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

Contents

• 1. Liquids analysis
  • 1.1. Sample pre-treatment
  • 1.2. Sample introduction
  • 1.3. Direct methods

• 2. Solids analysis
  • 2.1. Direct methods
  • 2.2. Indirect methods

• 3. Instrumentation, fundamentals and chemometrics
  • 3.1. Instrumentation
  • 3.2. Fundamentals
  • 3.3. Chemometrics

• 4. Laser-based atomic spectrometry
  • 4.1. Laser induced breakdown spectroscopy
  • 4.2. Laser induced fluorescence
  • 4.3. Laser atomic absorption spectroscopy
  • 4.4. Laser ablation ionisation mass spectrometry

• 5. Isotope analysis
  • 5.1. New developments
  • 5.2. Radiogenic isotope ratio analysis
  • 5.3. Geological studies
  • 5.4. Stable isotope ratio studies
  • 5.5. Nuclear forensics

• 6. Glossary of abbreviations

• 7. Conflicts of interest

• References

---

Abstract

This review of 177 references covers developments in ‘Atomic Spectrometry’ published in the twelve months from December 2022 to November 2023 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 integration of chip-based sample preparation and sample introduction devices has become more advanced, pointing the way to more highly automated methods for clinical analysis. Linked with miniaturisation this should also drive the development of field deployable devices. The analysis of nanoparticles and nanomaterials continues to be an important area of research given the topical concerns of these particles in the biosphere. The trend to nano-analysis is also reflected by developments in LA and LIBS analyses: e.g. spatially highly resolved bioimaging, analysis of nanoparticles or isotope ratios at extremely low concentrations and high precision.

---

1. Liquids analysis

1.1. Sample pre-treatment

Microfluidic sample pre-treatment systems have found new utility for single-cell manipulations prior to analysis. Ma et al.⁷ designed a flow-through chip using PDMS formed on a silicon photoresist template to form the channels, then bonded to a glass slide backplate. The channels, used for water, organic phase and focusing were 100, 60, and 80 μm wide, respectively. A macrophage suspension (water phase) was introduced into channel A, and n-hexanol containing 1% Span 80 (organic phase) into channel B. Under the action of the shear force of the organic phases, the cell suspension was separated into small homogeneous droplets that encapsulated single cells. These were then introduced into the nebuliser of ICP-MS via a quartz capillary of 75 μm i.d. for SF-ICP-MS analysis. The authors used this system to analyse cells (THP-1 cells) from a human leukaemia monocytic cell line. The purpose was to investigate the hydrolysis of thimerosal (THI) to ethyl-mercury in the macrophages. Hence, by analysing individual cells the authors found, amongst other things, that e.g. <5% of the cells exhibited a high uptake (>200 fg per cell) for THI, but most cells exhibited low uptake (<20 fg per cell).

Xu et al.⁸ developed a microfluidic chip system for the immunolabeling, magnetic separation and focusing of HepG2 cells coupled with SC-ICP-MS for quantitative analysis of the asialoglycoprotein receptor (ASGPR) on single HepG2 cells. The microfluidic chip was fabricated from PDMS using a silicon photoresist template to form the channels and then bonded to a glass backplate. Lanthanide-labelled anti-ASGPR monoclonal antibody and magnetic beads modified with anti-epithelial cell adhesion molecules were prepared as signal and magnetic probes, respectively. Target cells were labelled with signal and magnetic probes in the mixing zone of the microfluidic chip and then focused and sorted in the separation zone by magnetic separation to avoid matrix contamination. The i.d., length and flow-rate of the channels were optimised to achieve the best mixing, labelling and separation. Recovery of HepG2 cells was 94.1 ± 5.7% with high separation efficiency and purity. The sorted cells with signal probes were detected using SC-ICP-MS with an average of (1.0 ± 0.2) × 10⁵ ASGPR molecules per cell surface. According to the authors, this method can be used for absolute quantitative analysis of ASGPRs on the surface of single hepatocellular carcinoma cells.

Wei et al.⁹ developed a spiral microfluidic chip to focus and sort cells at low flow rate. The chip was fabricated in a similar manner to the two previous papers to yield a micrometer-scale spiral structure of 8 loops, ∼8 cm in length, 100 μm in width, 45 μm in height, 2300 μm radius of the outer ring, with channels containing micropillars. Four configurations with different widths but the same height of the channel were evaluated. Cultures of MCF-7 and LoVo cells were incubated with between 50 and 250 μg L⁻¹ (0.85 to 4.2 μmol L⁻¹) cobalt curcumin complex of [Co(tpa)(cur)](ClO₄)₂ suspended in ultrapure water and injected into the microchip at a flow rate of 50 μL min⁻¹. The focused single cells were analysed using SC-ICP-MS by monitoring ⁵⁹Co⁺. Single cell detection efficiency of 67.5 ± 4.6% was obtained for a cell suspension with a cell density of 3 × 10⁵ cells per mL.

1.1.1. Extraction methods. Sample extraction is used to either pre-concentrate the analyte, separate the analyte from the matrix, or both. It is a well-researched field and there are few novel developments. Azooz et al.¹⁰ reviewed LPE and SPE for the determination of Se, in particular the use of supramolecular solvents, hydrophobic deep eutectic solvents (DESs), and ionic liquids (ILs). Wang et al.¹¹ developed a battery-operated, SPME device for field analysis of Hg. The Au@W SPME fibre extraction device was fabricated from tungsten wire coil, electro-plated with gold and inserted into a sample vial. Digested samples were placed in the vial and reduced with NaBH₄ to yield volatile Hg⁰, which was adsorbed on the coil. The Hg was subsequently desorbed by heating the coil and determined using μPD-OES with an LOD of 0.008 mg kg⁻¹. A soil CRM and nine soil samples were analysed and recoveries of between 86% and 111% for Hg were obtained. The fibre could also be used for preservation of samples for later analysis, with <5% analyte loss after 30 days storage at room temperature.

Matrix volatilisation as a means of separating the analyte from the sample is not commonly used because of the danger of analyte co-losses. Medvedev et al.¹² used vacuum distillation to preconcentrate trace elements from tellurium materials. The vacuum distillation system consisted of a quartz cover, cup, and cup stand, fluoroplastic (PTFE) base, union nut, heater (tubular resistance furnace) and cooler (silicone hose coil). A sample of Te was placed in the quartz cup and heated, under argon, until it melted, whereupon air was introduced so that an oxide film formed. Vacuum distillation was then performed so that Te was volatilised and the trace elements concentrated in the remaining TeO₂ film. A 2g sample of Te yielded ∼0.01 g of concentrate which was then digested in acid and analysed using ICP-MS and ICP-OES. Recoveries for 49 trace elements were between 70% (Fe) and 120% (Li), with the majority in the range 80% to 100%. The obtained LODs were from 0.2 to 80 ng g⁻¹ using ICP-OES and from 0.05 to 90 ng g⁻¹ using ICP-MS.

1.1.2. Elemental tagging. Indirect quantitation of proteins, peptides and nucleic acids can be achieved by several means: 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 IDMS; labelling using an antigen–antibody immunoassay approach; hybridisation of target strands of DNA or RNA with complementary labelled probes; or variations which combine all of these.

Multiplex determination of proteins in single cells was investigated by Menero-Valdes et al.¹³ The proteins, hepcidin, metallothionein-2 and ferroprotein (expressed in ARPE-19 cells), were tagged with specific antibodies labelled with Ir-, Pt- and Au-nanoclusters (NCs). Anti-h-HP:IrNCs, Anti-h-MT2:PtNCs and Anti-h-FPN:AuNCs were used for tagging at concentrations of 4, 10 and 4 μg mL⁻¹, respectively. Ruthenium red staining, which binds to the cells surface, was used to determine the number of cell events and the relative volume of the cells, assuming that cells were spherical in suspension. Analysis was performed using SC-ICP-TOF-MS to determine the relative masses of proteins in the cells, making it possible to evaluate changes that occur in cells under different stress conditions such as, in this case, hyperglycaemia and oxidative stress.

Amplification using DNA and AuNPs has now become fairly routine. Zhang et al.¹⁴ developed a DNA amplification method to determine ochratoxin-A (OTA), a neurotoxin found in contaminated cereal. The experimental details were complicated, but involved a biotin-labelled hairpin, containing the OTA aptamer and rolling circle amplification (RCA) primer, attached onto streptavidin-coated magnetic beads. Thus, OTA, binding between the OTA molecule and its aptamer liberated the RCA primer. Subsequent hybridisation of a dumbbell DNA template with the free primer initiated RCA amplification and production of repeated sequences. A DNA-AuNP tag, complementary with the RCA products, was then hybridised and detection was performed using ICP-MS. The method was used for detection of OTA in flour, with recoveries between 91.6% and 107% for concentrations in the range 10 to 500 pM, being comparable with an ELISA method.

A related method of amplification was developed by Zhu et al.¹⁵ for the multiplex determination of carcinoembryonic antigen (CEA) and a-fetoprotein (AFP) in human serum. In this case the DNA bases at the 5′ ends of the aptamers were replaced with corresponding RNA bases to enhance amplification efficiency. Signal amplification using both Au- and Cu-NPs was used to amplify the signals measured by SP-ICP-MS. The dynamic ranges were from 0.01 to 100 ng mL⁻¹ and 0.01 to 200 ng mL⁻¹ and LODs were 0.6 and 0.5 pg mL⁻¹ for CEA and AFP respectively.

1.2. Sample introduction

1.2.1. Nebulisation. Capillary vibrating sharp-edge spray ionisation (cVSSI) was proposed as a nebulisation technique for ICP-MS.¹⁶ cVSSI acts as a low-flow, in-line, continuous nebulisation source and differs from conventional PN in that the nebulisation process is independent from the carrier gas used to transport nebulised aerosols into the plasma. The design, manufacture and operation of the sample introduction chamber used to integrate cVSSI with ICP-MS was described. Using this sample introduction system, sensitivity for Ce was found to be 1.98 × 10⁴ counts per s per ng per mL at a sample flow rate of 15 μL min⁻¹ with an absolute sensitivity of 5.24 × 10¹⁷ counts per g. Signal stability was better than 10% RSD for a 45 min measurement.

Janeda et al.¹⁷ developed a discrete sample introduction system (DSIS) for μ-volume nanomaterial suspension injection into MIP-OES, and simultaneous determination of Cd and Pb pre-concentrated onto oxidised multiwalled carbon nanotubes. Minimal dead volume, high nebulisation and suspension transport efficiency were obtained by using a miniaturised spray chamber and v-groove PN interface for 10 μL suspension injections in discontinuous mode (at 0.5 mL per min pump speed). LODs of 0.7 and 0.1 ng mL⁻¹ were obtained for Cd and Pb, respectively and precision ranged from 6 to 8%. Quantitative data obtained for water RMs (ERM-CA011b and SRM 1643e) were in good agreement with the corresponding reference values. Martinez¹⁸ coupled a high temperature torch integrated sample introduction system (hTISIS) to MIP-OES to allow total sample consumption. Operating conditions were optimised for sensitivity, LOQs and BECs for the determination of Ca, Cr, Cu Fe, K, Mg, Mn, Na, Pb and Zn, and compared with a conventional sample introduction system. Under optimum conditions, the hTISIS improved analytical figures of merit and shortened wash-out times when using a cyclonic spray chamber. Sensitivity was improved between 2 and 47-fold and LOQs between 0.9 and 360 μg kg⁻¹ were reported. Interference caused by acid matrices were also reduced compared with a conventional spray chamber. The result of the analysis of digested oil samples using this total consumption method were found to be comparable with those obtained using ICP-OES. Sánchez et al.¹⁹ also coupled a total sample consumption IR-heated system (IR-TSC) to ICP-MS-MS and the system performance evaluated for 27 elements (Ag, Al, As, B, Ba, Be, Bi, Cd, Co, Cr, Cu, Fe, K, Li, Mn, Mo, Ni, Pb, Rb, Sb, Se, Sn, Sr, Ti, Tl, V and Zn). Operating conditions were optimised to improve sensitivity, signal stability and plasma thermal conditions. Signal intensity was found to increase by a factor of between 2 and 4 compared to a conventional sample introduction system with LODs improved by a factor of 1.2 (Ba) to 3.3 (Cr) for a sample with HNO₃ added to a concentration of 2%. For a 10% ethanol sample the LOD improvement factor ranged from 1.4 (Ti) to 3.8 (Cr). The matrix effects from solutions containing inorganic acids or high sodium (2% HNO₃, 10% HNO₃, 20% HNO₃, 10% HCl; 500 mg per kg Na, 500 mg per kg Na + 10% HNO₃) were mitigated under optimum conditions (IR-TSC at 125 °C). It was not possible, however, to correct the signal enhancement effect of organic solvents (10% ethanol, 10% acetic acid, and 10% formic acid) even using an internal standard. A range of CRMs and liquid samples were analysed with good agreement to reference values.

Heating the entire sample introduction system of an ICP-MS instrument with an IR heater was found to improve analytical performance.²⁰ Sensitivities were improved by a factor of between 2 and 16 times for most analytes, with high mass elements showing the greatest improvement. LODs were improved by a factor of between 1.1 and 3.5 times. Oxide formation doubled and the formation of doubly charged species decreased 3-fold, suggesting a drop in plasma temperature. However, plasma robustness increased significantly, illustrated by the increase in the ⁹Be⁺ : ⁷Li⁺ signal ratio from 0.174 to 0.235 with IR heating. The IR-heated system enabled the accurate analysis of a drinking water SRM (for Be, Mg, Al, Cr, Mn, Co, Cu, Zn and As) and a waste water SRM (for Be, Mg, Cr, Mn, Co, Cu, Zn, Cd, and Pb) by simple external calibration.

1.2.2. Single particle analysis. Over 20 years since the first publications describing SP-ICP-MS, the technique has reached a level of maturity, with an increasing number of applications in a wide range of fields. Despite this trend, there are aspects related to the fundamentals of the technique that require further study. Laborda et al.²¹ provided a useful review describing the fundamentals of the technique along with the steps and processes involved (from sample introduction to signal processing), focusing on the benefits, current limitations and future prospects for the technique A review by Meng et al.²² also provided a description of recent advances in ICP-TOFMS instrumentation and methodology, followed by a discussion of the use of SP-ICP-TOFMS in the analysis of NPs in the environment. SP-ICP-TOFMS has the potential to identify and quantify both anthropogenic and natural NPs in the environment, providing valuable insights into their occurrence, fate, behaviour and potential environmental risks.

Metals have a fundamental role in microbiology, and accurate analytical methods are essential. The successful treatment of disease can be impeded by the inability to assess cellular heterogeneity. Unlike bulk approaches, single-cell analysis allows elemental heterogeneity across genetically identical populations to be related to specific biological events and therefore to the effectiveness of drugs. SP-ICP-MS can be used to analyse single cells in suspension and measure cellular heterogeneity. Davison et al.²³ reviewed advances in SP-ICP-MS, LIBS and LA-ICP-MS for the elemental analysis of tissues and single cells. The authors identified that the effect of pre-treatment of cell suspensions and cell fixation approaches require further study, and novel validation methods are needed because bulk measurements are unsatisfactory. SP-ICP-MS has the advantage that a large number of cells can be analysed; however, it does not provide spatial information. In contrast LA-based techniques can provide elemental mapping at the single-cell level. The sensitivity of commercial LIBS instruments restricts its use for sub-tissue applications, however, the capacity to analyse endogenous bulk components paired with developments in nano-LIBS technology shows potential for cellular research. LA-ICP-MS offers high sensitivity for the direct analysis of single cells, but standardisation also requires further development. The hyphenation of these trace elemental analysis techniques and their coupling with multi-omic technologies for single-cell analysis certainly show potential in answering fundamental biological questions.

Bastardo-Fernández et al.²⁴ reported a new method for the analysis of TiO₂ NPs in food stimulants using a high efficiency sample introduction system (HESIS) for SP-ICP-MS-MS. Mass-shift mode (O₂ + H₂ as a mixed reaction gas) was used to avoid isobaric interferences from ⁴⁸Ca⁺, allowing accurate characterisation of TiO₂ NPs in complex matrices such as foodstuffs. Two HESIS systems were compared with a conventional sample introduction system with respect to sensitivity and LOD for the determination of TiO₂ NPs in various (organic) food simulants. One system provided a reduction in matrix effects and lower size-detection compared with a conventional introduction system. To avoid saturation of the detector when detecting large TiO₂ particles (>300 nm), analysis was carried out using a forced analogue detection mode and accurate characterisation of TiO₂ NPs, both in terms of size and concentration, was achieved. An average size-detection of 20 nm over 3 days was determined. The results of this study confirm the feasibility of HESIS for the characterisation and quantification of TiO₂ NPs in complex matrices such as foods. Furutani et al.²⁵ used a new system to create a liquid aerosol, formed of ‘aeromicelles’, to deliver aqueous samples directly into the vacuum region of an SP-MS. The aeromicelles were generated by spraying an aqueous solution containing a surfactant at a concentration significantly lower than its critical micelle concentration (CMC). When the solution was sprayed, liquid droplets containing the surfactant were formed. As the droplets gradually dried the surfactant concentration in the droplet exceeded its CMC, so the surfactant molecules eventually fully covered the droplet surface, resembling a reverse micelle. The surface coverage helped reduce the evaporation of water, enhancing the residence time of the liquid droplet. Results demonstrated that the aeromicelles remained in liquid form for at least 100 s in air and were delivered into the vacuum region of an SP-MS, where they were ablated and mass analysed. Individual aeromicelles generated from an aqueous solution containing 10 nM CsCl yielded Cs⁺ ions peaks in the MS spectrum. The number of Cs atoms in each was estimated to be approximately 7 × 10³, which corresponds to 1.2 × 10⁻²⁰ mol (12 zmol). Tyrosine was also analysed, giving both positive and negative fragmentation ions in the mass spectrum. 4.6 × 10⁵ (760 zmol) tyrosine molecules were detected.

