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

E. Hywel Evans *a, Jorge Pisonero b, Clare M. M. Smith c and Rex N. Taylor d
aThe Open University, UK. E-mail: e.h.evans@open.ac.uk
bUniversity of Oviedo, Faculty of Science, Department of Physics, c/ Calvo Sotelo s/n, 33006 Oviedo, Spain
cSchool of Education, College of Social Sciences, University of Glasgow, University Avenue, Glasgow G12 8QQ, UK
dOcean and Earth Science, University of Southampton, NOC, Southampton, SO14 3ZH, UK

Received 17th March 2025

First published on 8th April 2025


Abstract

This review of 180 references covers developments in ‘Atomic Spectrometry’ published in the twelve months from December 2023 to November 2024 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. Automated chip-based methods for elemental tagging and amplification are being developed which promise to further bring this technique into routine operation. High resolution bioimaging using LA-ICP-MS is becoming more sophisticated due to finer control of the laser pulse and new methods developed to deal with the large amounts of data generated. Single particle analysis utilising ICP-MS continues to offer new developments, particularly when used in conjunction with elemental tagging and bioimaging; and a new hyphenated method for single particle analysis was reported which married optofluidic force induction, Raman spectroscopy and ICP-MS. New research on LIBS techniques continues to address the issues of matrix effects and calibration, liquid analysis and chemometric data treatment, driving the technique forward.


1. Liquids analysis

1.1. Sample pre-treatment

1.1.1. Extraction methods. Extraction methods are used to separate analyte and matrix, and for preconcentration of the analyte. Liquid extraction methods have been the subject of development over many years so new advances are rare. Botella et al.7 developed a solvent extraction method using a deep eutectic solvent (DES). Such solvents exhibit similar properties to ionic liquids, such as low conductivity, low vapour pressure and high thermal stability. However, they have lower toxicity and are more biodegradable. A DES composed of thymol and decanoic acid was used to extract SeIV after complexation with ammonium pyrrolidine dithiocarbamate (APDC). The extraction was performed into 50 μL of the DES (a 1[thin space (1/6-em)]:[thin space (1/6-em)]2 ratio of thymol and decanoic acid) followed by phase separation upon freezing. The SeVI remained in the aqueous phase and was determined by ICP-MS while total Se was determined by analysis of the acidic extract. Thus, SeIV was calculated by subtraction. An extraction efficiency of 99.6% for SeIV with LODs of 0.31 and 0.24 μg L−1 for SeIV and SeVI respectively.

Cheng and Su8 fabricated a digital light processing, 3D-printed photopolymer-based porous monolith in an SPE column. The integrated nature of the fabrication eliminated separate fabrication and packing steps. Porous structures were formed on 3D-printed, sub-millimetre spaced interlacing cuboids rather than as a compact monolith. In order to control the pore characteristics, they incorporated thermally expandable microspheres into the photocurable resins, which were subjected to a post-printing thermal foaming. This increased the surface roughness of the monolithic packing and allowed a flow rate of 1.0 mL min−1 through the column, thus allowing the extraction and determination of Mn, Co, Ni, Cu, Zn, Cd, and Pb by 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.

Hu et al.9 reviewed the label-free detection of biomolecules using ICP-MS. The review included: enzymatic cleavage of metal-labelled substrates; release of immobilised metal ions from DNA; and nucleic acid amplification-assisted aggregation and release of metal tags to achieve amplified detection. Song et al.10 claimed to have developed a label-free method for the determination of lymphoblastic leukaemia cells. The detection method was based on a competitive pre-reduction of ascorbic acid by partially complexed Cu2+ and free Hg2+ and the subsequent reduction in Hg signal using ICP-MS. However, the methodology was difficult to follow and it was unclear how reduction of Hg2+ to Hg0 could be reliably measured by ICP-MS if the two species were not separated somehow.

Metal nanoparticles form the basis for many biological assays, using them as elemental tags and as a signal amplification method, particularly when ICP-MS detection is used. Further amplification of nucleic acid analytes can also be achieved by polymerase chain reaction of the tagged molecules. Wei et al.11 developed a method whereby AuNPs, modified with a linker DNA and poly-T10, were used as the amplification tag. These AuNP–DNA nanoparticles were then conjugated with streptavidin functionalised magnetic beads (MBs) with diameters of 1 μm to form DNA tracks. The MB-AuNP-DNA conjugates were then mixed with target miRNA-200c, an aggressive prostate cancer biomarker, and subjected to polymerase chain reaction. The cleaved AuNPs were analysed using SP ICP-MS, after the MBs had been magnetically separated from them. Detection of miR-200c in human serum with an LOD of 0.93 pM was achieved. A similar, though not identical, amplification scheme was reported by Wang et al.12 for simultaneous detection of miR296 and miR16 using ICP-MS. AgNPs were the basis for detection with an LOD of ∼0.4 nM.

Torregrosa et al.13 evaluated two assays for the determination of β-conglutin. Two assay schemes were investigated: thiolated aptamers chemisorbed onto AuNPs; and biotinylated aptamers linked to streptavidin-AuNPs. In addition, both conventional and SP ICP-MS were compared. They found that the respective LODs for the thiolated and biotinylated aptamer assays were 200 and 2 pM using conventional ICP-MS, and 150 and 10 pM using SP ICP-MS.

Li et al.14 developed a microfluidic chip for labelling and determination of E. coli and Salmonella by ICP-MS. The microfluidic chip had a single channel with four inlet ports: for sample, washing buffer, elemental tag inlet and elution. A single outlet port was interfaced with ICP-MS via a micro-concentric nebuliser with a flow rate of 0.3 mL min−1. In practice, 100 μL of a capturing probe solution containing 5 mg L−1 of antigen probes, which had been conjugated with MBs (MB-anti-7C2 and MB-anti-8G3), were injected into the channel and immobilised for 5 min by high-intensity neodymium–iron–boron magnets. Next, 100 μL of sample containing E. coli and Salmonella was injected into the channel and held for 20 min to allow capture by the immobilised MB-antigens. This was then washed for 3 min with buffer, and then with 100 μL of a labelling probe solution containing antigen probes which had been conjugated with Ag and Au NPs (AgNP-anti-8B1 and AuNP-anti-5H12). This was allowed to stand for 20 min. After washing, the labelled analytes were eluted in 1.2 mol L−1 formic acid and analysed using ICP-MS. The system was fully automated. Method LODs of 200 CFU mL−1 for E. coli O157:H7 and 152 CFU mL−1 for Salmonella were obtained. Spiking experiments on human blood yielded recoveries between 86.8 and 106%.

An automated method for labelling was also developed by Sun and Zhang15 using a well-plate. Their labelling strategy utilised a biotin-antibody REE method for multiplex detection of several lung cancer markers using ICP-MS. After adding samples to the well-plate, washing, sampling, uptake and detection were automated to give an analysis time of 40 min. The LODs (2 SD) were within the range necessary for clinical analysis (∼1.4 to 400 pg mL−1).

1.2. Sample introduction

1.2.1. Nebulisation. A vibrating capillary nebuliser (VCN) for on-line coupling of CE with ICP-MS was evaluated by Taylor et al.16 The VCN created an aerosol independent of gas flows and did not produce a low-pressure region at the nebuliser orifice. The new device was compared to two commercially available PNs and was found to have the highest baseline stability with a standard deviation of 3500 counts per s, corresponding to an RSD of 2.7%.
1.2.2. Single particle analysis. The accurate and reliable characterisation of nanomaterials and biological samples at the individual particle and cell level continues to be of significant interest. Davison et al.17 reviewed recent advances in SP ICP-MS, LIBS and LA-ICP-MS for elemental analysis of tissues and single cells, highlighting the importance of single-cell analysis for understanding cellular heterogeneity and drug effectiveness. The authors noted that while SP ICP-MS allows for the analysis of a large number of cells, it does not provide spatial information, whereas laser-based techniques like LA-ICP-MS enable elemental mapping at the single-cell level. The sensitivity of commercial LIBS instruments currently restricts its use for sub-tissue applications. However, the capacity to analyse endogenous bulk components, combined with developments in nano-LIBS technology, shows great potential for cellular research. LA-ICP-MS offers high sensitivity for the direct analysis of single cells, but standardisation requires further development. The authors concluded that the hyphenation of these trace elemental analysis techniques and their coupling with multi-omic technologies for single-cell analysis have enormous potential for answering fundamental biological questions.

Two novel sample introduction systems have been proposed for SP ICP-MS. A miniaturised ultrasonic nebulisation system that achieved transport efficiency of approximately 80% for AgNPs was proposed by Dong et al.18 The system addressed the limitations of conventional PN systems with typical TEs of 1–5%. Kajner et al.19 developed a 3D-printed polymer sample introduction system for SP ICP-MS. Compared to standard systems, it had four-fold higher particle detection efficiency, significantly better SNRs, 20% lower size detection limit and allowed an extension of the upper limit of transportable particle diameters to about 25 μm.

Bolea and Laborda20 addressed the critical role of transport efficiency (TE) in SP ICP-MS. They highlighted discrepancies in TE determination and emphasised that the methods used, namely; particle frequency (TEF), particle size (TES), and dynamic mass flow (DMF) methods measured different efficiencies. TEF and TES measure the TE corresponding to the particles or the dissolved element, respectively, whereas the DMF method measures the solvent TE. Hence, different assumptions must be made when using each method. These assumptions, together with sources of bias associated with each method were critically discussed to provide a holistic view of TE in the context of SP ICP-MS metrology. Murphy et al.21 also considered the TEF, TES, and DMF methods for TE measurement. The three methods were compared using three different spray chambers operated at cooled (2 °C to 10 °C) and ambient (19 °C to 21 °C) temperature conditions, equipped with different nebulisers on different ICP-MS platforms. While agreement between the three measures of TE was achieved when operating the spray chambers at 2 °C, repeatability of the DMF method was poor. They found that the DMF method overestimated TE and suggested that the direct measures of TE (TEF and TES) yielded more reliable results over a wide range of conditions, especially at ambient temperature. In a critique of this conclusion, Infante22 argued that Murphy et al.21 applied non-optimal conditions in the use of DMF and did not adhere to recommended practices. Murphy et al.23 defended their approach by stating that they aimed to only evaluate those parameters under which TE measurements were accurate. Geiss et al.24 conducted an inter-laboratory comparison study of TE determination using six AuNP suspensions across seven laboratories. They found that while TE values did not deviate much under ideal conditions, the particle frequency method had more systematic variability.

Goodman et al.25 described particle generated spectral interferences in SP ICP-MS as ‘a roadblock to accurate nanometrology’. In the SP ICP-MS analysis of Y- and Nd-NP suspensions, interferences from YO+ and Nd2+ are often falsely identified as Pd, Ge, and As NPs. Using a DRC, the YO+ interferences were eliminated and Pd NPs were accurately measured in mixed suspensions of Pd- and Y-NPs. A strong correlation was observed between the magnitude of interferences in solution mode and SP mode, supporting a similar mechanism of interferent formation in the plasma for dissolved and NP-associated ions. In addition, interference formation was affected by changes in nebuliser gas flow in a manner similar to dissolved ions. Ion clouds from ablated NPs form spectral interferences in a manner similar to their dissolved counterparts.

1.2.3. Vapour generation. A review of pyrohydrolysis sample introduction methods for determination of halogens was conducted by Oliveira et al.26 The review covered the last 30 years of literature and included 147 references. The information was ordered in useful tables arranged by sample type and detection technique, including ICP-MS, ICP-OES and EDXRF. Rocks, minerals, soils and organic matter were amongst the sample types considered. A consideration of the different types of pyrolysis systems was also given.

Atomisation cells such as the heated QTA, DBD or miniature diffusion flames have insufficient thermal energy to effect thermal atomisation of hydrides, thus a free radical reaction pathway for formation of volatile hydrides seems likely. This radical theory of hydride atomisation was reviewed by Dedina.27 Using examples from the authors own research, the free radical theory was applied to each of these techniques. The theory was well corroborated for the QTA based on the existence of H˙ radicals at high concentrations compared with the analyte hydride. The example of Se was given where results of research suggested atomisation of its hydride in the H˙ cloud and fast decay of free Se atoms upon leaving the H˙ cloud. In contrast, there was negligible free atom decay in the case of Pb after the initial atomisation of hydride in the H˙. In the case of the DBD and miniature diffusion flame the evidence was less conclusive and more research is required. Campanella and D'Ulivo28 reviewed (36 references) the last 20 years of research into mechanisms of vapour generation by tetrahydridoborate (THB) and amine boranes. The mechanism of hydrolysis of aqueous THB amine boranes, and the effects of reaction conditions, interferences, and chemical additives were examined.

Buoso et al.29 investigated the interferences caused by selected transition metals on the generation of SbH3 and BiH3 by aqueous THB. They used hydridoboron intermediates (BH) as derivatisation reagents by generating them in a pre-hydrolysis coil. They found that, compared to NaBH4, the BH intermediates resulted in improved tolerance to interferences caused by of NiII, CoII and CuII up to 100 mg L−1 in the absence of masking agents, particularly when an acid concentration of 1 mol L−1 HCl and a 1 mL mixing coil were used. The authors noted the results indicated that the BH intermediates were more selective towards the analyte relative to the interfering metal ions, thus eliminating the requirement for masking agents.

Photochemical vapour generation (PVG) is a well-established technique whereby free radicals and aqueous electrons are generated from a sample solution exposed to UV radiation. These undergo reactions to from volatile species such as hydride, alkyl or carbonyl compounds of analytes if interest. The specificity and extent of reaction is influenced by the reaction medium and concomitant species, particularly the presence of organic acids and transition metals used as so-called ‘sensitisers’. Thus, much research has been directed towards the optimisation of the reaction conditions. Xu et al.30 investigated the mechanism of PVG of Se using HPLC-ICP-MS, GC-MS and AFS to identify the species generated. They generated R˙ and ˙COO radicals from low molecular weight organic acids, which were then successively reacted with SeIV to generate elemental Se0, which was further reduced to release volatile Se species. Specifically, they investigated: the conventional reduction of SeIV to SeH2 and (CH3)2Se; photo-oxidation of SeIV to SeVI; photo-reduction of SeVI to SeIV; the comproportionation reaction of SeH2 with SeIV; and the photo-oxidation of SeH2. Unfortunately, the authors were unable to come to any firm conclusions about the mechanism other than to conclude that it was a complicated process.

Deng et al.30investigated the PVG of As and the effect of the ‘sensitisers’ Sb and Cd in ethanoic acid, using ICP-MS, GC-MS UV-vis spectroscopy and XPS. Transition-metal ions can form complexes with carboxylic acids. So, based on the premise that this can promote UV absorption by the PVG medium and thus accelerate the ligand–metal charge transfer process in the photochemical reaction, an array of measurements were performed to try and identify increased UV absorption and the species generated. The authors drew no firm conclusion from these measurements but speculated that the addition of SbIII alone or with CdII resulted in an increase in the ratio of reductive/oxidative radicals, and the enhanced PVG efficiency of As might be attributed to a decrease in ˙OH radicals and the consequent increase in reductive radicals for photochemical reduction of As.

Dong et al.31 investigated the PVG mechanism for the formation of volatile Sb species, using EPR, in the presence of VIV sensitiser and formic/ethanoic acids. GC-MS confirmed the generation of (CH3)3Sb for both SbIII and SbV. They speculated that SbIII and SbV was first reduced to Sb0 by ˙H, ˙CO2 or electrons, formed by photochemical decomposition of formic acid, then formed volatile methylated (CH3)3Sb by reaction with ˙CH3. Hu et al.32 of the same group reported on a study the PVG of Sn and the influence of the sensitisers Cd2+ and Cl. They also studied the mechanism using EPR and observed the formation of ˙CO2, ˙CH3, and ˙OH free radicals. They found that more free radicals were generated in ethanoic acid when the Cd2+ and Cl sensitisers were present. Thus, they speculated, in strikingly similar language to the previous study, that Sn2+ was reduced by electrons or free radicals such as ˙CO2, then formed volatile methylated Sn compounds by reaction with ˙CH3.

