for most practitioners. A novel method that improved the sensitivity of CF-LIBS for the analysis of food samples was developed by Chen et al.⁹² The method was based on the acquisition of two spectra recorded with different delays after the excitation laser pulse. Major and minor element fractions were determined at short delays, when the plasma was in full LTE, while minor and trace elements were quantified from the spectra collected at longer delays, under conditions of partial LTE. In the latter case, analytes in organic matrices were out of equilibrium due to the reduced electron density, whereas the metal atoms had Boltzmann equilibrium distributions. The SNR of element emission lines was improved due to lower continuum radiation at the selected temporal acquisition window, as a consequence of the reduced collision rates of charged particles. The method was validated for the analysis of three types of most consumed Mediterranean seafood species, sampled from the bay of Marseille.

A self-absorption correction method for CF-LIBS, called blackbody radiation referenced self-absorption correction (BRR-SAC), was proposed by Li et al.⁹³ An iterative algorithm, utilising a comparison of the measured spectrum with the corresponding theoretical blackbody radiation, was employed to calculate the plasma temperature and the collection efficiency of the experimental set-up.

The temporal stability of LIBS signals was investigated by Fu et al.⁹⁴ using time-resolved spectroscopy and a fast photography technique. Plasma morphology was found to be much more repeatable at the early stage of plasma evolution. The LIBS spectral signal was more stable for delay times of ∼800 ns, with a deteriorating shot-to-shot signal fluctuation as the plasma continued to evolve. Based on the spectral data from a Ti alloy plasma, relative contributions of plasma temperature, electron density, and total number density on several Ti II lines were determined. Fluctuation of the number density was found to be the key parameter affecting the signal repeatability. The authors proposed that the creation of a big and stable plasma core (e.g. by better cavity design, shaping the laser beam, or designing a particular gas mixture) was fundamental to suppression of spectral intensity fluctuations.

3.1.2 Instrumentation. There has been a spate of reports of LIBS coupled with a variety of other techniques. Al Shuaili et al.⁹⁵ developed a pulsed-microwave system, operated at 2.45 GHz, to improve the sensitivity (92-fold) and LODs (8-fold) of ns-LIBS for the detection of Pd in solid samples at ambient conditions. The microwave radiation was directed to the sample surface using a near field applicator with a pointed tip, which was located 45° at a vertical and horizontal distance of 0.5 mm from the ablation site. Optimisation studies were performed showing that this MW-LIBS signal strength was more significantly affected by the vertical position of the applicator, obtaining maximum SNR at a microwave power of 750 W and laser pulse fluence of 157 J cm⁻² for the Pd I 340.46 nm line. The microwave pulse of 1 ms was triggered prior to the laser pulse. The LOD for Pd improved from 40 ppm to 5 ppm. The analytical potential of MW-LIBS was demonstrated by Wakil et al.⁹⁶ for quantitative detection of Cl on a cement surface, via molecular emission from CaCl. Following a 1.5 ms microwave pulse, an optimal gate-width and gate-delay was arrived at to collect emission spectra of CaCl at 617.9 nm. The LOD was a 10-fold improvement.

LIBS and RAMAN has been combined by Allen and Angel.⁹⁷ They used a spatial heterodyne spectrometer (SHS), employing Fresnel collection optics (∼100 mm diameter Fresnel lens), for combined LIBS and Raman remote analysis at 10 m distance. A fast collection lens (Fresnel lens) and a small SHS were both included in a 1 U CubeSat architecture (developed by NASA). Performance for the analysis of organic and mineral samples, including materials from deep ocean hydrothermal vents, was compared to that achieved with a high quality ∼100 mm aperture Questar long-range microscope under identical conditions. Spectral intensities were reduced by a factor of ∼4 using the Fresnel lens but the overall quality of spectra was similar, encouraging for the development of instruments designed for spacecraft and planetary landers. The main advantage of this spectrometer was that it is based on a fixed diffraction grating interferometer with no moving parts, but offers a very large field of view, high light throughput, and high spectral resolution in a small package.

LIBS and QMS has been coupled in an in situ K–Ar dating instrument (KArMars).⁹⁸ It was based on the use of ns-UV-LIBS under ultra-high vacuum conditions for the determination of K (by measurement of K emission lines at 766.49 and 769.89 nm), and combined with a used QMS for Ar determination in reference samples, for calibration of ablated mass. The sample chamber was isolated during ablation and the released gas was purified before introduction into the QMS where EI of the gas atoms was achieved using a thoriated iridium filament source. Ablated mass was successfully determined by QMS analyses of a set of reference samples for which the ⁴⁰Ar content had been previously precisely determined. The K–Ar ages obtained were used for dating a large compositional range of Martian rocks with uncertainties as low as 10%, and with high accuracy.

LIBS and acoustic monitoring at ambient pressure was applied by Chide et al.⁹⁹ to study the shot-to-shot evolution of crater morphology and plasma emission. The study was part of the SuperCam/Mars 2020 microphone investigation, and employed eight different geological samples with varying chemical and physical properties of interest. A decrease in the acoustic energy as a function of the number of shots was well correlated with the target hardness and density. It was demonstrated that the acoustic energy could be employed as a remote tracer for the ablated volume of the target.

Double pulse (DP)-LIBS using a combination of circular and annular nanosecond laser pulses, was developed by Hai et al.,¹⁰⁰ to enhance sensitivity through better coupling of laser energy with the target. An optimised delay between the two laser pulses of ∼10 μs, using first the annular laser pulse, provided signal enhancement of ∼ 4× and a ∼3× improvement in LODs for the determination of Li and Mg in aluminum alloy, compared to irradiation using both pulses simultaneously. This method might be effective for in situ and minimally destructive LIBS analysis, with potential applications for trace impurity determinations in walls of fusion devices.

3.1.3 Novel LIBS approaches. The selection of calibration samples for the LIBS Mars surface composition detection package, to be included in the first Mars Global remote sensing and regional survey mission of China, was discussed by Cai et al.¹⁰¹ They considered the importance of the calibrants in order to ensure reliability of the returned data. Babos et al.¹⁰² evaluated several univariate calibration strategies for Ca and P determination in mineral supplements, in order to reduce matrix effects usually observed in LIBS analysis. Multi-energy calibration, based on the analysis of two calibration standards for each sample (e.g. sample + blank and sample + standard using the same amount of unknown sample), and the measurement of several emission lines (e.g. 5 ion lines for Ca and 4 atom lines for P), provided good recoveries compared to ICP-OES. Successful results were also achieved using a one-point gravimetric standard addition method, which was based on the use of the same two calibration standards but only required the measurement of one emission line per element (e.g. Ca II at 396.85 nm and P I at 213.62 nm).

Robust and reliable results in LIBS imaging analysis require an effective signal extraction method. Motto-Ros et al.¹⁰³ tested the performance of three extraction methods with regard to linearity, dynamic range and operating speed. As a result, they proposed a new conditional data extraction procedure, which was able to provide the best line intensity extraction and to display only those pixels that were statistically significant in a robust, fast and unsupervised manner. This was based on the use of two spectral windows covering the line of interest and the surrounding background.

Palleschi et al.¹⁰⁴ used an extended Kalman filter, employed in robotics and automation for predicting the evolution of noisy systems, for the multivariate non-linear analysis of LIBS spectra. The method was considered to be more robust and simpler than those base on artificial neural networks. A method named particle swarm optimisation-support vector machine (SVM) was used by Chen et al.¹⁰⁵ for the quantitative determination of Cr in pork by analysing the spectral data between 424.77 and 429.99 nm. The optimisation method was said to be easy to implement, with fewer parameters to be adjusted and less chance of a local minima. Hence, it might be applied to solve non-linear, non-differentiable and multimodal problems. In the study it was used to optimise the hyper parameters of the SVM model (i.e. the regularisation and kernel parameters), which avoided over- or under-fitting. Good agreement to the results by AAS was achieved, and the model was demonstrated to improve robustness and prediction accuracy in comparison to other models.

A calibration model based on PLS and support vector regression correction for C analysis in coal was developed by Dong et al.¹⁰⁶ They studied the importance of the atomic spectra of C and the molecular spectra of CN and C₂ for the quantitative analysis of the C content, which helped reduce matrix effects from the coal samples, and improved accuracy.

Spectral interferences are a perennial problem regardless of the spectroscopic technique. Chappell¹⁰⁷ evaluated spectral line interferences using a statistical interference factor (SIF), which provided a statistical measurement of how much a line was interfered with. This was calculated for each spectral line in an experimental spectrum using an algorithm based on an existing database and Bayesian statistics. This approach could be applied with or without knowledge of the LIBS plasma and sample. The approach was demonstrated on spectra from a pure silicon sample and from NIST SRM610 glass, where it was possible to select reliable signals for elemental analysis.