Like most ICP-MS-based measurements, matrix effects can be a major challenge for accurate quantification in SP-ICP-MS. Online microdroplet calibration was proposed²⁶ to overcome the matrix effects observed for the analysis of NPs and microparticles in seawater. With online microdroplet calibration, particle-containing samples were introduced into the ICP along with monodisperse microdroplets containing known element mass amounts. The microdroplet standards, which experienced the same plasma conditions as the analyte particles, were used to measure matrix-matched absolute element sensitivities. A single multi-elemental standard could be used to determine the element mass amounts in diverse types of analyte particles independent of the sample matrix. Results obtained demonstrated mass recoveries of between 98 and 90% for the analysis of individual gold NPs in ultrapure water and 99% seawater. In the analysis of food-grade TiO₂ submicron particles, accurate Ti-mass per particle was determined with matrix-induced signal attenuation of up to 80% in a pure seawater matrix. Accurate diameter determinations of individual 3.4 μm polystyrene beads at concentrations of up to 80% simulated seawater were achieved in addition to the simultaneous and accurate quantification of rare-earth elements in polystyrene beads. Interest in the use of SP-ICP-MS for nanoparticle (NP) characterisation in organic matrices continues to grow. The effect of organic matrix interferences on nanoparticle characterisation (i.e. number concentration and size distribution) was investigated by Torregrosa et al.²⁷ Au-, Pt- and SeNPs were characterised in two carbon-containing matrices (6% w/w glycerol and 10% w/w ethanol). It was observed that 10% w/w ethanol gave rise to positive matrix effects on the number concentration due to changes in aerosol generation and transport compared to aqueous standards. Irrespective of the organic source employed, either positive or negative bias on the size distributions could be obtained and these effects depended on plasma operating conditions, instrument characteristics and NP composition. In addition to changes in transport efficiency, matrix effects on size distribution also depended on plasma characteristics and carbon-based charge transfer reactions (Au and Se). These non-spectral interferences were mitigated by means of internal standardisation. When operating with 10% w/w ethanol, two internal standards (ionic solution plus a NP suspension of known concentration) were required to correct for changes in analyte ionisation and transport. For 6% w/w glycerol, however, a single internal standard (ionic solution) was sufficient because this matrix was not observed to affect aerosol generation and analyte transport. Torregrosa et al.²⁸ also reported on the role of aerosol transport on NM characterisation by SP-ICP-MS. Determination of a suspension containing 70 nm Pt NPs was compared with that of a 10 ng mL⁻¹ ionic Pt solution under different operating conditions. The Pt NP event intensity and the Pt ionic signal depended on both aerosol transport and plasma operating conditions. Tertiary aerosol characterisation demonstrated that the transport efficiency of NMs, ionic Pt solution and solvent differed significantly. Irrespective of the nebulisation conditions, transport efficiency was found to follow the order; solvent > ionic Pt > Pt NPs.

1.2.3. Vapour generation. Vapour generation (VG), used as a method of sample introduction, has several advantages: the analyte is separated (as a vapour) from the matrix; analyte transport into the source approaches 100%; and analyte preconcentration is possible using techniques that allow adsorption and subsequent desorption of the vapour. Disadvantages include: interferences at the vapour generation stage caused by co-reacting matrix elements; incomplete reaction of the analyte; instability of the generated analyte vapour; and the frequent requirement to use specific reaction conditions for different analytes. Nevertheless, VG has been a popular method when applied to easy-to-generate, volatile hydrides of metalloids, and more recently for the generation of volatile transition metal species. Several techniques for the generation of volatile species have been developed over the years:

• chemical vapour generation (CVG) using e.g. sodium borohydride or tetraethyl borate, in the presence of various catalysts, is well established;

• photochemical vapour generation (PVG) using intense UV radiation is now also commonly used, where vapour generation of volatile species is thought to be promoted by photo-induced free radical reactions in the presence of suitable chemical agents;

•plasma induced vapour generation (PIVG) has recently proved popular, particularly when used in conjunction with miniaturised plasma sources.

Chemical vapour generation (CVG) is now a mature technique, so new developments are limited, however, the reaction mechanisms are still not fully understood. Pitzalis et al.²⁹ investigated the mechanism of PbH₄ generation by reaction of Pb^II with aqueous NaBH₄ both in the presence and in the absence of K₃Fe(CN)₆, using GC-MS, AFS and deuterium labelling. The authors observed that, in the absence of K₃Fe(CN)₆ and from pH 1 to 8, Pb²⁺ and/or Pb(OH)⁺ were initially converted to solid, non-volatile species as hydrides and/or hydrido-metal complexes containing Pb–Pb and Pb–H bonds. However, these unstable species then subsequently formed elemental lead NPs, without formation of volatile Pb hydrides. Above pH 12 inorganic Pb^II is mostly present as Pb(OH)₃⁻, which is unreactive towards BH₄⁻. However, at greater than nM concentrations, reaction rate increased because reactive polycations of Pb formed, and the formation of elemental lead NPs was observed. In the presence of [Fe(CN)₆]³⁻, Pb²⁺ and Pb(OH)⁺ were converted to PbH₄ without formation of solid species. The authors thought that the most likely mechanism involved the interaction between [Fe(CN)₆]³⁻ and Pb hydride-metal complexes (HMCs) and intermediates. They postulated that the selective oxidation of Pb–Pb bonds converted the solid, polynuclear Pb HMCs to mono-nuclear fragments, which were converted to PbH₄ by reaction with BH₄⁻.

Photochemical vapour generation (PVG) requires the presence of a low molecular weight organic acid and a transition metal ion to ‘sensitise’ the reaction of the analyte to form a volatile species. Research into the best combination of organic acid and sensitiser for various analytes is an active area. Musil et al.³⁰ found that 4 M formic acid, with 10 mg per L Co²⁺ and 25 mg per L Cd²⁺ as added promoters, was the ideal medium for PVG of Ir³⁺ and Ir⁴⁺, giving over 90% efficiency. They used a flow-through PVG reactor with a 19 W low-pressure mercury discharge lamp, and an irradiation time of 29 s. LODs were between 3 and 6 pg L⁻¹ (1.5 to 3 fg absolute) using ICP-MS detection. Suppression of the analyte signal caused by HNO₃ was significant above 5 mM (26% decrease in sensitivity), with a similar effect observed for NaNO₃. Suppression was less pronounced in the presence of H₂SO₄ (11% at 200 mM), HCl (10% at >500 mM) and NaCl (up to1 M).

Hu et al.³¹ reported PVG of Zn and Ga. They used formic acid (30% v/v) and 20 mg per L Cu²⁺ to affect the PVG reaction in a stopped flow (stopped for 7 s) photochemical reactor at a flow rate of 4.0 mL min⁻¹, to give a total irradiation time of 20 s. After passing through two tandem gas liquid separators, volatile species were detected using ICP-MS. The LODs were 8.8 ng mL⁻¹ and 0.77 ng mL⁻¹ for Zn and Ga respectively, reflecting the low generation efficiency of only 1%. Nanoparticles containing Cu, Zn or Ga were found in the liquid solution after UV irradiation, so the reaction seems to require further optimisation to be effective.

Dong et al.³² used PVG for the determination of Sb using the medium of 10% (v/v) formic acid + 10% (v/v) acetic acid + 80 mg L⁻¹ of V^IV. GC-MS was used to detect the formation of (CH₃)₃Sb both from Sb^III and Sb^V. The authors postulated a mechanism, based on previous studies using Te as a sensitiser, such that Sb^III and Sb^V were reduced to Sb⁰ by ˙H and ˙CO₂⁻ radicals or electrons formed by photochemical decomposition of formic acid. The Sb⁰ species then reacted with ˙CH₃ to form (CH₃)₃Sb under UV irradiation. The LOD was 4.7 ng L⁻¹ for Sb with ICP-MS detection, with 19-fold greater analytical sensitivity compared to direct solution nebulisation.

Su et al.³³ developed a solid phase, photothermal-induced PVG method for the determination of Hg by AFS. The analyte was first preconcentrated from solution using a SPME array of 20 nano-TiO₂-coated tungsten fibres. The fibres were inserted into an inner tube inside a UV lamp, and Hg⁰ was generated and desorbed in the presence of formic acid vapour. Interferences caused by a selection of cations and anions were absent below concentrations of 10 mg L⁻¹ for a 1 μg L⁻¹ concentration of Hg. An LOD of 2.3 ng L⁻¹ was achieved. The same group³⁴ developed a headspace sampling device for photochemically generated Hg, Ni and Se species. The system operated much like a GC autosampler. Glass sample vials with septa were arranged in a rotating carousel around a coiled quartz tube wrapped around a 15 W low pressure mercury vapor UV lamp (253.7 nm). Pre-digested samples were aliquoted into the vials together with formic acid to affect the PVG reaction and produce volatile species in the sealed vials. The vials were then pierced and sampled using a headspace sampling arrangement controlled with an electromagnetic valve, so that the generated volatile species in the headspace of the vials were swept, by an Ar gas flow, into an AFS or μPD-OES for detection. LODs of 0.002, 0.007, and 0.01 μg L⁻¹ were obtained for Hg^II, Ni^II, and Se^IV respectively using AFS, and 0.02 and 0.2 μg L⁻¹ for Hg^II and Ni^II respectively using μPD-OES. Interferences from a variety of transition metals were negligible (better than 90% recovery for 1 μg L⁻¹ analyte in most cases) up to concentrations of 1000 mol L⁻¹. Recoveries for Hg, Ni and Se in several waters and predigested fish sample CRMS (DOLT-5 and DORM-4) were between 92% and 106%.

Li et al.³⁵ reviewed (95 references) the use of PVG for speciation of As, Hg and Se. The review covered both speciation and non-speciation applications, with tables of data containing figures of merit for speciation analysis using chromatographic separation.

Plasma induced vapour generation (PIVG) was investigated by Liu et al.³⁶ for the determination of As using AFS. In order to improve the vapour generation step, they pre-irradiated a flowing sample (in 10% formic acid) using a 19 W flow-through low-pressure mercury lamp embedded with a high-permeability synthetic quartz tube (3 mm o.d. × 2 mm i.d.). This had the effect of generating reactive species prior, to vapour generation in a dielectric barrier discharge, which increased the As signal between 10- and 16-fold compared to using UV-PVG or PIVG alone, resulting in an LOD of 0.13 μg L⁻¹. They found that this method was more tolerant of interferences in some limited cases (e.g. NO₃⁻). However, it should be noted that what constitutes an interference in one study is classed as a ‘sensitiser’ in another – e.g. the addition of a transition metal ion such as Co²⁺ to enhance vapour generation.

Coelho et al.³⁷ used a DBD to generate volatile Hg⁰ from single sample drop containing inorganic mercury and/or methylmercury, using AAS for detection. The DBD electrodes were contained within a glass tube in a PVC body. A 2 μL drop was placed directly onto the lower, stainless-steel, grid electrode and the DBD generated using He (150 mL min⁻¹) and a HV power supply (23 W, 38 kV, 40 kHz). Conversion of both Hg²⁺ and MeHg⁺, to Hg⁰ was observed, with efficiencies of 87% ± 8% and 91% ± 10%, respectively. However, the LOD of 100 μg L⁻¹ (200 pg absolute) was high compared to CVG due to the small sample volume.

One of the impediments to the portability of field-deployable PIVG devices is the requirement for specialist gases to generate the plasma. Zheng et al.³⁸ developed a PIVG system which utilised in situ H₂ generation, using a cathode composed of cobalt-phosphorous nanomaterial, in order to address this problem. However, the experimental description includes the requirement for an Ar purge of the electrolytic cell, so safety rather than portability would seem to have been the main motivation. A point discharge (PD) plasma was coupled with CCD-OES for detection of Cd, Hg and Zn with LODs of 10, 0.8, and 14 μg L⁻¹ respectively.

1.3. Direct methods

Solution-cathode glow discharge (SCGD) is a useful plasma source for elemental analysis, combining atomisation/excitation/ionisation capabilities for aqueous solutions with relative simplicity. When combined with OES, ng per mL LODs were achieved for multiple elements. The plasma source remained stable in the presence of dissolved non-analyte ions. However, the mechanisms by which the SCGD can transfer the dissolved analytes from the solution into the plasma, and the mechanisms of how concomitant ions in the sample solution affect the performance of the analysis, are not fully understood. Orejas et al.³⁹ carried out fundamental plasma diagnostics of the SCGD, including electrical measurements, n_e, T_rot and T_exc measurements. Measurements were performed using solutions with NaCl concentrations up to 3.5 g L⁻¹ to understand how NaCl affected the atomisation/excitation efficiency and the analyte solution-to-plasma transfer processes of the source. Some initial conclusions were discussed and future work suggested. The use of Solution-Anode Glow Discharge (SAGD), operating in an H₂–He mixture, as an excitation source for the determination of As was evaluated by Cai and Wang.⁴⁰ Using an enclosed system, the H₂–He mixture was fed into the discharge chamber via a non-electrode hollow channel. The H₂–He plasma integrated vapour generation, atomisation and excitation of As, allowing determination without coupling any separation and/or enrichment technique, such as HG. The effects of the addition of small organic molecules were investigated. Propionic acid (0.1%, v/v) was found to show an amplification effect of up to 3.4×. The LOD for As was calculated as 28 μg L⁻¹, an improvement on the open to air system. Chemical interferences due to transition metals were observed to a higher degree than that due to alkali and alkaline earth metals, this is also observed with CVG. Direct determination of As was also reported using an H₂-doped SAGD.⁴¹ It was found that efficient vapour generation and excitation of As could be achieved simultaneously in an Ar–H₂ SAGD, contributing to the significant improvement in As signal sensitivity, compared to that in SAGDs without H₂ doping or SCGD, and eliminating the need for HG. Under optimal operating conditions, the LOD for As was found to be 1.4 μg L⁻¹ (193.7 nm), with good calibration linearity in the range of 5–500 μg L⁻¹. Ionic interference was found to significantly suppress the As signal and it was required to remove concomitant cations prior to analysis using cation exchange resin. The accuracy of the proposed method was validated by the analysis of a CRM [GSB 07-3171-2014] and real water samples.

Gorska et al.⁴² evaluated the impact of the material and diameter of delivery tubes used in flowing liquid cathode (FLC)-APGD. Tube materials tested were W, Mo, Ti, corundum, quartz and graphite. To estimate the impact of various combinations of tubes on the excitation conditions in the discharge, T_exc, T_rot, T_vib and n_e were determined. It was found that certain combinations of tubes affected the discharge stability; either the precision was poor or the discharge could not be maintained at all. LODs of the studied elements were assessed for some combinations of tubes in order to establish the impact of the discharge stability on the background fluctuations. Tungsten and corundum were found to be superior tube materials in terms of their resistance to the conditions under which the discharge was sustained. The analyte signals were found to be slightly affected by tube material and it was again shown that the W and corundum tubes provided better results in terms of signal intensity compared with the other materials tested. Interestingly, the commonly used combination of glass and graphite tubes resulted in the lowest analyte signals and relatively higher LODs.

2. Solids analysis

2.1. Direct methods

A review of the application of slurry sampling to elemental analysis included the fundamental processes, advantages, limitations and new trends in this established methodology.⁴³ Newer applications of slurry sampling, such as dispersive SPE with the direct introduction of the adsorbent into the equipment, coupling to flow injection analysis, the use of slurries in TXRF, and gas phase MAS and the performance of speciation analysis, were also discussed.

2.1.1. Arc & spark. Direct AES determination of F impurities in solid (powdered) carbon materials was proposed⁴⁴ using two simple radiation sources: DC arc (25A) and high-power low voltage spark (300 V, of 2348 μF). For each source, standard extra high purity graphite electrodes and an Ar discharge atmosphere were used. Spectra were recorded using a compact, tunable fibre-optic spectrometer with linear CMOS sensor and short-focus lens. Measures were taken to counter red spectral region interferences, such as CN molecular bands and thermal radiation. The excitation of F in both types of radiation source was studied. A set of artificial graphite and NaF based standards were prepared for calibration and evaluated by two independent approaches: indirect DC arc AES and water extraction photometry. LODs of 5 μg g⁻¹ and 20 μg g⁻¹ were obtained for arc and spark, respectively. It was noted that, as the power of DC arc is limited by generator capabilities, it is possible that a lower LOD could be achievable with higher discharge current.

SIBS is a qualitative and quantitative technique capable of real-time, in situ, on-line rapid detection, with relatively high sensitivity, small sample volume, low cost and simple maintenance. The basic principles of SIBS, recent technological innovations and the advantages and disadvantages of the method were described by Zheng et al.⁴⁵ Future developments were also discussed for the technique, with applications to environmental monitoring, industrial health, food safety, and biomedicine.

2.1.2. Glow discharge. GD-OES elemental mapping can provide large area maps directly and cost-effectively from solid samples within tens of seconds. The benefits of the approach to the characterisation of NPs in terms of mass, elemental composition and size/structure dimensions were demonstrated for Ag and Au NPs by Finch et al.⁴⁶ The effects of GD pulsed power, pressure, and sample substrate were studied, and optimised conditions resulted in LODs at single pg levels. While this is not at the level of single NP sensitivity, size differentiation of Ag and Au NPs was successfully demonstrated between 5 and 100 nm, while the internal dimensions of complex core–shell NPs were also identified through the optical emission changes as a function of time.

