Atomic spectrometry update: review of advances in the analysis of clinical and biological materials, foods and beverages

Contents

• 1. Reviews

• 2. Metrology, interlaboratory studies, reference materials and reference ranges

• 3. Sample collection and preparation
  • 3.1 Collection, storage and preliminary preparation
  • 3.2 Digestion, extraction and pre-concentration

• 4. Progress with analytical techniques
  • 4.1 Mass spectrometry
  • 4.2 Atomic absorption and atomic emission spectrometry
  • 4.3 Laser induced breakdown spectroscopy
  • 4.4 Vapour generation procedures and atomic fluorescence spectrometry
  • 4.5 X-ray spectrometry

• 5. Nanomaterials

• 6. Applications: clinical and biological materials
  • 6.1 Metallomics
  • 6.2 Imaging with LIBS, MS and X-rays
  • 6.3 Elements as tags for indirect determinations
  • 6.4 Multi-element applications
  • 6.5 Progress for individual elements

• 7. Applications: drugs and pharmaceuticals, traditional medicines and supplements

• 8. Applications: foods and beverages
  • 8.1 Progress for individual elements
  • 8.2 Single and multi-element applications in food and beverages

• 9. Abbreviations

• Conflicts of interest

• References

---

Abstract

This update covers publications from the second half of 2022 to the middle of 2023. Advances in analytical techniques and their applications relevant to clinical and biological materials, foods and beverages are reviewed in the text, highlighting their key-features. Four tables complement the text, aiming to summarise technical details of interest for sample preparation based on extraction procedures, applications to clinical and biological materials, analysis of food and beverages and studies of food authenticity or origin. Techniques with multielement capabilities are more and more popular as they allow maximisation of the information that can be obtained from a single sample and the results are often analysed by sophisticated statistical techniques. However, the interest in single-element methods has not decreased and it is supported by the development of smart extraction and pre-concentration methods that make the most of nanomaterials. Beside more conventional applications of ICP-MS, most of the developments in the biological field are devoted to exploring the capabilities for single cell and single particle measurements. Much work is in progress, but standardization of these procedures will require more time and effort. A number of papers also reported the use of ICP-MS as a detector to support measurement of biological macromolecules. The use of LIBS, MS and X-ray for imaging, alongside other techniques, continues to attract attention, in particular for applications in support of cancer diagnosis and bioanalytical research. Applications to food and beverage samples have focused both on achieving lower limits of detection and on multiple measurement capabilities, in particular, following the interest in documenting the authenticity and origin of food products, several reviews have addressed specific aspects of the process of measuring chemical elements in biological matrices, food and beverages. Attention was also given to various techniques for micro-sampling biological fluids that would facilitate monitoring programmes with less invasive procedures. The issuing of regulations for the control of chemical elements as impurities in medicines has driven commitment to develop or improve suitable analytical methods.

---

1. Reviews

This latest Update follows on from last year's¹ and is a critical review of the relevant literature published during the approximately 12 months period from the second half of 2022 through the first half of 2023. It should be read in conjunction with the five other Atomic Spectrometry Updates published during the same period.^2–6

This year, a number of reviews, covering various aspects of atomic spectrometry techniques applied to clinical, food and beverage samples, as well as pharmaceutical products, were published. However, we decided that it may be best to discuss the most relevant ones in the context of specific sections, and only briefly mention them here. This approach should allow readers to find information related to a particular subject in one place and should avoid repetitions, although some papers may include features that need to be discussed under different angles.