Background estimation and subtraction was the subject of a study by Kepes et al.,¹⁰⁸ who considered whether application of a numerical treatment of LIBS background signals, originating from various sources, might lead to loss of information or to the introduction of spectral artefacts. The effects of different background estimation and subtraction algorithms on LOD, background RSD and SBR were studied. Results showed that there was a threshold gate delay, above which the numerical treatment of spectral background must be applied cautiously. The optimal measurement parameters and feature selection (selection of the emission line that yields the best results) were also found to depend on the method of spectral processing.

Dell’Aglio et al.¹⁰⁹ analysed amyloid fibrils coated with gold NPs using so-called nanoparticle enhanced (NE) LIBS. Amyloid fibrils are used e.g. for water purification, medical applications, and investigating the performance of laser-matter interaction in biological systems. They used NE-LIBS to determine trace elements in low sample volumes (2–3 μL) of solution using amyloid protein fibril systems. Likewise, Hao et al.¹¹⁰ demonstrated that the combination of NE-LIBS with magnetic field confinement was effective for the enhancement of the emission signal and SNR in a copper sample.

Colloidal particle lens array (CPLA) assisted fs-LIBS was developed by Wang et al.¹¹¹ They undertook a multiscale visualisation of a silica colloidal particle lens array in the fs-LIBS of copper samples. The visualisation included time-resolved plasma expansion, shockwave propagation, plasma plume emission, and nanoparticle distribution. They found that increased plasma temperature was the main reason for enhanced signal in CPLA-fs-LIBS.

Improved optical emission of Si I at 390.55 nm was achieved by Guo et al.¹¹² They increased the temperature of samples of silicon from 25 to 250 °C in a 6 mm i.d. by 6 mm long cylindrical confinement cavity. This effect of temperature was also studied by Lednev et al.,¹¹³ who heated the sample surface in an argon atmosphere using a powerful continuous wave fiber laser, while a ns laser ablated the high temperature area. Experiments were performed over a wide temperature range, from 25 to 1600 °C. Plasma imaging showed that size and brightness of the expanding plume was not affected by solid sample temperature, but increased during molten sample ablation. They proposed that enhanced plasma emissivity was mainly attributed to greater ablation rate.

3.2 Laser atomic absorption spectroscopy (LAAS)

Pseudo-continuum source AAS was used by Merten et al.¹¹⁴ to map neutral atomic populations, and their evolution, in a LA plasma of ablated Ti. A clear resolution of the plasma’s spatial structure was obtained. The authors considered that this method could be very useful for measuring the ablated mass in samples where crater imaging is impossible (e.g. powders or samples prone to spallation) or measurements that do not allow repeated ablation required to generate measurable craters. The method was proposed to be particularly suited for understanding chemical reactions in the cooling plume and quenching in LA-LIF measurements.

3.3 Cavity ringdown spectroscopy (CRDS)

Pal et al.¹¹⁵ used a continuous wave external cavity quantum cascade laser coupled to HR-CRDS to simultaneously determine three stable isotopes of hydrogen sulfide containing ³²S, ³³S and ³⁴S. This was achieved in a single laser scan, with high sensitivity and molecular selectivity, and within a very narrow spectral range of 0.05 cm⁻¹. The method was considered to have great potential for understanding S isotope fractionation in geochemistry and environmental sensing applications.

3.4 Laser induced fluorescence (LIF)

Vacuum UV LIF of plumes is considered as an ideal technique for minimally destructive multielement analysis. De-noising is a key process that was investigated by Cheung,¹¹⁶ using a pre-processing scheme that involved three steps: rejection of dim, featureless spectra as outliers; removal of a wobbly baseline by mean subtraction; and reduction of fluctuations by normalising the spectral area to a constant. Signal uncertainty was reduced from 54 to 6.5% and clean sorting of ink samples based on single-shot spectra was demonstrated.

4. Isotope analysis

LA-MC-ICP-MS was the predominant technique to push the boundaries of IRA, with applications in nuclear forensics, geosciences and biosciences. Lowering LODs, increasing precision and accuracy, and measuring smaller samples were the focus of a number of studies on U, Os, Li and B. Another advance was the progress of high-impedance resistors (10¹² Ω and 10¹³Ω) from the research lab into useful applications, with solutions being found to address the slow response times of their amplifiers resulting in their useful incorporation into the multi-collector arrays for the detection of minor isotopes.

4.1 Reviews

Double-spiking or poly-spiking is now a frequently used method to accurately calculate the instrumental mass fractionation of a sample in situations where only one non-radiogenic isotope is available. The method can also be used to determine mass dependent isotopic changes generated by natural processes, or investigate the enhancement of particular isotopes generated by anthropogenic nuclear processes. A timely review of the practical considerations of the double spike technique was provided by Klaver and Coath,¹¹⁷ covering the subject from first principles. This included illustrating the double-spike deconvolution with schematic diagrams, and providing key equations for data reduction. The review also covered error propagation of the double-spike mixtures, optimisation of the double-spike composition, and the best strategy to calibrate a double-spike relative to a reference material. The authors touched on the possible utility of systems where more than three IRs are available, and the potential for regression of a higher number of dimensions to further improve spike performance.

Instrumentation for MC-ICP-MS has continued to develop as isotopic applications have diversified. Scott et al.¹¹⁸ reviewed the use of retardation energy filters for U-series IRA. They covered the theory and practical aspects of filters such as the retarding potential quadrupole (RPQ) lens deployed on a commercial instrument; particularly its effect on reducing the background caused by peak tailing from adjacent abundant isotopes on the nuclides ²³⁴U, ²³⁰Th, ²²⁸Ra and ²¹⁰Pb. In addition, the study presented a measured, compiled set of new recommended values for U, Th and Ra concentrations and ²³⁰Th : ²³²Th, ²²⁶Ra : ²³⁰Th ratios in rock reference materials.

4.2 New developments

Wang et al.¹¹⁹ measured Sr isotope ratios using total evaporation TIMS. They evaluated each factor within the total evaporation technique in terms of its contribution to the overall measurement of the error. Triple filament ionisation was found to provide the most stable total evaporation ratios regardless of the sample load. Key factors identified in accurate Sr isotope determinations were ion loss before the onset of data acquisition and variation in the ion transmission efficiency. It is noted that this study used an ⁸⁷Sr : ⁸⁶Sr reference value of 0.710340 for NBS 987 rather than the value used more commonly in the geosciences of 0.710248. A new emitter for Sr isotope analysis by TIMS was developed by Li et al.¹²⁰ This was instigated to yield precise ⁸⁷Sr : ⁸⁶Sr ratios for analyte levels of <100 pg Sr. Currently, the most commonly used emitter for Sr is one based on TaF₅ or TaCl₅ with phosphoric acid, which provides a sufficient ion beam to generate precisions of 0.004% (2s) RSD for as little as 500 pg of Sr. A 22 mg mL⁻¹ solution of silicotungstic acid on a Re filament pre-treated with H₃PO₄ was used. It was estimated that the emitter was around three times more efficient in ionisation than the TaF₅ based activator, with a 150 mV ⁸⁸Sr ion beam sustained for >20 min. This resulted in 0.0034% (2s) RSD, when using the ‘standard’ 10¹¹ Ω resistors with Faraday cups, for 100 pg Sr and 0.0094% for 30 pg. This suggests that further improvements in precision could be obtained using higher impedance amplifiers with this new emitter.

A new method for IDA for monoisotopic elements was presented by Narukawa et al.¹²¹ They used ICP-MS fitted with a dynamic reaction cell which was supplied with isotopically labelled O₂. This generated ⁸⁹Y¹⁶O⁺ and ⁸⁹Y¹⁸O⁺ molecular ions which were measured alongside ¹⁰⁷Ag⁺ which is isobaric with ⁸⁹Y¹⁸O⁺. The concentration of Y was calculated from the known abundances of ¹⁶O, ¹⁸O and ¹⁰⁷Ag with the measured intensities at m/z 105 and 107. It was noted that this technique has the potential to also be applied to poly-isotopic elements and additionally may permit simultaneous multi-element IDA.

A potential advance in the separation of Nd from difficult high-Mg rock matrices was presented by Pin and Gannoun.¹²² The method minimised residual fluorides, present after dissolution, by treating with boric acid, and chemically reducing Fe with ascorbic acid. Column chemistry was restricted to 6 steps, including two tandem-column separations: the first using TRU resin and DGA resin to isolate Nd from the matrix and La-Ce-Pr, the second eluting Nd from the DGA column to an Ln2 resin which further purified Nd from all adjacent REEs. This method was successfully tested on a series of mafic and ultramafic rock standards producing repeatable measurements of the stable ¹⁴²Nd : ¹⁴⁴Nd and radiogenic ¹⁴³Nd : ¹⁴⁴Nd ratios. Improvements in Nd ionisation efficiency in TIMS was the subject of a paper by Shao et al.¹²³ who developed a new porous ion emitter made from a slurry of Pt and Re powders. Dissected fragments of the resulting ‘cake’ were sintered onto Re filaments prior to sample loading. Results for Nd⁺ ionised from this emitter indicated that ion yields were an order of magnitude higher than most other published studies using between 1 and 100 ng Nd loads.