2.1.3. Secondary ion mass spectrometry. TOF-SIMS is an important technique for chemical imaging of NMs and biological samples with high lateral resolution. However, low ionisation efficiency limits the detection of many molecules at low concentrations or in very small volumes. A promising approach to increasing the sensitivity of the technique is by the addition of a matrix that promotes ionisation and desorption of important analyte molecules, an approach known as matrix-enhanced secondary-ion mass spectrometry (ME-SIMS). The effect of matrix acidity on molecular ion formation in three different biomolecules was investigated by Pohkrel et al.⁴⁷ A series of cinnamic acid-based matrixes that varied in acidity was employed to systematically investigate the influence of matrix acidity on analyte ion formation. The positive ion signal for all three biomolecules showed a strong increase for more acidic matrixes. This work indicated that proton donation plays an important role in the formation of molecular ions in ME-SIMS.

2.2. Indirect methods

2.2.1. Laser ablation. The first lasers were introduced around 60 years ago. The unique laser property of high peak power in short pulses has led to applications in which light energy replaces mechanical energy for removing mass, structuring surfaces, creating new materials, weapons, remote analysis, fusion, surgery, and many other applications that fall under the term laser ablation (LA). Russo⁴⁸ provided a review of LA research and development, including fundamental behaviour, some unique applications with emphasis on chemical analysis, and the current interest in measuring IRs in laser induced plasmas at atmospheric pressure.

Several processes affect the measured concentration of elements determined by LA-ICP-MS. Depending on the element of interest, a range of corrections are required to account for sensitivity drift and downhole effects, among other sources of inaccuracy. A method of calibrating LA-ICP-MS measurements taking into consideration time-dependent sensitivity changes to produce a 3D calibration surface through time was proposed by Paul et al.⁴⁹ These 3D calibration surfaces resulted in up to 20% improvements in sensitivity for some elements. To ensure that calibration surfaces created using multiple RMs were not degraded by matrix effects, a median yield correction factor was determined relative to a primary RM. Sensitivity modelling was used for elements without accepted values or where interferences might affect calibration. In addition to the correction of drift, a correction for downhole fractionation effects that can improve the precision of spot analyses by up to 20% in a sample dataset was demonstrated. Combined, these approaches were shown to improve the accuracy and precision of concentration determinations by LA-ICP-MS in a wide variety of applications.

A new ablation chamber was designed to be used with LA-MPT-OES for the direct determination of heavy metals in soil.⁵⁰ With this chamber, the washout time and RSD were almost one-third of those of the conventional chamber, indicating a smaller dead volume to provide effective transport of ablated sample particles. To ensure a high signal intensity during long exposure times, a moving sampling method was used to guarantee a sufficient injection volume. LODs of 0.075, 0.093, 0.068, 0.009 mg kg⁻¹ were obtained for Cu, Pb, Cr and Ag, respectively, representing improvements of 1 to 2 orders of magnitude compared to LIBS and XRF and similar to those obtained by techniques requiring prior sample digestion (e.g. ICP-OES and MPT-OES) and other LA-based techniques (e.g. LA-ICP-MS). The LA-MPT-OES method was used to analyse SRMs and the results found to be in agreement with reference values. Laser beam profiles in analytical LA-ICP-MS instruments are in general homogenised to produce a flat-top beam profile. However, in practice, they are mostly super-Gaussian in nature, and for small laser beam sizes (<5 μm) they approach a Gaussian profile. This implies that the amount of surface material sampled by the laser directly depends on the beam profile and ablation grid. Contraction of the ablation grid (to sub-pixel mapping) can produce more accurate surface sampling, a higher pixel density, an improved spatial resolution, and improved SNR although LA sampling is predominantly performed on an orthogonal grid, hexagonal sampling may further improve the image quality as regular hexagons are more compact than squares and suffer less from orientation bias. Due to the current limitations of LA stages in executing precise hexagonal sampling with small beam sizes, computational protocols were employed by Metarapi et al.⁵¹ to simulate LA-ICP-MS mapping. Simulation was performed by discrete convolution using the crater profile as the kernel, followed by the application of Poisson/Flicker noise related to the local concentration and instrumental sensitivity. A free, online app was developed (https://laicpms-apps.ki.si/webapps/home/) to study the effect of sampling grid contraction (orthogonal and hexagonal) on the image map quality (spatial resolution and SNR) by virtual ablation of phantoms. Comparison of experimental LA-ICP-MS maps obtained through orthogonal and hexagonal sampling methods was performed using a beam size of 150 μm and a macroscale inkjet-printed resolution target.

Schweikert et al.⁵² described a method for quantitative bioimaging by LA-ICP-TOFMS involving the design of standardisation methods based on robotic micro-droplet dispensing. The production of controlled and highly precise pL droplets was exploited to enable on-tissue isotope dilution and standard addition. Both strategies eliminated matrix effects and offered accurate, traceable quantification of μm-sized regions of interest in tissue samples, as defined by the extension of the deposited pL-volume droplet. Although IDA was restricted to single element quantification, multiplexed matrix-matched calibration was obtained by on-tissue standard addition by depositing a dilution series of certified multi-element standards. Comparison with external calibration by gelatin μ-droplets demonstrated their applicability as accurate, high-throughput standardisation. In fact, gelatin μ-droplet standards provided comparable results to matrix-matched on-tissue quantification, while being easier to use. They also mimicked the matrix well enough to be used as matrix-matched standards for a great variety of biological matrices.

The digital mapping of biomarkers in tissues based on desorption and counting intact Au NP tags using infrared (IR)-LA-SP-ICP-MS was described by Stiborek et al.⁵³ In contrast to conventional UV laser ablation, Au NPs are not disintegrated during the desorption process due to their low absorption at 2940 nm. The technique is demonstrated by mapping a proliferation marker, nuclear protein Ki-67, in 3D aggregates of colorectal carcinoma cells, and the results compared with confocal fluorescence microscopy and UV-LA-ICP-MS. Precise counting of 20 nm Au NPs with a single-particle LOD in each pixel by the new approach generated sharp distribution maps of a specific biomarkers in the tissue. Advantageously, the desorption of Au NPs from regions outside the tissue was strongly suppressed. The developed methodology suggested promise for the multiplex mapping of low-abundant biomarkers in biological and medical applications using multi-elemental mass spectrometers.

Methodology for the quantitative bioimaging of essential metals including Mg, Mn, Fe, Cu and Zn in tissue cryosections based on the use of high-resolution LA-ICP-TOFMS was developed by Strekopytov et al.⁵⁴ Cryosectioned gelatin standards spiked with the elements of interest were used for calibration. Analysis under ‘no gas’ conditions showed satisfactory selectivity and 3-fold improvement in LODs for Mn, Fe and Cu when compared with those obtained using 2.4 mL min⁻¹ H₂ in the collision/reaction cell of the ICP-TOFMS instrument. Absolute single-pixel LODs at 3 μm spatial resolution were in the range of 4–9 fg for Mn, Cu and Zn and approximately 40 fg for Mg and Fe. The impact of variable thickness of cryosectioned gelatin on the signal intensity was studied and a linear response was established for thicknesses between 10 and 30 μm. The developed method was used for the quantification of essential metals in kidney sections of rats.

LA-ICP-MS continues to be a popular technique for quantifying and tracking NPs, its quantitative calibration remains a major challenge, however, due to the lack of suitable standards and the uncertain matrix effects. Luo et al. describe an approach to preparing quantitative standards via precise synthesis of NPs, nanoscale characterisation, on-demand NP distribution, and deep learning-assisted NP counting.⁵⁵ AuNP standards were prepared to cover the mass range from sub-femtogram to picogram levels with sufficient accuracy and precision, thus establishing an unambiguous relationship between the sampled NP number in each ablation and the corresponding mass spectral signal. This strategy permitted the study of the factors affecting particulate sample capture and signal transductions in LA-ICP-MS analysis and facilitated the development of an LA-ICP-MS-based method for absolute NP quantification with single-NP sensitivity and single-cell quantification capability.

2.2.2. Thermal vaporisation. A DSS method was developed by Silva et al.⁵⁶ for determining Hg in bioresorbable Ca₃(PO₄)₂-based materials by ETV-ICP-MS. The ETV system was based on fast sample heating using a halogen lamp enclosed in a glass chamber. The heating program comprises a vaporisation step (10 s) followed by a cooling step (120 s) for solid samples. An additional step of 20 s at 80 °C was required when aqueous calibration standards were used. Hg was vapourised at a relatively low temperature (650 °C) without co-vapourizing the main matrix (more than 95% of the matrix remained). External calibration with aqueous reference solutions deposited on the sample residue was used for Hg quantification. Marine sediment CRM (MESS-3) was analysed and good agreement was found compared with reference values. Relatively high masses (up to 10 mg) were analysed without interferences. The limit of quantification (0.2 ng g⁻¹) by ETV-ICP-MS was lower than that obtained by NEB-ICP-MS (60 ng g⁻¹) with a sample throughput of 20 per hour.

3. Instrumentation, fundamentals and chemometrics

3.1. Instrumentation

3.1.1. Sources. A molten-salt-electrode microplasma source was developed by Cai et al.⁵⁷ The source effectively consisted of a metal T-piece with He plasma gas introduced through the side arm. A hollow stainless steel tube counter electrode was inserted into one end, axially in-line with a fibre optic inserted into the opposite end. The counter electrode was directed towards the molten salt sample and a He discharge generated between it and the sample (He 250 mL min⁻¹, discharge current 30 mA, discharge gap distance 11 mm). Spectral lines of Na, K, Li and Zn the molten salts LiCl–KCl–NaCl, NaCl–ZnCl₂, LiCl–KCl–CaCl₂, and LiCl–KCl were observed, and Li determined at the % level with comparable results to that obtained using ICP-OES.

The problem of inefficient ionisation of F in the Ar ICP-MS was addressed by Tanen et al.⁵⁸ who developed a post-ICP chemical ionisation method to form BaF⁺ ions. They achieved this using an interface composed of two chambers. The first chamber, directly behind a 4 mm nickel plasma sampling cone, was encased in an acrylic tube between two aluminium plates. The ICP was sampled ∼10.5 mm downstream of the load coil, passing through the sampler and then into a quartz tube just inside the first chamber, where plasma cooling and recombination reactions occurred. The end of the tube was fitted with a cylindrical (or plate) brass electrode held at 400 V. Reagent ions were introduced into the first chamber perpendicular to the sampled plasma flow using a nano-spray micro-capillary (∼1250 V), located ∼1 cm downstream of the brass electrode and 7 mm from the central axis of the quartz tube. Sample solutions containing the analyte fluconazole were nebulised into the ICP in 50 : 50 water : acetonitrile with 0.1% v/v formic acid at a flow rate of 50 μL min⁻¹. Thus it was necessary to introduce O₂ into the carrier gas to prevent carbon deposition. Post ICP chemical ionisation was achieved by introducing BaCH₃COOH reagent via the nanospray to promote the formation of BaF⁺ ions. The ions were further sampled into the second chamber, which was vacuum pumped at a rate of 4 L min⁻¹, and thence into a triple-quadrupole MS system. Excess plasma O₂ resulted in a high abundance of HNO₃ in the post-plasma flow, which caused BaNO₃⁺ formation to predominate, rather than BaF⁺. However, this situation was ameliorated by using alternative nanospray electrolytes, with their effectiveness being in the order AlNO₃F(H₂O)_n⁺ ∼ ScNO₃F(H₂O)_n⁺ > LaNO₃F(H₂O)_n⁺ > MgF(H₂O)_n⁺ > BaF(H₂O)_n⁺. The authors commented that this corresponded to fluoride ion association constants with the metals in aqueous solution. The Sc-based reagent gave the best compromise between sensitivity and robustness in the presence of HNO₃ and Cl-containing matrices.

Kuonen et al.⁵⁹ coupled a 1.45 kW nitrogen MIP with MS and evaluated its suitability for trace metals analysis. The effects of operating conditions and cone orifice diameter were investigated. Contour plots of signal versus power and nebuliser gas flow revealed three groups of analyte. Atomic ions of elements with low first IE and low M–O bond strength exhibited maxima at high nebuliser gas flow rate and low power; elements with intermediate IE and higher M–O bond strength required higher power settings for maximum signal; and elements with the highest IE required the highest power and lowest nebuliser gas flow rate. Signals for high IE elements were suppressed compared with elements of similar mass. The authors attributed this to the high abundance of NO species in the plasma source, which is known to dominate in N₂ plasmas. Metal oxide ions were observed at similar or higher abundances than in a conventional Ar ICP. Different sampler cone orifice sizes (0.8 or 1.1 mm) and sampler–skimmer distances (0.5 or 1.0 mm reductions from standard), were investigated and the interface pressure was lowered through an additional pump. Ion transmission was unaffected by these modifications but lowering the interface pressure reduced the relative abundance of MO⁺ species.

Komin and Pelipasov⁶⁰ evaluated the effect of O₂ introduction into a 1.2 kW N₂ MIP. A quartz ICP torch with an injector tube of 1.5 mm i.d. was used. Gas flow rates were: outer, 10 L min⁻¹; intermediate, 0.5 L min⁻¹; and nebuliser, 0.4 L min⁻¹. A concentric nebuliser was used for aerosol sample introduction. Increasing O₂ content was observed to: reduce T_exc from 3.9 × 10³ to 3.3 × 10³ K; reduce the Mg(II) : Mg(I) ratio from 1.02 to 0.18; and reduce n_e from 6.6 × 10¹² to 4 × 10¹¹ cm⁻³. Intensities of molecular emission of N₂, N₂⁺ and NH decreased on addition of O₂ whereas NO, and OH intensities increased up to 20% then decreased (NH) or remained constant. Intensities of atomic and ionic emission lines with E_sum between 1.6 and 14.7 eV were suppressed. The LODs were increased by between 1.2× and 4× for lines with E_sum between 1.5 and 4 eV, and between 50× and 60× for lines with E_sum between 4 and 15 eV.

Koppenaal and Marcus⁶¹ provided a historical, tutorial description of the interfacing and the evolution of their liquid sampling-atmospheric pressure glow discharge (LS-APGD) with Orbitrap-MS for IRMS of U. They described the principles and challenges of the research, including: the problems associated with coupling the source to the MS; the use of collisional dissociation to simplify spectra; optimisation of the MS; data acquisition and digital signal processing. Using the optimised instrumental set-up they benchmarked their U IRMS method against target values of measurement uncertainty of the ²³⁵U/²³⁸U ratio set by the IAEA, and compared with other methods such as TIMS, ICP-SFMS and ICP-QMS.

Štádlerová et al.⁶² evaluated and compared EDL and boosted HCL sources for use in AFS of B at the 222.8 and 223.1 nm lines. Using HG and a miniature diffusion flame, use of the EDL resulted in sensitivity an order of magnitude higher than with the HCL, resulting in LODs of 1.5 pg and 11 pg, respectively.

3.1.2. Miniaturisation. The quest for more compact, portable instruments continues. While there have been great advances in miniaturisation of spectrometers, both sample introduction systems and excitation/ionisation sources still present some challenges. Xu et al.⁶³ addressed the source issue by developing a piezoelectric transformer-driven microplasma (PTM). This consisted of a PTFE block (30 mm × 30 mm × 10 mm) with an 8 mm² hole covered with two quartz pieces to form a discharge cavity. A 0.2 mm o.d. wire electrode (attached to the PT) and a 0.6 mm o.d. syringe needle electrode pierced the PTFE block from opposite ends to form a discharge gap of between 1 and 3 mm from the centre of the discharge cavity. The discharge was formed using a low voltage (from 12 to 20 V) supply to the PT with Ar gas (from 50 to 400 mL min⁻¹) present. The sample was introduced by HG for the determination of Hg by OES, observed using a portable CCD spectrometer. The LOD was 2.8 μg L⁻¹ using the Hg 253.65 nm line. The main advantage of the system was that it did not require a HV power supply. This clearly increased portability, however, Ar carrier gas and a gas–liquid separation system were still required.

Liu et al.⁶⁴ partially addressed these problems with an ultrasonic nebulisation-accelerated gas-phase enrichment (GPE) system, followed by in situ desorption and DBD-OES. The significant part of the system was a batch-mode HG reaction which was performed in, and accelerated by, an ultrasonic nebulisation chamber. The volatile heavy metal species thus generated were separated from the matrix by ultrasonic nebulisation, adsorbed onto the surface of an activated carbon electrode tip, then desorbed and excited by a DBD micro-plasma for multielement OES analysis. An array nebuliser plate with 10 samples was used to speed up the batch mode operation (10 samples in 40 s). LODs for Hg, Cd, Cu, and Sn were 0.005, 0.01, 0.03, and 0.04 μg L⁻¹, respectively. The system was relatively compact but still required HV power, an Ar gas supply, and intervention to change the batch sample plate. An array point discharge was developed by Zhang et al.⁶⁵ for use as a compact excitation source with a miniaturised spectrometer. The source contained three pairs of point discharges arranged in sequence to construct the array, which the authors claimed improved excitation. Samples were introduced by HG and a carrier gas was required. The LODs for As, Ge, Hg, Pb, Sb, Se, and Sn were 0.7, 0.4, 0.05, 0.7, 0.3, 2, and 0.08 μg L⁻¹, respectively.

3.1.3. Spectrometers. A spatial heterodyne spectrometer (SHS) was tested in conjunction with FAAS by Li et al.⁶⁶ for use in high-resolution AAS. Using interferometry and by exploiting modern computational power to enable high-speed real-time Fourier transformation of large data-sets, they were able to extract analytical information from ‘invisible’ features to resolve narrow-band absorption interferograms. Thus, a single SHS absorption interferogram contained both illumination background and absorption information, which could be distinguished by computation. Results demonstrated the construction of Na absorbance spectra from a single image, an approach which can be used to investigate highly dynamic systems.