The review by El Hosry et al.⁷ covered both sample preparation and analytical techniques for the determination of trace elements in food. The paper includes a discussion of the criticalities of the initial steps required to produce a representative laboratory sample (Section 3.1), as well as an overview of techniques for sample pre-treatment, e.g. ashing, digestion, separation, pre-concentration (Section 3.2). Sample preparation, including pre-concentration and extraction approaches, has been and still is a topic of great interest (Section 3.2). Out of concern for the increasing release of Tl into the environment by human activities, Shi et al.⁸ provided an overview of current progress for its determination in biological and environmental samples, covering both atomic and molecular spectrometry techniques and a detailed discussion of sample preparation methods. Ibourki et al.⁹ considered the importance of sample preparation for improving both the LODs and the speed of determinations of chemical elements in food and provided a general overview of both strategies and methods. Unsurprisingly, several papers explored the state-of-the-art of specific sample preparation techniques. An extensive review¹⁰ evaluated the use of deep eutectic solvents (DESs) for LLME, whereas Lanjwani et al.¹¹ summarised the application of DESs for the extraction and pre-concentration of both organic and inorganic species in water and food samples. The application of CPE in methods for the determination and speciation of Se in food and beverages was reviewed by Hagarová and Nemček.¹² Ullah and Tuzen¹³ discussed the trends and future perspectives of the miniaturisation of conventional extraction methods, unfortunately mostly related to environmental samples. The use of magnetic graphene oxide (GO) for the pre-concentration and/or speciation of several elements in environmental and food samples was examined by Morales-Benítez et al.¹⁴

Progresses in the application of atomic spectrometry techniques to samples of clinical interest, as well as food and beverage samples, are regularly reported in these Updates, but making the most of this overload of information is not an easy task. Reviews related to advances in ICP-MS are discussed in Section 4.1. The team of Clases provided two comprehensive reviews^15,16 of the applications of ICP-MS in the biomedical field, discussing both current practices and future directions, whereas Silva et al.¹⁷ devoted their efforts to cover all aspects of single cell ICP-MS (scICP-MS) for biomedical applications. Vladitsi et al.¹⁸ summarised the recent trends and criticalities of analytical methodologies for the characterisation and quantification of NPs in biological matrices (Section 5).

Two review papers cover the use of atomic spectrometry techniques for imaging of biological samples (Section 6.2). Davison et al.¹⁹ discussed the state-of-the-art and perspectives of single cell analysis using spICP-MS, LIBS and LA-ICP-MS, whereas the work of Peng et al.²⁰ was devoted to explore the use of multiple laser imaging techniques and nanomaterials to improve cancer diagnosis in vivo.

Methods for the evaluation of elemental impurities in medicines, required by regulatory authorities, were reviewed by Aleluia et al.²¹ (Section 7). In addition, Lindenmayer et al.²² discussed the use of atomic spectrometry to support national and international regulations related to the presence of trace elements in seaweeds (Section 8.2.6).

Another two papers^23,24 explored the analytical tools available to assess authenticity and origin of food products (Section 8.2.8).

2. Metrology, interlaboratory studies, reference materials and reference ranges

Good laboratory practices, now detailed in international standards related to the accreditation of analytical laboratories and other organisations, are of the utmost importance for the production of reliable measurement results and their future use. In this section, we cover recent publications addressing the quality of analytical measurements obtained by atomic spectrometry on clinical specimens, food and beverages. In addition, we report updates of reference ranges for chemical elements or their species, to be used in clinical practice or to assess background levels.

The availability of measurement standards as well as CRMs in appropriate matrices, traceable to the SI, is a key issue to ensure the comparability of the measurement results across space and time. During this Update's period, three papers addressed this issue. In our last year's Update,²⁵ we reported the assignment of SI-traceable values to in-house protein standards, to be applied for proteomic studies. This was achieved through the determination of their S content by means of ICP-MS.²⁶ The same approach was explored by Escudero-Cernuda et al.²⁷ to certify the mass purity of human cytokines, aiming to support studies of the role of these proteins as biomarkers for carcinoma or inflammatory diseases. They applied a capillary chromatographic separation coupled with parallel detection by means of ICP-MS and ESI-MS, thus allowing both quantification and identification of the protein species. As part of the assessment of candidate CRMs, studies are performed under accelerated degradation conditions, in order to estimate the expected stability of the certified values over a period of time. Since the production of CRMs requires considerable resources as well as a service to customers, monitoring studies of the certified values are carried out at stated frequencies. The team of Narukawa²⁸ reported data on the stability of three As species aqueous standard solutions, monitored over a period of 13 years. These CRMs (CRM 7901-a for AB, CRM 7912-a for As^V, and CRM 7913-a for DMA) are currently supplied by the National Metrology Institute of Japan (NMIJ). The authors applied LC-ICP-MS to confirm the certified values and related uncertainties, obtained in 2009 by using several independent analytical techniques – GFAAS, ICP-OES (optical emission spectrometry), ICP-MS, LC-ICP-MS and SF-ICP-MS. They noted the importance of the data collected for these As species over long-term monitoring studies, since predictions of stability based on a limited set of data (up to 2 years) may lead to overestimated instability. Rice is a staple food for many populations and therefore international and local standards set limits for the levels of toxic chemical elements in this crop, as the presence of such contaminants may have a large impact on the dietary intake for regular consumers. Basmati rice is a valuable rice variety, largely grown in India and other Asian countries. Considering the need to ensure reliable analytical results from control plans for this cereal, Yadav et al.,²⁹ at the India Council of Scientific & Industrial Research-National Physical Laboratory (CSIR-NPL), set up to develop a rice flour CRM for the content of As, Cd, Hg and Pb, based on cultivated Basmati rice. To achieve the expected levels, they supplemented the crop with these elements at various time intervals during cultivation. Preparation and characterisation of the candidate rice flour CRM was carried out according to the ISO Guide 35:2017 and analytical measurements were obtained using different analytical techniques (AAS, ICP-OES and ICP-MS).