4.3 Radiogenic isotope ratio analysis

Ce isotope measurement was investigated by Bonnand et al.¹²⁴ This recent study provided a comprehensive examination and deconvolution of the effects of tailing of ¹⁴⁰Ce into ¹³⁶Ce and ¹³⁸Ce using 10¹³ Ω amplifiers, and the oxide corrections required for isobaric interference from Ba, Nd and La. The new methodologies utilising a ¹³⁶Ce–¹³⁸Ce–140Ce triple-spike were used to deconvolve mass fractionation caused by chemical separation and instrumental bias. The radiogenic ratio ¹³⁸Ce : ¹⁴²Ce was measured with a reproducibility of ±25 ppm during a static collector configuration, and was possible on ∼700 ng mass of Ce. Application of the triple-spike method to correct for mass dependent Ce isotope variations produced accurate ∂¹⁴²Ce measurements, providing that Ba overlap was minimised by chemical separation.

Hf isotope measurements in geoscience samples has been considerably refined by the use of MC-ICP-MS but lacks calibrated measurement standards for the IRs. Tong et al.¹²⁵ used calibrated Re isotopic material analysed with a series of Hf standards to independently determine mass fractionation. The study proposed a ¹⁷⁹Hf : ¹⁷⁷Hf value of 0.7334 ± 0.0006 for JMC-475 which contrasted with the arbitrary fractionation correction value of 0.7325 originally proposed in 1983.

Os isotope ratios in geological reference materials were determined using MC-ICP-MS by Wang et al.¹²⁶ They compared 10¹³Ω amplifiers and secondary electron multipliers for measuring the ¹⁸⁷Os : ¹⁸⁸Os ratio. For samples with >80 pg Os, the Faraday collectors gave the best precision, but for <36 pg the electron multiplier was superior. The standards WPR-1 and BIR-1 were found to be the best Os isotopic reference materials because they gave the most consistent results. The ¹⁸⁷Os : ¹⁸⁸Os ratio was also investigated by Zhu et al.,¹²⁷ in sulfides trapped in mantle peridotites. Sulfide grains and inclusions were analysed using LA-MC-ICP-MS with an array of ion counting detectors. These were calibrated for their different efficiencies by dynamic peak jumping while monitoring Au isotopes. Sample-standard bracketing was used to correct for signal drift and mass fractionation. Uncertainty was primarily influenced by the proportion of Re interference, such that for ¹⁸⁷Re : ¹⁸⁸Os ratios <0.055 and <0.75 precisions were better than 0.8% and 5% respectively. This was satisfactory for discriminating between the evolution of the sub-continental lithospheric mantle.

An assessment of utilising small LA spot sizes for fine spatial resolution zircon geochronology was presented by Mukherjee et al.¹²⁸ who used a single-collector SF-ICP-MS for their measurements of the ²⁰⁶Pb : ²³⁸U. This was observed to increase during ablation pit deepening, and varied differently between the assorted zircon standards used. This, along with matrix-induced instrumental mass fractionation, was considered the key source of uncertainty in U–Pb ages. It was found that if only data from the first 10 to 15 s of ablation was evaluated, the restricted down-hole fractionation resulted in accurate and precise ages of better than 1.4%.

4.4 Stable isotope ratio analysis

Advances in analytical innovation have a history of being driven by the needs of explorative planetary science. In the case of Avice et al.,¹²⁹ the requirement was for precise heavy noble gas isotope measurement in a compact and light instrument suitable for space craft deployment. Using a quadrupole ion trap mass spectrometer (QIT-MS), their study demonstrated that ⁸⁶Kr : ⁸⁴Kr and ¹²⁹Xe : ¹³⁰Xe could be measured with precision better than 1‰ and 1% respectively given several hours of acquisition. Such a system would have the sensitivity to measure noble gases collected in the Venus atmosphere.

Studies of stable K isotopes by ICP-MS are generally hampered by Ar⁺ and ArH⁺ isobaric interferences generated by the plasma source. One method of minimising these effects is to use a collision gas (He) or a reaction gas (H₂ or D₂) to dissociate or react with molecular ions. Another method is the ‘cold plasma’ technique which reduces isobaric ArH⁺. Chen et al.¹³⁰ used this latter method to attenuate and resolve ⁴¹K from ⁴⁰Ar¹H⁺ using MC-ICP-MS, thus achieving a long-term reproducibility of ∂⁴¹K of better than ±0.11‰.

High-precision Mg isotope determinations are now routine by MC-ICP-MS. Gou et al.¹³¹ studied the effect of type of sampling and skimmer cone on the analytical systematics. They concluded that the cone combination did not affect the peak shapes of Mg ion beams, but ‘Jet’ and ‘X’ cones enhanced beam intensities by 1.9 relative to standard cones. Different cone combinations were found to affect the mass fractionation factor, with the ‘Jet’ plus ‘H-cone’ combination generating the least mass discrimination but with good sensitivity.

Hirata et al.¹³² measured Pt isotope ratios in Pt–Au NPs of between 10 and 100 nm, using MC-ICP-MS combined with an “LA in liquid” technique. They also used high-speed data integration, with approaching 100% time-integration efficiencies from high-gain ion collectors. The ¹⁹⁵Pt : ¹⁹⁴Pt ratio could be measured with an uncertainty equivalent to the counting statistic error. Hu and Teng¹³³ also assessed precision attainable using MC-ICP-MS. Their study focussed on Li, Mg, Cu, Zn and Fe measurement, and examined the behaviour of isotopes in measured and sample-standard corrected ratios. Fluctuations in stable IRs were examined in response to changing matrix acid strength, analyte concentration and elemental matrix variations. A key conclusion of the study was that ‘wet’ plasma conditions and normal RF power provided the most repeatable data for all of the isotope systems examined. Matching acid molarity and concentration between the sample and standards was important for most systems except Li. Matrix elements affected IRs, with Na causing the most significant mass independent Mg isotope fractionation.

Huang et al.¹³⁴ used a binary mixing technique to determine Hg IRs in samples with <1 ng mL⁻¹ Hg. This was tested by the standard addition of the CRM NIST SRM3133 to natural samples and extrapolation from the measured mixture ratio to the natural ratio. Results of the binary mixing were compared to direct analyses of the natural samples. Following uncertainty calculations for the binary mixing scheme, it was estimated that precise δ²⁰²Hg (±0.1‰) could be achieved on samples with 0.9 ng mL⁻¹ Hg, which was ∼3× lower than using direct measurements. Queipo-Abad et al.¹³⁵ also measured Hg isotopes, but with GC coupled with MC-ICP-MS. They proposed a new method for correcting the time lag effects of transient signals entering Faraday cups, which is a known problem for transient sample introduction such as GC, LC and LA. Application of the method to narrow GC peaks resulted in the same precision and accuracy for Hg isotope ratios (<0.4‰ (2s) RSD) as wide peaks, opening the possibility of shorter analysis times and higher chromatographic resolution.

An interesting assessment of double spike measurement and calibration was made by Zhang et al.,¹³⁶ who used theoretical calculations to assess how necessary it is to have a precise knowledge of the mass fractionation factor of the spike. They found that if the double spike was not corrected for fractionation it did not introduce any systematic bias into the measurement of δ¹³Cr, providing all measurements and standards were completed using the same instrument and measurement protocol. Double-spiking and measurement of Cd isotope ratios to a high precision was reported by Li et al.¹³⁷ They used MC-ICP-MS and sample-standard bracketing with double-spike approach to correct for mass fractionation. A one-stage anion exchange resin purification resulted in a high Cd yield of ∼98% that eliminated the spectral interferences caused by Sn in the sample. Intermediate precision for NIST SRM3108 over a period of three years determined for the ∂¹¹⁴Cd : ¹¹⁰Cd ratio was ±0.09‰ (2s) RSD.

Stable isotope fractionation in the geological environment, and its potential for systematic mass fractionation correction of isotope ratios, has been a subject of significant investigation since the introduction of MC-ICP-MS. Lee and Tanaka¹³⁸ investigated the potential for Eu isotopic fractionation in natural materials. Chromatographic isolation of Eu from matrix and other REE elements caused column fractionation and a progressively lighter Eu isotopic fraction through the course of elution, with ∂¹⁵¹Eu : ¹⁵³Eu changing by >4‰. Samarium was added to the post-separation Eu fraction to assess mass fractionation within the sample run using the ¹⁵⁰Sm and ¹⁵⁴Sm isotopes to bracket the Eu isotopes. Samarium was also used in bracketing solutions to compare the internal and external corrections. Results of the analysis of geological standards produced comparable ∂¹⁵¹Eu : ¹⁵³Eu values, but slight differences were observed between the results for different rock groups using the internal Sm correction technique, indicating that Eu fractionation may occur in the geological environment.