3.2. Fundamentals

3.2.1. Fundamental constants. Reliable fundamental constants are necessary to perform diagnostic studies. Irvine et al.⁶⁷ used LIBS to determine the radiative transition probabilities of neutral and singly ionised Eu. They used a Saha–Boltzmann plot method after spectral background correction and fitting using the Voigt function. The five transition probabilities for Eu I ranged from 0.172 to 7.38 × 10⁷ s⁻¹ and three for Eu II ranged from 1.56 to 6.75 × 10⁷ s⁻¹. In a subsequent paper⁶⁸ the same group reported 967 values for transition probabilities for La, Ce, Pr, Nd, Sm, Gd, Tb, Dy, Ho, Er, Tm, Yb and Lu determined using the same method. The simplicity of the method allowed them to rapidly acquire replicate data and obtain transition probabilities with uncertainties <10%.

Stark widths for 16 W II lines were determined by Dojić et al.⁶⁹ using an LIP. Using He as the ambient gas they were able to minimise Doppler splitting but still have minimal self-absorption and n_e high enough to produce broadening of spectral lines. They used the Saha–Boltzmann plot method. Stark widths were measured for certain radial position and n_e normalised to 1 × 10²³ m⁻³. Fikry et al.⁷⁰ determined the Stark broadening of the Al I line at 305.007 nm with configuration 3s3p(³P°)4s in an atmospheric LIP. The n_e was estimated (22.433 × 10¹⁷ to 0.870 × 10¹⁷ cm³) using the H_α line at 656.28 nm and the Boltzmann plot was thus used to calculate T_e (3156 to 9981 K) with the assumption of LTE. Values of the Stark broadening parameters varied from 0.244 to 2.156 pm and were inversely proportional to n_e and T_e.

Excitation energies, term energies and effective quantum numbers for the odd-parity 5s5d²D₁ → 5snf ¹F₃ Rydberg states of cadmium from n = 14 to 45, were determined by Khan et al.⁷¹ using multi-step laser excitation. Three dye lasers were used, simultaneously pumped by the second (532 nm) and third (355 nm) harmonics of a Q-switch Nd:YAG laser in conjunction with an atomic beam setup. A two-parameter fit to the transition energies of the 5snf ¹F₃ series yielded the binding energy of the 5 s5d ²D₁ level as 13 320.33 ± 0.03 cm⁻¹ and the first IE of Cd to be 72 540.06 ± 0.15 cm⁻¹. This was in agreement with the value reported in the NIST database.

Isotope shifts and hyperfine structure of Xe were determined by Bounds et al.⁷² using Doppler-free saturated absorption spectroscopy. A tunable narrow-line Ti:sapphire laser was used to resolve hyperfine structure in the spectra of Xe for near-IR transitions of the 6ns–6np series in the 820–841 nm spectral interval. By observing the spectra in an ICP discharge it was possible to overcome Doppler broadening and resolve the separate lines. Lines of the even Xe isotopes (136, 134, 132, 130, 128) were fully resolved for the transitions near 820.860 nm and 841.150 nm. Weak lines of peaks of ¹²⁴Xe and ¹²⁶Xe with low natural abundances were not observed. Liu et al.⁷³ determined the isotope shifts and hyperfine structure of 80 levels in Re I and three levels in Re II by fitting FT spectra obtained using a HCL with Re cathode, running in either Ar or Ne–N₂ buffer gases. The hyperfine structure constants for 80 levels from 0 to 54 729.52 cm⁻¹ and transition isotope shifts for 161 transitions were determined for Re I. Hyperfine structure constants for three levels from 14 352.2 to 22 544.7 cm⁻¹ and isotope shift transition of four transitions in Re II were determined.

Electron impact cross sections for the transitions from the ground state of Cu 3d¹⁰4s ²S_1/2 to the excited fine structure energy levels of the configurations 3d⁹4s², 3d⁹4s4p, 3 d¹⁰4l, (l = p, d, f), 3d¹⁰5l, (l = s, p, d), 3d¹⁰nl, (n = 6, 7; l = s, p) were determined by Agrawal et al.⁷⁴ using a collisional-radiative model for a Cu LIBS plasma. Rate coefficients with calculated cross-sections were obtained and incorporated into a particle balance equation, which was solved for a range of T_e to obtain the level population densities of Cu, which were then used to obtain line emission intensities. For diagnostics, the intensities of three atomic emission lines of Cu (510.5, 515.3, and 521.8 nm) were observed in the absence and presence of a magnetic field at different axial lengths (z = 0.5–6.5 mm) from the Cu target. In the absence of magnetic field, T_e varied from 1.32 to 0.66 eV as distance was increased (0.5–6.5 mm), whereas in a static magnetic field of 0.3 T, T_e decreased from 1.47 to 0.72 eV.

3.2.2. Diagnostics.
3.2.2.1. Plasmas. Diagnostic studies of plasma sources offer insights into the mechanisms of atom formation, subsequent excitation and ionisation, and the formation of molecular species. Pupyshev⁷⁵ reviewed (108 references) the effect of interferences caused by ArM⁺ ions in ICP-MS. As well as a discussion of published studies and the effect of operating parameters, the review contains a very useful table of published values for the theoretical and experimental dissociation energies of ArM⁺ ions.

Kazama⁷⁶ investigated the ion–molecule reactions of Np⁺, Am⁺ and Cm⁺ with reactive CO₂ and O₂ gases in an ICP-MS-MS reaction cell. They developed a gas reaction model to evaluate reaction constants (k) and reaction efficiencies (k_eff) and compared them with values reported using FTICR-MS. Reaction constants were found to be in the order Np⁺ > Cm⁺ > Am⁺. The k_eff values for Am⁺ were much smaller than Np⁺, thought to be due to the high promotion energy from ground to reactive state for Am⁺. The k_eff values were not directly comparable with those reported using FTICR-MS so ratios to the values obtained for Np⁺ were used. The k_eff(Am⁺)/k_eff(Np⁺) was 1/10 of that obtained by FTICR-MS, but the k_eff(Cm⁺) : k_eff(Np⁺) ratios were comparable. These discrepancies were explained by the fact that thermalisation with Ar was used in FTICR-MS but not in ICP-MS-MS. The ratios of the rate constants measured by the two methods were close for Am⁺ and Cm⁺, and the reaction was more efficient with O₂ than that with CO₂.

Hou et al.⁷⁷ developed a self-absorption correction method based on intensity self-calibration of doublet lines belonging to the same multiplet. They defined a K/Δλ₀ parameter ratio, which was a function of various fundamental constants, and used this in the calculation of a self-absorption coefficient of doublet lines of an analytical element, based on a measured line intensity ratio and the K/Δλ₀ parameter ratio. The authors claimed that this method reduced the influence of laser energy, plasma fluctuations and the non-uniform element distribution, and did not require Stark broadening coefficients. The R² values for calibration curves for Al increased from 0.971 to 0.997, with relative errors of 3.89% and 1.33% without and with self-absorption correction, respectively.

Temperature measurement has long been a staple of diagnostic studies. The Boltzmann plot method has been most widely used even though it must be used with caution, particularly in the absence of LTE and/or optical thinness. Bousquet et al.⁷⁸ evaluated the Boltzmann plot method with respect to the effect of SNR, the difference between average temperature and time and space integrated temperature, and sensitivity. They focussed on the determination of T_exc in a LIBS plasma and assumed that the instrument response function was perfectly determined and corrected, and that the plasma was optically thin. They evaluated how SNR affected the temperature values using a numerical approach and by processing experimental data recorded under various SNR conditions. In summary, their recommendations were: LTE and optical thinness should not be assumed; corrections for an optically thick plasma should be made; uncertainties in plot fitting reflect precision rather than accuracy; emission lines should be selected to give the highest SNRs possible and lines with SNR <3 should not be used; lines of different elements will have different SNRs, thus yielding different T_exc for the same plasma, so comparison should only be made between values obtained using similar SNRs; the energy interval range used in the Boltzmann plot should be as large as possible, whilst also taking into account SNR of the lines; and lines with Einstein coefficients of spontaneous emission known with acceptable accuracy (>90%) should be selected.

Gao et al.⁷⁹ determined T_gas using a new optical temperature measurement technique for gas flow fields using ns-LIBS atomic spectral line broadening. The authors cite several advantages for this technique which are, in summary: it is an experimentally simple, single-pulse technique which does not require complex data processing; it can be performed simultaneously with other LIBS measurements, so can thus be used to improve the reliability of other LIBS applications; it has a strong signal so is more robust; it is independent of the laser pulse energy, pulse width, and focal spot size when using laser energy above the breakdown threshold and it is relatively sensitive.

Chemin et al.⁸⁰ investigated LTE in a LIBS plasma ablated from an alumina target. They used LIF to probe the population of the rotational levels in the ground electronic state X²Σ⁺ of AlO, 5 μs and 30 μs after ablation. The T_rot and T_vib temperatures thus determined were different, depending on the excitation channel. This indicated that the molecules did not reach full internal equilibrium, or equilibrium with the plasma kinetic temperature. The authors concluded that collisions were insufficient to ensure excited-state equilibrium. In comparison, AlO molecules formed in the excited state B²Σ⁺ reached equilibrium before emission. Merten et al.⁸¹ used AAS to determine V and Ti stoichiometry in the plasma of a laser ablated Ti alloy target. Measurements were performed at delays between 2 and 20 μs and T_exc calculated using time-resolved Boltzmann plots, in pure helium and an 80 : 20 He : O₂ mixture. The Boltzmann plots included the ground term which the authors claimed produced high-quality estimates of the plasma composition. Using T_exc it was possible to calculate the absolute population of any one level in the atomic state distribution function, thus allowing estimation of the total neutral mass of the plasma, and of Ti and V. While acknowledging the potential systematic error in the oscillator strengths, the authors concluded from temporal trends that there was evidence of post-ablation fractionation in the neutral atomic population by 3.5 μs, possibly due to differential oxidation. Significantly, the influence of the sample oxygen content on plasma chemistry highlighted the importance of matrix-matching.

Wang et al.⁸² used a collisional radiative and hyperfine splitting model to determine n_e and T_e in a miniaturised electron-cyclotron-resonance ion thruster along its radial direction. They first determined the metastable density by the self-absorption effect, then used this in rate-balance equations for other excited species. Hence, the contribution of metastable states was considered under the influence of metastable transport.

The T_e and n_e in a low pressure ICP were determined by Kornev et al.⁸³ The aim of the research was to optimise the conversion of SiF₄ into Si in a H₂–Ar ICP at 3 Torr pressure but the results may have some use in atomic spectroscopy. Langmuir probe, OES, MS and IR were all used to evaluate the plasma conditions. The study confirmed earlier published results that H₂ reduction of SiF₄ proceeded with ˙SiF_x radicals as intermediates. The energy required for H₂ dissociation led to lower n_e (∼10⁹ cm⁻³), reduced efficiency of the H₂–SiF₄ gas discharge and thus to ∼53% conversion of SiF4 into Si. When Ar was added n_e increased (∼10¹¹ cm⁻³), resulting in 90%. Results of IR and MS analyses of exhaust gases showed that they contained 10% of unreacted SiF₄ and ∼5% of various fluorosilanes. Langmuir probe measurements were also made by Senturk et al.⁸⁴ in an argon DC plasma used for magnetron sputtering, operated at between 9 and 60 mTorr and from 300 to 500 V. The n_e decreased and T_e increased from 5.5 to 11 eV with increasing pressure.

Hallwirth et al.⁸⁵ addressed the hoary problem of matrix effects caused by EIEs in the plasma. They set out to establish the effect of 18 major matrix elements on 105 emission lines of 42 elements using an axially viewed MICAP-OES system operated using compromise conditions. Significantly, they observed up to 25% signal suppression of some analytical lines for matrix elements concentrations as low as 20 mg L⁻¹, particularly for alkaline matrix elements Li, Na, K and Cs. Equimolar concentrations of these elements had the same effect, indicating that the absolute number of atoms in the plasma was more significant than IE.

3.2.2.2. Graphite furnaces. An obituary to Boris L'vov, the author of many of the publications featured in this section of ASU over the decades, was published in 2023.⁸⁶ His outstanding contributions to the development and understanding of ETAAS, driven by his great interest in fundamental studies of kinetics and mechanisms of solid-state reactions, inspired work in the field for decades. A recent review of ETAAS by Butcher⁸⁷ includes a summary of its principles, recent fundamental developments and applications of the technique. Future developments in ETAAS, including the continued development of ‘green’ protocols and speciation analyses were discussed. Acar⁸⁸ provided a useful review of chemical modifiers used in ETAAS.

3.3. Chemometrics

Most of the current research into chemometrics is focused on calibration strategies for techniques such as LIBS, so the reader should refer to Section 4 for a discussion of the latest developments. Romanov⁸⁹ made a mathematical assessment of the effect of uncertainty components on measurement results using the concentration ratio calibration (CRC) method with ICP-OES. The most significant uncertainty component was the ratio of line intensities between the analyte and the main matrix component of the sample. Unlike traditional calibration methods CRC did not require the evaluation of uncertainty components for weighing and dilution procedures steps, although these are usually extremely small in any case.

4. Laser-based atomic spectrometry

Key fundamental studies and instrumental developments (published in 2023 and at the end of 2022) in laser-based atomic spectrometry are highlighted in this section. Progress in this area during the previous years can be followed here.¹ Atomic spectrometry techniques where the laser is used as either an intense energy source or a source of precise wavelength (e.g. LIBS, LIF or LIMS) 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.

4.1. Laser induced breakdown spectroscopy

This section describes the latest instrumental developments and fundamental studies related to LIBS. Reviews that cover detailed LIBS applications in imaging, classification, or quantitative analysis of different kinds of samples can be found elsewhere. Recent advances and future perspectives of LIBS imaging for material and biomedical applications were discussed by Gardette et al.,⁹⁰ while novel methodologies and technological developments were summarised by Thomas et al.⁹¹ A tutorial describing in detail how to statistically compare predictive models for LIBS quantitative analysis and classification was published by Duponchel et al.⁹² Recent trends in sample classification using LIBS, providing a comprehensive introduction and overview of the steps to go from recorded data to a well-performing classifier, was reviewed by Brunnbauer et al.⁹³ The development of calibration-free LIBS (CF-LIBS) to improve accuracy in different applications was critically reviewed by Hermann et al.⁹⁴ and Poggialini et al.⁹⁵

In 2023 there were several international conferences dedicated to discuss the recent progress in LIBS, such as the 5th Asian Symposium on LIBS held in Tokushima (Japan) from June 26th to June 30th (https://j-libs.org/CSI2023_ASLIBS2023/index.html), and the 12th Euro-Mediterranean Symposium on LIBS (EMSLIBS) held in Porto (Portugal) from September 4th to 7th (https://emslibs2023.inesctec.pt/).

4.1.1. Fundamental studies. A calibration-free approach for boron isotopic composition calculation using molecular spectra was developed by Mohan et al.⁹⁶ The authors calculated isotopic concentrations using a database of plausible ro-vibrational bands of B²Σ⁺ → X²Σ⁺ transitions, and molecular spectra simulator under Born–Oppenheimer approximation and a Boltzmann equilibrium environment (MAHADEV) algorithm. The boron isotopic concentrations were measured at ambient conditions without any standard sample and with reasonable accuracy and precision.

Dependence of the self-absorption coefficient on the plasma temperature, optical path length and element concentration in a sample was investigated by John et al., ⁹⁷ using a numerical procedure to simulate optically thick emission spectra. Results indicated that self-absorption decreases with increased plasma temperature, whereas it increases with increasing optical path length and analyte concentration. Moreover, an improved method for self-absorption correction, based on the use of Planck function and an iterative numerical calculation of the plasma temperature, was investigated by Völker et al.⁹⁸ The method was compared to other methods that require knowledge of different parameters, such as line shape function, linewidth, or electron density. The authors employed for this study synthetic spectra that fully correspond to the assumed conditions of a homogeneous isothermal plasma at LTE.

Different AuNP shapes, including nanospheres and nanorods, deposited on a Ti target, were investigated by Dell'Aglio et al.⁹⁹ using LIBS. The aim was to explore effects that might be caused by different coupling between the incoming laser and the NPs. Results showed that plasmon-enhanced ablation was only moderately affected by the variation in NP shape, providing similar enhancement values for the different types of AuNPs. Burgos-Palop et al.¹⁰⁰ used ultrafast (ps) laser excitation to investigate optically trapped single NPs in LIBS. They found a higher number of analyte-related lines and reduced presence of air components in the spectra. The authors claimed that this was a consequence of the reduced thermal effects of ps-excitation, which favoured a closer plasma–particle interaction. Chen et al.¹⁰¹ investigated the use of single, micro-size, suspended particles, and AgNPs to enhance emission signals. They also used machine learning to improve calibration models for the quantification of Cu in carbon black particles.

Effects of laser polarisation on both atomic and ionic emission lines in LIBS spectra of metals (Cu, Ag) and non-metals (polyethylene) were thoroughly investigated by Adarsh et al.¹⁰² Changes in emission line intensities were observed after inducing changes in the polarisation direction of a linearly polarised laser beam or using a circularly polarised laser beam. Enhanced emission from ionic lines was observed when the polarisation angle was 45°. Further theoretical interpretations are required to explain this polarisation-dependent effect.

The influence of ambient humidity on LIBS spectra of pure copper was investigated by Liu et al.¹⁰³ At high ambient humidity two opposing effects occurred: sample ablation increased but the lifetime of the laser-induced plasma decreased, resulting in enhanced or depleted Cu I emission signal as a function of the laser energy. Higher emission signals were obtained at pulse laser energies above 15 mJ. The authors claimed that this effect should be considered for long-term or open-air LIBS measurements.