Beside implementing traceability of measurement results in one's laboratory, the independent assessment of laboratory results submitted in interlaboratory comparisons and PTs is an important tool to achieve and maintain good performances as well as to support the validation of new methods. This is particularly important in new fields of testing, where analytical capabilities are not yet fully established. As the determination of NPs is an example of such a situation, we report the interlaboratory comparison (ILC) organised by a group of researchers³⁰ within the framework of the Horizon 2020 project ACEnano. The aim of the study was to validate a standard operating procedure, based on spICP-TOF-MS for the characterisation of Pt NPs. Nine laboratories analysed two different PtNP suspensions to determine particle mass, particle number concentration and isotopic composition. The evaluation of the performances was based on the average and SD of the participants' results. The authors concluded that all participants showed capabilities to determine the parameters under study. However they also highlighted discrepancies between the values assigned by the test item manufacturer and the average of participants' results, as well as a high interlaboratory variation for particle number concentration (53%), that was, however, similar to previous reports. This work seems to contribute substantial new information for further advancement in a rapidly expanding field of testing and future legislation. The determination of Pb in blood is a well-established measurement. However, a large study³¹ reported the progress achieved in China for blood lead measurements, evaluated from the results of EQA schemes carried out over a period of 6 years (2017–2022). The 4283 participants applied methods of their choice and a variety of analytical techniques. Assigned values and acceptability criteria for performance evaluation were based on consensus among all participants and by different method groups. The authors observed better performance for GFAAS and tungsten boat AAS compared to differential potentiometric stripping, ICP-MS and FAAS, whereas ASV had the lowest performance. Both overall performance and interlaboratory RSD improved over time, thus supporting the role of EQA schemes for increasing the level of harmonisation. However, significant differences may still exist among the measurement values obtained with different analytical techniques in these programmes and the comparability of these assessments with those performed elsewhere may be questionable.

In this year's review, developments in the establishment of population based reference values for elements in biological samples were limited to two publications of note. Liao et al.³² reported on a large human biomonitoring survey looking at urinary reference values for 15 metals in a general Taiwanese population. Urine samples were collected from 1871 participants, aged from 7 to 79 years, recruited from the Taiwan Environmental Survey for Toxicants (TEST). The samples were analysed for As, Cd, Co, Cr, Cu, Fe, Ga, In, Mn, Ni, Pb, Se, Sr, Ti and Zn via ICP-MS. Metal exposure was found to vary with age, sex, region and urbanisation level. Age-related reference values were determined for all metals in the following age groups: 7–11 years (n = 343); 12–17 years (n = 262); 18–39 years (n = 356); 40–64 years (n = 556) and ≥65 years (n = 354). In contrast, a much smaller study interrogated elemental concentrations in hair samples from 419 children (3–12 years) from a non-exposed population in the Mediterranean.³³ Quantification of 28 elements in hair was achieved by ICP-MS following optimised acid digestion of the samples. Reference values for each of the 28 elements were proposed, based on both the 5th–95th percentile data alone and on the 5th–95th percentile data modulated considering statistical distribution and other published works. Whilst the authors acknowledged the limitations of this small pilot study, they suggested these derived reference values may be provisionally applicable to wider populations until larger datasets are available.