Photochemical vapour generation is suited to the generation of volatile species of elements which are not traditionally associated with traditional vapour generation methods such as CVG. Dong et al.33 investigated the use of Cu2+, Co2+ and Cd2+ as sensitisers for the PVG of Ta and Nb, with Cu2+ proving to be the best, resulting on in 76- and 91-fold enhancements for Nb and Ta using ICP-MS detection. Using TEM and XPS, they also identified the formation of nanoparticles of Nb and Ta the gaseous and liquid phases after photochemical reduction. The authors speculated that carbonyls of Nb and Ta were generated but then rapidly decomposed to form the nanoparticles. Nanoparticles of Cu, Zn and Ga were also observed by Hu et al.34 during the PVG of those elements in the presence of formic acid, and Cu2+ used as a sensitiser.

The dielectric barrier discharge (DBD) can be used both as a vapour generator and as an atomisation/excitation source. When used as a vapour generator, some of the mechanisms are probably similar to those that occur during PVG. Wang et al.35 used SPME coupled directly with a DBD for vapour generation of Hg and detection using AFS. An SPME comprised of multiwalled carbon nanotubes coated onto a glass tube was used to extract Hg2+ from seawater, which was then desorbed and reduced to Hg0 vapour directly in the DBD. The discharge was operated at 250 V with a 600 mL min−1 flow of Ar directly into AFS for detection.

Electrochemical vapour generation (EVG) has some particular advantages, but also serious drawbacks. The advantages are selectivity and the ability to preconcentrate a particular analyte, but it does suffer from interferences from a complex sample matrix. Gao et al.36 tested different electrode structures for the selective electroreduction of AsIII. They constructed several modified foam electrodes by chronoamperometry, i.e. by plating a nickel coated foam electrode with Co, Fe, Cu or Mn. Of these, the Fe modified foam electrode proved to be the most effective in terms of selectivity and electroreduction efficiency for AsIII and AsV. The authors postulated that conversion of AsV/III to AsH3 involved a multi-step, electron proton transfer process. Binding between Fe–AsIII was thought to promote the forward reaction probably due to the increase in surface area and decrease in charge transfer resistance, likely by a catalytic mechanism. Thus, it was possible to distinguish between the valence states of arsenic through simple surface modification.

1.2.4. Direct methods. Janeda et al.37 described a new sample injection system for the direct analysis of 10 μL micro-slurries using an MIP-OES. It provided rapid and direct sample injection, with no prior sample decomposition. Analytical figures of merit (LODs, absolute LODs and precision) were presented for Ba, Cd, Cu, Mn, Pb, Sr and Zn and the elements determined successfully in four CRMs and four real samples.

Beauchemin38 described a liquid microjunction (LMJ), that allowed for the injection of 1 μL of sample into ICP-MS. The system functioned at a continuous flow and was self-cleaning, enabling sampling of minute liquid volumes and surface leaching of solids. Simply touching the sample to the dome of liquid at the inlet of the LMJ allowed for sample introduction similar to the mechanism of sample transport in flow injection analysis. The approach was found to significantly reduce matrix effects and oxide interferences.

The application of OES to gas samples has been rare owing to the limited sensitivity. A microplasma with a long life span, generated using a high AC drive voltage, was used as a source for a home-built compact GD-OES system by Kim et al.39 Determination of N2 and Ar in ambient O2, provided detection sensitivities of <1 and <0.2 ppm (3δ), respectively, within 10 s, suggesting that ppb-level detection is possible within a few minutes.

2. Solids analysis

2.1. Direct methods

2.1.1. Glow discharge. The continued interest in GD as a source for OES and MS confirms the versatility of the technique. Different GD configurations, such as SCGD, SAGD, and Grimm-type discharges, offer unique capabilities and advantages for a range of applications.

Cai et al.40 studied the effect of different salt matrices on the signals of Zn, Cd, Ni, Cu, Cr, and Pb using SCGD-OES. They reported significant matrix effects caused by common cations, such as Na+, K+, Mg2+, and Ca2+, with these ions inhibiting analyte emission when present at concentrations exceeding 200 mg L−1. Anions such as Cl and SO42− generally showed little interference at the same concentrations in SCGD-OES. Formic acid (HCOOH) was found to play a positive role in the formation of volatile species of metals such as Pb, regardless of the presence of Na salts. The same modifier was found to exhibit inhibitory effects on elements that are more difficult to evaporate and stimulate, such as Cr. Zhang et al.41 also observed complex interactions when using HCOOH as an additive. The emission intensity for certain analytes was observed to decline with increasing HCOOH concentration under certain sample matrix conditions. This intensity depression mainly appeared when the sample matrix contained alkali cations, regardless of the anion type, a finding that was in agreement with Cai et al.40

Khanh et al.42 demonstrated the application of artificial neural networks (ANN) in conjunction with SCGD-OES for Cu determination. The LOD for Cu was reduced from 1.2 mg L−1 by linear regression to 0.3 mg L−1 using ANN models. To interpret the ANN model, ANN's weight characterisation was analysed, and its results show that ANN could recognise the critical emission lines that affect the prediction results and separate the spectral line even in spectrum superposition. This work demonstrated the potential of machine learning to enhance analytical performance and interpretation of complex spectral data.

Liu et al.43 developed a droplet cathode GD for serum analysis and achieved LODs for Li, Ca and K of 0.002 mg L−1 (20 pg), 0.078 mg L−1 (780 pg), and 0.005 mg L−1 (50 pg), respectively. The system showed potential for rapid, low-volume analysis of trace metals in biological samples.

The determination of relative sensitivity factor (RSF) plays a crucial role in achieving accurate quantitative analysis in GD-MS. The RSFs of typical elements in 14 metal and metal oxide matrices were determined, and the correlations with discharge parameters, matrix effects, first ionisation energies and sample shapes were investigated.44 Matrix effects were found to be more pronounced in metal oxides, with the RSDs of RSFs for typical elements ranging from 25.17% to 101.04%. This indicated the necessity to mitigate matrix-related signal variations for accurate quantification. A correlation was found between RSF values for the same element across different metal oxides and their lattice binding energies.

The performance of GD systems is heavily influenced by parameters like gas type, gas flow, discharge current, voltage, tube material, and the design of the discharge cell. Gorska and Pohl45 found that tubes made from W and corundum showed superior resistance to the conditions under which the discharge was sustained, and that they provided better signals than other available tube materials. Significant improvements in plasma stability and LODs were found when He was used rather than Ar for HG-APGD by Gręda et al.46 Additional system modifications, i.e. operating the gas nozzle (used to introduce carrier He) as a cathode, and forcing the cathode spot to cover most of the surface of the gas nozzle, also contributed to these improvements. LODs of 0.6, 0.14, and 1.2 μg L−1 for As, Sb, and Se, respectively, were obtained using this HG-APGD system.

A new method to correct the elemental depths in thin film analysis using GD-MS was described by Xu and Hang.47 The sputtering rates from high-purity samples, representing the main element of the substrate and the main element of the film, were used to perform weighting ratio correction for each pulse. By leveraging the advantages of microsecond pulse GD-MS, a precise depth result was achieved, allowing for accurate GD-MS depth profiling.

2.2. Indirect methods

2.2.1. Laser ablation. LA-ICP-MS is a powerful analytical technique that enables the direct analysis of solid samples. It is used to determine the elemental and isotopic composition of a wide variety of materials. Advances focus on improvements in hardware and software in addition to some novel applications and calibration methods.

Mervič et al.48 provided a review of the calibration methods used in LA-ICP-MS bioimaging applications. The advantages and drawbacks of quantification approaches in terms of analytical capabilities were critically discussed and application to biological samples described. Recent developments and future directions for the field were discussed. The calibration materials used to quantify trace elements in different biological samples such as soft and hard tissues by LA-ICP-MS were reviewed by Martinez and Baudelet.49 The authors concluded that more fundamental research is required to obtain good matrix-matched reference materials to further promote the use of LA-ICP-MS and LIBS as reference quantitative techniques for biological analysis in fields such as forensics, medicine, archaeology, and anthropology.

This need for matrix-matched standards for accurate quantification in LA-ICP-MS was addressed by Kobayashi et al.50 They developed a screen-printing technique to prepare film-like samples that include internal standards, allowing for internal standardisation without needing an element of known concentration in the sample. Rua-Ibarz et al.51 introduced a multi-signal calibration strategy that varies the laser repetition rate or beam diameter, enabling a calibration curve to be constructed using a single SRM, addressing the shortage of SRMs for quantitative LA-ICP-MS. They also explored a solution-based calibration that allowed quantitative multi-element analysis without a solid standard.

Traditional LA-ICP-MS suffers from thermal instability during ablation, causing problems in the analysis of solid samples. A study by Chen et al.52 showed that cryogenic cooling during ablation reduced signal fluctuation, improved data distribution to normal distribution and resulted in more precise elemental mapping, especially in metal coatings on conductive glass. Li et al.53 used a Peltier-cooled ablation cell and found significant reductions in melting, vapour redeposition, and heat-affected zones. They found a significant improvement in signal precision (RSDs improved from between 28 and 39% to between 11 and 24%) and signal intensity (enhanced by between 5 and 70%) for different elements when using the cryogenic setup. In considering laser parameters, Pisonero et al.54 successfully used a femtosecond laser (both IR and UV) for multi-elemental mapping of otoliths, highlighting high spatial resolution with minimal thermal effects. A study by Hola et al.55 demonstrated that the degree of laser pulse overlap significantly impacted the analytical signal and aerosol properties.

Mervič et al.56 developed a method to correct for variations in ablation rates by normalising element maps using the associated ablation volume per pixel. This volume correction means that element concentrations were no longer defined as mass per mass concentrations (in μg g−1) but by mass per volume concentrations (in μg cm−3), which can be interconverted if matrix densities are known. Accurate element concentrations in 2D LA-ICP-MS maps for a decorative glass with highly varying elemental concentrations were determined using this volume-aided calibration. van Elteren et al.57 investigated the potential benefits of adapting the ablating grid in 2D and 3D mapping. Their objectives were: to enhance the accuracy of surface sampling of element distributions; improve the control of depth-related sampling; smooth the post-ablation surface for layer-by-layer sampling; and increase image quality. To emulate the capabilities of currently unavailable LA stages, a computational approach using geometrical modelling was employed to compound square or round experimentally obtained 3D crater profiles on variable orthogonal or hexagonal ablation grids. These grids were optimised by minimising surface roughness as a function of average ablation depth, followed by simulating the post-ablation surface and related image quality. Chernonozhkin et al.58 demonstrated that LA-ICP-TOFMS can provide quantitative 2D element distribution maps for micrometeorites with high accuracy and precision, with LODs ranging from 0.1 to 10 μg g−1 for a 5 μm × 5 μm laser spot. They also reported intermediate precision down to 3% RSD.

The demand for high-resolution imaging, and high sample throughput in LA-ICP-MS has led to the development of rapid-response ablation cells with low dispersion. These cells can achieve short transient signals called single pulse responses (SPRs), whose width is in the single-digit millisecond range at 1% of the peak maximum. However, coupled with ICP-QMS, recording those short signals poses a problem due to the sequential measurement of selected m/z ratios. Podsednik et al.59 demonstrated the optimisation of SPRs for multi-element depth profiling using LA-ICP-QMS. Depth profiles on 30 different sample locations in approximately 2 minutes were obtained with a depth resolution of 55 nm. The measurement of SPRs was optimised to allow the analysis of more than one m/z ratio. ICP-TOFMS enabled rapid acquisition of the entire mass spectrum, which allowed for high-speed 2D elemental mapping when coupled with LA.

Bioimaging applications in LA-ICP-TOFMS require comprehensive elemental analysis spanning the entire mass range. The application of CRC mode for simultaneously analysing elements across the complete mass spectrum was described by Theiner et al.60 The CRC mode was observed to outperform the standard/no gas mode for LA-ICP-TOFMS measurements, particularly in quantifying endogenous elements susceptible to interferences. Through the analysis of picolitre-volume micro-droplet standards and serum reference material (deposited as micro-droplets), accurate quantification of Fe and Se was achieved, with isotope ratios closely resembling natural compositions. The CRC mode also eliminated the need for post-data processing and provided enhanced sensitivity (factors of 1.5–2) for elements in the higher mass range compared to the standard mode, without compromising the sensitivity for endogenous elements.

Mello et al.61 studied the effect of immunolabelling on endogenous and antibody-conjugated elemental concentrations during LA-ICP-MS imaging. Human muscle biopsy cryosections were washed with distilled water and air-dried. Separate samples were labelled either with Gd-conjugated anti-dystrophin primary antibodies or left untreated before imaging analysis by LA-ICP-MS. They found that the labelled samples showed a decrease in Zn and an increase in Cu compared to the untreated samples. It was hypothesised that more Zn was removed from the cytoplasm but remained in the cell membrane of the labelled samples. The increase in Cu was observed primarily in the cell membrane. Furthermore, addition of a secondary antibody and colour visualisation with a DAB chromogen altered the concentration of the metal conjugate, as did the process of coverslipping with an aqueous mounting media. The significance of these findings is that multiplexed imaging of biological samples for both endogenous metals and metals used in immunolabelling should be approached with caution to avoid misinterpretation of the results. Clearly, further studies of different tissue types and immunolabelling methods are necessary.

3. Instrumentation, fundamentals and chemometrics

3.1. Instrumentation

New instrumental techniques come on the scene infrequently. More commonly, hyphenated techniques provide information which is difficult to obtain using independent methods. Neuper et al.62 hyphenated optofluidic force induction (OF2i) with Raman spectroscopy and ICP-MS for comprehensive characterisation of a single particle. The OF2i instrument used by the authors was based around a 2D optical trap, in a cylindrically shaped microfluidic flow channel with an inner diameter of 1.3 mm. A linearly polarised laser beam was generated by a 532 nm laser with a maximum power of 2 W. Individual particles were trapped by this weakly focused laser beam by aligning the microfluidic flow antiparallel to the laser propagation direction. The position of particles depended on the hydrodynamic diameter which enabled size calibration. The scattered laser light was analysed using SP Raman spectroscopy to identify particulate species and phases prior to sample introduction into ICP-TOFMS. The OF2i microfluidic system was interfaced with ICP-TOFMS, using a single cell introduction kit, to enable elemental composition and size/mass distribution determinations. Using the OF2i-SP ICP-TOFMS system, it was possible to trap and release particles more than once, and to re-analyse the same particle prior to ICP-MS. However, large number of particles are required to generate representative models based on mass and size distributions, so OF2i-Raman was first used to identify the dominant particle species, and sizes and masses were determined using OF2i-SP ICP-TOFMS. So, using the examples of polystyrene and TiO2 particles, analysis of 12C and 48Ti enabled their identification as polystyrene and anatase respectively, and thus enabled particle mass calibrations and size modelling. In the case of polystyrene, a 5 μm standard was analysed for a proof of concept and accurate mean sizes were calibrated. Calibration using a 21 nm anatase particle standard was complicated by agglomeration and background contamination, so the trapping range was tuned to retain particle agglomerates above 100 nm.
3.1.1. Sources. A review of atmospheric pressure plasmas used for OES was published by Geng et al.63 The review (179 references) includes sections on DBDs, LE-GDs, CCPs, PDs, and MIPs. There is a useful summary table listing the advantages and disadvantages of the sources for elemental analysis. They highlight the suitability of these sources for use in portable instruments while also acknowledging their limitations with respect to interferences and optimisation of sample introduction systems.