4.1.2. Instrumentation. A novel calibration approach based on online nebulisation of different concentrations of an aqueous solution of NaF onto a CaCO₃ pellet was developed and optimised by Méndez-López et al.¹⁰⁴ to obtain F concentration calibration curves from LIBS molecular emission of CaF. The method provided LODs for F of 10 mg kg⁻¹ with a linear response range up to 900 mg kg⁻¹. The same research group further improved their method.¹⁰⁵ by evaluating the use of paper as a Ca-containing solid target instead of a CaCO₃ pellet. Different sources of F were considered including inorganic fluoride and fluorocarbon-containing samples. Results demonstrated the capabilities for the determination of F and Cl in liquid samples, with LODs of 5 and 192 ppm, respectively.

Microwave-enhanced LIBS was evaluated for the analysis of nuclear fuel debris by Ikeda et al.¹⁰⁶ Attention was paid to Zr emission intensities, which were observed to increase, possibly due to re-excitation and re-ionisation through increased collisions. Zr ion emission signals were sustained for more than 1.0 ms (microwave application period). An increase in intensity ratios of ion and atomic emission lines was measured, e.g., Zr II (449.69 nm) : Zr I (450.71 nm) ratio.

Khan et al., ¹⁰⁷ developed and evaluated a compact, transverse magnetic field confined LIBS system for the analysis of Al alloys. The pulse energy of a Nd:YAG laser operated at its fundamental wavelength (up to 160 mJ) and magnetic field strengths (up to 3.5 kG) were optimised to achieve up to a 7× enhanced emission signal, improved LODs and improved repeatability.

A modulated discharge-assisted LIBS system was developed by Lei et al.¹⁰⁸ in order to improve LODs while reducing the energy costs and environmental hazards. The modulation approach, where a short-time spark discharge and a long-time arc discharge were regulated in a proper sequence, trapped and confined numerous charged particles in an effective discharge space. This maintained their electric-drift dynamics in the ms timescale, thereby improving coupling between the electric energy and the laser plasma and prolonging plasma lifetime.

A novel, highly sensitive detection method for dissolved trace heavy metals in water was developed by Shang et al.¹⁰⁹ The method was based on the transformation of water samples into micrometer droplets, using a micro-hole array injection device. Droplets were then sprayed onto a rotating polypropylene organic film and let dry before LIBS analysis. The authors claimed that the technique was a feasible method to improve the sensitivity and speed of the detection of trace heavy elements in liquid samples and could facilitate wider application of LIBS in water quality monitoring. Improved LIBS sensitivity was also achieved by Robledo-Martinez et al.¹¹⁰ by pre-heating the sample surface up to 300 °C through the application of a focused, CW diode laser. The authors concluded that the intensification in the pre-heated sample surface was not related to increased electronic plasma density or temperature but to an increased amount of ablated material.

A system based on the combination of LIBS and an ion enrichment chip was developed by Xu et al.¹¹¹ to rapidly and sensitively detect the total Cr and Cr^VI in water and soil. Total Cr was enriched using an activated carbon fiber, and Cr^VI was enriched using an anion exchange resin. The authors claimed that this method was simple, sensitive, and environmentally friendly, and that detection sensitivity met the Cr environmental quality standards in China.

A confocal controlled Raman-LIBS hybrid microscope with high spatial resolution and anti-drift properties was developed by Zhao et al.¹¹² to improve the analysis of unknown minerals. The approach combined Rayleigh scattering, reflected light, LIBS signal, and Raman spectra to simultaneously determine geometrical topography and elemental and molecular structure. The method was validated through 3D high-resolution topological and hybrid spectral maps of the Northwest Africa 13 323 meteorite.

Jia et al.¹¹³ studied the early stage (<2 μs) temporal evolution of an underwater LIP. Cavitation bubble dynamics were investigated by using fast imaging and shadowgraph techniques. The bubbles were observed to experience a transition phase, between the moving breakdown phase and the thermal expansion phase, after 20–50 ns, at which point both the expansion speed and pressure of the bubbles dropped drastically, with concomitant severe plasma fluctuations in this period. These results highlight the critical role of bubble-plasma interaction on the characterisation of a LIP in water.

Wang et al.¹¹⁴ developed various external normalisation strategies utilising PLS of plasma acoustic signals, plasma images, and acoustic-image combinations to improve underwater LIBS analysis. The PLS generated calibrations of Mn, Sr and Li showed that the acoustic-image combination worked most efficiently.

An innovative LIBS methodology based on the collection of extreme UV (XUV) spectra was developed by Bleiner et al.¹¹⁵ Plasma emission was detected at very short delays (ps or ns), when the plasma was still very hot (several eV's) and dense (10¹⁸–10²⁰ cm⁻³). This resulted in increased sensitivity, lower noise and higher precision. Under these conditions, emission in the XUV spectral region was considered to be mainly due to highly ionised elements, in isoelectronic closed-shell configuration.

Ji et al.¹¹⁶ found that shaping the laser beam profile from Gaussian distribution to flat-top resulted in a higher plasma temperature and similar n_e. This resulted in an improvement in the method to determine U in ores, due to less line broadening and less line overlap or interferences. In particular, a 5× higher U II (409.013 nm) emission signal was obtained using the flat-top beam shape. An LOD of about 21 μg g⁻¹ for U in ores was obtained.

4.1.3. Novel LIBS approaches. Emission signals from diatomic molecules, which are frequently employed for various purposes in LIBS, are characterised by broad spectra that are prone to spectral interference. Fernández-Menéndez et al.¹¹⁷ developed a method to critically assess and eliminate CaO and Na interferences from CaCl emission (at around 593 nm), allowing a proper quantification of the Cl in cements. The study also addressed challenges related to the absence of a Cl-blank cement sample and the variability of Na concentration in the analysed samples. Additionally, the generation of a plasma in atmospheres enriched with Ar and He was explored, with the conclusion that the use of Ar is recommended to enhance Cl determination due to its higher sensitivity while maintaining a lower contribution of Na interferences. The efficacy of the developed protocol was demonstrated by successfully determining Cl content in real cement samples, with concentrations ranging between 0.23 and 1.5% wt. of Cl. Likewise, diagnosis and spectral correction methods for LIBS imaging were developed by Duponchel et al.¹¹⁸ The authors were able to diagnose the presence of a potential spectral interference and to correct it using PCA and multivariate curve resolution (MCR)-ALS methods, respectively. This method was applied to LIBS imaging dataset obtained from the analysis of a complex rock. An automatic background correction method for LIBS, based on the ideas of window functions and differentiation combined with PCHIP, was investigated by Chen et al.¹¹⁹ and evaluated it for the determination of Mg in aluminum alloys. The authors claimed that the method outperformed ALS and model-free background correction methods, in terms of improved linear correlations and smaller errors in the obtained calibration curves.

Bai et al.¹²⁰ developed a linear regression model denominated “elastic net”, which combines stable ridge regression and least absolute shrinkage and selection operator (LASSO) models. This was successfully applied in quantitative analysis of major elements in geological samples, using LIBS datasets from Martian samples. The authors claimed that elastic net is a candidate model for the quantitative analysis of the Zhurong MarSCoDe LIBS spectra. Machine learning-based LIBS spectra classification is an emerging method, whose major drawbacks are to be considered a complicated black box model with difficult to interpret spectroscopic terms and to be prone to overfitting. In this context, a hierarchical modeling approach for selecting spectral features, specifically wavelengths, was developed by Huffman and Sobral¹²¹ The whole distribution of intensities of a single spectral point was modelled, accounting for the structure nature of LIBS data (e.g., shots within spots within samples within classes). The authors considered that this hierarchical modelling was closer to a spectroscopically explainable classification mode(l) (i.e., grey box instead of a black box model). A classification dataset for LIBS, that was prepared for the 2019 EMSLIBS conference contest, was employed to successfully validate the performance of this approach.

The determination of the LODs is not considered an easy task for LIBS analytical applications. Progress towards the new official definition of LOD, recently adopted by IUPAC, and methods to extend its calculation in LIBS multivariate analysis were discussed by Poggialini et al.¹²² A statistical definition of LOD was proposed for CF-LIBS by Villas-Boas and Borduchi¹²³ based on the density of emitters, plasma parameters and measurement uncertainties. The method was validated using C and Ca in sodium chloride samples.

A novel method to improve long-term repeatability of LIBS measurements was developed by Liu et al.¹²⁴ This was based on the modification of the spectral intensity in relation to the laser beam intensity distribution, in particular the relationship between relative deviations of beam and spectral intensities modelled using PLS. The model was successfully validated using Cu and Si samples that were measured for more than 30 days, significantly reducing long-term RSD. An alternative method, based on the use of Kalman filtering, was proposed by Lu et al.¹²⁵ to improve long-term reproducibility of LIBS quantitative analysis. The method was validated by the quantitative determination of elemental concentrations in low-alloy steel samples over 10 days. Moreover, a spectral standardisation method based on plasma image-spectrum fusion (SS-PISF) was developed by Nie et al.¹²⁶ to correct fluctuations in total number density, plasma temperature and n_e (based on the effective information in plasma images and spectra), resulting in improved spectral stability.

4.2. Laser induced fluorescence

Ablation process, excitation process, and coupling behavior of multi-element detectable laser excited atomic fluorescence (LEAF) were investigated by Kou et al.¹²⁷ Using a brass sample, LEAF was conducted by employing a 193 nm laser to intercept the ablated plasma generated with a Nd:YAG laser at 1064 nm. In particular, multi-element signals were significantly enhanced using low LA energy (1.49 mJ), which resulted in minimal sample destruction, and using an inter-pulsed delay of 200 ns. From the spectra measured at different delay times and from time-resolved plasma images, it was concluded that photoexcitation contributed to the enhanced signal other than the electron thermal excitation, as signal gained its higher intensity when the re-excited plasma was less bright.

4.3. Laser atomic absorption spectroscopy

The effects of long and short pulses on underwater Fraunhofer-type absorption (absorption peaks in the continuum radiation spectrum) of a submerged Cu target were investigated by Li et al.¹²⁸ using single pulsed LIBS. This absorption was considered to take place at the edge of the plasma where particles in ground or low excitation states absorb the continuum radiation emitted from the plasma core. Clear absorption lines were only found using long pulses (160 ns). Moreover, they were generated at the beginning of the plasma and gradually weakened with delay time, with a duration of about 300 ns. The authors also claimed to observe distinct Fraunhofer-type absorption lines without time-gated detection, providing another approach for the study of underwater single-pulse LIBS.

4.4. Laser ablation ionisation mass spectrometry

The ablation of non-conductive samples might produce electric charging of the surface that can influence the spread of the ablation plume, reducing the spectral quality. Gruchola et al.¹²⁹ employed a method, based on the sputtering of a thin Au layer on the sample, to prevent charge build-up. This improved the analytical performance of fs-LIMS for the analysis of non-conductive samples. Furthermore, a novel LIMS method for high spatially resolved mapping of F was developed by Ma et al.,¹³⁰ who employed a cryogenic laser ionisation source to eliminate the interference of water-derived species at m/z 19. The Cryo-LI-TOFMS was validated mapping the distribution of F concentration in human teeth, with a spatial resolution of 20 μm.

5. Isotope analysis

5.1. New developments

Double spike analysis to determine the stable isotope ratios of a sample expressed relative to a standard requires a single analysis of a sample-spike mixture. This mixture must be optimised such that errors generated by extreme sample-spike ratios are minimised, which requires sufficient analyte to achieve minor isotope signals sufficiently above the background to give good counting statistics. Lu et al.¹³¹ developed a new technique designed to improve the analysis of stable isotope ratios in samples where concentrations of the analyte element are low. The study introduced a double spike standard addition technique which for sample analysis, was combined with a double spiked secondary standard solution. This spike-standard mixture had a precisely known spike/standard concentration ratio which was at the high-end of the optimal range. When combined with the sample solution, the complete mixture was proportioned to remain within the optimal uncertainty range. Deconvolution of the deviation from the sample isotope ratio of the primary standard solution was completed using the isotope binary mixing equation and standard double spike deconvolution. The ability to provide high precision analysis on around 20% of the normal sample size means this methodology is likely to become a go-to technique for ng-level isotope analyses of elements such as Cd and Mo.

A comprehensive assessment of Sr isotope measurement by LA-MC-ICP-MS was made by Mulder et al.¹³² This focussed on producing a robust analytical procedure and creating a suite of plagioclase RMs suitable for distribution. The study detailed the potential isobaric interferences from Rb, Kr, Ca and doubly charged rare earth elements. The doubly charged elements affected both Sr isotopes, e.g.¹⁷⁴Yb²⁺ on ⁸⁷Sr, and on half masses, e.g.¹⁷³Yb²⁺ on m/z 86.5. Mass bias procedures were found to be most effective when the ⁸⁷Sr/⁸⁶Sr ratio was corrected by external standard isotope measurements rather than using the stable Sr isotope ratios with their added interference complexities. A recommended data reduction scheme suitable for use with Iolite software was presented in this study. However, a key component is the characterisation of three new plagioclase RMs, with Sr content ranging from ∼200 to3000 μg g⁻¹ and ⁸⁷Sr/⁸⁶Sr ratio of between 0.7031 and 0.7076. The RMs are available for distribution from the Swedish Natural History Museum.

Some initial results from a new MC-ICP-MS-MS instrument were presented by Télouk et al.¹³³ The instrument used in their study had a prefiltering arrangement of a double-Wien filter within an ion lens stack, succeeded by a collision/reaction cell. This system was tested for removal of isobaric ⁴⁰Ar²³Na⁺ on ⁶³Cu⁺, an interference which hinders the precise measurement of stable Cu isotope composition. Na was found to be reduced by the electric prefiltering whereas the ⁴⁰Ar²³Na⁺ was removed by the reaction cell. Results indicated that precise ⁶⁵Cu/⁶³Cu ratios could be achieved with Na:Cu of 10, which was ten times higher than the level at which standard MC-ICP-MS analyses degrade.

The ⁴¹Ca isotope system could potentially be used to date the burial age of biological or geological samples. It is produced by capturing cosmic-ray-induced thermal neutrons a few metres below the Earth's surface. After several half-lives (∼99 ka) of exposure the ⁴¹Ca/Ca ratio approaches a saturation value. Surface samples – biological or geological – when buried deeper than this level will be isolated from further ⁴¹Ca growth and hence it will decay: which means that low levels of ⁴¹Ca reflect longer burial. Xia et al.¹³⁴ used an atom-trap trace analysis method which captures individual ⁴¹Ca atoms and counts them by detecting their fluorescence. This method was found to push the detection limit of ⁴¹Ca/Ca down to 10⁻¹⁷, and produced a precision of 12% at 10⁻¹⁶. Hence, this system may provide a dating mechanism for samples too old for ¹⁴C but too young for ¹⁰Be.

5.2. Radiogenic isotope ratio analysis

Grotti et al.¹³⁵ re-evaluated the analysis of Pb isotopes using non-multicollector ICP-MS. Their procedure used an external correction for mass fractionation and routine corrections for ²⁰⁴Hg, blank and dead time. The aim of their work was to determine if useable ^20xPb/²⁰⁴Pb ratios could be generated for environmental studies. The results indicated that precision on ²⁰⁶Pb/²⁰⁴Pb was ∼±0.4% (2S) with solutions of 10 ng mL⁻¹ (of Pb). In comparison, the highest precision analysis using double spike MC-ICP-MS results in ²⁰⁶Pb/²⁰⁴Pb ± 0.02% (2S). Hence, these results indicate that a balance has to be considered between the significantly reduced precision of the ICP-MS data, the requirements for discrimination between sample sets and the cost/rapidity/availability of the analytical technique.

A new emitter for the analysis of Pb isotopes by TIMS was evaluated by Li et al.¹³⁶ The emitter was a colloidal solution of β-Si₃N₄ (10 μg in 10 μL) added to a Ta filament pre-treated with phosphoric acid. The system was found to provide similar or slightly lower beam intensities to the familiar Si-gel activators, yet was able to do so with Ta filaments rather than the significantly more expensive Re substrate.

5.3. Geological studies

Redaa et al.¹³⁷ investigated Rb–Sr dating using geological RMs and their application using LA-ICP-MS. Their study assessed nano-powder pellets and laser-fused glass disks of mica, glauconite and feldspar standards. Results indicated that, of the materials studied, Mica-Mg and glauconite-O, as nanopowders, were suitable and robust for use in dating applications.

Chu et al.¹³⁸ examined the use of uranium double spikes in dating of single zircons by ID-TIMS. Typically, a ²⁰⁵Pb–²³³U–²³⁵U spike is used to calculate instrumental mass fractionation and concentrations of U and Pb. However, more precise U–Pb ages can be achieved using ²³⁶U instead of ²³⁵U. A problem with the ²³⁶U spike is that it generates polyatomic uranium oxide ions such as ²³⁶U¹⁶O¹⁸O⁺ on ²³⁸U¹⁶O₂⁺. The study found that the effects of ²³⁶U oxide interferences were insignificant on the final dating result if ²³⁸U/²³⁶U was >0.5. overall, and the proposed method was estimated to generate analytical precision better than 0.05% (2SE).