Advances in the measurement of boron isotopes has opened a pathway for detailed investigations into their behaviour during mantle recycling and in volcanic systems. Li et al.¹³⁹ measured B isotopes in volcanic rocks which had undergone variable degrees of alteration by seawater and hydrothermal systems. Their precision for ∂¹¹B using MC-ICP-MS was reported as better than 0.5‰ (2s) RSD. Leaching experiments on the rock powders indicated that the acid attack did not change the B isotope composition of the rock. When acid leaching was applied to altered samples it was found that 6 M HCl efficiently eliminated the effects of low degrees of alteration and adsorption from fluids. Restite minerals, predominantly plagioclase and clinopyroxene, were found to have pre-eruptive B isotope compositions after leaching.

Boron isotopes are also important in providing information about the composition and evolution of deep water masses, in particular in reconstructing paleo-pH. Deep-sea corals and benthic foraminifera were a taken for B analysis by Sadekov et al.,¹⁴⁰ chosen because they exhibit millimetric-scale biogenic carbonate growth during their lifetime. Using 10¹³Ω amplifier technology and careful monitoring of the on-peak interferences (e.g. Ar⁴⁺) and response time of the collectors, their LA-MC-ICP-MS system produced ±0.15% (2s) RSD precision. Interesting examples of B isotope patterns across biogenic carbonates showed the potential to resolve monthly changes in ocean pH. In a similar study, Standish et al.¹⁴¹ investigated the use of matrix-matched standards, the effect of isobaric Ca⁴⁺ and the characterisation of the slope of Ca interference. This was applied to the LA-MC-ICP-MS measurement of the B isotope systematics of corals from Southern Belize and foraminifera from Ocean Drilling Program Site 999.

Steinmann et al.¹⁴² also used LA-MC-ICP-MS, in their case to determine the optimum methodology for Li isotope analysis at low concentrations in geological materials. The laser was strategically focused some 130 μm beneath the sample surface to avoid thermal effects, melting the crater rim. Ablated material in an He carrier gas was passed through a 50 cm³ mixing tube to enhance signal stability. This was found to be beneficial when using the 10¹³Ω amplifiers to capture the smaller ⁷Li⁺ ion beam. Overall precision and accuracy on δ⁷Li was estimated as 2‰ at μg g⁻¹ Li concentrations. The potential of this was demonstrated by analysing the ∼15‰ isotopic changes across a zoned olivine crystal. Zhu et al.¹⁴³ used a tandem-column ion-exchange method to separate Li for isotopic analysis. Using a small upper column and a longer lower column to separate Li from Na, the need for an evaporation step was eliminated and matrix-induced shifting of the Li elution was prevented. The authors also optimised the signal stability of the Li⁺ rather than the signal intensity; the best zone for Li IRA being further from the sampling inlet than the point of maximum sensitivity.

4.5 Nuclear forensics

Nuclear non-proliferation testing requires a rapid, field-portable instrument to make an accurate initial assessment of the proportion of enriched uranium. Bartlett and Castro¹⁴⁴ demonstrated the potential for absorption spectroscopy to measure U isotopes. They heated U₃O₈ or UF₄ with Lu in a Ta crucible and then produced atomic beams of uranium down to enrichment levels of 0.7%. Minimal sample preparation and chemical waste of this technique indicated that this method could be a method of choice for field assessment.

Krachler et al.¹⁴⁵ determined the ²³⁵U : ²³⁸U ratio within single UO₂ crystals. They used LA to direct a series of 1 Hz laser shots at ∼10 s intervals into the UO₂, then time-resolved MC-ICP-MS to monitor the U ion current. Measurement of the ²³⁵U : ²³⁸U ratio with progressive deepening of the hole indicated there was little isotopic fractionation during 80 shots. Precision of ∼3–6% was achieved for the enriched surface region, and ∼12% in the deeper depleted uranium. However, a well-defined depth versus pulse rate calibration was not possible due to variable crystallinity with depth affecting the LA rate. Small-scale U isotope analysis was also investigated by Strashnov et al.¹⁴⁶ used resonance excitation TOF-MS on ∼20 μm spots, with better than 7% precision for depleted U samples with <80 fg of U.

Uranium isotope ratios were determined on sub-micrometre-sized particles (280–600 nm), using LA-MC-ICP-MS, by Ronzani et al.¹⁴⁷ Uncertainties reported for ²³⁵U : ²³⁸U were better than 2% for 90% of the particles, and better than 20% for the ²³⁴U : ²³⁸U ratio. A rigorous cleaning regime and ‘wet’ plasma conditions contributed to an LOD <1 ag for ²³⁴U, ²³⁵U and ²³⁶U. Uranium IRA was also the subject of a study by Forbes and Szakal,¹⁴⁸ who evaluated a hybrid QTOF mass spectrometer for the analysis of U solutions. They work concluded that the method had the potential to achieve accurate ²³⁵U : ²³⁸U IRs for UO₂⁺ down to tens of pg of solubilised uranium oxide, by using small, discrete injections of highly charged species which may enhance ion transport and reduce matrix effects.

Traditional TIMS measurements of U isotopes at low concentrations or in small samples is limited by the ionisation efficiency from the TIMS filament. Trinquier et al.¹⁴⁹ demonstrated >5% efficiency of a recently developed cavity source TIMS. In this technique a resin bead loaded with sub-nanogram amounts of uranium was mounted in a conical Re sample holder, capped by a Re cavity tube heated by electron impact. Initial results suggested that, after further developments in the ion pathway from the cavity source to detectors, the reported order of magnitude improvement in sensitivity will ultimately lead to precision of ²³⁵U : ²³⁸U of <0.5% (2s) RSD for 100 pg U.

Fenske et al.¹⁵⁰ developed an online LC-ICP-MS separation and analysis method for concentration and isotopic determination of a number of elements, including the REE, Cs, Sr and Zr. Deviations of between 1 and 2% from modelled fission product yields were achieved in the presence of a U matrix of post-irradiation debris. This was expected to be beneficial in reducing analytical time and operator dose.

Radioactive contamination from the Fukushima Daiichi nuclear power plant can be traced through Cs isotope ratios. Bu et al.¹⁵¹ measured the ¹³⁵Cs : ¹³⁷Cs ratio by peak-jumping masses and secondary electron multiplier detection with TIMS. This was preceded by a 3-step chemical separation process (AMP-PAN resin, anion exchange and cation exchange) to remove isobaric and matrix interferences. The measurement precision for 1 pg and 10 fg ¹³⁷Cs was ∼1% and better than 10%, respectively. Tests on IAEA-372 and IAEA-330 reference materials yielded ¹³⁵Cs : ¹³⁷Cs isotope ratios of 0.296 and 0.297 respectively, which were similar to previously published ratios. Isotope ratios determined in these materials were similar to those taken from a Chernobyl-contaminated farm near Kiev some four years after the nuclear accident.

The use of AMS is becoming more routine. Measurement of U and Pu isotopes using AMS was reported by Hotchkis et al.¹⁵² Extended running times were used to provide an overall detection efficiency of better than 1% on 64 × 10⁶ atoms, translating to sub-ag detection of Pu isotopes. Kazi et al.¹⁵³ also used AMS to measure the ²³⁶U isotope, which was suppressed in UO⁻ ion production by ion source poisoning and a large background signal. These problems were circumvented by the addition of Si powder to the UO_x–Fe₂O₃ sample, which resulted in doubling of the 5 nA beam current.

Characterisation of nuclear materials can be extended to the isotopic characterisation of Pu which can provide information on the irradiation history, reactor conditions and U enrichment of the material. Pu isotopes can also be used as chronometers. For example, ²³⁶U is produced by the decay of ²⁴⁰Pu, and hence the ²³⁶U : ²⁴⁰Pu ratio is governed by the time since Pu purification. Comparison of four different Pu isotope chronometers by Mathew et al.,¹⁵⁴ across a range of laboratories and mass spectrometric techniques, provided an assessment of the precision achievable on the age of production. Certified reference materials were age dated by these chronometers and resulted in predicted ages for CRM 126 A of early 2001, and for CRM 138 of late 1963, which corresponded well with the manufacturers production dates. Maassen et al.¹⁵⁵ used TIMS to measure ²³⁸Pu at sub-pg levels by using a resin bead loading technique. It was noted that Pu ionised at considerably lower temperatures than U, which enabled a total evaporation technique to be used for Pu measurement prior to a significant isobaric overlap of U with ²³⁸Pu. This contrasts with MC-ICP-MS, where U is more difficult to separate from Pu as they have similar levels of ionisation.

Conflicts of interest

There are no conflicts to declare.

Glossary of abbreviations

Whenever suitable, elements may be referred to by their chemical symbols and compounds by their formulae. The following abbreviations may be used without definition. Abbreviations also cover the plural form.