Portability and field deployment are frequently cited as reasons for miniaturisation of analytical instrumentation. For example, Xu et al.64 developed a piezoelectric transformer-driven microplasma as an OES source for the determination of Hg, with the aim of eliminating a bulky power supply. However, a power supply and signal generator were still required, and the CVG sample introduction system also required reagents, a pump, a GLS and argon carrier gas. Many other systems have been reported but all of them require various add-on parts. Hence, the future of such devices seems to be when they are developed to meet a very specific analytical requirement. One such example of this is a novel adaption reported by He et al.65 They integrated a capillary LE-GD with OES by 3D printing. The discharge power supply was a 3.7 V lithium battery and a high-voltage pulse generator (50 kV). The discharge was designed so that the capillary LE eliminated the requirement for a sample introduction system. Between 100 to 1000 mL min−1 of Ar flow was required to maintain the discharge and act as a sheath gas. The system included an integrated immunogold-labelled, silver-amplified amplification method which was used to determine carcinoembryonic antigen at detection levels similar to that of ICP-OES, with an LDR from 1 to 500 ng mL−1 and an LOD of 0.9 ng mL−1. In a similar vein, Li et al.66 developed a compact system for the determination of urea in dried blood spots using PD-OES. Urea was converted to CO2 by catalytic CVG and detected by OES at the C 193.0 nm line with an LOD of 0.03 mM. The same group67 also developed a compact system for the determination of Li in dried blood spots, this time by using a miniature ultrasonic nebuliser for sample introduction. Blood samples (3 μL) were placed on filter papers and dried, on top of which 10 μL of 1% (v/v) formic acid and 0.05% (v/v) Triton-X was added. The papers were placed onto the vibrating steel membrane of the ultrasonic nebuliser and the subsequent aerosol was directly introduced into the μPD-OES. An LOD of 1.3 μg L−1 at the 670.8 nm line was well below the Li minimum therapeutic concentration (2800 μg L−1).

The quest for a ‘universal’ ion source continues. Boillat et al.68 utilised an argon MIP for detection of both atomic and protonated molecular ions. This type of approach has a long history, and the novelty of this latest iteration resides in the fact that a desolvated liquid aerosol was used for sample introduction and an MIP was coupled with ion trap (IT) MS. However, advances in multiple CRC/CID-MS/MS instruments and the sophistication of interpretive software means that this approach struggles to gain traction in the ‘organic’ MS field, and on the flip-side, while instruments used for elemental analysis might still only do one job, they do it extremely well. Lin et al.69 approached the question from the opposite end, by modifying an ESI source so that it could be used with CID-MS for both molecular and elemental analysis. However, the same caveats apply. Hirata et al.70 took a slightly different approach by introducing VOCs into the collision cell region of an ICP-MS/MS instrument. Benzene, toluene, ethyl acetate, methyl butyrate, ethyl butyrate, pentyl acetate, pyridine, and 2-methylfuran were detected as molecular ions (M+), protonated ions ([M + H]+), or deprotonated ions ([M − H]+). The LODs were calculated based on the ‘counting statistics of the background three times’, and were 0.3, 0.56, 75, 34, 1.0, 400, 46 and 1.7 fg for the aforementioned analytes respectively. Dimethyl diselenide fragmentation was also observed, resulting in Se+, SeH+, Se22+, Se–CH3+, Se-dimer (Se2+), and Se–Se–CH3+ fragment ions and the molecular ion (H3C–Se–Se–CH3+). Again, the major advantage of this technique was the possibility for both molecular and elemental analysis. But the aforementioned caveats apply once again.

Burhenn et al.71 developed a novel DBD for elemental detection consisting of a T-shaped quartz tube with line-shaped co-planar electrodes. The configuration of the electrodes was such that the discharge was forced close to the quartz wall, thus releasing analyte deposited in the surface. The DBD was operated using He at a flow of 500 mL min−1 and a 2.7 kV square wave voltage at 20 kHz. Temporally and spatially resolved OES measurements revealed that the plasma propagated as a bent structure, close to the wall between the electrodes. The He metastable density was determined to be ∼4.5 × 1010 cm−3.

3.1.2. Spectrometers. Li et al.72 constructed a high resolution coma-free spectrometer for measurement of Fe emission between 170 and 600 nm. They eliminated the coma effect (comatic aberration) over the entire spectral region by making the off-axis angle equal to zero, i.e. α = β = 0 (where α and β are the angles of the incidence and diffraction light paths). They also used a diffraction grating that consisted of 20 sub-gratings, thereby dividing the spectral region from 170 to 600 nm into 20 sub-spectral regions. The 20 subgratings were vertically distributed along the direction of the incidence slit and set at specific diffraction angles to form an integrated grating module. The diffraction image of the sub-spectral regions was then imaged onto the focal plane of a CMOS detector. An Fe HCL operated at 20 mA was used as a radiation source and 2451 known Fe lines were identified. A further 1100, previously unidentified, lower intensity lines were also observed.

Hu and Gundlach-Graham73 investigated the potential of a frequency-scanned digital quadrupole mass filter (QMF) as a high resolution analyser for ICP-MS. In normal operation of a QMF the RF voltage and duty cycle are fixed and the RF frequency is scanned, thus allowing different m/z to pass through the mass filter. However, in digital QMF mass filtration is realised by scanning RF and DC voltages at a known ratio with a fixed frequency. Simulations suggested that, by choosing appropriate duty cycles, the mass resolution could be increased while retaining high transmission. For example, a resolution at 10% peak width of 10[thin space (1/6-em)]420 was achieved for m/z 80, with transmission ∼2.5% for all isoenergetic ions. Most ion loss was due to the defocusing effects of the fringe field. Clearly, practical tests are required in order to prove the potential.

3.1.3. Interferences. The removal of interferences is one of the perennial battles. This can be achieved either by removing them at the sample preparation stage, by using an instrumental method, or by data treatment.

Most of the recent focus on interference removal for ICP-MS has been on the use of collision/reaction cells to either remove the interference entirely, convert it to a polyatomic ion with a different m/z or do likewise to the analyte ion. French et al.74 investigated the effect of ion kinetic energy on gas phase ion reactivity in CRC-ICP-MS/MS. Ion kinetic energy is important because it affects ion transmission through the ion lenses and mass filter, and also influences reactivity with reaction gases. Specifically, they wanted to determine the ion kinetic energy dependence of gas phase ion interactions of 48 ions with CO2, N2O, and O2, with the aim of establishing selective ion-gas reactivity at a given range of ion kinetic energies. Kinetic energies of the incident ions were varied by adjusting the octopole bias (Voct), and the effect on primary product ion production of MO+, MO2+, MN+ and MC+ was observed with different reaction gases. Practical application of the results was demonstrated to show how the ion kinetic energy could be varied to achieve an ion product distribution for interference resolution. For example, separation of the interference of 151Eu+ on 151Sm+ was possible by adjusting the KE to positive Voct where the formation of EuO+ (compared to Eu+) and Sm+ (compared to SmO+) were favoured. This separation was much less pronounced at negative Voct. The same group75 used density functional theory (DFT) calculations of reaction enthalpy to predict whether M+ would react with carbonyl sulfide in CRC-ICP-MS/MS. For the 46 elements studied the dominant product was MS+, when a reaction was observed, with some formation of MO+. The DFT prediction was accurate for 90% of ions when predicting MS+ formation, and 77% accurate for prediction of the minor and higher order products MO+ and MS2+.

Atmospheric gases entrained into the plasma and through the sampling cone can cause polyatomic ion interferences in ICP-MS. Jiang et al.76 reduced such interferences by using a shielding device which they called an atmosphere-induced interference reduction device. The device employed a quartz cap on the ICP torch and an arrangement whereby a shielding gas was introduced around the interface region between the end of the torch and the sampling cone. Computer simulations of gas flow indicated that the streamline velocity in the outlet was high enough to prevent air from entering this region. Its effectiveness in reducing interferences caused by atmospheric gases such as H, C, N and O was evaluated using Ar and He as shielding gases at 10 L min−1. Helium was most effective, resulting in reductions of 32%, 51%, 56%, 54%, 51% and 42% for 12C+, 15N+, 16O+, 16O1H1H+, 40Ar14N+ and 40Ar16O+ respectively. Whether this is sufficient to improve analytical utility is another question.

3.2. Fundamentals

3.2.1. Fundamental constants. Fundamental constants are necessary to perform diagnostic studies and for plasma modelling.

Popov et al.77 determined Stark broadening parameters for Ca I lines within the singlet transitions image file: d5ja90013a-t1.tif. They used a so called ‘long spark’ LIP created by using a cylindrical lens to focus the laser and create an elongated laser plasma. The stated advantage of this approach was to improve the signal intensity to allow short signal acquisition times, thereby reducing problems associated with line profile asymmetry. The Te and ne were varied in the ranges 3500 to 10[thin space (1/6-em)]000 K and 1 × 1016 to 16 × 1016 cm−3 respectively. Recommended ion broadening parameters, Stark widths (FWHM) and shifts in these ranges were reported.

Zengin et al.78identified Ho I and Ho II lines in the near IR region from 700 nm to 1750 nm. The spectrum of an Ho HCL was examined using an FT spectrometer. In total, 589 Ho I transitions, and 31 Ho II transitions were identified. Of these, 216 previously unreported lines were observed in the 700 to 1200 nm wavelength range, with the rest of the lines in the 1200 to 1750 nm wavelength range. In a separate paper, Barka et al.79 identified new energy levels and their hyperfine structures by studying 35 Ho I lines in the NIR range from 698 nm to 833 nm. The same group also published80 results of a hyperfine structure analysis of two Ho I lines, at 775.21 and 755.50 nm.

Kodangil et al.81 determined transition probabilities for 32 La II lines which are not available in the NIST database. They used LIBS and a Hartree–Fock procedure for the calculations and found reasonable agreement with other published work. There were some inconsistencies which they attributed to ‘branching fractions’ and ‘blending of other lines’.

Suresh et al.82 calculated electron impact excitation cross sections of Ga atoms. They used a relativistic distorted wave approximation method together with multiconfigurational Dirac–Fock wavefunctions. The cross sections for the excitation from 4s24p(2P1/2) to 4s24f, 4s4p2(4P1/2, 4P3/2, and 4P5/2), 4s25f, 4s26f, 4s27[d, f], and 4s28[s, p, d, f] and all the excitation cross sections from 4s24p(2P3/2) were reported for a range of electron energies. Oscillator strengths and transition probabilities were compared with previous calculations and found to be in good agreement. Electron impact cross sections were calculated by Agrawal et al.83 for a Kr plasma. The fully relativistic distorted wave method was used to calculate electron impact cross-sections for transitions from the ground state, four metastable states of 4p44d and a quasi-metastable state of 4p45s to the fine structure levels of 4p4n1s(7 ≤ n1 ≤ 9), 4p4n1p, 4p4n2d(6 ≤ n2 ≤ 9) and 4p4n3f(4 ≤ n3 ≤ 9) excited states. These constant were incorporated into a collisional-radiative (CR) model and used with experimental data from a pulsed Kr+ discharge at a pressure of 3.3 × 103 Pa to obtain Te (2.0 eV at 40 μs and 1.2 eV at 90 μs) and ne (1 × 1022 m−3) which agreed with experimental results.

Sobolewski et al.84 determined hyperfine constants and isotopic shifts for six Pb I lines (between 569 and 612 nm) and two Pb II lines (between 560 and 608 nm) using laser optogalvanic spectroscopy. They also performed theoretical calculations using the multiconfigurational Dirac–Hartree–Fock method and compared these with experimental results, with good agreement.

Transition probabilities for Ti II lines were calculated by Mejia et al.85 using Hartree–Fock relativistic (HFR) methods. These were used to determine Te and ne in a time-resolved LIBS Ti plasma. The Te was determined to be in the range from 11[thin space (1/6-em)]700 to 7700 K, with an uncertainty of 15%, for laser pulse energies of 48, 38, 28, and 18 mJ. The ne was estimated to be between 9 × 1016 and 18 × 1016 cm−3 for a laser energy of 48 mJ using Stark broadening of the Ti II 368.50 nm line.

3.2.2. Diagnostics.
3.2.2.1. Plasmas. Plasma modelling studies are useful in a variety of contexts. With respect to analytical atomic spectroscopy, a reliable model is useful for the prediction of excitation and ionisation conditions (ideally in the presence of matrix species), subsequent line emission intensities (in the case of OES), or polyatomic ion formation (in the case of MS).

Arellano et al.86 coupled one-dimensional particle-in-cell/Monte Carlo collision (PIC/MCC) simulations with a collisional-radiative model (CRM) to simulate emission spectra of a low pressure Ar RF CCP in the range 2 to 100 Pa. The simulations were compared with experimental results and found to agree well up to 20 Pa but deviated above this pressure. The authors noted that the calculated 1s5 increased with pressure whereas the measured value reached saturation, a discrepancy which they attributed to the neglect of the loss processes of this long-lived metastable state in the PIC/MCC simulation. Since the model included only electron-impact ionisation and excitation of ground-state Ar atoms to reduce computing time, the authors proposed that inclusion of metastable states might improve the model at higher pressures.

There have been numerous previous studies to calculate plasma temperature, predominantly using the Boltzmann plot method or a variation thereof. In contrast, Képeš et al.87 used an ANN to predict temperature in a LIBS plasma. They used post-hoc model interpretation to confirm that the ANNs learned meaningful spectroscopic features from synthetic spectra. They concluded that ANNs are capable of effective learning with regard to emission spectra, but this is largely dependent on the availability of accurate ‘ground truth’ values (temperatures) on which to train the models. Hence, the authors proposed that future models should be trained using methods based on Thomson and Rayleigh scattering rather than the Boltzmann plot, presumably to account for variations in TE conditions.

Srikar et al.88 used random forest and deep neural network (DNN) machine learning methods, in conjunction with OES and a collisional radiative (CR) model, to predict Te in an atmospheric argon plasma jet. The models were tested against a data-set retained separately to the training data-set. The DNN model yielded a prediction accuracy with an R2 value of 0.9964, and the random forest model yielded an R2 value of 0.9869.

Jovović et al.89 determined Trot in a pulsed atmospheric pressure needle-to-cylinder gas discharge source, with a graphite cathode in an Ar and Ar–H2 mixture. Values of Trot were between 2890 and 3615 K and between 4100 and 5690 for the R and P branches respectively of C2, CN and CH spectra.

Pupyshev et al.90 used a thermodynamic model to predict background ions in an ICP under ‘cold plasma’ conditions (i.e. between 2000 and 5000 K) with respect to ICP-MS. The model only applied to the analytical zone (AZ), about which several assumptions were made: LTE applied; it was homogeneous; the aerosol was completely vaporised; there was a 0.1013 MPa constant pressure; temperature may vary across the AZ; there was no mixing between the AZ and other gas flows from the torch; surrounding air did not enter; matrix effects due to concomitant elements were negligible; double ionisation was only significant for atoms with low second IE; ion extraction into the interface was representative of the AZ. A good correlation between predicted and experimental intensities of background ions was observed at 3000 K, but not very good outside this temperature. Given the numerous assumptions of the model this is hardly surprising, and it is quite likely that any perturbation of the system would cause significant disagreement. A separate study91 from the same group used a thermodynamic model to predict the behaviour of background ArH+, ArN+, ArO+ and Ar2+ ions in ICP-MS. Results were compared with experimental data obtained at different flow rates and plasma powers.

French et al.92 investigated the effect of gas impurities in NO reaction gas introduced into a collision reaction cell (CRC) used with ICP-MS. They deliberately chose a less pure (99.5%) grade of NO and monitored the effect on reactivity of 60 elements with and without a gas purifier. The gas purifier was primarily used to reduce the H2O content to below 100 ppt (V). When the purifier was used oxy-hydride interference were reduced and ion sensitivity was increased. The reduction in oxy-hydride ions favoured the production of MN+ ions, presumably (according to the authors) because of the removal of competitive reactions with H2O and reduced dispersion of M+ reactant ion kinetic energy.

In a continuation of previous work, Komin and Pelipasov93 studied the effect of increasing oxygen concentration up to 80% in a nitrogen MIP. The H011 cavity MIP was operated using an ICP torch at 1.2 kW power and with nitrogen outer, intermediate and nebuliser gas flow rates of 10, 0.5 and 0.4 L min−1. An aqueous aerosol was introduced using a concentric nebuliser and double-pass spray chamber at 1.2 mL min−1. The Texc was observed to decrease from 3.9 × 103 to 3.3 × 103 K as the oxygen content (present in all flows) was increased from 0 to 80%. The Mg II[thin space (1/6-em)]:[thin space (1/6-em)]Mg I ratio decreased from 1.02 to 0.18, and ne from 6.6 × 1012 to 4 × 1011 cm−3. Most significantly, oxygen addition increased the LODs between 1.2 to 4× for lines with Esum between 1.5 and 4 eV, and between 50 to 60× for lines with Esum between 4 and 15 eV. However, it was possible to run the plasma on atmospheric air which may be attractive in some circumstances.