A potential alternative to zircon for U–Pb dating is the mineral apatite (Ca₅PO₄)₃(F,Cl,OH). Zircon is ideal because it includes little Pb during crystallisation yet has significant U in its structure. This allows the vast majority of Pb measured in a zircon to be generated by radiogenic growth and makes dating calculations simpler and more precise. However, zircon is not present in all rock types and can be sparsely present in others. Apatite has a similar preference for U over Pb in its structure, but has slightly more of the inherent “common” Pb: generally ²⁰⁶Pb/²⁰⁴Pb is ∼500 in an apatite versus >5000 in a zircon. Duan et al.¹³⁹ sourced and analysed an apatite (MAP-3) as a gemstone originally from Myanmar and aimed to use this as a standard in U–Pb analysis of apatites. It was found to be isotopically and elementally homogenous on a scale suitable for LA analysis and had suitably low levels of common Pb with ²⁰⁶Pb/²⁰⁴Pb ∼850. The age of the apatite was reproducibly found to be 800.5 ± 0.9 Ma, making this suitable to be distributed and used as an apatite dating standard material. Another mineral system that can be used for U–Pb dating is columbite-tantalite. Xiang et al.¹⁴⁰ analysed three potential RMs for in situ U–Pb dating. These columbite-tantalite samples yielded internally consistent ages that were comparable to TIMS ages. Only one of the three samples (SN3) was found to sufficiently homogenous to be used as a primary RM for use in LA-ICP-MS dating.

Liu et al.¹⁴¹ developed a freely available program (Isoclock) to perform common Pb corrections in U–Th–Pb systems. LA-ICP-MS analysis of various accessory minerals such as apatite, titanite, wolframite and cassiterite can be processed by this procedure. A facet of the program was the calculation of common Pb from RMs in terms of both U and Th, followed by the estimation of mass fractionation.

A new U–Pb/Pb–Pb data reduction software for LA-ICP-MS was also developed by Silva et al.¹⁴² This program “U–Pb Saturn” was designed to process hundreds of spot analyses, from a variety of mass spectrometers, for interferences, drift and mass bias. Data is plotted and visualised on concordia diagrams. As this software is non-commercial, easy to run and very interactive it is an attractive option for analysts starting off in laser U–Pb geochronology.

Analysis of U–Pb ages and rare earth abundances at depth within zircon crystals was considered by Nakazato et al.¹⁴³ Their study utilised an ICP-TOFMS with a dry plasma cone system. To achieve a “clean” analysis at a particular depth within the zircon, a “moat” was ablated around the analytical location to remove the edge effects from the overlying layer prior to ablation of the target. Analysis of a Himalayan zircon demonstrated the effectiveness of the depth profiling, with identification of three distinct growth events. The outermost layer was characterised by a younger age, potentially reflecting a later overgrowth event.

Niki et al.¹⁴⁴ examined the ²³⁸U–²³⁰Th dating method using an ICP-CCMS. The collision cell was used in kinetic energy discrimination mode combined with a medium mass resolution to reduce interfering ions generated by zircon matrices at m/z 230. Minimising these interferences while maximising Th signal was found to be key in providing accurate dates for Quaternary zircons. Rare earth mineralisation is often dated using the mineral bastnasite (Ce(CO₃)F). However, Tang et al.¹⁴⁵ noted that the only reliable primary standard for in situ U–Th–Pb dating, bastnasite K-9, is inhomogeneous and in short supply. Hence, they proposed a new standard xenotime XN01 (Y(PO₄)), which has similar U–Th–Pb isotope ratios at given ablation settings. In particular, the study recognised that N in the ablation recovery system combined with low laser sampling rates helped to minimise any differential matrix effects between xenotime and bastnasite.

Kerber et al.¹⁴⁶ examined a method to simultaneously measure U and Th via MC-ICP-MS for U-series dating of geological materials. Combining measurement of the two systems meant no chemical separation was required. However, background measurements had to be carefully monitored to constrain the tailing from ²³⁸U across both the electron multiplier and Faraday detectors. In addition, the effects of “ghost signals” from ²³⁸U on masses 229 and 230 m/z needed correction on samples with very low ²³⁰Th abundance.

Wang et al.¹⁴⁷ presented a study of the measurement of U stable isotopes of 29 geological RMs available from China. This involved a ²³³U–²³⁶U double spike and measurement by MC-ICP-MS. Results indicated that igneous and metamorphic rocks amongst the standards were found to have δ²³⁸U/²³⁵U within ± 0.2 of continental crustal values (−δ = 0.29). However, some sedimentary rocks and several sediments were found to have significant positive and negative ratios respectively.

Harrison et al.¹⁴⁸ examined the Ca isotope fractionation between calcite-fluid and aragonite-fluid. The experiments used fluids enriched in ⁴³Ca, which were equilibrated with synthetic calcite and aragonite for around 5 months. Results indicated that calcite did not increase in size during the experimental period, whereas the aragonite developed crystalline needles. Isotopically, the aragonite was found to have a significantly more rapid exchange with the fluid compared to calcite. However, both CaCO₃ polymorphs could have modified Ca isotope compositions without any overt signs of diagenesis.

Wang et al.¹⁴⁹ examined the possibility of simultaneously measuring three Os isotope ratios (¹⁸⁴Os/¹⁸⁸Os, ¹⁸⁶Os/¹⁸⁸Os, ¹⁸⁷Os/¹⁸⁸Os) in low concentration samples. Using negativeTIMS equipped with 10¹² Ω, 10¹³ Ω Faraday detectors and two ion counting detectors, this study carefully documented the signal to noise ratio of each detector and the potential isobaric interferences across the OsO₃ masses. Oxygen isotopes were found to vary between different filaments, but were constant during each measurement, permitting correction on a sample-by-sample basis. High precision analysis was achieved with sample sizes of down to 76 pg g⁻¹, which is around 5 times lower than previous methods.

5.4. Stable isotope ratio studies

An et al.¹⁵⁰ assessed the role of reaction gases used in K isotope measurement by MC-ICP-CCMS. An issue with the use of H₂ in the collision cell is the formation of ⁴⁰CaH⁺ which is an isobaric interference with ⁴¹K⁺. To circumvent this, their study used D₂ rather than H_2, which allowed accurate ∂⁴¹K to be achieved in solutions with Ca:K concentration ratios up to ∼40 compared to previous studies where analyses were significantly degraded with Ca:K >4. He et al.¹⁵¹ also completed a K isotope study using MC-ICP-MS but were focussed on techniques suitable for plant materials. Their method included purification utilising one or two passes though a AG50Wx8 cationic resin, eluting K with 0.5 M HNO₃. Mass spectrometric methodology involved a reduced forward power cool plasma, low mass resolution and no collision cell activation to achieve precision of δ⁴¹K better than ±0.05‰ (2SD) for both plant materials and rock RM.

Multicollector-ICP-CCMS was deployed by Lewis et al.¹⁵² to investigate Ca isotope measurement. The prototype instrument “Proteus” had the collision cell charged with H₂, to react and effectively remove the ⁴⁰Ar⁺, supplemented with He to thermalise and increase transmission of the ion beam to enable accurate determination of ⁴⁰Ca. Radiogenic ⁴⁰Ca/⁴⁴Ca was normalised to ⁴²Ca/⁴⁴Ca and was reproducible to <40 ppm, with geological RMs (e.g. BIR-1 and BHVO-2) found to have slightly unradiogenic compositions relative to the NIST SRM 915a. Mass dependent δ⁴⁴Ca/⁴⁰Ca isotope variations were determined with a double spike that produced long term precision of ±0.035‰ (2S).

Collision cell MC-ICP-MS was also used to examine the need for concentration matching between sample and standards in K isotope measurement by Li et al.¹⁵³ Their study examined the hexapole control parameters. The voltage of the RF alternating current applied to hexapole rods was found to be the key function that effects the δ⁴¹K variability with sample/standard ratios. Careful optimisation of the RF current was found to significantly increase the tolerance of the measurement to sample-standard mismatches.

Bai et al.¹⁵⁴ utilised Eichrom TODGA resin to isolate Sm from Nd and other REE using a procedure combining AG50W-X12 resin with TODGA. This allowed for a high-precision measurement of δ¹⁵²Sm/¹⁴⁹Sm with a long-term intermediate precision of ±0.04‰ (2SD) using MC-ICP-MS, which was ∼4 times better than other MC-ICP-MS and double spike TIMS methods. Such improvement in Sm isotope analysis is likely to generate advances in our understanding of the processes of chemical weathering and ocean circulation.

Liu et al.¹⁵⁵ introduced a method for the simultaneous determination of stable and radiogenic Nd isotopes using double spike TIMS following a single chromatographic isolation procedure. Column separation using DGA Eichrom resin was found to preferentially retain the heavy isotopes, but this fractionation was found to follow a mass-dependent systematics. Radiogenic ¹⁴³Nd/¹⁴⁴Nd was deconvoluted by a double spike process to give ±5 ppm (2SE) precision and accurate values in the JNdi-1 RM, while δ¹⁴⁶Nd/¹⁴⁴Nd was determined to ±0.016‰ (2SD).

Ti isotopes in rutiles were the subject of an LA-MC-ICP-MS study by Liu et al.¹⁵⁶ This experiment combined a high sensitivity cone arrangement combined with 1 Hz ablation and a signal smoothing device. Adjustment of the laser spot size and fluence along with signal mismatch between sample and standard were found to generate higher δ⁴⁹Ti values under dry plasma conditions. However, a wet plasma was found to suppress this effect. A key finding of the study is that some rutile crystals were found to have significant internal δ⁴⁹Ti variations which developed through different conditions prevailing during their growth.

Cai et al.¹⁵⁷ investigated the significant issue of memory effects between samples in measuring B isotopes using ICP-MS. The effect was identified as stemming from the ICP torch, in particular the tip which was found to incorporate B from sample solutions into its silica structure at high temperatures. Simple background subtraction produced systematic differences in measured values related to the isotope ratios of the previous solution. Using a subtraction that involved B isotope ratios and improving the rigour of the H₂O–HNO₃ – wash cycles were found to produce accurate B isotope values even with high SBRs. Fietzke and Anagnostou¹⁵⁸ assessed sources of inaccuracy in B isotope measurement. Their study focussed on LA-MC-ICP-MS. A key finding was that a source of baseline inconsistencies was the reflection of Ca and Ar ions within the flight tube of the mass spectrometer. Furthermore, these issues were found to be present in instruments from three different manufacturers. A potential solution to these baseline inconsistencies could be the use of deflectors to guide ion beams, which are typical fitments to instruments with ion counting detectors. Paul et al.¹⁵⁹ examined the complete procedure for B isotope measurement in silicate samples. A problem with such samples is the need to involve HF in silicate dissolution. Boron is known to volatilise during HF evaporation, effectively loosing mass and potentially fractionating the B isotope ratio. Consequently, the proposed dissolution was designed to dry down samples at low temperature (<60 °C). Evaporation was further controlled by an upturned beaker placed on top, but slightly offset from, the heated sample vessel. A chromatographic separation involving a three-stage separation with cation–anion-cation columns was used for purification. Analysis using MC-ICP-MS demonstrated a ±0.6‰ (2SD) measurement precision for B concentrations of <2 μg g⁻¹.

Measurement of Li isotope compositions at sub-nanogram levels was the subject of a study by He et al.¹⁶⁰ They used MC-ICP-MS in combination with a membrane desolvation sample introduction system. They achieved precision better than 0.5‰ (2SD) for δ⁷Li using ∼2 ng Li for analyses of geological RMs. This makes their technique highly applicable to samples with low Li concentrations such as seawater, rainwater, limestones and meteorites. Li was also the subject of a study by Liu et al.¹⁶¹ who used MC-ICP-CCMS but opted to bypass the collision cell for these analyses. Results of repeat measurements of rock and solution RMs resulted in Li isotope ratios with a precision better than ± 0.23‰ (2SD).

The double spike standard addition technique introduced by Lu et al.¹³¹ (see New Developments section above) was applied to Cd isotopes by Chang et al.¹⁶² Their study demonstrated that δ¹¹⁴Cd/¹¹⁰Cd could be measured with a precision of ±0.022‰ using a mass as low as 2 ng of Cd. Fang et al.¹⁶³ developed a standard addition technique to measure Sb isotopes. Their method involved a proof of concept by mixing two solution standards and measuring isotope ratios using MC-ICP-MS. Mass fractionation was assessed using a combination of Sn addition and correction using ¹¹⁹Sn/¹¹⁷Sn with Sb sample-standard bracketing. True Sb isotope ratios were calculated using standard isotope mixing laws. As with the other standard addition techniques, the overarching benefit is the ability to measure low concentration samples: in the case of Sb below 5 ng. Overall precision was determined to be better than ± 0.2 for δ¹²³Sb.

Te isotope measurements were examined by Morton et al.¹⁶⁴ who used a double spike methodology and MC-ICP-MS. Their aim was to establish a technique suitable for meteorites and terrestrial materials. Te was separated from Fe, Mo and other transition metals using a four-stage column process involving an initial AG1-X-8 anion exchange followed by a TRUSpec separation, then two more stages of anion exchange to remove all matrix elements. A¹²⁵Te–¹²⁸Te double spike was prepared and error-optimised and added to samples and the London Te reference standard. Typical reproducibility of δ¹³⁰Te/¹²⁵Te was found to be ± 0.07‰ (2SD) with samples containing 20 ng mL⁻¹. If δ¹³⁰Te is measured at this level of analytical precision, it was concluded that terrestrial samples were unlikely to have a significant impact from mass-independent Te isotope effects in meteorites and terrestrial samples.

Sn isotopes have not been extensively investigated but She et al.¹⁶⁵ analysed cassiterite (SnO₂) using LA-MC-ICP-MS. This involved a natural homogenous cassiterite crystal which was used as a matrix-matched bracketing standard. Sb-doping was also used to constrain the mass bias correction with the laser-generated aerosol being combined with an Sb aerosol in a wet plasma spray chamber. Repeated analysis estimated that the intermediate precision for δ¹²²Sn/¹¹⁸Sn was ± 0.12‰. Zhang et al.¹⁶⁶ took the approach of synthesising a homogenous cassiterite RM. Their study involved sintering a micrometre powder to produce a material suitable for LA analysis. Measurement of the resulting material determined an increase in the isotopic homogeneity from the original cassiterite, which suggests that, where appropriate, the sintering technique may be used for other minerals for microanalysis. Cassiterite was also the focus of a study by Yang et al.¹⁶⁷ who designed a protocol for in situ measurement of Hf isotopes by LA-MC-ICP-MS. This study also recommended the use of a homogenous primary RM; in this case the cassiterite Rond-A.

Hg isotopes can be used as tracers of biochemical processes. Shi et al.¹⁶⁸ used CVG for Hg sample introduction. Their results showed that this was strongly affected by the design of the vapour separator, the sample matrix and sample uptake rate. Importantly, the Hg isotope fractionation induced by restricted CVG could not be fully corrected by matrix matching and sample-standard bracketing. Ensuring 100% Hg generation efficiency by matrix separation was therefore recommended as a pathway to high precision and accuracy of Hg isotope data.

Wang et al.¹⁶⁹ compared MC-ICP-MS with double spike TIMS for the measurement of Sr isotopes. In particular, for the simultaneous analysis of both the radiogenic ⁸⁷Sr/⁸⁶Sr and the stable δ^88/86Sr ratios. Plasma source analyses of stable Sr isotopes generally involve doping the samples with Zr to correct for mass bias. Their study examined the importance of matching the Sr and Zr concentrations and the acid molarity of samples and standards through the run. In addition, the effects of contaminant cations such as Ba, Ca and K were found to cause large Sr isotope ratio variations that could not be corrected by Zr doping. The results indicated that a complete chromatographic recovery of Sr and removal of interferants was able to produce ⁸⁷Sr/⁸⁶Sr ± 0.000015 and δ^88/86Sr ± 0.03‰ (2SD).

Sb can pose serious health risks if present in high concentrations in the environment. Processes such as reduction, adsorption, biochemical, evaporation and precipitation have been found to cause significant Sb isotopic fractionation. Hence, the Sb system could be used to trace sources of pollution, in particular in the aquatic environment. Xia et al.¹⁷⁰ addressed the preparation of freshwater for Sb isotopic analysis, which has been difficult due to the low concentrations involved. Their study designed a novel preconcentration and purification system in which water samples were processed using KI-ascorbic acid Sb reduction prior to a vacuum-driven thiol resin ion exchange system. Further oxidation and digestion of samples used BrCl. External precision of ε¹²³Sb was estimated to be ± 0.4 (2SD) using MC-ICP-MS analysis.

Zhu et al.¹⁷¹ detailed a method for determining mass-independent Cr isotope variation. They used TIMS in total evaporation analysis mode following sample purification with a three stage column chromatography and subsequent heating with H₂O₂ to remove residual organics. Between 10 and 15 ng Cr resulted in ε⁵³Cr of ± 0.05 (2S). It was noted that samples measured by this total evaporation technique were found to have systematically lower ε⁵³Cr and ε⁵⁴Cr than were measured by the normal (non-total evaporation) TIMS method. It was suggested that this difference was caused by fractionation differences between the Cr metal and oxide species present during ionisation. Potentially, total evaporation may minimise this effect by ionising all species during the run.

The Cu isotope reference material SRM976 is no longer available so Yang et al.¹⁷² developed a new RM: HICU-1. This was prepared using a gravimetric isotope mixture approach using two high-purity enriched Cu isotopes. The ⁶³Cu/⁶⁵Cu ratio determined for this mixture was 2.2423 ± 0.0009 (2SD) which was nearly identical to that of SRM976.

Zhao et al.¹⁷³ investigated the mass independent fractionation of Mo isotopes during MC-ICP-MS measurement. They found that mass independent fractionation occurs during Mo isotope measurement. This has implications for the double spike technique, which has a basic assumption of mass dependent isotope behaviour. The recommendation from the study was that a secondary sample standard bracketing method is used alongside the double spike methodology for MC-ICP-MS analyses.