2D	two dimensional
3D	three dimensional
AA	atomic absorption
AAS	atomic absorption spectrometry
ADI	ambient desorption ionisation
AES	atomic emission spectrometry
AF	atomic fluorescence
AFM	atomic force microscopy
AFS	atomic fluorescence spectrometry
ANN	artificial neural network
APDC	mmonium pyrrolidine dithiocarbamate
APGD	atmospheric pressure glow discharge
APS	aerodynamic particle sizer
BGA	buffer gas assisted
CCD	charge coupled detector
CCP	capacitively coupled plasma
CDSS	continuous direct solid sampling
CID	collision-induced discociation
CNN	convolutional neural network
CPLA	colloidal particle lens array
CRM	certified reference material
CRS	cavity ringdown spectroscopy
CS	continuum source
CV	cold vapour
CVG	chemical vapour generation
DBD	dielectric barrier discharge
dc	direct current
DC-GD-QMS	direct current glow discharge quadrupole mass spectrometry
DDW	distilled deionised water
DIHEN	direct injection high efficiency nebuliser
DL	diode laser
DLTV	diode laser thermal vaporisation
DMA	dimethylarsenic
DOF-MS	distance of flight mass spectrometry
DOTA	1,4,7,10-tetraazacyclo-dodecane n,n′,n′′,n′′′-tetra acetic acid
ECAD	electrolyte cathode atmospheric pressure discharge
EDB-LIBS	electrodynamic balance trap laser induced breakdown spectroscopy
ED-XRF	energy dispersive X-ray fluorescence
ES-MS	electrospray mass spectrometry
ETV	electrothermal vaporisation
ETV-AAS	electrothermal vaporisation atomic absorption spectrometry
ETV-ICP-MS	electrothermal vaporisation inductively coupled plasma mass spectrometry
EVG	electrochemical vapour generation
FAPA	flowing atmospheric pressure afterglow
FF	flame furnace
FFPN	flow focussing pneumatic nebuliser
FI	flow injection
FMPS	fast mobility particle sizer
fs	femto second
FWHM	full width at half maximum
GD	glow discharge
GD-MS	glow discharge mass spectrometry
GD-OES	glow discharge optical emission spectrometry
GFAAS	graphite furnace atomic absorption spectrometry
HG	hydride generation
HG-AAS	hydride generation atomic absorption spectrometry
HG-AFS	hydride generation atomic fluorescence spectrometry
HPLC-ICP-MS	high performance liquid chromatography inductively coupled plasma mass spectrometry
HR	high resolution
HR-CS-GFAAS	high resolution continuum source graphite furnace atomic absorption spectrometry
hTISIS	heated torch integrated sample introduction system
HV	high voltage
ICHPN	in-column high-pressure nebuliser
ICP-AES	inductively coupled plasma atomic emission spectrometry
ICP-MC-MS	inductively coupled plasma multicollector mass spectrometry
ICP-MS	inductively coupled plasma mass spectrometry
ICP-MS/MS	triple quadrupole inductively coupled plasma mass spectrometry
ICP-OES	inductively coupled plasma optical emission spectrometry
ICP-QMS	inductively coupled plasma quadrupole mass spectrometry
ICP-QQQ-MS	inductively coupled plasma triple quadrupole mass spectrometry
ICP-TOF-MS	inductively coupled plasma time-of-flight mass spectrometry
ID	isotope dilution
IDA	isotope dilution analysis
ID-ICP-MS	isotope dilution inductively coupled plasma mass spectrometry
ID-MRM-MS	isotope dilution multiple reaction monitoring mass spectrometry
ID-MS	isotope dilution mass spectrometry
IL	ionic liquid
IR	isotope ratio
IRA	isotope ratio analysis
IRMS	isotope ratio mass spectrometry
IT-SPME	in-tube solid-phase microextraction
LA	laser ablation
LAAS	laser atomic absorption spectroscopy
LA-ICP-MS	laser ablation inductively coupled plasma mass spectrometry
LA-MC-ICP-MS	laser ablation multicollector inductively coupled plasma mass spectrometry
LAMIS	laser ablation molecular isotopic spectrometry
LC	liquid chromatography
LCGD	liquid cathode glow discharge
LDR	linear dynamic range
LEP	liquid electrode microplasma
LI	laser ionisation
LIBS	laser induced breakdown spectroscopy
LIF	laser induced fluorescence
LOD	limit of detection
LOQ	limit of quantification
LOV	lab-on-a-valve
LPME	liquid phase microextraction
LS-APGD	liquid sampling-atmospheric pressure glow discharge
MALDI	matrix-assisted laser desorption ionisation
MC-ICP-MS	multicollector inductively coupled plasma mass spectrometry
MIP	microwave induced plasma
MIP-OES	microwave induced plasma optical emission spectrometry
MIS	monochromatic imaging spectrometer
MMA	monomethylarsenic
MNP	magnetic nanoparticle
MP	microsecond pulsed
MPT	microwave plasma torch
MSIS	multimode sample introduction system
MW-LIBS	microwave assisted LIBS
Nd:YAG	neodymium doped:yttrium aluminum garnet
n e	electron number density
NIST	National Institure of Standards and Technology
NP	nanoparticle
ns	nanosecond
o.d.	outer diameter
OPO	optical parametric oscillator
PCA	principal components analysis
PCR	polymerase chain reaction
PDMS	polydimethylsiloxane
PF	particle flow
PFH	perfluorohexane
PIL	plasma induced luminescence
PLS	partial least squares
PN	pneumatic nebuliser
PNC	particle number concentration
ppb	parts per billion (10-9)
ppm	parts per million (10-6)
ppq	parts per quadrillion (10-15)
ppt	parts per trillion (10-12)
PSD	particle size distribution
PVG	photochemical vapour generation
PVP	polyvinylpyrrolidone
Q	quadrupole
QD	quantum dot
QMS	quadrupole mass spectrometry
QTF	quartz tube furnace
REE	rare earth element
RF	radiofrequency
RIMS	resonance ionisation mass spectrometry
RSD	relative standard deviation
RVM	relevance vector machine
SALD	substrate assisted laser desorption
SBR	signal-to-background ratio
SCGD	solution cathode glow discharge
SEM	scanning electron microscopy
SF-ICP-MS	sector-field inductively coupled plasma mass spectrometry
SHS	spatial heterodyne spectrometer
SIBS	spark-induced breakdown spectroscopy
SIF	statistical interference factor
SIMS	secondary ionisation mass spectrometry
SMPS	scanning mobility particle sizer
SNR	signal to noise ratio
SP-ICP-MS	single particle inductively coupled plasma mass spectrometry
SPME	solid phase microextraction
SRM	standard reference material
STM	scanning tunnelling microscopy
SVR	support vector regression
TDLAAS	tunable diode laser atomic absorption spectroscopy
T e	electron temperature
TE	transport efficiency
TEM	transmission electron microscopy
TGA	thermogravimetric analysis
T gas	gas temperature
TIMS	thermal ionisation mass spectrometry
T ion	ionisation temperature
TOF	time-of-flight
TOF-MS	time-of-flight mass spectrometry
T rot	rotational temperature
UV	ultraviolet
VG	vapour generation
XPS	X-ray photoelectron spectroscopy
XRF	X-ray fluorescence

Acknowledgements

Jorge Pisonero acknowledges support from Ministerio de Economía y Competitividad (Spain) through the I+D+i project referenced MINECO-17-CTQ2016-77887-C2-1-R.