3.2.2.2 Graphite furnaces. The recent fundamental developments and applications of GFAAS were reviewed by Butcher.94 Advances in the understanding of atomisation mechanisms and applications of GFAAS in speciation analysis were discussed. Instrumental developments, including the development of direct solid sampling techniques as convenient, rapid, and green methodologies were also covered. The article concluded with a discussion of future directions for GFAAS, including the development of more environmentally friendly protocols for trace elemental analysis.

3.3. Chemometrics

Single particle analysis using ICP-TOFMS generates a considerable amount of data which can be time consuming to interpret effectively. Luckily, Lockwood et al.95 have developed an open source software programme called ‘SPCal’ which should make this task easier. The software can be used with csv, txt and TOFWERK HDF5 data files and includes various functions such as: screening tools to pinpoint particulate elements in unknown samples; calibration of size; mass distributions; cluster analysis (PCA, HAC) to study groups of particles based on their composition and specific features. The code for SPCal is available at https://github.com/djdt/spcal. A previous version for SP ICP-QMS is also available (https://doi.org/10.1039/D1JA00297J).

4. Laser-based atomic spectrometry

Key fundamental studies and instrumental developments (published in 2024 and at the end of 2023) in laser-based atomic spectrometry are highlighted in this section. 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/OES, and to the use of lasers for fundamental studies of the properties of atoms or for thin film deposition are not reviewed in this section.

4.1. Laser induced breakdown spectroscopy (LIBS)

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 in the ASU applications reviews. For instance: geo-applications;96,97 space applications;98 nuclear applications;99 analyses of water samples for the detection of heavy metals;100 analysis of atmospheric particulate matter;101 spectral tandem technology;102 calibration strategies for LIBS bio-applications;49 and the application of LIBS for monitoring fuel thermal conversion processes.103

The critical aspects and emerging trends in LIBS, in relation to calibration challenges, integration of various techniques (data fusion), and application of data science for spectral analysis were reviewed by Ferreira et al.104 Duponchel et al.105 also published a practical tutorial on the statistical comparison of predictive models for quantitative LIBS analysis and classification. Progress in machine learning application in LIBS was summarised by Hao et al.,106 highlighting pre-processing data, machine learning methods and issues that researchers need to address.

It should be also remarked that in 2024 there were several international conferences dedicated to discuss the recent progress in LIBS, such as the combined XIII LIBS Conference and II Latin American Meeting on LIBS held in Iguazu (Argentina) from September 2nd to 6th (https://libs2024-ar.sciencesconf.org/); and the Nordic LIBS held in Tampere (Finland) from March 5th to 6th (https://events.tuni.fi/nordiclibs2024/).

4.1.1. Fundamental studies. A new temperature iterative self-absorption correction method based on the plasma thermal equilibrium radiation model was developed by Hou et al.107 The authors claimed that this method showed simple programming, high computational efficiency and independence from the availability and accuracy of Stark broadening coefficients. The method was successfully applied to improve the quantification of Cu in steel samples.

A detailed collisional radiative model for a laser-produced Cu-plasma was developed by Agrawal et al.,108 including solution of a detailed particle balance equation and validation of the results using spectroscopic measurements. This model considered different processes such as, electron-impact excitation and de-excitation, radiative absorption and decay, ionisation, and two- and three-body recombination. Required sets of electron-impact excitation cross-sections were calculated using the fully relativistic distorted wave theory. The authors claimed that Boltzmann plots (based on the LTE assumption) to obtain the Te of the plasma might not be very reliable so instead they suggested the use of their model.

Favre et al.109 developed a model, denominated MERLIN (MultiElemental Radiative equiLibrIum emission), which is based on resolution of the radiative transfer equation. They used the model to simulate LIBS emission spectra in LTE conditions and achieve calibration free LIBS. The model provided simulated spectra for any element mixture and was validated using the alloy Eurofer97. The authors claimed that they were working on extending the capabilities of this code to molecular spectra calculations, and on building a spectral and Stark database built in SQL (Structured Query Language) with the aim of reducing dependence on the availability of online data.

Ciniglia et al.110 investigated the effects of concomitant elements on analyte signal in an aerosol LIBS plasma. Matrix elements Mg, Cr and Na were used with Cu as the analyte of interest. The effect on the Cu I emission signals was investigated by increasing the matrix element concentrations with respect to Cu concentration. The authors demonstrated that specific physical properties of the matrix affected plasma conditions, by changing the temperature and the plasma temporal evolution, and ultimately affecting the LIBS signal.

The structure of a fs-LIP at atmospheric pressure was spatially and spectroscopically characterised by Méndez-López et al.111 The general behaviour of the plasma plume was evaluated for different target materials, showing differences between metal and dielectric samples. In particular, the plasma plume in a Cu sample was observed to be composed of a fast-displacing upper component, with higher excitation and ionisation conditions, and a slow or static component that remained near the sample surface throughout the plasma evolution. Conversely, the plasma plume in a PVC sample was only composed of a single fast-displacing component, which was not homogeneous in terms of excitation. A spatio-temporal plasma-particle characterisation from dry NP in suspension was investigated by Latty et al.112 They used different laser conditions that included: Q-switched ns-pulses; short-focus fs-pulses; and fs-filaments. The study highlighted the influence of early air breakdown behaviour and subsequent plasma evolution. Additionally, the authors performed conditional analysis of LIBS measurements to determine associated sampling probabilities.

4.1.2. Instrumentation. Abdel-Harith et al.113 described how the wavefront of an IR-laser beam was shaped with a flat top energy distribution using the transmission of the beam through a thin (100 μm) crystalline silicon wafer (placed before the focusing lens). This configuration was proven to enhance sensitivity up to 3-fold and to improve LODs for the analysis of metallic samples. The same group investigated114 the use of a crystalline quartz slide instead of the crystalline silicon wafer to achieve wavefront enhanced-LIBS. The capabilities of this experimental set-up were demonstrated using three laser wavelengths, including the IR (1064 nm), green (532 nm), and UV (355 nm) regions, for the improved analysis of bronze samples.

Álvarez-Llamas et al.115 developed a LIBS system operated at 1 kHz and 500 μJ per pulse that enabled ultra-rapid compositional mapping of geological samples at high spatial resolution (∼10 s μm). Multielement information was obtained by detecting different spectral intervals during sequential mapping analysis. The authors demonstrated the effectiveness of this approach by characterising mineralogical compositions at various scales in a shorter experimental time. Moreover, the large amount of data generated using these operating conditions (∼107 pixels) was handled using binary image masks to identify the main mineral phases of samples and to provide a rough estimation of the quantitative composition of the sample.

Batsaikhan et al.116 developed a novel AW-mLIBS system to simultaneously analyse the elemental composition and surface imaging of a sample, using optical and acoustic signals. The LIP produced was generated using a monolithic microchip laser crystal (Nd:YAG/Cr:YAG), pumped using a fiber-coupled laser diode, while the acoustic waveforms from the plasma were simultaneously recorded using a condenser microphone with a preamplifier. The authors applied this set-up to detect Gd and surface images in homogeneous and heterogeneous surrogate nuclear fuel debris. Moreover, a data fusion method integrating spectral and acoustic signals generated by LIP was developed by Chai et al.117 (LIPs). This methodology was employed to improve the measurement accuracy of carbon concentrations in post-combustion flue gas. In particular, the high sensitivity of acoustic signals (from both the time-domain waveforms and frequency spectra) to variations in gas temperature was employed to correct the effects of gas temperature variations.

It is known that there is a correlation between NE-LIBS and the structure of NP–protein conjugates in terms of protein content absorbed on the NP surface. In this context Dell'Aglio et al.118 investigated the plasma parameters (electron density and temperatures) in the presence of NP–protein conjugates. Ultrapure AuNPs, produced by pulsed LA of a liquid and human serum albumin protein were used to form AuNP–protein conjugates further deposited on a titanium target. The artificial implementation of NP labels also enabled the use of LIBS as a readout technique for immunochemical assays. Farka et al.119 developed an optimised DP-LIBS system based on the use of photon up-conversion nanoparticle (UCNP) droplets deposited on the microlitre plate wells, for the indirect detection of human serum albumin. The method was validated using luminescence of UCNPs and absorbance, and it was successfully applied to the analysis of urine samples.

The presence of different concentrations of Ag NPs was evaluated on the LIBS emission signals of Eu and Yb, present in metal-loaded polymers conjugated to antibodies, by Safi et al.120 The authors demonstrated that an optimised concentration of AgNPs could provide enhanced emission signals for both elements (by one order of magnitude); however, for Yb only its ionic emission signals were increased. It should be highlighted that these results are promising to further enhance biomarkers detection sensitivity. The time evolution of nanoparticle enhanced molecular LIBS (NEM-LIBS) of graphite samples was investigated by Soumyashree121 using 10 nm AuNPs. Higher ne and slightly higher Tvib were measured at high delays (up to 100 μs), concluding that NEM-LIBS enhancements at these delays were due to the increase of atomic species in the plasma, which were antecedents of the molecules formed.

A large-area, highly homogeneous ordered nano-array substrate with a tunable NP size and inter-particle distance was developed by Nan et al.122 to be employed for NE-LIBS applications. It was successfully applied to the analysis of liquid samples, demonstrating signal enhancements of one order of magnitude and high reproducibility for selected analytes (e.g. Pb I 405.78 nm and Cr I 425.43 nm). These results were achieved with optimised particle sizes and inter-particle distances using ordered nano-array substrates from different manufacturing batches.

Simultaneous application of Raman and LIBS in gas phase analysis was demonstrated by Kiefer,123 using a single laser and a single detector. The Raman signal was initially measured using an intensified camera after the application of a first prolonged pulse, which was achieved using two types of optical pulse stretchers: a ring cavity with a beamsplitter; and a polarisation controlled optical ring cavity. Afterwards, the intensifier was switched off during the LIBS pulse. It should be noted that an optical delay line was installed to obtain the LIBS pulse as a part of the original laser pulse. Once the unspecific plasma continuum emission signal decayed, the intensifier was switched on again and LIBS signals were collected with similar amplitudes to those from Raman. The proposed technique allowed comprehensive chemical characterisation of gas samples, providing simultaneous information about all molecular and atomic species. Moreover, the author claimed that Raman and LIBS signals were generated only about 60 ns after each other, which was the actual time scale of the measurement.

The influence of magnetic-spatial confinement was investigated in detail by Li et al.124 using a fiber-optic LIBS system. Multiple parameters were evaluated and compared including the temporal evolution of plasma morphology, emission spectra and signal uncertainties. Magnetic-related enhancements and confinement-related enhancement were found not to match up optimally. Nevertheless, their combination provided higher sensitivity, more signal stability and reduced self-absorption effects due to the higher temperatures achieved and less plasma fluctuations.

4.1.3. Novel LIBS approaches. A high correlation between simultaneously measured plasma image features and spectral line intensities of matrix and trace elements in steel samples was observed by Chen et al.125 The plasma average intensity, which demonstrated the highest correlation with the spectral line intensity, was used as a reference index in combination with the minimum distance and averaging method to eliminate invalid data. The authors claimed that this method could effectively improve the spectral stability and quantitative analysis performance of LIBS.

Chen et al.126 demonstrated that the combination of cavity confinement-enhanced LIBS with chemometrics provided a powerful analytical tool for the classification and identification of samples. The authors combined optimised cavity-constrained LIBS and used a gray wolf optimisation algorithm with an optimised bidirectional long short-term memory network for classifying and identifying cigarette samples of 10 different brands.

Using a LIBS imaging dataset obtained from the analysis of a complex rock sample Duponchel et al.127 developed a novel method that combined the use of PCA to diagnose the potential presence of a spectral interference, which might generate a biased distribution image; and the use of Multivariate Curve Resolution-Alternating Least Squares to correct it.

Feng et al.128 developed a novel LIBS method for the analysis of liquid samples based on the use of sprayed charged microdroplets generated by a paper spray. This avoided effects observed in the direct analysis of liquids such as splashing and surface ripples. The authors claimed that this was a simple and robust method, with potential in situ applications, which allowed for the direct loading of liquid samples onto paper. Nevertheless, it was observed that the performance of the method was significantly affected by parameters such as paper shape, tip angle, spray voltage or solvent composition.

Ju et al.129 evaluated the use of an electrostatic-assisted LIBS system, combined with machine-learning algorithms, to improve the quantitative determination of trace metal elements in aqueous solutions. The electrostatic field was generated using two polished parallel copper plates, placed at both ends of a liquid jet and connected to a high-voltage DC power supply.

Gaft et al.130 investigated the potential application of LAMIS for isotopic shift analysis of GdO, making used of a high resolution double echelle monochromator with a spectral resolution of 0.001 nm. Vibrational and rotational transitions A(1–0), A(1–2), and A(2–1) of GdO were detected showing differences between natural GdO, 156GdO and 158GdO emissions.

Harilal et al.131 evaluated fs-LIBS for the determination of 1H, 2H and 3H isotopes after proper resolution of their isotope shifted emission lines. The plasma was generated in an Ar atmosphere at different pressure conditions. The methodology was successfully applied to determine the isotopes during depth profiling of neutron-irradiated Zircaloy-4 samples. The same group investigated132 the use of orthogonal DP-LIBS, employing two ns Nd:YAG lasers, to enhance 2H α-emission, while minimising line broadening and self-absorption. Critical parameters such as interpulse delay, ambient gas pressure and heating laser energy were evaluated. Similarly, maximum line widths for the determination of H isotopes using LIBS was investigated by Traparic et al.,133 using heavy water-doped graphite/silica gel targets in both Ar and He atmospheres. In their study laser pulse energy, gas pressure, delay and gate times were optimised to achieve fully resolved lines.

A novel algorithm for semi-automatic identification of elemental lines in LIBS spectra was developed by Gajarska et al.134 The algorithm was based on the use of spectral fingerprints of individual elements, which were configured into comb-like filters and correlated with measured spectra. Spectral variations were accommodated by adjusting the micro-parameters of the comb filter. The algorithm only required two parameters: the spectral window size, which regulated the fit of the decision threshold; and the upper percentile from baseline deviation, which regulated the sensitivity of the detection. It was validated using spectra of two SRMs.

A new method to obtain precise information about the temporal evolution of emission line intensities in a LIBS plasma was developed by Lellouche et al.135 The authors claimed that the method was inspired by the Bredice 3D-Boltzmann plot formalism, and employed a series of time-integrated measurements obtained at different delay times after the laser pulse. The method was successfully validated for the quantitative analysis of alloy CRMs.

LIBS spectra collected on >2500 unique geological targets, using three different instruments, and a range of collection protocols was processed by Lepore et al.136 to evaluate their potential in training multivariate models for prediction of rock and mineral compositions. The authors demonstrated that reasonable predictions could be achieved by binning peak areas prior to training the calibration models and using calibration transfer algorithms to reduce element prediction uncertainties when the resolution of training and test spectra did not match. The study concluded that large-scale LIBS calibration databases could be used to boost LIBS accuracy, employing matched plasma conditions and post-processing.

Lu et al.137 measured LIBS spectra at 60 MPa of He gas pressure by simulating the method of draining the water from the sample surface in deep ocean LIBS applications. The authors observed a nonlinear decrease in the spectral intensities and gradual increase in the spectral broadening during gas pressure increase. The results concluded that LIBS signal intensities in a high-pressure gas environment could be improved by increasing the laser energy and shortening the laser transmission distance.

A micro-hole array sprayer (MHAS) device was developed by Pan et al.138 to rapidly generate uniformly distributed droplets from a sample solution onto an Al substrate. These droplets dried quickly and separately into particles with micrometer diameters. The method was found to provide high sensitivity and rapid quantitative capabilities, for the determination of REEs in aqueous solutions. The LODs for La, Nd and Pr ranged between 6 and 22.3 mg L−1, meeting the sensitivity requirements for REE extraction and recovery processes.

4.2. Laser induced fluorescence (LIF)

A hybrid LIBS system assisted with laser-induced fluorescence (LIBS-LIF), which employed a gated PMT as detector, was developed by Ouyang et al.139 and applied to the determination of Pb concentration in aerosols. This hybrid system improved sensitivity by >3 orders of magnitude and reduced signal fluctuations by up to 90%. An LOD of 0.0035 μg m−3 for Pb in aerosols was achieved. The authors claimed that this innovative approach provided significant improvements for rapid and in situ pollution monitoring.