3. Sample collection and preparation

3.1 Collection, storage and preliminary preparation

Microsampling techniques for biological fluids for trace elements analysis have gained popularity over the past years, as a less invasive and more flexible approach than venepuncture for the collection of blood samples. To clarify doubts regarding drawbacks related to potential contamination or alteration of the sample, that may affect its representativeness, sampling devices alternative to traditional venepuncture were investigated in three papers. The first one, by Wikstrom et al.,³⁴ aimed to assess the use of volumetric (10 μL) dry blood spots (DBS) for home-sampling of capillary blood in the monitoring of Li treatment of bipolar disorder. Lithium is usually measured in serum, since the concentrations are higher than in whole blood. Although there is no conclusive evidence in favour or against the clinical use of whole blood as the sample, one main drawback is that whole blood cannot be analysed by means of colorimetric methods. In addition, the measurements need to have a low uncertainty, since the action limits are narrow (0.5–0.9 mmol L⁻¹). This, in turn, calls for an exact volume of blood to be sampled. The study was performed on 39 patients, using both a volumetric DBS device, for capillary blood samples, and parallel venepuncture to obtain serum specimens. The DBS was placed in a centrifugal filter tube, with 300 μL of 0.56 mmol L⁻¹ NaCl, gently shaken at 700 rpm, for 30 min, then centrifuged for 10 min at 12 000g. The resulting supernatant was analysed by FAAS, using a 100 mm burner, a narrower sample inlet tubing, and monitoring the absorbance signal in transient area mode. Lithium in the serum samples was determined using a colorimetric method on an automated clinical chemistry analyser. Serum Li concentrations ranged from 0.41 to 1.22 mmol L⁻¹. As expected, those measured in capillary blood sampled by DBS were lower (0.26–0.99 mmol L⁻¹). The ratio between Li measured concentrations in paired DBS and serum samples was 0.78 and the two parameters showed a better linear correlation than previously reported (Pearson's R = 0.95). According to these findings, volumetric DBS shows potential for a future use in the monitoring of Li therapy. However, given the small size of the study, its application in routine clinical practice would require further studies, including investigations of potential confounding factors that may affect the reliability of this sampling method when practiced at home. Perrais et al.³⁵ also explored the use of DBS in comparison with microtubes, for the determination, by means of ICP-MS, of 12 trace elements (Al, total As, Cd, Co, Cr, Cu, Mn, Mo, Ni, Pb, Se and Zn) in whole blood. To this aim, they used CRMs at three concentration levels (TEs in Whole Blood ClinChek® ref. 8840–8843, Recipe Quality Controls, Germany). Preparation of DBS consisted in applying a 10 μL drop of the sample on a filter paper card, followed by overnight drying, extraction with 150 μL of 0.5% (v/v) HNO₃, for 30 min at 120 rpm on a horizontal shaker and dilution of the extract (1 + 9). A 150 μL aliquot of CRM was sampled with the microtube, of which 110 μL were diluted (1 + 9), centrifuged and the supernatant used for analysis. Both types of samples were diluted with a 1% (v/v) HNO₃–0.5% (v/v) n-butanol–0.1% (v/v) Triton X-100 solution, containing 1 μg L⁻¹ of rhodium and 1 μg L⁻¹ of indium as ISs. Five replicates were prepared for each type of sampling and for each concentration level. The LODs, determined on blanks prepared in the same way, were <1 μg L⁻¹ for microtube sampling, except for Al (15 μg L⁻¹) but ranged from 0.03 μg L⁻¹ to 14.6 μg L⁻¹ for DBS (204 μg L⁻¹ for Al). Samples taken with both devices were stable for up to three months when stored at 4 °C or −20 °C, but not at room temperature. The comparison of measured concentrations in the CRMs highlighted a higher variability between measured and assigned values for DBS than microtube sampling. The authors concluded that microtubes were more suitable than DBS for trace elements analysis in whole blood. However, since only CRMs were analysed, the findings of this study may need to be confirmed in a clinical setting. In a third paper, Breton et al.³⁶ described the development of a method for the quantification of Pb in whole blood by means of ICP-MS/MS, using Mitra®, a small, portable device with an absorbent VAMS® tip for capillary sampling. They validated their method by comparison with the one in use, based on conventional sampling, and reported no evidence of significant differences between the measured results obtained with the two methods. However, given the widespread presence of Pb in the environment, caution would still be necessary, as well as further investigations of different operational settings, for a wider use of this method of sampling.