References

1. E. H. Evans, J. Pisonero, C. M. M. Smith and R. N. Taylor, J. Anal. At. Spectrom., 2019, 34(5), 803–822.
2. J. R. Bacon, O. T. Butler, W. R. L. Cairns, J. M. Cook, R. Mertz-Kraus and J. F. Tyson, J. Anal. At. Spectrom., 2019, 34(1), 9–58.
3. R. Clough, C. F. Harrington, S. J. Hill, Y. Madrid and J. F. Tyson, J. Anal. At. Spectrom., 2019, 34(7), 1306–1350.
4. A. Taylor, N. Barlow, M. P. Day, S. Hill, N. Martin and M. Patriarca, J. Anal. At. Spectrom., 2019, 34(3), 426–459.
5. C. Vanhoof, J. R. Bacon, A. T. Ellis, U. E. A. Fittschen and L. Vincze, J. Anal. At. Spectrom., 2019, 34(9), 1750–1767.
6. S. Carter, R. Clough, A. Fisher, B. Gibson, B. Russell and J. Waack, J. Anal. At. Spectrom., 2019, 34(11), 2159–2216.
7. C. Basheer, M. Kamran, M. Ashraf and H. K. Lee, TrAC, Trends Anal. Chem., 2019, 118, 426–433.
8. M. Á. Aguirre, P. Baile, L. Vidal and A. Canals, TrAC, Trends Anal. Chem., 2019, 112, 241–247.
9. V. A. Lemos, R. V. Oliveira, W. N. L. dos Santos, R. M. Menezes, L. B. Santos and S. L. C. Ferreira, TrAC, Trends Anal. Chem., 2019, 110, 357–366.
10. M. He, B. B. Chen, H. Wang and B. Hu, Appl. Spectrosc. Rev., 2019, 54(3), 250–263.
11. Z. Chen, B. Chen, M. He, H. Wang and B. Hu, Microchim. Acta, 2019, 186(2), 107.
12. M. Ghorbani, M. Aghamohammadhassan, M. Chamsaz, H. Akhlaghi and T. Pedramrad, TrAC, Trends Anal. Chem., 2019, 118, 793–809.
13. P. P. Chen, K. Huang, R. Dai, E. Sawyer, K. Sun, B. W. Ying, X. W. Wei and J. Geng, Analyst, 2019, 144(8), 2797–2802.
14. R. Dai, P. Y. Hu, X. Wang, S. X. Wang, X. M. Song, K. Huang and P. P. Chen, Analyst, 2019, 144(14), 4407–4412.
15. R. Liu, J. Y. Hu, Y. X. Chen, M. Jiang and Y. Lv, Anal. Chem., 2019, 91(13), 8691–8696.
16. G. Y. Xiao, B. B. Chen, M. He, X. T. Li and B. Hu, Talanta, 2019, 202, 207–213.
17. Y. He, S. L. Chen, L. Huang, Z. W. Wang, Y. N. Wu and F. F. Fu, Anal. Chem., 2019, 91(1), 1171–1177.
18. Y. Liang, Q. Liu, Y. Zhou, S. Chen, L. M. Yang, M. Zhu and Q. Q. Wang, Anal. Chem., 2019, 91(13), 8341–8349.
19. F. J. Alonso-Garcia, E. Blanco-Gonzalez and M. Montes-Bayon, Talanta, 2019, 194, 336–342.
20. T. Vaneckova, J. Bezdekova, M. Tvrdonova, M. Vlcnovska, V. Novotna, J. Neuman, A. Stossova, V. Kanicky, V. Adam, M. Vaculovicova and T. Vaculovic, Sci. Rep., 2019, 9(1), 11840.
21. C. F. Ren, C. E. Bobst and I. A. Kaltashov, Anal. Chem., 2019, 91(11), 7189–7198.
22. C. X. Ji, Y. Liang, F. C. Ge, L. M. Yang and Q. Q. Wang, Anal. Chem., 2019, 91(11), 7032–7038.
23. J. Masone, R. Brennan, G. Russell and T. Hettipathirana, Spectroscopy, 2019, 6–7.
24. A. Al Hejami and D. Beauchemin, J. Anal. At. Spectrom., 2019, 34(1), 232–238.
25. P. Louvat, M. Tharaud, M. Buisson, C. Rollion-Bard and M. F. Benedetti, J. Anal. At. Spectrom., 2019, 34(8), 1553–1563.
26. G. L. Scheffler, D. Pozebon and D. Beauchemin, J. Anal. At. Spectrom., 2019, 34(5), 891–898.
27. J. Patocka, T. Cernohorsky, A. Krejcova and S. Slang, Talanta, 2019, 204, 555–560.
28. A. R. M. Bustos, K. P. Purushotham, A. Possolo, N. Farkas, A. E. Vladar, K. E. Murphy and M. R. Winchester, Anal. Chem., 2018, 90(24), 14376–14386.
29. L. Hendriks, B. Ramkorun-Schmidt, A. Gundlach-Graham, J. Koch, R. N. Grass, N. Jakubowski and D. Gunther, J. Anal. At. Spectrom., 2019, 34(4), 716–728.
30. M. Loula, A. Kana and O. Mestek, Talanta, 2019, 202, 565–571.
31. F. H. Lin, S. Miyashita, K. Inagaki, Y. H. Liu and I. H. Hsu, J. Anal. At. Spectrom., 2019, 34(2), 401–406.
32. E. Bolea-Fernandez, D. Leite, A. Rua-Ibarz, T. Liu, G. Woods, M. Aramendia, M. Resano and F. Vanhaecke, Anal. Chim. Acta, 2019, 1077, 95–106.
33. X. X. Yu, B. B. Chen, M. He, H. Wang and B. Hu, Anal. Chem., 2019, 91(4), 2869–2875.
34. X. Wei, D. H. Zheng, Y. Cai, R. Jiang, M. L. Chen, T. Yang, Z. R. Xu, Y. L. Yu and J. H. Wang, Anal. Chem., 2018, 90(24), 14543–14550.
35. Z. Wang, X. Wang, Q. Wang, X. Xiong, H. Luo and K. Huang, Microchem. J., 2019, 149, 104052.
36. A. D’Ulivo, J. Anal. At. Spectrom., 2019, 34(5), 823–847.
37. E. Pitzalis, D. Angelini, M. C. Mascherpa and A. D’Ulivo, J. Anal. At. Spectrom., 2018, 33(12), 2160–2171.
38. G. P. Ma, S. Liu, J. S. Sun and X. C. Duan, Spectrochim. Acta, Part B, 2018, 150, 38–42.
39. M. Slachcinski, J. Anal. At. Spectrom., 2019, 34(2), 257–273.
40. Z. R. Zou, J. Hu, F. J. Xu, X. D. Hou and X. M. Jiang, TrAC, Trends Anal. Chem., 2019, 114, 242–250.
41. J. Hu, H. J. Chen, X. D. Hou and X. M. Jiang, Anal. Chem., 2019, 91(9), 5938–5944.
42. S. Burhenn, J. Kratzer, F. D. Klute, J. Dedina and J. Franzke, Spectrochim. Acta, Part B, 2019, 152, 68–73.
43. J. Kratzer, S. Musil and J. Dedina, J. Anal. At. Spectrom., 2019, 34(1), 193–202.
44. L. Juhaszova, S. Burhenn, L. Sagapova, J. Franzke, J. Dedina and J. Kratzer, Spectrochim. Acta, Part B, 2019, 158, 105630.
45. X. Liu, Z. L. Zhu, Z. Y. Bao, D. He, H. T. Zheng, Z. F. Liu and S. H. Hu, Anal. Chem., 2019, 91(1), 928–934.
46. I. J. Doh, C. Gondhalekar, V. Patsekin, B. Rajwa, K. Hernandez, E. Bae and J. P. Robinson, Appl. Spectrosc., 2019, 73(6), 698–708.
47. A. Ganeev, A. Gubal, V. Chuchina, Y. Lyalkin, O. Glumov, V. Yakobson and N. Solovyev, J. Anal. At. Spectrom., 2019, 34(3), 588–597.
48. C. Gonzalez-Gago, P. Smid, T. Hofmann, C. Venzago, V. Hoffmann, W. Gruner, J. Pfeifer, S. Richter and H. Kipphardt, J. Anal. At. Spectrom., 2019, 34(6), 1109–1125.
49. G. Paudel, M. Kasik and M. Di Sabatino, J. Anal. At. Spectrom., 2019, 34(9), 1829–1837.
50. K. Wagatsuma, Appl. Spectrosc. Rev., 2019, 55(1), 76–86.
51. T. J. Williams and R. K. Marcus, J. Anal. At. Spectrom., 2019, 34(7), 1468–1477.
52. K. A. Hall, H. W. Paing, M. R. Webb and R. K. Marcus, Spectrochim. Acta, Part B, 2019, 154, 33–42.
53. J. Pisonero, D. Bouzas-Ramos, H. Traub, B. Cappella, C. Alvarez-Llamas, S. Richter, J. C. Mayo, J. M. Costa-Fernandez, N. Bordel and N. Jakubowski, J. Anal. At. Spectrom., 2019, 34(4), 655–663.
54. J. T. van Elteren, V. S. Selih and M. Sala, J. Anal. At. Spectrom., 2019, 34(9), 1919–1931.
55. A. Rodionov, E. Lehndorff, C. C. Stremtan, W. A. Brand, H. P. Konigshoven and W. Amelung, Anal. Chem., 2019, 91(9), 6225–6232.