4.3. Laser atomic absorption spectroscopy (LAAS)

A self-designed open multi-pass cell structure was developed by Qi et al.140 to obtain Rb isotope absorption spectra, based on tunable diode laser absorption spectroscopy (TDLAS), with high sensitivity and high signal-to-noise ratio. Rubidium atoms were obtained via a redox reaction, and the reaction device was designed with a micro-channel array structure to provide a highly collimated Rb atomic vapour beam. The optical path of the open cell had a compact structure that reduced the complexity of system adjustments and allowed for a complete isotopic analysis in 4 minutes. The authors highlighted that TDLAS was expected to become an effective complement to Rb isotope ratio detection due to its advantages of low cost, fast measurement and in situ detection.

5. Isotope analysis

5.1. Reviews

A comprehensive critical review of IDA of radionuclides was produced by Quemet et al.141 It focused on mass spectrometric determinations involving the elements Np, Pu, Am, Th, U, Pa, I, Cs and Tc, provided an excellent review of the major fundamental aspects of IDA, and was applicable to any elemental or isotopic system. Many of the fundamental equations used in IDA were provided in the article, including a rigorous documentation of error propagation determination with graphic examples for a variety of elements. Subjects covered included single and double IDA methodology, LOD estimation, weighing calculation strategy and sample–spike mixture optimisation. These analytical fundamentals were followed by coverage of applications including nuclear forensics, safeguards, irradiated samples and biological systems. Overall, the review compiled a knowledge base from a group that has been at the forefront of radionuclide measurement and standardisation for several decades. It provided a go-to summary for any analytical scientist either embarking on quantitative isotopic measurements or for experienced analysts looking to improve the quality of their procedures and laboratory practices.

Wu142 produced a comprehensive review of V isotope geochemistry and isotopic analysis. This started with some fundamental aspects of V geochemistry and cosmochemistry, including the effects of redox conditions, fractionation between the Earth's mantle, crust, sediments and oceans. Isotopic variation of δ51V amongst these systems was usefully summarised in a comparative diagram. This depicted the more +ve δ51V values found in sediments, seawater and fractionated igneous rocks, relative to mantle-derived magmas (∼0) and lunar and meteoritic material (−ve). Chemical separation procedures were reviewed, along with the requirements to remove matrix elements which influence isotope measurement. MC-ICP-MS measurement of V isotopes was summarised, which highlighted isobaric interference from Ti, Cr and potential correction protocols. Future developments of V were highlighted as the incoming development of CC-MC-ICP-MS along with in situ sampling for V isotopes by LA-MC-ICP-MS.

5.2. Radiogenic isotope ratio analysis

Buzenchi et al.143 investigated the use of 1013 Ω amplifiers to increase precision of Sr isotope measurement by laser ablation. Their study used an MC-ICP-MS in conjunction with an excimer laser to sample apatite crystals and determine 87Sr/86Sr. A limitation of 1013 Ω amplifiers is their inability to cope with signals >5 nA, so to control the 88Sr ion beam the amount of ablated material was constrained by masking to control the size of the ablation pit. Signal instability can induce imprecision and bias in multi-collector results but in this study this was negated by τ (tau) correction for each amplifier. Results indicated that external reproducibility of 87Sr/86Sr was better than ±2‰ for the Durango reference apatite, which is around 4 times better than analyses using 1011 Ω-equipped amplifiers.

Schneider and Kleine144 examined Faraday cup deterioration with usage through several years. As with many other studies of Faraday degradation, it was found that the cups, in particular the ‘centre’ or axial detector, showed signs of drift and scatter in 87Sr/86Sr mass fractionation corrected data. This was found to only affect static data, i.e. where the mass spectrum is fixed relative to the detector array. Multi-dynamic measurements, as previous studies have identified, were unaffected due to the cancellation of individual cup efficiencies during the data reduction. Interestingly, Cr isotopes measured on the same instruments as the Sr, showed a similar degradation to mass fractionation corrected ratios on similarly aged collectors. Faraday degradation was observed in static measurements a few months after installation.

Télouk and Balter145 used CRC-MC-ICP-MS coupled with LA for radiogenic Sr isotope measurements. If N2O was used in the reaction cell, it generated 14N and 18O isobaric interferences on Sr16O. Instead, SF6 was used as a reactant, and was found to produce fewer interferences because F is mono-isotopic. In solution mode, 87SrF/86SrF in liquid samples without prior separation produced equivalent analyses to those with prior Sr isolation. In LA analysis, when SF6 was used it produced a slight improvement despite a significant reduction in ion transmission.

For small sample sizes (3 mg) Xu et al.146 developed a single column protocol to separate Sr, Nd and Sm. They used a TOGDA-normal resin (DN-B100-S, 50–100 μm, supplied by TrisKem International) eluting with 3 M HNO3 to remove the matrix, 4 M to elute Sr, then 2.5 M HCl to rinse Ce, La and Pr, followed by 1.2 M HCL to progressively elute Nd and Sm. Some issues were found in samples with extremely low concentrations of these elements, as increasing the sample size to ≫3 mg was found to overload the resin column, causing a loss of Sr with the matrix elution. In this case a second column of Sr Spec was needed. Sufficient precision for Nd and Sr was achieved using 1012 and 1013 equipped amplifiers, using a TIMS instrument.

Yu et al.147 investigated the potential effects of non-exponential as opposed to the conventionally assumed exponential law mass fractionation during MC-ICP-MS analyses. They chose Nd isotopes as their tool to pinpoint non-exponential effects because it is a well-characterised system with a wide range of isotope ratios. A key aspect of their study was using the normalised argon index (NAI) developed by Fietzke and Frische (2016) (https://doi.org/10.1039/C5JA00253B). The NAI (2 38Ar/40Ar2+) represents an independent index of plasma temperature and was compared by to different exponentially corrected Nd mass ratios. It was found that some Nd isotope ratios were significantly correlated with changes in NAI and approached expected exponential corrected values with higher NAI (which? Equated to higher temperatures). Two solutions to this issue were highlighted. Firstly, increasing the NAI such that the non-exponential component is negated, or secondly a secondary regression correction between two exponentially corrected isotope pairs such as 143Nd/144Nd and 142Nd/144Nd.

A laser RIMS instrument was developed and applied to determine Pu isotope ratios by Zhang et al.148 A three-color-three-photon resonance ionisation, using photons of different wavelengths (586.654 nm, 606.181 nm and 642.722 nm, respectively) was employed, allowing a very selective ionisation of 238Pu. Moreover, the authors deposited a graphene oxide solution on the Pu sample surface to achieve up to 3 orders of magnitude more effective electrothermal atomisation of Pu. The developed method provided accurate and precise 238Pu/239Pu isotopic ratios, avoiding isobaric interferences from 238U.

5.3. Geological studies

Cruz-Uribe et al.149 evaluated the potential for single-spot Rb–Sr isochron dating of biotite using LA-ICP-MS. The technique involved statically measuring the four Sr isotopes alongside 87Rb to generate an 87Sr/86Sr vs.87Rb/86Sr isochron. A key difference of this analysis was that each spot was ablated for ∼40 s and then time-slices within the ablation were dated. It was found that different spots from the same crystal had different isochron ages, and indeed some spots were found to have different ages within their ablation depth profile. The host rock for this biotite is a schist with an estimated metamorphic age of ∼370 Ma, but which has undergone further heating and fluid-alteration events in the following 100 Ma. These events were identified within and between the ablation spots of a single crystal, implying that typical isochrons that integrate a number of spots simply give an overall average age of the metamorphic events that the rock has experienced.

Li et al.150 examined the potential of an ultrasound nebulisation VG-DBD system as an introduction system for MC-ICP-MS to analyse Nd isotopes. The study centred on determining Nd isotopes in seawater samples following Nd isolation. It was found that the introduction system could reduce the amount of seawater required to <1 L. Analysis of ∼1 μg per L Nd produced a precision of ±0.000033 on 143Nd/144Nd.

An alternative LA-ICP-MS method of ablating zircons for U–Pb age determination was used by Lu et al.151 Their technique involved a reverse depth profiling of the zircon, which meant instead of ablating down through the rim of the zircon towards the core, the zircon was cut such that the ablation started in the core and progressed through to the rim. This method was found to provide more effective data from an equivalent number of grains and better spatial resolution compared to the traditional ablation depth profiling. Results indicated that core-rim ages in Himalayan zircons could be clearly and statistically distinguished.

U–Pb dating of ilmenite was re-appraised by Tang et al.152 who used an LA-ICP-SFMS for their analyses. It has previously been found that using rutile (TiO2) as a calibration reference material is not satisfactory because of matrix differences between the minerals. This study tested rutile, garnet and reference zircon 91[thin space (1/6-em)]500 as potential calibrants for PL-57 ilmenite. Results showed that the zircon had both similar U–Pb fractionation during ablation and average normalised Pb/U to the ilmenite. Hence, it was proposed that by reducing matrix differences by optimising ablation settings, zircon 91[thin space (1/6-em)]500 could be deployed as a reliable reference for U–Pb dating of ilmenite. Zirakparvar et al.153 used LA-MC-ICP-MS to measure U–Pb in zircons. Commonly, an electron multiplier or similar detector is used in conjunction with Faraday cups to detect large and small abundance ion beams. In this study, only a Faraday cup array was used, including 1013 Ω equipped amplifiers. It was found that these resistors proved to have sufficient resolving power to provide meaningful ages for zircons as young as ∼55 Ma, or down to ∼28 Ma when 206Pb/238Pb was considered. The key advantage of this technique was that it avoided the need to complete a complex calibration of the electron multiplier detector relative to the Faraday array.

Wang et al.154 tackled the U–Th dating of gypsum (CaSO4·2H2O) which, because of difficulties in fully dissolving this sulfate mineral, is a challenging technique. Their dissolution method used (NH4)2CO3, which exchanged sulfate for more readily dissolved carbonate. Following digestion and chromatographic separation 230Th and δ234U were measured on gypsum standards using MC-ICP-MS. These data provided high precision ages for six gypsum reference materials, which demonstrated the validity of this method for samples <400 ka.

5.4. Stable isotope ratio studies

Rodriguez-Diaz et al.155 used MC-ICP-MS to measure B isotopes. They analysed nanogram quantities of B following a micro-distillation isolation procedure. The study was focused on marine biogenic carbonates with the aim was to provide an analytical procedure for B isotope determination on coral and foraminiferal material with <50 ng B available for measurement. 10B and 11B were measured using 1011 Ω and 1012 Ω -equipped amplifiers, respectively. Results for biogenic carbonate standards demonstrated that long-term reproducibility for δ11B was ±0.23‰ (2SD).

Coenen et al.156 provided an assessment of interferences and biases during solution and LA MC-ICP-MS measurement of B isotopes. They examined the effects of Ca- and Mg-based solutions on the baseline around m/z 10. Mg-solutions had little influence on baseline whereas it was significantly and proportionally elevated with increased Ca concentrations. It was also noted that calibration of instrumental mass bias was significantly affected by differences in mass loading of the plasma ion source. This was generated by variations in the amount of sample ablated which, in turn, was related to differences in the nature of the material. Hence, if standards and unknowns had different properties, then plasma loading affected sample-standard bias corrections: making the need for plasma loading to be routinely assessed during B isotope determination.

High-precision K isotope measurements by ICP-MS are generally impeded by interferences caused by proximity to 40Ar. Recent developments in CRC interfaces for MC-ICP-MS led Albalat et al.157 to investigate how K isotope measurements could be enhanced by such a system. Their study utilised a continuous-flow autosampler with a desolvator coupled to MC-ICP-MS/MS. They found that sample-standard mismatches for matrix elements, e.g. Ca and Mg, produced offsets in isotope ratios but acid molarity did not. Differences in concentration between the sample and standard were adjusted using injection flow rate and this resulted in a long-term reproducibility of 0.07‰ (2SD, n = 500).

Gerritzen et al.158 examined the potential for high-throughput simultaneous measurement of δ88Sr and 87Sr/86Sr by MC-ICP-MS. Their study used Zr spiking with a matched Sr/Zr ratio? Between samples and standards to provide a combined mass bias correction using both 92Zr/90Zr and 91Zr/90Zr. The key advantage of this system was the high sample throughput of around 30 samples per day, around a factor of five more than methods where the stable and radiogenic isotopes are measured separately.

Silver isotopes were measured to high-precision in silicate rocks by Fang et al.159 who used MC-ICP-MS in solution mode. Their procedure involved a four-stage chromatographic purification of Ag, followed by addition of Pd for synchronous mass bias correction. Repeated measurements of δ109Ag over a three-year period resulted in an external precision better than 0.05‰ (2SD). Analysis of USGS rock standards demonstrated differences in δ109Ag of up to 0.54, with continental and granitic rocks having ratios lower than most basaltic rocks and bulk silicate earth.

Gu et al.160 defined a method to measure Ca isotopes to high precision using MC-ICP-MS without using a collision cell. Isobaric 40Ar interference on 40Ca is a significant issue in an Ar plasma, but this study minimised this issue by using ‘cold plasma’ conditions. This involved a step-wise reduction in RF power from a normal (hot) 1200 W to around 600 W, which had the effect of reducing the 40Ar signal from >1000 V to <300 mV. Overall, the long term precision was found to be similar for the δ44/40Ca and δ44/42Ca at ±0.1‰ (2SD), which was predicted to open up high-precision Ca isotope measurements to a wider range of applications.

Using double spike methods and MC-ICP-MS Kashiwabara et al.161 measured W and Mo isotopes. Both elements were separated from a single sample aliquot via a two-stage anion-exchange procedure, critically removing the Ta and Hf interferences. Recovery of both elements was found to be quantitative, and the precision of δ186W was ±0.02‰ and δ98Mo was ±0.03‰. Twenty seven geological reference materials were analysed in the study for both isotope systems and results compared favourably with previously published values. The data was presented as a series of specific charts for each element showing the difference between each standard. However, an omission from the paper is a plot showing the relationship between the two isotope systems for the rock standards. Although not the specific point of the paper analytically, it would have been good, given the brief discussion in the paper of geological processes, to highlight the general covariation of these isotope systems.

King-Doonan et al.162 investigated the effects of concentration mismatch between sample and bracketing standard during Fe isotope measurement. Their study used three MC-ICP-MS instruments of different designs to rigorously evaluate δ56Fe deviations in relation to the Fesample/Festandard concentration ratio. Results indicate that a maximum of ±10% difference in the concentration between sample and standard is required to minimise the self-induced matrix effect mass bias and keep effects on δ56Fe to <0.05‰.

et al.163 ablated iron oxide minerals and examined the effects of Cr isobaric interference when measuring Fe isotope compositions by LA-MC-ICP-MS. Correction for 54Cr on 54Fe is known to break down when 54Cr/54Fe > 0.0001 which then significantly affects δ56Fe accuracy. This is thought to be due to inequivalence of the mass bias between Cr and Fe and potential matrix mismatches between Cr standard and samples. Their study used a pure Cr metal as a reference point for 54Cr interference correction and was measured before and after the sample ablation. Experiments were conducted on synthetic materials where Cr metal and hematite (Fe2O3) were simultaneously ablated. The results indicated that δ56Fe was consistent with 54Cr/54Fe up to 0.5, which suggested that LA-ICP-MS Fe isotope studies can be extended to Cr-rich iron oxides using this methodology. Xu et al.164 measured Fe isotopes by LA-MC-ICP-MS without using directly matrix-matched standards. Attempts were made to eliminate bias known to be caused by sample position within the ablation cell by use of a double-volume ablation cell and addition of water vapour during the measurement. A solution containing Ni was admitted after the ablation cell to both suppress matrix effects by adding a large water matrix, and used the 60Ni/58Ni ratio to correct instrumental mass bias. Isobaric interference of 54Cr on 54Fe was corrected by using the 53Cr signal intensity, the 54Cr/53Cr ratio and the exponential law, assuming βCr = βFe. Glass standard BCR-2G was used a reference glass. A number of mineral and rock glass standards were measured for δ56Fe and δ57Fe by both ablation and in solution-mode. Comparison of the accuracy of the results between in situ and solution analyses was very favourable. In terms of precision, the measurement (around ±0.2‰ for δ56Fe) was 4 to 8 times less precise than when solutions were used. Imprecision in some measurements was ascribed to heterogeneity within the minerals, in particular YS18 garnet and YS-31 biotite.