5.5. Nuclear forensics

Jaegler and Gourgiotis¹⁷⁴ utilised an ICP-MS-MS technique with the aim of improving the detection limit of ²³⁶U/²³⁸U. Advances in the ion transmission and the range of measurable signals, coupled with desolvation-nebulisation and a reaction cell primed with oxygen provided the potential for the determination of low levels of ²³⁶U. Analysis of RMs and environmental samples achieved a precise determination of a²³⁶U/²³⁸U ratio of 10⁻¹¹, which is a 10-fold improvement on previous ICP-MS-based analyses.

Wimpenny¹⁷⁵ devised a rapid measurement procedure for isotopic analysis of uranium particles using LA-MC-ICP-MS, which was trialled on two solid-sample RMs. Results from laser raster analysis of the particles was found to produce accurate ²³⁴U/²³⁸U, ²³⁵U/²³⁸U, ²³⁶U/²³⁸U ratios, suggesting that this may be a suitable technique where rapid analysis is required.

Nd is often used as a “burn-up” monitor of fast reactor fuels. Burn-up is defined as the number of fissions per 100 initial heavy element atoms (mass >225 amu) and is quantified by the proportion of heavy elements relative to the monitor, which can be ¹⁴⁸Nd for thermal reactor fuels. A problem with ¹⁴⁸Nd is an isobaric interference with ¹⁴⁸Sm, which generally requires chemical separation prior to mass spectrometric measurement. A study by Janardhanan et al.¹⁷⁶ examined the potential for direct measurement by LIMS. This involved tuning laser parameters to generate NdO⁺ while leaving Sm as an elemental ion which eliminated the isobaric interference. Overall, the LIMS method was found to significantly reduce radiation exposure, active waste and costs of burn-up determination.

The Cs isotope ratio is important in analysis of nuclear discharges into the natural environment from weapons tests or accidents. The ¹³⁵Cs/¹³⁷Cs ratio varies according to the flux and exposure time in the reactors with nuclear power station discharge having ratios of <0.52 and weapons tests >2.6. Magre et al.¹⁷⁷ conducted an ICP-CC-MS study of Cs isotopes using N₂O, He and NH₃ in the collision cell to minimise interference on ¹³⁵Cs and ¹³⁷Cs. Gas conditions were adjusted to negate interference while maximising Cs signal intensity. Results were calibrated using solutions previously analysed by TIMS and produced reference ratios for sediment samples from the Fukushima fallout.

Pu isotope measurement by TIMS was advanced by Zhang et al.¹⁷⁸ They used graphene oxide, which was bonded to the surface of the Rh filament, as an ionisation enhancement additive. It was found that the graphene oxide produced an overall ionisation efficiency of 0.44% for Pu, which is around 5 times more effective than standard graphite coating methods, making this technique suitable for a number of applications in nuclear forensics and environmental monitoring.

6. 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.

AAS	atomic absorption spectrometry
AES	atomic emission spectrometry
AFS	atomic fluorescence spectrometry
BEC	background equivalent concentration
CC	collision cell
CCD	charge coupled detector
CMOS	complementary metal oxide semiconductor
CRM	certified reference material
CVG	chemical vapour generation
DBD	dielectric barrier discharge
DC	direct current
EDL	electrodeless discharge lamp
ETAAS	electrothermal atomic absorption spectrometry
ETV	electrothermal vaporisation
FAAS	flame atomic absorption spectrometry
FT	Fourier transform
GC	gas chromatography
GD	glow discharge
GF	graphite furnace
HCL	hollow cathode lamp
HG	hydride generation
HV	high voltage
IAEA	International Atomic Energy Agency
ICP	inductively coupled plasma
ICR	ion cyclotron resonance
ID	isotope dilution
IDA	isotope dilution analysis
IE	ionisation energy
IR	infra-red
IRMS	isotope ratio mass spectrometry
IUPAC	International Union of Pure and Applied Chemistry
LA	laser ablation
LI	laser ionisation
LIBS	laser induced breakdown spectroscopy
LIF	laser induced fluorescence
LIMS	laser ionisation mass spectrometry
LIPS	laser induced plasma spectroscopy
LOD	limit of detection
LOQ	limit of quantitation
LPE	liquid phase extraction
LTE	local thermodynamic equilibrium
MC	multi collector
MIP	microwave induced plasma
MPT	microwave plasma torch
MS	mass spectrometry
n e	electron number density
NIST	National Institute of Standards and Technology
NP	nano particle
OES	optical emission spectroscopy
PCA	principal components analysis
PD	point discharge
PDMS	polydimethylsiloxane
PIVG	plasma induced vapour generation
PLS	partial least squares
PN	pneumatic nebulisation
PTFE	poly(tetrafluororthylene)
PVC	poly(vinyl chloride)
PVG	photochemcial vapour generation
QMS	quadrupole mass spectrometry
REE	rare earth element
RF	radiofrequency
RM	reference material
RSD	relative standard deviation
SBR	signal to background ratio
SIBS	spark induced breakdown spectrometry
SIMS	secondary ion mass spectrometry
SNR	signal to noise ratio
SP	single particle
SPE	solid phase extraction
SPME	solid phase microextraction
SRM	standard reference material
SRMS	sector field mass spectrometry
T e	electron temperature
T exc	excitation temperature
TIMS	thermal ionisation mass spectrometry
T ion	ionisation temperature
TISIS	total sample consumption system
TOF	time-of-flight
TOFMS	time-of-flight mass spectrometry
T rot	rotational temperature
T vib	vibrational temperature
TXRF	total reflection X-ray fluorescence
UV	ultra-violet
VG	vapour generation
XRF	X-ray fluorescence

7. Conflicts of interest

There are no conflicts of interest to declare.