56. E. S. Steenstra, J. Berndt, S. Klemme and W. van Westrenen, J. Anal. At. Spectrom., 2019, 34(1), 222–231.
57. R. Weiskirchen, S. Weiskirchen, P. Kim and R. Winkler, J. Cheminf., 2019, 11(1), 16.
58. A. Al Hejami and D. Beauchemin, J. Anal. At. Spectrom., 2019, 34(7), 1426–1432.
59. A. Bierstedt, Y. You, S. van Wasen, G. Bosc-Bierne, M. Weller and J. Riedel, Anal. Chem., 2019, 91(9), 5922–5928.
60. S. A. Xia, A. Q. Leng, Y. Lin, L. Wu, Y. F. Tian, X. D. Hou and C. B. Zheng, Anal. Chem., 2019, 91(4), 2701–2709.
61. K. Swiderski, P. Pohl and P. Jamroz, J. Anal. At. Spectrom., 2019, 34(6), 1287–1293.
62. C. Tao, C. S. Li, Y. C. Li, H. X. Wang, Y. R. Zhang, Z. H. Zhou, X. F. Mao, Z. Y. Ma and D. Tian, J. Anal. At. Spectrom., 2018, 33(12), 2098–2106.
63. E. D. Hoegg, S. Godin, J. Szpunar, R. Lobinski, D. W. Koppenaal and R. K. Marcus, J. Anal. At. Spectrom., 2019, 34(7), 1387–1395.
64. E. D. Hoegg, B. T. Manard, E. M. Wylie, K. J. Mathew, C. F. Ottenfeld and R. K. Marcus, J. Am. Soc. Mass Spectrom., 2019, 30(2), 278–288.
65. A. Alonso-Medina, Appl. Spectrosc., 2019, 73(2), 133–151.
66. J. R. dos Santos, J. J. Neto, N. Rodrigues, M. G. Destro, J. W. Neri, P. Bueno and B. Christ, Appl. Spectrosc., 2019, 73(2), 203–213.
67. S. Raeder, H. Heggen, A. Teigelhofer and J. Lassen, Spectrochim. Acta, Part B, 2019, 151, 65–71.
68. D. A. Fink, K. Blaum, V. N. Fedosseev, B. A. Marsh, R. E. Rossel and S. Rothe, Spectrochim. Acta, Part B, 2019, 151, 72–82.
69. A. R. Vernon, J. Billowes, C. L. Binnersley, M. L. Bissell, T. E. Cocolios, G. J. Farooq-Smith, K. T. Flanagan, R. F. G. Ruiz, W. Gins, R. P. de Groote, A. Koszorus, K. M. Lynch, G. Neyens, C. M. Ricketts, K. D. A. Wendt, S. G. Wilkins and X. F. Yang, Spectrochim. Acta, Part B, 2019, 153, 61–83.
70. I. Urbina, D. Carneiro, S. Rocha, E. Farias, F. Bredice and V. Palleschi, Spectrochim. Acta, Part B, 2019, 154, 91–96.
71. S. Kobayashi, N. Nishimiya and M. Suzuki, Spectrochim. Acta, Part B, 2019, 152, 30–37.
72. S. Gupta, R. K. Gangwar and R. Srivastava, Spectrochim. Acta, Part B, 2018, 149, 203–213.
73. Z. Weiss, Spectrochim. Acta, Part B, 2019, 158, 105637.
74. H. Matsuta, Spectrochim. Acta, Part B, 2018, 149, 150–155.
75. T. Ariga, Y. B. Zhu and K. Inagaki, J. Anal. At. Spectrom., 2019, 34(9), 1865–1874.
76. X. M. Guo, S. Alavi, E. Dalir, J. M. Dai and J. Mostaghimi, J. Anal. At. Spectrom., 2019, 34(3), 469–479.
77. X. M. Guo, S. Alavi and J. Mostaghimi, J. Anal. At. Spectrom., 2019, 34(10), 2126–2135.
78. A. Brandt and K. Leopold, Spectrochim. Acta, Part B, 2018, 150, 26–32.
79. J. A. Carter, J. T. Sloop, T. McSweeney, B. T. Jones and G. L. Donati, Anal. Chim. Acta, 2019, 1062, 37–46.
80. M. Burger, L. Hendriks, J. Kaeslin, A. Gundlach-Graham, B. Hattendorf and D. Gunther, J. Anal. At. Spectrom., 2019, 34(1), 135–146.
81. M. Sargent, H. Goenaga-Infante, K. Inagaki, L. Ma, J. Meija, A. Pramann, O. Rienitz, R. Sturgeon, J. Vogl, J. Wang and L. Yang, Metrologia, 2019, 56(3), 034005.
82. M. Fujiwara, K. Hirosawa, N. Nonose, S. Nishida and N. Furuta, Accredit. Qual. Assur., 2019, 24(5), 329–339.
83. C. B. Williams, B. T. Jones and G. Donati, J. Anal. At. Spectrom., 2019, 34(6), 1191–1197.
84. L. Jolivet, M. Leprince, S. Moncayo, L. Sorbier, C. P. Lienemann and V. Motto-Ros, Spectrochim. Acta, Part B, 2019, 151, 41–53.
85. G. S. Senesi, J. Cabral, C. R. Menegatti, B. Marangoni and G. Nicolodelli, TrAC, Trends Anal. Chem., 2019, 118, 453–469.
86. J. Moros and J. Laserna, Appl. Spectrosc., 2019, 73(9), 963–1011.
87. V. C. Costa, A. S. Augusto, J. P. Castro, R. C. Machado, D. F. Andrade, D. V. Babos, M. A. Speranca, R. R. Gamela and E. R. Pereira, Quim. Nova, 2019, 42(5), 527–545.
88. S. Buckley, Spectroscopy, 2019, 34(4), 16–20.
89. J. J. Camacho, J. Vrabel, S. Manzoor, L. V. Perez-Arribas, D. Diaz and J. O. Caceres, J. Anal. At. Spectrom., 2019, 34(6), 1247–1255.
90. I. Carrasco-Garcia, J. M. Vadillo and J. J. Laserna, J. Anal. At. Spectrom., 2019, 34(10), 2119–2125.
91. A. Safi, S. H. Tavassoli, G. Cristoforetti, E. Tognoni, B. Campanella, S. Legnaioli, S. Pagnotta, F. Poggialini and V. Palleschi, Anal. Chem., 2019, 91(13), 8595–8601.
92. C. T. Chen, D. Banaru, T. Sarnet and J. Hermann, Spectrochim. Acta, Part B, 2018, 150, 77–85.
93. T. Q. Li, Z. Y. Hou, Y. T. Fu, J. L. Yu, W. L. Gu and Z. Wang, Anal. Chim. Acta, 2019, 1058, 39–47.
94. Y. T. Fu, Z. Y. Hou, T. Q. Li, Z. Li and Z. Wang, Spectrochim. Acta, Part B, 2019, 155, 67–78.
95. A. A. Al Shuaili, A. M. Al Hadhrami, M. A. Wakil and Z. T. Alwahabi, Spectrochim. Acta, Part B, 2019, 159, 105666.
96. M. A. Wakil and Z. T. Alwahabi, J. Anal. At. Spectrom., 2019, 34(9), 1892–1899.
97. A. Allen and S. M. Angel, Spectrochim. Acta, Part B, 2018, 149, 91–98.
98. F. Cattani, P. Y. Gillot, X. Quidelleur, A. Hildenbrand, J. C. Lefevre, C. Boukari and F. Courtade, Chem. Geol., 2019, 506, 1–16.
99. B. Chide, S. Maurice, N. Murdoch, J. Lasue, B. Bousquet, X. Jacob, A. Cousin, O. Forni, O. Gasnault, P. Y. Meslin, J. F. Fronton, M. Bassas-Portus, A. Cadu, A. Sournac, D. Mimoun and R. C. Wiens, Spectrochim. Acta, Part B, 2019, 153, 50–60.
100. R. Hai, L. Y. Sun, D. Wu, Z. L. He, H. Sattar, J. M. Liu, W. N. Tong, C. Li, C. L. Feng and H. B. Ding, J. Anal. At. Spectrom., 2019, 34(10), 1982–1987.
101. T. N. Cai, C. L. Li, X. Ren, B. Liu and D. W. Liu, Spectrosc. Spectral Anal., 2019, 39(5), 1623–1629.
102. D. V. Babos, A. I. Barros, J. A. Nobrega and E. R. Pereira, Spectrochim. Acta, Part B, 2019, 155, 90–98.
103. V. Motto-Ros, S. Moncayo, F. Trichard and F. Pelascini, Spectrochim. Acta, Part B, 2019, 155, 127–133.
104. A. Palleschi and V. Palleschi, Spectrochim. Acta, Part B, 2018, 149, 271–275.
105. T. B. Chen, L. Y. Zhang, L. Huang, M. H. Liu, J. Y. Chen and M. Y. Yao, J. Anal. At. Spectrom., 2019, 34(5), 884–890.
106. M. R. Dong, L. P. Wei, J. D. Lu, W. B. Li, S. Z. Lu, S. S. Li, C. Y. Liu and J. H. Yoo, J. Anal. At. Spectrom., 2019, 34(3), 480–488.
107. J. Chappell, M. Martinez and M. Baudelet, Spectrochim. Acta, Part B, 2018, 149, 167–175.