Europium has two naturally occurring isotopes, and their ratio can be measured by MC-ICP-MS. Lee165 investigated natural variations in Eu isotope ratio, expressed as δ153/151Eu, by measuring a series of international rock standards. A range of Sm isotope pairs were used to correct for instrumental mass bias, but these can be affected by Gd and BaO+ isobaric effects. These interferences were monitored to determine if an Eu isotopic variation was detectable in natural materials. The study found that complete purification of Eu was required to determine δ153/151Eu accurately and concluded that significant Eu isotope fractionation was present in Si-rich igneous rocks and feldspars.

Previous studies of Cr isotopes by TIMS have identified a residual isotopic fractionation not accounted for following standard mass fractionation correction procedures. Yokoyama et al.166 investigated this issue and found that this fractionation could be reduced by using a W instead of a Re filament. The lower temperature of the ionising filament significantly suppressed the formation of CrO+ which had the effect of improving precision of Cr isotope ratios by two to three times.

Improving Cd isotope measurement by TIMS was the aim of a study by Li et al.167 They used a new filament loading technique deploying molybdenum silicide (MoSi2) as an ionisation-enhancing emitter. The increase in ionisation efficiency enabled between 3 and 20 ng of Cd to be isotopically characterised using double spike deconvolution with an external precision of <±0.05‰ (2SD) for δ114/110Cd. Li et al.168 described a method which also aimed to improve Cd isotope analysis, but using MC-ICP-MS. The study used a 106Cd–111Cd double spike for mass bias correction and used δ113/110Cd rather than δ114/110Cd to demarcate the isotopic differences. The advantage of using these Cd isotopes was to avoid the isobaric interference of 114Sn. The equivalent value of δ114/110Cd was calculated as 1.33 × δ113/110Cd.

López-Urzúa et al.169 examined the influence of matrix effects on Si isotope measurements by MC-ICP-MS. They investigated issues related to chloride, sulfate and dissolved organic carbon present in the analyte. A standard Si isotope solution, NIST SRM 8546, was doped with variable amounts of sulfate and chloride to elucidate the effects on δ30Si.

Luu et al.170 used CRC-MC-ICP-MS to test the removal of sample matrix-based 40Ar23Na+ interference on Cu isotopes. Their study used a He–H2 gas mix in the reaction cell which efficiently removed the Na-based polyatomic isobaric interference. The method was found to be effective for Na[thin space (1/6-em)]:[thin space (1/6-em)]Cu concentration ratios of up to ∼65 when analysing 100 ppb Cu solutions. A key use of this methodology is in biological samples which can have high Na[thin space (1/6-em)]:[thin space (1/6-em)]Cu even after processing. Tests were performed on a tuna fish standard IRMM ERM-CE 464 and a human serum standard and produced results consistent with previously reported data.

Wang et al.171 investigated the effects of double spike proportions in analyses of Zn isotopes using MC-ICP-MS. Previous studies of isotope ratios of elements such as Nd, Ba and Mo have found a component of instrumental isotopic fractionation that varied independently of mass. The study, which measured variable double spike and sample proportions, resulted in correlations between δ66Zn and these mixtures. After eliminating the possibility of isobaric interference, the authors concluded that non-systematic variations in the exponential mass fractionation factor β across the spread of Zn isotopes was caused by mass independent fractionation. A correction protocol for this effect using the β factors measured in a bracketed standard was found to eliminate the effects on δ66Zn. This study highlighted the potential for double spike analyses of stable isotopes to be affected by mass independent fractionation: an effect that is negligible in sample-standard bracketing methodology.

Qu et al.172 devised a new chemical purification method to separate Sn from potentially interfering matrix elements during isotopic analysis by MC-ICP-MS. Their procedure used a two-stage chromatographic procedure; an anion exchange column of? AG1-X8 with? HNO3 elution to isolate Sn from the major and some trace elements, followed by a cation exchange column AG50W-X12 which sequentially used HNO3 and HF to isolate Sn from Mo. This method was demonstrated to produce precise isotopic values for both odd Sn isotope ratios, δ117Sn and δ119Sn, and the even ratios, δ120Sn and δ122Sn. Results for geological reference materials demonstrated significant differences between a sedimentary standard with 0.04‰ δ120Sn compared with a granite standard with 0.23‰. Song et al.173 used MC-ICP-MS to produce high-precision Sn isotopic measurements with a double spike. After selection of 117Sn and 122Sn for the double spike, the sample–spike mixture system was evaluated by a Monte Carlo simulation of the error propagation. Results of intermediate precision (over 6 months) for the NIST SMR 3161a were δ124/116Sn values ±0.08 (2SD, n = 52). Measurements of geological reference materials produced similar results using two separate instruments. Notably, large differences between the Sn isotope values of the standards were recorded. The basaltic standard materials had the highest δ124/116Sn ∼+0.7, the granites ∼+0.45, soils ∼−0.1 and the sulfide standard ∼−0.8.

Tupys et al.174 examined a new correction procedure for instrumental mass fractionation that used pairs of mono isotopic elements 93Nb/89Y, 165Ho/159Tb and 209Bi/197Au, model isotopic systems were studied (e.g.87Sr/86Sr–93Nb/89Y and 208Pb/207Pb–209Bi/197Au) with the mono isotopic pairs produced by gravimetric mixtures of the elements. Fractionation corrected results from MC-ICP-MS analysis indicated that significant and variable offsets were found relative to recommended values for the analyte. The conclusion was that the two mono isotopic elements had sufficient differences in their ionisation energies and other physical properties that generated offsets from their true isotopic ratios. Despite the lack of fractionation coordination between the mono isotopic pair and the target element, it was proposed that it may be possible in the future to develop a mathematical model that accounts for the fractionation differences between mono isotopic pairs, just as Russell's law constrains the internal fractionation correction.

Mass fractionation correction of Tl isotopes was investigated by Wang et al.175 They compared different element systems, W, Ir, Hg and Pb, to generate internal corrections for mass fractionation using suitable isotope pairs. An issue found in using a Pb standard to dope the sample solutions was residual Pb within the sample after separation. This, or blank contamination, was found to influence the normalising 208Pb/206Pb ratio and hence influence the Tl isotope measurement. Of the other elements, 186W/184W produced a strong linear correlation with β Tl and was found to give the most accurate results for standards and a series of geological reference materials.

Zhang et al.176 produced a comprehensive examination of Rb isotope ratios in geological reference materials. This study detailed a purification using Sr-Spec resin but switched to a two-column procedure with AGMP-50 resin for samples with low Rb content, such as seawater. Instrumental mass bias of MC-ICP-MS was corrected using sample-standard bracketing and Zr internal standard normalisation, which resulted in an overall intermediate measurement precision for δ87Rb of ±0.06‰ (2SD). The igneous rock standards produced interesting results in that it was observed that those that have experienced the most crystallisation, such as andesites and granites, had generally lighter δ87Rb (∼−0.2‰) than the more primitive basaltic rocks (∼−0.1‰). This was suggested as being related to the preferential uptake of 87Rb in the crystalline phase. Seawater had the heaviest Rb at 0.14‰. It was observed that sediments, soils and products of low-temperature processes generally had a more variable δ87Rb than high-temperature igneous systems.

5.5. Nuclear forensics

Determination of 238Pu/239Pu by low-resolution MS is hampered by the isobaric interference of 238U on 238Pu. Goswami et al.177 developed a TIMS method to measure 238Pu/239Pu following Pu isolation using a multi-stage extraction with a supported liquid membrane (DOHA-RTL@PP). Purification resulted in U/Pu between 0 and 1000. Plutonium isotope measurement was made on Re filaments. A key part of the TIMS methodology was to experiment with the vaporisation current (i.e. filament temperature) and observe 238Pu/239Pu. At lower currents U was preferentially ionised relative to Pu, and essentially holding the temperature would ‘burn off’ the U and hence reduce the 238U isobar. Beyond these filament currents the Pu signal decreased as the plutonium load became depleted. External precision on 238Pu/239Pu for the NIST-SRM-947 was ±0.8% (2SD).

Jaegler and Gourgiotis178 attempted to reduce the measurable 236U/238U using ICP-MS/MS. To minimise the 235UH+ interference on 236U, U was measured as UO2+, formed by adding O2 in the reaction cell; hence 236U, as UO2+, was measured at m/z 268. The result was a UH+/U+ of 7.2 × 10−10 which meant a 235UH+:236U+ ∼1. From this it was calculated that a new low measurable 236U/238U of 5 × 10−12 was feasible, provided that purification of natural samples with high U/Pu could be achieved.

Reinhard et al.179 used the ATONA amplifier system and examined it's efficacy in measuring U isotopes. The ATONA amplifier (atto-to-nano amp) for TIMS uses capacitance-based amplification instead of the high-ohmage resistor used on traditional Faraday detector systems. Advantages include a capacity to cope with a large dynamic range, short response time and very stable long-term response requiring infrequent gain calibration. The study made a direct comparison by measuring 234U/238U using: resistor amplifiers; a combination of resistor and secondary electron multiplier detectors; ATONA amplifiers. Results indicated that the ATONA-based system produced a precision of ∼0.15% (2SD) compared to the 1.5% achieved by the combined resistor-multiplier detection. A highlighted benefit of the ATONA system was found to be the short response time of the detection which significantly reduced the length of measurement and hence the load of U required for precise analysis.

Characterisation of uranium ore samples is an important facet of nuclear forensic investigations. Wang et al.180 used MC-ICP-MS in both solution and LA mode and characterised the U and Th isotopic compositions of a number of uranium ores. The IRMM3636 233U–236U double spike was used to correct for mass fractionation of the ore solutions and to confirm measurements by LA. The results indicated that a three-way isotopic discrimination of ore samples was possible using the combination of 234U/238U, 235U/238U and 230Th/234U.

6. Abbreviations

AASatomic absorption spectrometry
ACalternating current
AESatomic emission spectrometry
AFSatomic fluorescence spectrometry
ANNartificial neural network
APGDatmospheric glow discharge
AZanalytical zone
BCRbureau of community reference
CCcollision cell
CCDcharge coupled detector
CCPcapacitively coupled plasma
CEcapillary electrophoresis
CIDcharge injection detector
CMOScomplementary metal oxide semiconductor
Cr:YAGchromium doped: yttrium aluminium garnet
CRCcollision reaction cell
CRMcertified reference material
CVGchemical vapour generation
DBDdielectric barrier discharge
DCdirect current
DESdeep eutectic solvent
DLPdigital light processing
DMFdynamic mass flow
DNNdeep neural network
DRCdynamic reaction cell
EDXRFenergy dispersive X-ray fluorescence
EIelectron ionisation
ESIelectrospray ionisation
EPRelectron paramagnetic resonance
FTFourier transform
FWHMfull width at half maximum
GCgas chromatography
GDglow discharge
GFgraphite furnace
GLSgas liquid separator
HCLhollow cathode lamp
HGhydride generation
HPLChigh performance liquid chromatography
ICPinductively coupled plasma
IDisotope dilution
IDAisotope dilution analysis
IDMSisotope dilution mass spectrometry
IEionisation energy
IRinfra-red
IRMMInstitute for reference materials and measurements
ITion trap
KEkinetic energy
LAlaser ablation
LAASlaser atomic absorption spectroscopy
LAMISlaser ablation molecular isotopic spectrometry
LDRlinear dynamic range
LEliquid electrode
LIBSlaser induced breakdown spectroscopy
LIFlaser induced fluorescence
LIMSlaser ionisation mass spectrometry
LIPSlaser induced plasma spectrometry
LODlimit of detection
LTElocal thermodynamic equilibrium
MBmagnetic bead
MCmulti-collector
MIPmicrowave induced plasma
MSmass spectrometry
Nd:YAGneodymium doped: yttrium aluminium garnet
NEnanoparticle enhanced
n e electron number density
NEMnanoparticle enhanced molecular
NIRnear infra-red
NISTNational Institute of Standards and Technology
NPnanoparticle
OESoptical emission spectroscopy
PCAprincipal components analysis
PCRprincipal components regression
PDpoint discharge
PMTphotomultiplier
PNpneumatic nebuliser
PVGphotochemical vapour generation
QMFquadrupole mass filter
QMSquadrupole mass spectrometry
QTAquartz tube atomiser
REErare earth element
RFradiofrequency
RIMSresonance ionisation mass spectrometry
RSDrelative standard deviation
RSFrelative sensitivity factor
SAGDsolution anode glow discharge
SCGDsolution cathode glow discharge
SDstandard deviation
SFMSsector field mass spectrometry
SNRsignal to noise ratio
SPsingle particle
SPEsolid phase extraction
SPMEsolid phase micro-extraction
SPRsingle pulse response
SRMstandard reference material
TEtransport efficiency
T e electron temperature
T exc excitation temperature
TIMSthermal ionisation mass spectrometry
T ion ionisation temperature
TISIStotal sample consumption system
TOFtime-of-flight
TOFMStime-of-flight mass spectrometry
T rot rotational temperature
T vib vibrational temperature
USGSUnited States Geological Survey
UVultra-violet
VCNvibrating capillary nebuliser
VGvapour generation
VOCvolatile organic compound
XPSX-ray photoelectron spectroscopy

7. Conflicts of interest

There are no conflicts to declare.