References

1. E. H. Evans, J. Pisonero, C. M. M. Smith and R. N. Taylor, J. Anal. At. Spectrom., 2023, 38(5), 974–999.
2. J. R. Bacon, O. T. Butler, W. R. L. Cairns, O. Cavoura, J. M. Cook, C. M. Davidson and R. Mertz-Kraus, J. Anal. At. Spectrom., 2023, 38(1), 10–56.
3. M. Patriarca, N. Barlow, A. Cross, S. Hill, A. Robson and J. Tyson, J. Anal. At. Spectrom., 2023, 38(3), 496–577.
4. R. Clough, C. F. Harrington, S. J. Hill, Y. Madrid and J. F. Tyson, J. Anal. At. Spectrom., 2023, 38(7), 1339–1371.
5. R. Clough, A. Fisher, B. Gibson and B. Russell, J. Anal. At. Spectrom., 2023, 38(11), 2215–2279.
6. C. Vanhoof, J. R. Bacon, U. E. A. Fittschen and L. Vincze, J. Anal. At. Spectrom., 2023, 38(9), 1730–1743.
7. J. Ma, M. He, B. Chen and B. Hu, Metallomics, 2023, 15(4), mfad013.
8. Y. Xu, B. Chen, M. He, Z. Cui and B. Hu, Anal. Chem., 2023, 95(37), 14061–14067.
9. X. Wei, Y.-J. Wei, X. Zhang, C.-X. Wu, M.-L. Chen and J.-H. Wang, Microchem. J., 2023, 190, 108635.
10. E. A. Azooz, M. Tuzen, W. I. Mortada and N. Ullah, Crit. Rev. Anal. Chem., 2022, 1–15.
11. Y. Wang, Y. Lin, T. Ren, Y. Yang, Z. He, Y. Deng and C. Zheng, Anal. Chem., 2023, 95(29), 10873–10878.
12. N. S. Medvedev, At. Spectrosc., 2023, 44(1), 430–436.
13. P. Menero-Valdes, M. I. Chronakis, B. Fernandez, C. D. Quarles Jr, H. Gonzalez-Iglesias, B. Meermann and R. Pereiro, Anal. Chem., 2023, 95(35), 13322–13329.
14. J. Zhang, Y. Lu, W. Gao, P. Yang, N. Cheng, Y. Jin and J. Chen, Microchem. J., 2023, 186, 108365.
15. Y. L. Zhu, J. K. Wang, Z. P. Chen, Y. J. Zhao and R. Q. Yu, Mikrochim. Acta, 2023, 190(1), 20.
16. T. L. Taylor and A. Gundlach-Graham, J. Anal. At. Spectrom., 2023, 38(3), 721–729.
17. M. Janeda, M. Krawczyk-Coda and M. Ślachciński, Spectrochim. Acta, Part B, 2023, 201, 106614.
18. S. Martinez, R. Sanchez and J. L. Todoli, Anal. Chim. Acta, 2023, 1249, 340948.
19. R. Sánchez and J.-L. Todolí, Spectrochim. Acta, Part B, 2023, 208, 106779.
20. D. Beauchemin, At. Spectrosc., 2022, 43(05), 371–377.
21. F. Laborda, I. Abad-Álvaro, M. S. Jiménez and E. Bolea, Spectrochim. Acta, Part B, 2023, 199, 106570.
22. Z. Meng, L. Zheng, H. Fang, P. Yang, B. Wang, L. Li, M. Wang and W. Feng, Processes, 2023, 11(4), 1237.
23. C. Davison, D. Beste, M. Bailey and M. Felipe-Sotelo, Anal. Bioanal. Chem., 2023, 415(28), 6931–6950.
24. I. Bastardo-Fernández, R. Chekri, J. Noireaux, P. Fisicaro and P. Jitaru, Spectrochim. Acta, Part B, 2023, 208, 106782.
25. H. Furutani, K. Kato, T. Hinoue, T. Kimoto and M. Toyoda, Talanta, 2023, 260, 124616.
26. S. Harycki and A. Gundlach-Graham, Anal. Chem., 2023, 95(41), 15318–15324.
27. D. Torregrosa, C. Gómez-Pertusa, G. Grindlay, L. Gras and J. Mora, J. Anal. At. Spectrom., 2023, 38(2), 403–413.
28. D. Torregrosa, G. Grindlay, L. Gras and J. Mora, J. Anal. At. Spectrom., 2023, 38(9), 1874–1884.
29. E. Pitzalis, B. Campanella, R. Bonini, M. Onor and A. D'Ulivo, Anal. Chim. Acta, 2023, 1269, 341427.
30. S. Musil, E. Jenikova, J. Vyhnanovsky and R. E. Sturgeon, Anal. Chem., 2023, 95(7), 3694–3702.
31. J. Hu, Y. Yu, Z. Xiao and Y. Gao, Microchem. J., 2023, 193, 109178.
32. L. Dong, W. Wang, Y. Ning, X. Deng and Y. Gao, Anal. Chim. Acta, 2023, 1251, 341006.
33. Y. Su, Y. Lin, T. Ren, Y. Li, Y. Deng and C. Zheng, Anal. Chem., 2023, 95(37), 14036–14042.
34. Y. Su, Y. Lin, R. Yang, L. He, J. Yang, X. Jiang and C. Zheng, Anal. Chem., 2022, 94(46), 16265–16273.
35. L. Li, C. Jiang, J. Xiao, H. Luo, S. Zhang, Z. Zou and K. Huang, Spectrochim. Acta, Part B, 2023, 199, 106579.
36. X. Liu, G. Cheng, C. Yang, G. Wang, S. Li, Y. Li, H. Zheng, S. Hu and Z. Zhu, Talanta, 2023, 257, 124339.
37. G. D. S. Coelho Junior, D. L. G. Borges, M. Svoboda, J. Dědina and J. Kratzer, Spectrochim. Acta, Part B, 2023, 200, 106596.
38. C. Zheng, At. Spectrosc., 2023, 44(02), 76–83.
39. J. Orejas, Y. Zhang, C. Soto-Gancedo, L. J. Fernández-Menéndez, J. Pisonero and N. Bordel, Spectrochim. Acta, Part B, 2023, 209, 106786.
40. Z. Cai and Z. Wang, Spectrochim. Acta, Part B, 2023, 199, 106578.
41. Y. Liu, C. Yang, P.-J. Xing, X. Liu, J.-Z. Liu and Z.-L. Zhu, J. Anal. At. Spectrom., 2023, 38(3), 742–750.
42. M. Gorska and P. Pohl, Spectrochim. Acta, Part B, 2023, 208, 106778.
43. U. M. F. M. Cerqueira, M. A. Bezerra, V. A. Lemos, J. J. Coutinho, C. G. Novaes, S. A. Araújo and L. P. d. Nascimento, TrAC, Trends Anal. Chem., 2023, 167, 117277.
44. D. Antonov, E. Silkis, D. Shilo and B. Zuev, Spectrochim. Acta, Part B, 2023, 199, 106598.
45. L. N. Zheng, P. Xuan, J. Huang and J. L. Li, Spectrosc. Spectral Anal., 2023, 43(3), 665–673.
46. K. Finch, A. Hernandez and G. Gamez, Anal. Chem., 2023, 95(4), 2269–2277.
47. Y. Pohkrel, T. Adolphs, R. E. Peterson, U. Allebrod, B. J. Ravoo, H. F. Arlinghaus and B. J. Tyler, J. Am. Soc. Mass Spectrom., 2023, 34(2), 218–226.
48. R. E. Russo, Appl. Phys. A: Mater. Sci. Process., 2023, 129(3), 168.
49. B. Paul, J. Petrus, D. Savard, J. Woodhead, J. Hergt, A. Greig, C. Paton and P. Rayner, J. Anal. At. Spectrom., 2023, 38(10), 1995–2006.
50. L. Zeng, M. Wu, S. Chen, R. Zheng, Y. Rao, X. He, Y. Duan and X. Wang, Talanta, 2022, 246, 123516.
51. D. Metarapi and J. T. van Elteren, Talanta, 2023, 263, 124699.
52. A. Schweikert, S. Theiner, M. Sala, P. Vician, W. Berger, B. K. Keppler and G. Koellensperger, Anal. Chim. Acta, 2022, 1223, 340200.
53. M. Stiborek, L. Jindrichova, S. Meliorisova, A. Bednarik, V. Prysiazhnyi, J. Kroupa, P. Houska, B. Adamova, J. Navratilova, V. Kanicky and J. Preisler, Anal. Chem., 2022, 94(51), 18114–18120.
54. S. Strekopytov, K. Billimoria and H. Goenaga-Infante, J. Anal. At. Spectrom., 2023, 38(3), 704–715.
55. W. Luo, F. Dong, M. Wang, T. Li, Y. Wang, W. Dai, J. Zhang, C. Jiao, Z. Song, J. Shen, Y. Ma, Y. Ding, F. Yang, Z. Zhang and X. He, Anal. Chem., 2023, 95(15), 6391–6398.
56. J. S. Silva, G. M. Heidrich, B. O. Poletto, J. N. G. Paniz, V. L. Dressler and E. M. M. Flores, J. Anal. At. Spectrom., 2023, 38(5), 1000–1006.
57. Z. Cai, H. Chen, M. Gao and Z. Wang, Talanta, 2024, 266(Pt 2), 125111.
58. J. L. Tanen, S. R. White, D. Ha and K. Jorabchi, J. Anal. At. Spectrom., 2023, 38(4), 854–864.
59. M. Kuonen, G. Niu, B. Hattendorf and D. Gunther, J. Anal. At. Spectrom., 2023, 38(3), 758–765.
60. O. V. Komin and O. V. Pelipasov, Spectrochim. Acta, Part B, 2023, 207, 106742.
61. D. W. Koppenaal and R. K. Marcus, Appl. Spectrosc., 2023, 77(8), 885–906.
62. B. Štádlerová, J. Dědina and S. Musil, Spectrochim. Acta, Part B, 2023, 205, 106692.
63. F. Xu, Y. Liu, X. Liu, J. Tan, H. Xia and M. Li, Microchem. J., 2023, 193, 109109.
64. S. Liu, Q. Y. Yang, S. Chen, Y. L. Yu and J. H. Wang, Anal. Chem., 2022, 94(48), 16549–16554.
65. M. Zhang, Q. Tang, P. Li, L. He, X. Hou and X. Jiang, Anal. Chem., 2023, 95(11), 5151–5158.
66. X. Li, J. Riedel and Y. You, J. Anal. At. Spectrom., 2023, 38(5), 1088–1096.
67. S. Irvine, H. Andrews, K. Myhre, K. Goldstein and J. Coble, J. Quant. Spectrosc. Radiat. Transfer, 2022, 286, 108184.
68. S. Irvine, H. Andrews, K. Myhre and J. Coble, J. Quant. Spectrosc. Radiat. Transfer, 2023, 297, 108486.
69. D. Dojić, N. V. Nedić and S. Bukvić, Spectrochim. Acta, Part B, 2023, 209, 106795.
70. M. Fikry, W. Tawfik and M. M. Omar, J. Mol. Struct., 2023, 1276, 134796.
71. F. A. Khan, M. Shah, S. U. Haq and A. Nadeem, Spectrochim. Acta, Part B, 2023, 199, 106574.
72. J. Bounds, A. Kolomenskii, R. Trainham, M. Manard and H. Schuessler, Spectrochim. Acta, Part B, 2023, 202, 106635.
73. M. Liu, H. Ma, Y. Xu, D. Fang, H. Fu and Z. Dai, Spectrochim. Acta, Part B, 2023, 199, 106590.
74. A. Agrawal, S. S. Baghel, L. Sharma and R. Srivastava, Spectrochim. Acta, Part B, 2023, 208, 106764.
75. A. A. Pupyshev, J. Anal. Chem., 2023, 78(9), 1125–1145.
76. H. Kazama, K. Konashi, T. Suzuki, S.-I. Koyama, K. Maeda, Y. Sekio, T. Onishi, C. Abe, Y. Shikamori and Y. Nagai, J. Anal. At. Spectrom., 2023, 38(8), 1676–1681.
77. J. Hou, D. Zhang, Z. Feng, J. Zhu and L. Zhang, Opt. Express, 2023, 31(21), 34404–34412.
78. B. Bousquet, V. Gardette, V. M. Ros, R. Gaudiuso, M. Dell'Aglio and A. De Giacomo, Spectrochim. Acta, Part B, 2023, 204, 106686.
79. E. Gao, R. Wei, D. Zhang, Z. Zhu, Q. Gao and B. Li, J. Anal. At. Spectrom., 2023, 38(5), 1116–1124.
80. A. Chemin, A. Ross, S. Hermelin, P. Crozet, V. Motto-Ros, G. Ledoux, C. Dujardin and D. Amans, Spectrochim. Acta, Part B, 2023, 205, 106685.
81. J. Merten, E. Nicholas, S. Ethridge, H. Bariola, S. Chestnut, A. Anders, J. Brees and M. Foster, Spectrochim. Acta, Part B, 2023, 200, 106600.
82. Y.-F. Wang and X.-M. Zhu, Spectrochim. Acta, Part B, 2023, 208, 106777.
83. R. A. Kornev, I. B. Gornushkin, V. V. Nazarov, V. E. Shkrunin and A. A. Ermakov, Spectrochim. Acta, Part B, 2022, 195, 106502.
84. K. Senturk, T. Sen, T. Coruhlu, A. Shahzad and N. Aslan, Radiat. Eff. Defects Solids, 2023, 178(7–8), 912–928.
85. F. Hallwirth, M. Wolfgang and H. Wiltsche, J. Anal. At. Spectrom., 2023, 38(8), 1682–1690.
86. G. Schlemmer and I. Thomassen, Spectrochim. Acta, Part B, 2023, 207, 106753.
87. D. J. Butcher, Appl. Spectrosc. Rev., 2023, 1–29.
88. O. Acar, Appl. Spectrosc. Rev., 2022, 1–15.
89. S. N. Romanov, J. Anal. Chem., 2023, 78(5), 635–643.
90. V. Gardette, V. Motto-Ros, C. Alvarez-Llamas, L. Sancey, L. Duponchel and B. Busser, Anal. Chem., 2023, 95(1), 49–69.
91. J. Thomas and H. Chandra Joshi, Appl. Spectrosc. Rev., 2024, 59(1), 124–155.
92. L. Duponchel, C. Fabre, B. Bousquet and V. Motto-Ros, Spectrochim. Acta, Part B, 2023, 208, 106776.
93. L. Brunnbauer, Z. Gajarska, H. Lohninger and A. Limbeck, TrAC, Trends Anal. Chem., 2023, 159, 116859.
94. J. Hermann, C. Gerhard, M. Burger, V. Craciun and F. Pelascini, Spectrochim. Acta, Part B, 2023, 200, 106595.
95. F. Poggialini, B. Campanella, B. Cocciaro, G. Lorenzetti, V. Palleschi and S. Legnaioli, J. Anal. At. Spectrom., 2023, 38(9), 1751–1771.
96. A. Mohan, A. Banerjee and A. Sarkar, J. Anal. At. Spectrom., 2023, 38(8), 1579–1591.
97. L. M. John and K. K. Anoop, RSC Adv., 2023, 13(42), 29613–29624.
98. T. Völker and I. B. Gornushkin, J. Anal. At. Spectrom., 2023, 38(4), 911–916.
99. M. Dell'Aglio, C. Di Franco and A. De Giacomo, J. Anal. At. Spectrom., 2023, 38(3), 766–774.
100. C. Burgos-Palop, P. Purohit, F. J. Fortes and J. Laserna, Anal. Chem., 2023, 95(39), 14541–14550.
101. T. Chen, T. Zhang, H. Tang, X. Cheng and H. Li, Anal. Chem., 2023, 95(10), 4819–4827.
102. U. K. Adarsh, V. K. Unnikrishnan, P. Vasa, S. D. George, S. Chidangil and D. Mathur, Appl. Phys. B: Lasers Opt., 2023, 129(12), 185.
103. J. Liu, Z. Hou and Z. Wang, J. Anal. At. Spectrom., 2023, 38(12), 2571–2580.
104. C. Méndez-López, L. J. Fernández-Menéndez, C. González-Gago, J. Pisonero and N. Bordel, J. Anal. At. Spectrom., 2023, 38(1), 80–89.
105. C. Mendez-Lopez, L. J. Fernandez-Menendez, C. Gonzalez-Gago, J. Pisonero and N. Bordel, Opt. Laser Technol., 2023, 164, 109536.
106. Y. Ikeda, J. K. Soriano, K. Akaoka and I. Wakaida, Spectrochim. Acta, Part B, 2023, 203, 106651.
107. M. R. Khan, S. U. Haq, Q. Abbas and A. Nadeem, Spectrochim. Acta, Part B, 2023, 200, 106612.
108. B. Lei, B. Xu, J. Wang, X. Mao, J. Li, Y. Wang, W. Zhao, Y. Duan, V. Zorba and J. Tang, Cell Rep. Phys. Sci., 2023, 4(2), 101267.
109. J. Shang, Y. Ling, L. Ping, K. Pan, Y. Xue, Y. Wang, J. Li, S. Zhang and S. Zhong, Talanta, 2023, 264, 124780.
110. A. Robledo-Martinez, A. Garcia-Villarreal, H. Sobral and O. H. Magos, Spectrochim. Acta, Part B, 2023, 199, 106588.
111. F. Xu, S. Ma, H. Tian, Z. Xing, C. Zhao, Q. Feng, X. Zhao and D. Dong, J. Anal. At. Spectrom., 2023, 38(7), 1442–1449.
112. W. Zhao, C. He, A. Li, L. Qiu and K.-M. Xu, J. Anal. At. Spectrom., 2023, 38(4), 792–799.
113. Z. Jia, D. Li, Y. Tian, H. Pan, Q. Zhong, Z. Yao, Y. Lu, J. Guo and R. Zheng, Spectrochim. Acta, Part B, 2023, 206, 106713.
114. B. Wang, W. Song, Y. Tian, Y. Lu, Y. Li, J. Guo, W. Ye and R. Zheng, J. Anal. At. Spectrom., 2023, 38(2), 281–292.
115. D. Bleiner, D. Qu, K. Kraft and O. Shlyakhtun, Spectrochim. Acta, Part B, 2023, 204, 106668.
116. J. Ji, W. Song, Z. Hou, L. Li, X. Yu and Z. Wang, Anal. Chim. Acta, 2022, 1235, 340551.
117. L. J. Fernández-Menéndez, C. Méndez-López, C. González-Gago, J. Pisonero and N. Bordel, J. Anal. At. Spectrom., 2023, 38(2), 325–332.
118. L. Duponchel, A. Cugerone, M. Muñoz and V. Motto-Ros, Spectrochim. Acta, Part B, 2023, 207, 106758.
119. H. Chen, X. Shi, Y. He and W. Zhang, Spectrochim. Acta, Part B, 2023, 208, 106763.
120. H. Bai, P. Liu, X. Fu, L. Qiao, C. Liu, Y. Xin and Z. Ling, Spectrochim. Acta, Part B, 2023, 199, 106587.
121. C. Huffman and H. Sobral, Spectrochim. Acta, Part B, 2023, 199, 106573.
122. F. Poggialini, S. Legnaioli, B. Campanella, B. Cocciaro, G. Lorenzetti, S. Raneri and V. Palleschi, Appl. Sci., 2023, 13(6), 3642.
123. P. R. Villas-Boas and L. C. L. Borduchi, Spectrochim. Acta, Part B, 2023, 205, 106690.
124. J. Liu, W. Song, W. Gu, Z. Hou, K. Kou and Z. Wang, Anal. Chim. Acta, 2023, 1251, 341004.
125. Y. Lu, L. Liu, Z. Wu, Z. Xu, Z. Zhao, Z. Hao, J. Shi and X. He, J. Anal. At. Spectrom., 2023, 38(12), 2619–2624.
126. J. Nie, Y. Zeng, X. Niu, D. Zhang and L. Guo, J. Anal. At. Spectrom., 2023, 38(11), 2387–2395.
127. K. Kou, W. Song, W. Gu, J. Liu, Y. Song, J. Ji, Z. Hou and Z. Wang, J. Anal. At. Spectrom., 2023, 38(4), 873–882.
128. N. Li, N. Nishi, R. Zheng, Y. Zheng, J. Guan, C. Xue, Z. Zhang and T. Sakka, J. Anal. At. Spectrom., 2023, 38(10), 1908–1913.
129. S. Gruchola, A. Riedo, P. K. Schmidt, C. P. de Koning, L. N. Knecht, M. Tulej, F. Westall and P. Wurz, J. Anal. At. Spectrom., 2023, 38(7), 1372–1378.
130. S. Ma, Z. Zhang, Z. Xu, X. Qiu and W. Hang, J. Anal. At. Spectrom., 2023, 38(3), 595–602.
131. Z. Lu, J. M. Zhu, D. Tan, T. M. Johnson and X. Wang, Anal. Chem., 2023, 95(4), 2253–2259.
132. J. Mulder, G. Hagen-Peter, T. Ubide, R. Andreasen, E. Kooijman, M. Kielman-Schmitt, Y. X. Feng, B. Paul, A. Karlsson, C. Tegner, C. Lesher and F. Costa, Geostand. Geoanal. Res., 2023, 47(2), 311–336.
133. P. Télouk, E. Albalat, B. Bourdon, F. Albarède and V. Balter, J. Anal. At. Spectrom., 2023, 38(10), 1973–1983.
134. T. Y. Xia, W. W. Sun, S. Ebser, W. Jiang, G. M. Yang, H. M. Zhu, Y.-C. Fu, F. Huang, G. D. Ming, T. Xia and Z. T. Lu, Nat. Phys., 2023, 19(6), 904–908.
135. M. Grotti, M. A. Vecchio, D. Gobbato, M. Mataloni and F. Ardini, J. Anal. At. Spectrom., 2023, 38(5), 1057–1064.
136. C. F. Li, Z. Y. Chu and P. Peng, Talanta, 2023, 257, 124390.
137. A. Redaa, J. Farkaš, S. Gilbert, A. S. Collins, S. Löhr, D. Vasegh, M. Forster, M. Blades, T. Zack, A. Giuliani, R. Maas, A. Baldermann, M. Dietzel and D. Garbe-Schönberg, Geostand. Geoanal. Res., 2023, 47(1), 23–48.
138. Z. Chu, J. Xu, X. Ling, P. Zhang, H. Cheng, J. Guo and X. Zhou, Acta Petrol. Sin., 2022, 38(12), 3695–3702.
139. L.-J. Duan, L.-L. Zhang, D.-C. Zhu, Y.-H. Yang, J.-C. Xie, Q. Wang, S.-T. Wu, C. Huang, C. Li, W.-T. Xu, S. Kamo, L.-J. Xu, C.-X. Pan and G.-H. Shi, J. Anal. At. Spectrom., 2023, 38(7), 1478–1493.
140. L. Xiang, R. Wang, R. L. Romer, X. Che, H. Hu and Z. Tang, Geostand. Geoanal. Res., 2023, 47(3), 609–628.
141. G.-Q. Liu, K.-D. Zhao, T. Ulrich, W. Chen, D. Zhang, Q. Li, H.-D. Zhao, R.-Q. Zhang and F. Xia, J. Anal. At. Spectrom., 2023, 38(10), 2007–2018.
142. J. P. A. Silva, C. Lana, A. Mazoz, I. Buick and R. Scholz, Geostand. Geoanal. Res., 2023, 47(1), 49–66.
143. M. Nakazato, H. Asanuma, S. Niki, H. Iwano and T. Hirata, Geostand. Geoanal. Res., 2022, 46(4), 603–620.
144. S. Niki, S. Kosugi, H. Iwano, T. Danhara and T. Hirata, Geostand. Geoanal. Res., 2022, 46(4), 589–602.
145. Y. Tang, N. Liu, J. Yang, G. O. Gonçalves, L. Liu, T. Lan, J. Gao and J. Han, J. Anal. At. Spectrom., 2022, 37(12), 2599–2614.
146. I. K. Kerber, J. Arps, R. Eichstädter, F. Kontor, C. Dornick, A. Schröder-Ritzrau, A. Babu, S. Warken and N. Frank, Nucl. Instrum. Methods Phys. Res., Sect. B, 2023, 539, 169–178.
147. X. Wang, X. Wang, L. Wang, S. Wu, D. Xue, W. Duan, Z. Ma, J. Xiao and X. Li, Geostand. Geoanal. Res., 2023, 47(4), 945–955.
148. A. L. Harrison, A. Heuser, V. Liebetrau, A. Eisenhauer, J. Schott and V. Mavromatis, Earth Planet. Sci. Lett., 2023, 603, 117985.
149. G. Wang, Y. Zeng, L. Qi, W. Liu and J. Xu, Anal. Chim. Acta, 2023, 1278, 341721.
150. S. An, J. Chen, S. Boschi and W. Li, Anal. Chem., 2023, 95(4), 2140–2145.
151. M.-Y. He, T. X. Ren, Z. D. Jin, L. Deng, H. J. Liu, Y. Y. Cheng, Z. Y. Li, X. X. Liu, Y. Yang and H. Chang, Spectrochim. Acta, Part B, 2023, 209, 106781.
152. J. Lewis, T.-H. Luu, C. D. Coath, H. Wehrs, J. B. Schwieters and T. Elliott, Chem. Geol., 2022, 614, 121185.
153. W. Li, Y. Zhao, B. Su, B. Gao, J. Wang and S. Liu, J. Anal. At. Spectrom., 2023, 38(3), 603–608.
154. J.-H. Bai, M. Lin, S.-X. Zhong, Y.-N. Deng, L. Zhang, K. Luo, H. Wu, J.-L. Ma and G.-J. Wei, J. Anal. At. Spectrom., 2023, 38(3), 629–637.
155. F. Liu, X. Li, H. Yang, Q. Peng, J. Wu and Z. Zhang, J. Anal. At. Spectrom., 2023, 38(12), 2581–2589.
156. H. Liu, W. Zhang, Z. Deng, Z. Hu, M. Schiller, M. Bizzarro, Y. Liu, T. Luo, Y. Feng and L. Feng, Spectrochim. Acta, Part B, 2023, 202, 106646.
157. Y. Cai, H. Chen, B. Li, E. T. Rasbury and A. N. Halliday, Rapid Commun. Mass Spectrom., 2023, 37(13), e9526.
158. J. Fietzke and E. Anagnostou, Geostand. Geoanal. Res., 2023, 47(3), 481–492.
159. A. N. Paul, J. A. Stewart, S. Agostini, L. A. Kirstein, J. C. M. De Hoog, I. P. Savov, K. Walowski, J. W. B. Rae and T. Elliott, Geostand. Geoanal. Res., 2023, 48(1), 91–108.
160. M. Y. He, L. Deng, J. Liu, Z. D. Jin and T. Ren, RSC Adv., 2023, 13(46), 32104–32109.
161. S. Liu, W. Li, B. Su, B. Gao, J. Wang, C. Wang, Y. Luo, L. Yan and Y. Zhao, J. Anal. At. Spectrom., 2023, 38(3), 656–666.
162. H. Chang, J.-M. Zhu, X. Wang and T. Gao, J. Anal. At. Spectrom., 2023, 38(4), 950–962.
163. Y. Fang, K. Chen, Z. Bao, C. Zong, H. Yuan and N. Lv, Anal. Chem., 2022, 94(48), 16746–16751.
164. E. M. Morton, K. Kreissig, B. J. Coles, C. S. Jaffe, R. Martins, G. M. Poole and M. Rehkämper, J. Anal. At. Spectrom., 2023, 38(6), 1192–1204.
165. J.-X. She, W. Li, S. An, T. Yang and R. Zhang, J. Anal. At. Spectrom., 2023, 38(5), 1043–1056.
166. D. Zhang, Z. Bao, P. Liu, G. Brügmann, W. Yang, K. Chen, P. Liang and H. Yuan, J. Anal. At. Spectrom., 2023, 38(1), 204–211.
167. M. Yang, Y.-H. Yang, R. L. Romer, S.-T. Wu, T. Wu and H. Wang, J. Anal. At. Spectrom., 2023, 38(2), 437–448.
168. M. Shi, B. A. Bergquist, A. Zhou, Y. Zhao, R. Sun, J. Chen and W. Zheng, J. Anal. At. Spectrom., 2023, 38(5), 1076–1087.
169. J. Wang, Y. Di, D. Asael, N. J. Planavsky and L. G. Tarhan, Chem. Geol., 2023, 621, 121365.
170. Y. Xia, Q. Deng, G. Sun, Y. Wu, S. Qiao, J. Ali, X. Fu and X. Feng, J. Anal. At. Spectrom., 2023, 38(2), 359–368.
171. K. Zhu, At. Spectrosc., 2023, 44(03), 142–152.
172. L. Yang, B. Methven, Z. Mester and J. Meija, J. Anal. At. Spectrom., 2023, 38(10), 2080–2086.
173. Y. Zhao, W. Yang and C. Leng, J. Anal. At. Spectrom., 2023, 38(4), 785–791.
174. H. Jaegler and A. Gourgiotis, J. Anal. At. Spectrom., 2023, 38(10), 1914–1919.
175. J. Wimpenny, K. M. Samperton, P. Sotorrio, M. S. Wellons, S. M. Scott, D. Willingham and K. Knight, J. Anal. At. Spectrom., 2023, 38(4), 827–840.
176. N. Janardhanan, U. K. Maity, M. Joseph, P. Manoravi and N. Sivaraman, J. Anal. At. Spectrom., 2023, 38(11), 2324–2331.
177. A. Magre, B. Boulet, H. Isnard, S. Mialle, O. Evrard and L. Pourcelot, Anal. Chem., 2023, 95(17), 6923–6930.
178. L. Zhang, L. Chen, J. Xu, H. Zhang, D. Zhang, Z. Qin and J. Liao, Anal. Chem., 2013, 85(2), 1106–1114.

---