108. E. Kepes, P. Porizka, J. Klus, P. Modlitbova and J. Kaiser, J. Anal. At. Spectrom., 2018, 33(12), 2107–2115.
109. M. Dell’Aglio, Z. Salajkova, A. Mallardi, R. Mezzenga, L. Van’t Hag, N. Cioffi, G. Palazzo and A. De Giacomo, Spectrochim. Acta, Part B, 2019, 155, 115–122.
110. X. J. Hao, H. J. Tang and X. T. Hu, Spectrosc. Spectral Anal., 2019, 39(5), 1599–1603.
111. M. M. Wang, L. Jiang, S. M. Wang, Q. T. Guo, F. Tian, Z. Y. Chu, J. Zhang, X. Li and Y. F. Lu, Anal. Chem., 2019, 91(15), 9952–9961.
112. J. Guo, T. F. Wang, J. F. Shao, A. M. Chen and M. X. Jin, J. Anal. At. Spectrom., 2018, 33(12), 2116–2123.
113. V. N. Lednev, M. Y. Grishin, P. A. Sdvizhenskii, R. D. Asyutin, R. S. Tretyakov, A. Y. Stavertiy and S. M. Pershin, J. Anal. At. Spectrom., 2019, 34(3), 607–615.
114. J. Merten and B. Johnson, Spectrochim. Acta, Part B, 2018, 149, 124–131.
115. M. Pal, S. Maithani, A. Maity and M. Pradhan, J. Anal. At. Spectrom., 2019, 34(5), 860–866.
116. N. H. Cheung, J. Anal. At. Spectrom., 2019, 34(3), 616–622.
117. M. Klaver and C. D. Coath, Geostand. Geoanal. Res., 2019, 43(1), 5–22.
118. S. R. Scott, K. W. W. Sims, M. K. Reagan, L. Ball, J. B. Schwieters, C. Bouman, N. S. Lloyd, C. L. Waters, J. J. Standish and D. L. Tollstrup, Int. J. Mass Spectrom., 2019, 435, 321–332.
119. S. Wang, J. Wang, H. Lu, X. Fang, P. S. Song and T. X. Ren, Rapid Commun. Mass Spectrom., 2019, 33(9), 857–866.
120. C. F. Li, Z. Y. Chu, X. C. Wang, J. H. Guo and S. A. Wilde, Anal. Chem., 2019, 91(11), 7288–7294.
121. T. Narukawa, O. Shikino and K. Chiba, J. Anal. At. Spectrom., 2019, 34(3), 517–525.
122. C. Pin and A. Gannoun, J. Anal. At. Spectrom., 2019, 34(10), 2136–2146.
123. X. Shao, W. Bu, K. Long, L. Tang, X. Liu, H. Yan and F. Hao, Spectrochim. Acta, Part B, 2019, 159.
124. P. Bonnand, C. Israel, M. Boyet, R. Doucelance and D. Auclair, J. Anal. At. Spectrom., 2019, 34(3), 504–516.
125. S. Tong, J. Meija, L. Zhou, Z. Mester and L. Yang, Metrologia, 2019, 56(4), 044008.
126. G. Q. Wang, H. Vollstaedt, J. F. Xu and W. G. Liu, Geostand. Geoanal. Res., 2019, 43(3), 419–433.
127. L. Y. Zhu, Y. S. Liu, S. Y. Jiang and J. Lin, J. Anal. At. Spectrom., 2019, 34(8), 1546–1552.
128. P. K. Mukherjee, A. K. Souders and P. J. Sylvester, J. Anal. At. Spectrom., 2019, 34(1), 180–192.
129. G. Avice, A. Belousov, K. A. Farley, S. M. Madzunkov, J. Simcic, D. Nikolic, M. R. Darrach and C. Sotin, J. Anal. At. Spectrom., 2019, 34(1), 104–117.
130. H. Chen, Z. Tian, B. Tuller-Ross, R. L. Korotev and K. Wang, J. Anal. At. Spectrom., 2019, 34(1), 160–171.
131. L. F. Gou, Z. D. Jin, A. Galy, H. Sun, L. Deng and Y. Xu, Rapid Commun. Mass Spectrom., 2019, 33(4), 351–360.
132. T. Hirata, S. Yamashita, T. Suzuki and M. Ishida, Bunseki Kagaku, 2019, 68(2), 81–88.
133. Y. Hu and F. Z. Teng, J. Anal. At. Spectrom., 2019, 34(2), 338–346.
134. S. Y. Huang, Q. Y. Song, Y. B. Zhang, D. X. Yuan, L. M. Sun, Y. J. Chen, R. G. Jiang and H. Lin, Anal. Chem., 2019, 91(11), 7063–7069.
135. S. Queipo-Abad, P. Rodriguez-Gonzalez and J. I. G. Alonso, J. Anal. At. Spectrom., 2019, 34(4), 753–763.
136. Q. Zhang, J. Liu, Y. N. Zhang, H. M. Yu, L. P. Qin and J. Shen, Rapid Commun. Mass Spectrom., 2019, 33(17), 1390–1400.
137. D. D. Li, M. L. Li, W. R. Liu, Z. Z. Qin and S. A. Liu, Geostand. Geoanal. Res., 2018, 42(4), 593–605.
138. S. G. Lee and T. Tanaka, Spectrochim. Acta, Part B, 2019, 156, 42–50.
139. X. Li, H. Y. Li, J. G. Ryan, G. J. Wei, L. Zhang, N. B. Li, X. L. Huang and Y. G. Xu, Chem. Geol., 2019, 505, 76–85.
140. A. Sadekov, N. S. Lloyd, S. Misra, J. Trotter, J. D’Olivo and M. McCulloch, J. Anal. At. Spectrom., 2019, 34(3), 550–560.
141. C. D. Standish, T. B. Chalk, T. L. Babila, J. A. Milton, M. R. Palmer and G. L. Foster, Rapid Commun. Mass Spectrom., 2019, 33(10), 959–968.
142. L. K. Steinmann, M. Oeser, I. Horn, H. M. Seitz and S. Weyer, J. Anal. At. Spectrom., 2019, 34(7), 1447–1458.
143. Z. Y. Zhu, T. Yang and X. K. Zhu, J. Anal. At. Spectrom., 2019, 34(7), 1503–1513.
144. J. H. Bartlett and A. Castro, Spectrochim. Acta, Part B, 2019, 155, 61–66.
145. M. Krachler, A. Bulgheroni, A. I. M. Ferri, Y. Ma, A. Miard and P. Garcia, J. Anal. At. Spectrom., 2019, 34(10), 1965–1974.
146. I. Strashnov, I. Izosimov, J. D. Gilmour, M. A. Denecke, J. Almiral, A. Cannavan, G. Chen, C. Dissanayake, I. Doroshenko, T. Elghali, E. Enston, B. R. Fernando, G. Kasozi, S. Kelly, M. Maqsood, S. A. Muhammad, C. Muryn, A. L. Pomerantsev, D. K. Singh, G. Smith, F. Taous, C. Webb, D. Williamson, Z. Z. Xu, S. M. Yangg and A. Zitek, J. Anal. At. Spectrom., 2019, 34(8), 1630–1638.
147. A. L. Ronzani, A. Hubert, F. Pointurier, O. Marie, N. Clavier, A. C. Humbert, J. Aupiais and N. Dacheux, Rapid Commun. Mass Spectrom., 2019, 33(5), 419–428.
148. T. P. Forbes and C. Szakal, Analyst, 2019, 144(1), 317–323.
149. A. Trinquier, C. Maden, A. L. Faure, A. Hubert, F. Pointurier, B. Bourdon and M. Schonbachler, Anal. Chem., 2019, 91(9), 6190–6199.
150. E. K. Fenske, B. D. Roach, D. C. Glasgow, R. H. Ilgner and J. M. Giaquinto, J. Radioanal. Nucl. Chem., 2019, 320(1), 153–163.
151. W. T. Bu, L. Tang, X. M. Liu, Z. T. Wang, M. Fukuda, J. Zheng, T. Aono, S. Hu and X. L. Wang, J. Anal. At. Spectrom., 2019, 34(2), 301–309.
152. M. A. C. Hotchkis, D. P. Child, M. B. Froehlich, A. Wallner, K. Wilcken and M. Williams, Nucl. Instrum. Methods Phys. Res., Sect. B, 2019, 438, 70–76.
153. Z. H. Kazi, C. R. J. Charles, X. L. Zhao, R. J. Cornett and W. E. Kieser, Nucl. Instrum. Methods Phys. Res., Sect. B, 2019, 456, 218–221.
154. K. Mathew, T. Kayzar-Boggs, Z. Varga, A. Gaffney, J. Denton, J. Fulwyler, K. Garduno, A. Gaunt, J. Inglis, R. Keller, W. Kinman, D. Labotka, E. Lujan, J. Maassen, T. Mastren, I. May, K. Mayer, A. Nichol, C. Ottenfeld, T. Parsons-Davis, D. Porterfield, J. Rim, J. Rolison, F. Stanley, R. Steiner, L. Tandon, M. Thomas, R. Torres, K. Treinen, M. Wallenius, A. Wende, R. Williams and J. Wimpenny, Anal. Chem., 2019, 91(18), 11643–11652.
155. J. Maassen, J. D. Inglis, A. Wende, T. M. Kayzar-Boggs, R. E. Steiner and A. Kara, J. Radioanal. Nucl. Chem., 2019, 321(3), 1073–1080.

---