References

  1. E. H. Evans, J. Pisonero, C. M. M. Smith and R. N. Taylor, J. Anal. At. Spectrom., 2024, 39(5), 1188–1211 Search PubMed.
  2. M. Patriarca, N. Barlow, A. Cross, S. Hill, A. Robson and J. Tyson, J. Anal. At. Spectrom., 2024, 39(3), 624–698 CAS.
  3. R. Clough, C. F. Harrington, S. J. Hill, Y. Madrid and J. F. Tyson, J. Anal. At. Spectrom., 2024, 1629–1664 Search PubMed.
  4. C. Vanhoof, J. R. Bacon, U. E. A. Fittschen and L. Vincze, J. Anal. At. Spectrom., 2024, 2152–2164 Search PubMed.
  5. E. Bolea-Fernandez, R. Clough, A. Fisher, B. Gibson and B. Russell, J. Anal. At. Spectrom., 2024, 39(11), 2617–2693 Search PubMed.
  6. 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., 2024, 39(1), 11–65 Search PubMed.
  7. M. B. Botella, M. N. Oviedo, P. Pacheco and R. G. Wuilloud, Spectrochim. Acta, Part B, 2024, 211, 106832 Search PubMed.
  8. Y. H. Cheng and C. K. Su, Microchim. Acta, 2024, 191(10), 598 CrossRef CAS.
  9. J. Hu, X. Yan and X. Chris Le, Anal. Bioanal. Chem., 2024, 416(11), 2625–2640 Search PubMed.
  10. X. M. Song, Y. Wang, X. H. Zeng, K. Huang and P. P. Chen, Sens. Actuators, B, 2025, 423, 136772 CrossRef CAS.
  11. Y. Wei, Y. Hu, C. Zhang, R. Liu and Y. Lv, Anal. Chem., 2024, 11566–11571 CrossRef CAS.
  12. Y. Wang, L. Shao, Z. Zhao, C. Huang, Y. Jiao, D. Sun, R. Liu, D. Jiang and X. Gao, Anal. Chim. Acta, 2024, 1294, 342272 CrossRef CAS.
  13. D. Torregrosa, M. Jauset-Rubio, R. Serrano, M. Svobodova, G. Grindlay, C. K. O'Sullivan and J. Mora, Anal. Chim. Acta, 2023, 1252, 341042 CrossRef CAS.
  14. X. T. Li, X. Li, G. Y. Wang, C. Wang and L. L. Huang, At. Spectrosc., 2024, 45(4), 309–315 Search PubMed.
  15. G. Sun and S. Zhang, At. Spectrosc., 2023, 44(06), 392–400 Search PubMed.
  16. T. L. Taylor, D. Tukhmetova, T. P. T. Duong, A. M. Bowe, B. Meermann and A. Gundlach-Graham, Anal. Bioanal. Chem., 2024, 416(7), 1613–1621 CrossRef CAS PubMed.
  17. C. Davison, D. Beste, M. Bailey and M. Felipe-Sotelo, Anal. Bioanal. Chem., 2023, 415(28), 6931–6950 CrossRef CAS PubMed.
  18. J. H. Dong, Z. L. Zhu, L. J. Li, P. J. Xing, S. Y. Li, L. Ouyang, X. Liu, W. Guo, H. T. Zheng and R. Qian, J. Anal. At. Spectrom., 2024, 39(11), 2791–2798 RSC.
  19. G. Kajner, A. Belteki, M. Cseh, Z. Geretovszky, T. Ajtai, L. Barna, M. A. Deli, B. Pap, G. Maroti and G. Galbacs, Nanomaterials, 2023, 13(23), 3018 CrossRef CAS PubMed.
  20. E. Bolea and F. Laborda, Spectrochim. Acta, Part B, 2024, 216, 106941 Search PubMed.
  21. K. E. Murphy, A. R. Montoro Bustos, L. L. Yu, M. E. Johnson and M. R. Winchester, Spectrochim. Acta, Part B, 2024, 212, 106841 Search PubMed.
  22. H. G. Infante, Spectrochim. Acta, Part B, 2024, 216, 106876 CrossRef CAS.
  23. K. E. Murphy, A. R. M. Bustos, L. L. Yu, M. E. Johnson and M. R. Winchester, Spectrochim. Acta, Part B, 2024, 216, 106902 CrossRef CAS.
  24. O. Geiss, I. Bianchi, G. Bucher, E. Verleysen, F. Brassinne, J. Mast, K. Loeschner, L. Givelet, F. Cubadda, F. Ferraris, A. Raggi, F. Iacoponi, R. Peters, A. Undas, A. Mueller, A.-K. Meinhardt, B. Hetzer, V. Graef, A. R. Montoro Bustos and J. Barrero-Moreno, Nanomaterials, 2022, 12(4), 725 CrossRef CAS PubMed.
  25. A. J. Goodman, S. Kanapilly, A. J. Bednar and J. F. Ranville, J. Anal. At. Spectrom., 2024, 39(2), 364–369 RSC.
  26. D. K. Oliveira, V. H. Cauduro, E. L. Moraes Flores and E. M. M. Flores, Anal. Chim. Acta, 2024, 1288, 342054 CrossRef CAS.
  27. J. Dedina, Spectrochim. Acta, Part B, 2024, 221, 107058 CrossRef CAS.
  28. B. Campanella and A. D'Ulivo, Appl. Spectrosc. Rev., 2024, 59(6), 732–779 CrossRef CAS.
  29. M. Buoso, B. Campanella, M. Onor, E. Pitzalis and A. D'Ulivo, J. Anal. At. Spectrom., 2024, 39(1), 66–70 RSC.
  30. X. Deng, L. Dong, H. Chen, W. Wang, Y. Yu and Y. Gao, Anal. Chem., 2024, 96(2), 652–660 Search PubMed.
  31. L. Dong, W. Wang, Y. Ning, X. Deng and Y. Gao, Anal. Chim. Acta, 2023, 1251, 341006 CrossRef CAS PubMed.
  32. J. Hu and X. Hou, At. Spectrosc., 2023, 44(06), 385–391 CrossRef.
  33. L. Dong, Y. Ning, J. Hu, W. Wang, Y. Yu and Y. Gao, J. Anal. At. Spectrom., 2024, 39(3), 791–799 RSC.
  34. J. Hu, Y. Yu, Z. Xiao and Y. Gao, Microchem. J., 2023, 193, 109178 CrossRef CAS.
  35. R. Y. Wang, S. S. Chen, Q. He and S. R. Xu, Anal. Chem., 2024, 96(43), 17405–17412 Search PubMed.
  36. Y.-Y. Gao, X.-A. Yang and W. B. Zhang, Talanta, 2024, 275, 126127 Search PubMed.
  37. M. Janeda, P. Pawłowski and M. Ślachciński, J. Anal. Chem., 2024, 78(12), 1741–1751 CrossRef.
  38. D. Beauchemin, At. Spectrosc., 2023, 44(04), 191–197 Search PubMed.
  39. H. Kim, M.-K. Oh, J.-u. Kim and G.-H. Kwak, J. Anal. At. Spectrom., 2024, 39(2), 384–388 RSC.
  40. Z. Cai, Z. Ni, M. Gao and Z. Wang, Microchem. J., 2023, 195, 109501 CrossRef CAS.
  41. Y. Zhang, J. Orejas, J. Pisonero and N. Bordel, J. Anal. At. Spectrom., 2024, 39(3), 808–819 RSC.
  42. H. B. Khanh, N. L. Duy, N. A. Tien and N. H. D. Khang, Spectrochim. Acta, Part B, 2024, 219, 107000 CrossRef CAS.
  43. J. Liu, J. Dong, S. Han, J. Zhang, X. Liu, H. Zheng and Z. Zhu, J. Anal. At. Spectrom., 2024, 39(5), 1343–1352 RSC.
  44. A. Q. Liu, W. Zeng, Z. L. Zhu, S. J. Zhuo, X. C. Liu, C. G. He, J. H. Fang, H. He and R. Qian, Microchem. J., 2024, 205, 111249 CrossRef CAS.
  45. M. Gorska and P. Pohl, Spectrochim. Acta, Part B, 2023, 208, 106778 CrossRef CAS.
  46. K. Gręda, J. Kratzer, N. Vlčková and P. Pohl, Spectrochim. Acta, Part B, 2024, 213, 106879 CrossRef.
  47. Z. Xu and W. Hang, At. Spectrosc., 2023, 44(06), 434–439 CrossRef.
  48. K. Mervič, M. Šala and S. Theiner, TrAC, Trends Anal. Chem., 2024, 172, 117574 CrossRef.
  49. M. Martinez and M. Baudelet, Anal. Bioanal. Chem., 2024, 416(26), 5853 CrossRef CAS PubMed.
  50. M. Kobayashi, H. Namiki and H. Hayashi, Anal. Sci., 2023, 39(10), 1757–1762 CAS.
  51. A. Rua-Ibarz, T. Van Acker, E. Bolea-Fernandez, M. Boccongelli and F. Vanhaecke, J. Anal. At. Spectrom., 2024, 39(3), 888–899 CAS.
  52. M.-L. Chen and J.-H. Wang, At. Spectrosc., 2023, 44(06), 401–408 Search PubMed.
  53. H. Li, F. Li, W. Guo, L. Jin and S. Hu, J. Anal. At. Spectrom., 2024, 39(6), 1493–1502 RSC.
  54. J. Pisonero, E. Avigliano, C. Soto-Gancedo, A. Méndez-Vicente, C. Méndez-López, J. Orejas and N. Bordel, J. Anal. At. Spectrom., 2024, 39(2), 601–609 CAS.
  55. M. Hola, Z. Salajkova, F. Gregar, J. Ondracek, D. Pavlinak, P. Sperka, A. Hrdlicka, J. Kaiser and V. Kanicky, Spectrochim. Acta, Part B, 2024, 219, 106999 CAS.
  56. K. Mervič, J. T. van Elteren, M. Bele and M. Sala, Talanta, 2024, 269, 125379 Search PubMed.
  57. J. T. van Elteren, D. Metarapi, K. Mervic and M. Sala, Anal. Chem., 2023, 95(26), 9863–9871 CAS.
  58. S. M. Chernonozhkin, T. Van Acker, S. J. M. Van Malderen, J. Belza, S. Goderis and F. Vanhaecke, J. Anal. At. Spectrom., 2024, 39(4), 1050–1056 CAS.
  59. M. Podsednik, F. Fahrnberger, D. K. Gibbs, B. Achleitner, S. Larisegger, M. Nelhiebel, H. Hutter and A. Limbeck, J. Anal. At. Spectrom., 2024, 1903–1909 CAS.
  60. S. Theiner, E. Foels and G. Koellensperger, Anal. Chim. Acta, 2024, 1332, 343345 CAS.
  61. M. G. Mello, M. T. Westerhausen, T. E. Lockwood, P. Singh, J. Wanagat and D. P. Bishop, Anal. Bioanal. Chem., 2024, 416(11), 2725–2735 CAS.
  62. C. Neuper, M. Simic, T. E. Lockwood, R. G. de Vega, U. Hohenester, H. Fitzek, L. Schlatt, C. Hill and D. Clases, Anal. Chem., 2024, 96(21), 8291–8299 Search PubMed.
  63. C. Geng, Z. Dong, T. Zhang, Z. Yang, Z. Xu, S. Liang and X. Ding, Talanta, 2024, 270, 125634 CrossRef CAS PubMed.
  64. F. Xu, Y. Liu, X. Liu, J. Tan, H. Xia and M. Li, Microchem. J., 2023, 193, 109109 CrossRef CAS.
  65. J. He, M. Hu, L. Yang, J. C. Yin, Z. X. Zhang, L. Y. Tu, Y. X. Heng, L. Tang, J. He, X. D. Hou and X. Jiang, Anal. Chem., 2024, 96(49), 19196–19201 CrossRef CAS PubMed.
  66. Y. Li, P. Chen, Y. Su, T. Ren, Y. Deng, Y. Lin and C. Zheng, Sens. Actuators, B, 2023, 395, 134534 Search PubMed.
  67. Y. Li, J. Tang, Y. Xiao, T. Ren, J. Yang, Y. Lin and C. Zheng, Anal. Chem., 2024, 96(18), 7187–7193 Search PubMed.
  68. M.-A. Boillat and P. C. Hauser, Analyst, 2024, 4213–4221 RSC.
  69. L. Lin, S. Sun and Q. Yu, Spectrochim. Acta, Part B, 2024, 217, 106959 CrossRef CAS.
  70. T. Hirata, K. Kobayashi, H. H. Khoo, O. Shikino and H. Asanuma, Analyst, 2024, 149(21), 5174–5183 RSC.
  71. S. Burhenn, J. Golda, J. Kratzer, S. Brandt and J. Held, Spectrochim. Acta, Part B, 2024, 213, 106884 CrossRef CAS.
  72. C. Li, G.-Q. Xia, A.-Q. Jiang, J.-C. Ren, E.-T. Hu, J.-K. Chen, Q.-C. Zhang, L. Yu, O. Yoshie, X. D. Xiang, H.-B. Zhao, Y.-X. Zheng, S.-Y. Wang, Y.-M. Yang, W. Wei, Y.-P. Lee, J. Guo, Y.-H. Jia and L.-Y. Chen, J. Anal. At. Spectrom., 2024, 39(2), 414–422 RSC.
  73. R. Hu and A. Gundlach-Graham, Rapid Commun. Mass Spectrom., 2024, 38(12), e9753 CAS.
  74. A. D. French, K. M. Melby, K. P. Hobbs, R. M. Cox, G. Eiden, E. W. Hoppe, I. J. Arnquist and K. Harouaka, Talanta, 2024, 272, 125799 CrossRef CAS PubMed.
  75. A. D. French, K. M. Melby, R. M. Cox, E. Bylaska, G. C. Eiden, E. W. Hoppe, I. J. Arnquist and K. Harouaka, Spectrochim. Acta, Part B, 2023, 207, 106754 CAS.
  76. X. Jiang, Y. S. Liu, W. G. Liu, J. Lin, Z. Y. Liu, L. F. Chen, X. Zhu, W. Zhang and Z. C. Hu, J. Anal. At. Spectrom., 2024, 39(10), 2452–2460 Search PubMed.
  77. A. M. Popov, I. N. Krylov and T. A. Labutin, Spectrochim. Acta, Part B, 2024, 211, 106837 CAS.
  78. M. Zengin, S. K. Barka, I. K. Ozturk, I. Klincare, S. Kroger and G. Basar, Spectrochim. Acta, Part B, 2024, 217, 106950 Search PubMed.
  79. S. K. Barka, G. Basar, S. Kroger and G. Basar, Spectrochim. Acta, Part B, 2024, 216, 106946 CAS.
  80. L. Windholz, G. Basar, S. Kroeger and G. Basar, Spectrochim. Acta, Part B, 2024, 214, 106900 Search PubMed.
  81. S. Kodangil, N. Domoto, M. Tanaka, D. Kato, G. Gaigalas, H. Tanuma and N. Nakamura, J. Quant. Spectrosc. Radiat. Transfer, 2024, 322, 109011 Search PubMed.
  82. I. Suresh, Priti, R. Srivastava and R. K. Gangwar, Spectrochim. Acta, Part B, 2024, 213, 106860 Search PubMed.
  83. A. Agrawal, S. Gupta, L. Sharma and R. Srivastava, Spectrochim. Acta, Part B, 2024, 217, 106953 CAS.
  84. Ł. M. Sobolewski, S. Rathi, L. Sharma, L. Windholz and J. Kwela, J. Quant. Spectrosc. Radiat. Transfer, 2024, 316, 108901 Search PubMed.
  85. R. E. Mejia, M. Raineri, R. Sarmiento, J. C. Álvarez and P. Pacheco, IEEE Trans. Plasma Sci., 2023, 51(10), 3035–3052 Search PubMed.
  86. F. J. Arellano, M. Gyulai, Z. Donkó, P. Hartmann, T. V. Tsankov, U. Czarnetzki and S. Hamaguchi, Plasma Sources Sci. Technol., 2023, 32(12), 125007 CAS.
  87. E. Képeš, H. Saeidfirozeh, V. Laitl, J. Vrábel, P. Kubelík, P. Pořízka, M. Ferus and J. Kaiser, J. Anal. At. Spectrom., 2024, 39(4), 1160–1174 Search PubMed.
  88. P. S. N. S. R. Srikar, I. Suresh and R. K. Gangwar, Spectrochim. Acta, Part B, 2024, 215, 106909 Search PubMed.
  89. J. Jovović and G. L. Majstorović, Spectrochim. Acta, Part B, 2024, 211, 106836 Search PubMed.
  90. A. A. Pupyshev, P. V. Zaitceva, M. Y. Burylin, M. A. Maltsev, I. V. Morozov and E. L. Osina, J. Anal. Chem., 2024, 79(8), 1038–1048 Search PubMed.
  91. A. A. Pupyshev, P. V. Zaitseva, M. Y. Burylin, A. G. Abakumov and P. G. Abakumov, J. Anal. Chem., 2024, 79(9), 1242–1251 Search PubMed.
  92. A. D. French, K. P. Hobbs, R. M. Cox and I. J. Arnquist, Analyst, 2024, 149, 5812–5820 RSC.
  93. O. V. Komin and O. V. Pelipasov, Spectrochim. Acta, Part B, 2023, 207, 106742 Search PubMed.
  94. D. J. Butcher, Appl. Spectrosc. Rev., 2024, 59(2), 247–275 CrossRef CAS.
  95. T. E. Lockwood, L. Schlatt and D. Clases, J. Anal. At. Spectrom., 2024, 130–136 Search PubMed.
  96. O. A. Al-Najjar, Y. S. Wudil, U. F. Ahmad, O. S. Baghabra Al-Amoudi, M. A. Al-Osta and M. A. Gondal, Appl. Spectrosc. Rev., 2022, 58(10), 687–723 CrossRef.
  97. R. S. Harmon and G. S. Senesi, Geostand. Geoanal. Res., 2024, 48(4), 763–792 CrossRef.
  98. H. Saeidfirozeh, P. Kubelik, V. Laitl, A. Krivková, J. Vrabel, K. Rammelkamp, S. Schröder, I. B. Gornushkin, E. Kepes, J. Zabka, M. Ferus, P. Porízka and J. Kaiser, TrAC, Trends Anal. Chem., 2024, 181, 117991 CrossRef CAS.
  99. E. H. Kwapis, J. Borrero, K. S. Latty, H. B. Andrews, S. S. Phongikaroon and K. C. Hartig, Appl. Spectrosc., 2024, 78(1), 9–55 CrossRef CAS PubMed.
  100. I. Goncharova, D. Guichaoua, S. Taboukhat, A. Tarbi, T. Chtouki, H. Erguig and B. Sahraoui, Spectrochim. Acta, Part B, 2024, 217, 106943 CrossRef CAS.
  101. Z. Sun, C. Yu, J. Feng, J. Zhu and Y. Liu, J. Anal. At. Spectrom., 2024, 39(5), 1212–1224 Search PubMed.
  102. S. C. Yao, Z. Y. Yu, Z. Y. Hou, L. B. Guo, L. Zhang, H. B. Ding, Y. Lu, Q. Q. Wang and Z. Wang, TrAC, Trends Anal. Chem., 2024, 177, 117795 CrossRef CAS.
  103. M. Dong, J. Cai, H. Liu, J. Xiong, G. Rao, S. Yao and J. Lu, Spectrochim. Acta, Part B, 2023, 210, 106807 Search PubMed.
  104. D. S. Ferreira, D. V. Babos, M. H. Lima, H. F. Castello, A. C. Olivieri, F. M. V. Pereira and E. R. Pereira, J. Anal. At. Spectrom., 2024, 39(12), 2949–2973 Search PubMed.
  105. L. Duponchel, C. Fabre, B. Bousquet and V. Motto-Ros, Spectrochim. Acta, Part B, 2023, 208, 106776 CAS.
  106. Z. Hao, K. Liu, Q. Lian, W. Song, Z. Hou, R. Zhang, Q. Wang, C. Sun, X. Li and Z. Wang, Front. Phys., 2024, 19(6), 62501 Search PubMed.
  107. J. Hou, D. Zhang, Z. Feng, J. Zhu and L. Zhang, Spectrochim. Acta, Part B, 2024, 215, 106925 CAS.
  108. A. Agrawal, S. S. Baghel, L. Sharma and R. Srivastava, Spectrochim. Acta, Part B, 2023, 208, 106764 CrossRef CAS.
  109. A. Favre, A. Bultel, V. Morel, M. Lesage and L. Gosse, J. Quant. Spectrosc. Radiat. Transfer, 2025, 330, 109222 Search PubMed.
  110. D. Ciniglia, F. Migliorini, R. Dondé, P. Diwakar and S. De Iuliis, Spectrochim. Acta, Part B, 2023, 208, 106784 Search PubMed.
  111. C. Méndez-López, C. Gonzalez-Gago, J. Pisonero and N. Bordel, Spectrochim. Acta, Part B, 2024, 216, 106949 CrossRef.
  112. K. S. Latty and K. C. Hartig, Appl. Spectrosc., 2023, 77(8), 848–859 Search PubMed.
  113. M. Abdel-Harith, R. H. El-Saeid, R. M. Abdelazeem, Z. Abdel-Salam and A. Elhassan, J. Anal. At. Spectrom., 2024, 39(2), 582–591 Search PubMed.
  114. A. Elhassan, R. H. El-Saeid, R. M. Abdelazeem, Z. Abdel-Salam and M. Abdel-Harith, J. Anal. At. Spectrom., 2024, 39(9), 2319–2329 RSC.
  115. C. Álvarez-Llamas, A. Tercier, C. Ballouard, C. Fabre, S. Hermelin, J. Margueritat, L. Duponchel, C. Dujardin and V. Motto-Ros, J. Anal. At. Spectrom., 2024, 39(4), 1077–1086 RSC.
  116. M. Batsaikhan, H. Ohba, T. Karino, K. Akaoka and I. Wakaida, J. Anal. At. Spectrom., 2024, 39(2), 423–432 RSC.
  117. S. Chai, J. Ren, S. M. Jiang, A. C. Li, Z. Q. Zhao, H. M. Peng, Q. W. Zhang and W. D. Wu, J. Anal. At. Spectrom., 2024, 39(11), 2756–2766 Search PubMed.
  118. M. Dell'Aglio, A. Mallardi, R. Gaudiuso and A. Giacomo, Appl. Spectrosc., 2023, 77(11), 1253–1263 CrossRef.
  119. Z. Farka, K. Vytiskova, E. Makhneva, E. Zikmundova, D. Holub, J. Buday, D. Prochazka, K. Novotny, P. Skladal, P. Porizka and J. Kaiser, Anal. Chim. Acta, 2024, 1299, 342418 Search PubMed.
  120. A. Safi, J. E. Landis, H. G. Adler, H. Khadem, K. E. Eseller, Y. Markushin, S. Honarparvaran, A. De Giacomo and N. Melikechi, Talanta, 2024, 271, 125723 CrossRef CAS PubMed.
  121. S. Soumyashree and P. Kumar, J. Anal. At. Spectrom., 2024, 39(9), 2270–2277 RSC.
  122. C. Nan, S. Xinjian, W. Jiahui, W. Jian, L. Bin and L. Yande, J. Anal. At. Spectrom., 2024, 39(2), 401–407 RSC.
  123. J. Kiefer, Appl. Spectrosc., 2024, 78(4), 438–441 CrossRef CAS PubMed.
  124. J. Li, J. Wu, M. Shi, Y. Qiu, Y. Zhou, H. Sun, X. Guo, D. Wu, Y. Hang, H. Yang and X. Li, J. Anal. At. Spectrom., 2024, 39(5), 1235–1247 RSC.
  125. G. Chen, P. Zheng, J. Wang, B. Li, X. Liu, Z. Yang, Z. Sun, H. Tian, D. Dong and L. Guo, J. Anal. At. Spectrom., 2024, 1235–1247 Search PubMed.
  126. J. J. Chen, X. J. Hao, B. M. Mo, S. J. Li, J. J. Ma, X. D. Liang, Z. Wang and H. Zhang, J. Anal. At. Spectrom., 2024, 39(10), 2382–2394 RSC.
  127. L. Duponchel, A. Cugerone, M. Muñoz and V. Motto-Ros, Spectrochim. Acta, Part B, 2023, 207, 106758 CrossRef CAS.
  128. J. Feng, Y. H. Yan, Y. Z. Liu and J. Lengyel, J. Anal. At. Spectrom., 2024, 114–121 Search PubMed.
  129. P. Ju, X. Gao, H. Yu, Q. Wang, Y. Dou and J. Lin, J. Anal. At. Spectrom., 2024, 39(3), 712–722 RSC.
  130. M. Gaft, L. Nagli, J. Fernandes, V. Gardette, Y. Raichlin and V. Motto-Ros, Spectrochim. Acta, Part B, 2024, 218, 106993 CrossRef CAS.
  131. S. S. Harilal, A. K. Shaik, E. J. Kautz, A. Devaraj, A. M. Casella and D. J. Senor, J. Anal. At. Spectrom., 2024, 39(3), 699–703 RSC.
  132. E. J. Kautz, M. P. Polek, E. C. E. Ronnebro and S. S. Harilal, Spectrochim. Acta, Part B, 2024, 217, 106952 CrossRef CAS.
  133. I. Traparic, D. Rankovic, B. D. Stankov, J. Savovic, M. Kuzmanovic and M. Ivkovic, Spectrochim. Acta, Part B, 2024, 221, 107050 CAS.
  134. Z. Gajarska, A. Faruzelová, E. Képes, D. Prochazka, P. Porízka, J. Kaiser, H. Lohninger and A. Limbeck, J. Anal. At. Spectrom., 2024, 39(12), 3151–3161 RSC.
  135. N. Lellouche, B. Cocciaro, S. Legnaioli, G. Lorenzetti, V. Palleschi, F. Poggialini, S. Raneri and S. M. Aberkane, Spectrochim. Acta, Part B, 2024, 215, 106903 CAS.
  136. K. H. Lepore, M. D. Dyar and C. R. Ytsma, Spectrochim. Acta, Part B, 2024, 211, 106839 CAS.
  137. G. Lu, L. Sun, Z. Cong and T. Chen, Talanta, 2024, 270, 125531 CrossRef CAS PubMed.
  138. K. Pan, J. Li, S. Zhong, C. Zhang, Y. Wang, Y. Xue and G. Xue, J. Anal. At. Spectrom., 2024, 1895–1902 CAS.
  139. Z. Y. Ouyang, C. Niu, Y. C. Liu, Z. L. Hu and L. B. Guo, Talanta, 2025, 283, 127091 CAS.
  140. G. Qi, Y.-B. Huang, J. Huang, X.-J. Lu, T. Yang and Z.-S. Cao, J. Anal. At. Spectrom., 2024, 39(3), 935–941 CAS.
  141. A. Quemet, A. Hubert, A. Gourgiotis, A. M. S. Hernandez, M. Crozet, G. Bailly, A. Dobney, G. Duhamel, J. Hiess, U. Repinc, S. Mialle, B. Boulet, R. Escoube, C. Bouvier-Capely, F. Pointurier and S. Picart, J. Anal. At. Spectrom., 2024, 1665–1699 CAS.
  142. H. Wu, J. H. Bai, X. R. Liang, X. B. Lu, Y. N. Deng, M. Li, J. L. Ma and G. J. Wei, Anal. Chem., 2024, 96(38), 15102–15107 CAS.
  143. A. Buzenchi, H. Moreira, E. Bruand, O. Bruguier and B. Dhuime, Chem. Geol., 2024, 666, 122306 CAS.
  144. J. M. Schneider and T. Kleine, J. Anal. At. Spectrom., 2024, 1910–1918 CAS.
  145. P. Télouk and V. Balter, J. Anal. At. Spectrom., 2024, 39(3), 879–887 Search PubMed.
  146. Y.-M. Xu, G.-Q. Wang, Z. Yang, Y.-L. Zeng and F. Guo, J. Anal. At. Spectrom., 2024, 39(1), 86–98 Search PubMed.
  147. Y. Yu, E. Hathorne, C. Siebert, M. Gutjahr, J. Fietzke and M. Frank, Chem. Geol., 2024, 662, 122220 CAS.
  148. P.-F. Zhang, Z.-M. Li, W.-L. Wang, L. Feng, J. Xu, M.-C. Zhang, X.-P. Shen, X.-L. Yuan, F.-T. Li, H. Deng and L.-H. Zhai, Int. J. Mass Spectrom., 2024, 499, 117236 CAS.
  149. A. M. Cruz-Uribe, G. Craig, J. M. Garber, B. Paul, C. Arkula and C. Bouman, Geostand. Geoanal. Res., 2023, 47(4), 795–809 CAS.
  150. D. Li, G. Cui, S. Chen, Y. Qiao, Q. Liu, J. Zhang and Q. He, J. Anal. At. Spectrom., 2024, 2395–2401 Search PubMed.
  151. Y. Lu, L.-L. Zhang, L. Liu, D.-C. Zhu, J.-C. Xie and Q. Wang, J. Anal. At. Spectrom., 2024, 39(3), 829–840 Search PubMed.
  152. Y. Tang, T. Lan, J. Gao, Z. Bai, X. Huang, J. Han and N. Liu, J. Anal. At. Spectrom., 2024, 39(1), 109–120 RSC.
  153. N. A. Zirakparvar, B. T. Manard, D. Dunlap and C. R. Hexel, Rapid Commun. Mass Spectrom., 2024, 38(16), e9781 Search PubMed.
  154. X. Wang, L. Wang, Z. Ma, Y. Liu, W. Duan, M. Tan and J. Xiao, J. Anal. At. Spectrom., 2024, 39(2), 447–454 Search PubMed.
  155. C. N. Rodriguez-Diaz, E. Paredes, L. D. Pena, I. Cacho, C. Pelejero and E. Calvo, Talanta, 2024, 269, 125473 CrossRef CAS PubMed.
  156. D. Coenen, D. Evans, H. Jurikova, M. Dumont, J. Rae and W. Mueller, J. Anal. At. Spectrom., 2024, 2409–2420 RSC.
  157. E. Albalat, P. Telouk and V. Balter, J. Anal. At. Spectrom., 2024, 1002–1011 Search PubMed.
  158. C. T. Gerritzen, S. Goderis, H. F. James and C. Snoeck, Spectrochim. Acta, Part B, 2024, 217, 106955 CrossRef CAS.
  159. Y. Fang, Q.-Y. Wen, Z.-C. Xiao, H.-M. Yu and F. Huang, J. Anal. At. Spectrom., 2024, 1750–1758 RSC.
  160. H.-O. Gu, X. Liu, H. Sun, G. Sun, D. Li, F. Wang, C. Ge, Y. Fan and T. Zhou, Anal. Chem., 2024, 96(16), 6131–6138 CrossRef CAS PubMed.
  161. T. Kashiwabara, Y. Fukami, S. Kubo, A. Watakabe, M. Kurisu, S. Tokeshi, T. Iizuka and K. Suzuki, J. Anal. At. Spectrom., 2024, 1759–1777 RSC.
  162. E. K. King-Doonan, L. D. Bilenker, D. Weis, K. Gordon, A. Fourny, G. M. Patton and Y. Zhao, Geostand. Geoanal. Res., 2024, 48(2), 309–330 CrossRef CAS.
  163. X. Lü, D. He, Y. Liu, M. Li, J. Lin, W. Chen, L. Zhu, A. Yang, Y. Feng, Z. Liu, X. Zeng and Z. Hu, Chem. Geol., 2024, 648, 121935 CrossRef.
  164. L. Xu, J.-H. Yang, H. Wang, H. Ye, L.-W. Xie, Y.-H. Yang, C. Huang and S.-T. Wu, J. Anal. At. Spectrom., 2024, 39(3), 900–916 RSC.
  165. S. G. Lee, Spectrochim. Acta, Part B, 2024, 218, 106991 Search PubMed.
  166. T. Yokoyama, Y. Ohkuma, K. Nishikawa, K. Sumiya and I. Gautam, Geostand. Geoanal. Res., 2023, 47(2), 415–435 CrossRef CAS.
  167. C. F. Li, Z. Y. Chu and P. Peng, Anal. Chem., 2024, 96(38), 15436–15445 Search PubMed.
  168. J. Li, S. H. Tang, X. K. Zhu, J. X. Ma, Z. Y. Zhu and B. Yan, J. Anal. At. Spectrom., 2024, 39(12), 3106–3115 Search PubMed.
  169. S. López-Urzúa, T.-H. Luu and L. Derry, J. Anal. At. Spectrom., 2024, 39(4), 1033–1042 Search PubMed.
  170. T. H. Luu, D. Peters, E. Lahoud, Y. Gerard and F. Moynier, Anal. Chem., 2024, 96(5), 2199–2205 CrossRef CAS PubMed.
  171. C. Wang, X.-K. Zhu, J. Li, Z. Zhu, J. Ma and Y. Chen, J. Anal. At. Spectrom., 2024, 39(6), 1529–1539 Search PubMed.
  172. Q. Qu, W. Liu, W. Zheng, B. Chetelat, Q. Liu and J. Chen, J. Anal. At. Spectrom., 2024, 39(9), 2258–2269 Search PubMed.
  173. W. Song, C. Zhu, H. Wen, Y. Zhang, Z. Liu and M. Zhou, Geostand. Geoanal. Res., 2024, 49(1), 245–258 Search PubMed.
  174. A. Tupys, K. Tetfejer, L. Halicz, E. Bulska and J. Karasinski, J. Anal. At. Spectrom., 2024, 39(12), 3142–3150 RSC.
  175. Z.-Y. Wang, F.-Y. Dong, L. Yin, J.-J. Liu, Q.-H. Zhong and J. Li, J. Anal. At. Spectrom., 2024, 2443–2451 RSC.
  176. Z. Zhang, J. Ma, L. Zhang, Y. Liu and G. Wei, Geostand. Geoanal. Res., 2023, 47(3), 697–712 Search PubMed.
  177. P. Goswami, S. Paul, K. S. Bhushan and P. G. Jaison, J. Anal. At. Spectrom., 2024, 39(2), 500–507 Search PubMed.
  178. H. Jaegler and A. Gourgiotis, J. Anal. At. Spectrom., 2023, 38(10), 1914–1919 Search PubMed.
  179. A. A. Reinhard, J. D. Inglis, M. J. Singleton, A. H. Kara, R. E. Steiner and S. P. Lamont, Anal. Chem., 2024, 96(19), 7585–7593 CAS.
  180. W. Wang, J. Xu, R. Xi, S. Guo, Y. Su, S. Fang, H. Zhang, Y. Wang, J. Fan, L. Feng, Y. Wang and Z. Li, J. Anal. At. Spectrom., 2024, 39(1), 178–189 Search PubMed.

This journal is © The Royal Society of Chemistry 2025
Click here to see how this site uses Cookies. View our privacy policy here.