Long-term storage of biological specimens attracts increasing interest, for example in view of future retrospective studies. However, appropriate storage of samples of biological origin for trace element determinations poses a number of problems, related to both the structure and the composition of the matrix as well as the potential interactions between analytes and matrix, that may induce changes in the analyte chemical forms or alterations in their levels, due to losses or contamination. Two studies considered the merits of formalin-fixed paraffin embedded (FFPE) tissues as a valuable resource for storage. Coyte et al.³⁷ compared trace element concentrations in 22 paired FFPE and fresh frozen human placenta samples. For this study, 5 g of only foetal villus tissue were collected from the maternal side of the placenta, blotted, frozen in liquid nitrogen and then stored at −80 °C. The rest of the placenta was treated according to the routine clinical protocol. After storage in 10% buffered formalin for a minimum of 3 days, tissue sections were taken and processed, using a Shannon Citadel Tissue Processor, with EtOH first, then xylene, and finally embedded in paraffin blocks. For trace element determination, sample treatment was carried out in a Class 1000 trace-metal-free clean laboratory facility. Aliquots (1 g) of the samples were weighed, dried, pulverised, then digested with ultrapure (UP) HNO₃. The FFPE samples were then decanted into clean vessels to eliminate paraffin. Digests for all samples were heated at 50 °C until dry, to remove the excess of HNO₃, residues digested again with ∼0.5 mL of UP HNO₃, then diluted to achieve a solution with a final concentration of 2% (v/v) HNO₃ 0.4% (v/v) HCl and 5 ng g⁻¹ of each of the ISs (Bi, In and Ir). Thirteen elements (Al, As, Ba, Cd, Cr, Cu, Fe, Gd, Hg, Mn, Pb, Sr, and Zn) were determined in the digests using DRC-ICP-MS. The range of measured concentrations was consistent with those reported in other studies. Statistical analysis by either a paired t-test or a sign test, depending on data normality, did not show significant differences between the paired specimens for any of the elements. The authors concluded that FFPE tissues from archived pathology specimens may offer a valuable resource for studies of trace metals in placenta. However, considering the complexity of the sample preparation used and the small size of the study, these findings may need to be confirmed by analysis of a larger number of samples, performed at different locations, before the use of FFPE placenta samples can be fully validated for clinical research. Another paper, by Copeland-Hardin and other researchers³⁸ also considered the potential of FFPE tissues from animals and human atomic-bomb survivors exposed to radioactive particulates as a resource to aid the understanding of the molecular effects of radiation exposure. As a proof of concept, the authors applied SXRF to 35 year-old, canine FFPE lung and lymph node specimens, obtained as part of a life-span study to investigate the effects of inhalation exposure to insoluble radionuclides, in an attempt to identify the nature and distribution of formerly radioactive micro-particulates.

Storage at low temperature (−80 °C) is probably the method most commonly applied to biological samples, however, the stability of both matrices and analytes is still a matter of concern. The work of Beauval et al.³⁹ was devoted to the investigation of the impact of long-term storage on the concentrations of ten trace elements in urine (As, Cd, Co, Cr, Mn, Ni, Pb, Sb, Tl, and Zn). To this aim, they re-analysed, by ICP-MS, 48 urine samples, collected to evaluate the environmental exposure to trace elements of the general population and kept stored in a biobank at −80 °C for a period of 11–13 years. A Wilcoxon matched-pairs signed rank test was applied to compare paired differences between the results of this study and those obtained at the time of collection, also by ICP-MS. To compensate for water loss, creatinine-adjusted data were also evaluated, based on the creatinine measurements performed at both times. Values <LOQ were excluded from statistical analysis. This affected in particular Mn, for which results for only 6 samples were available. Recoveries of the original creatinine-adjusted concentrations were close to 100% (89–102%) for As, Cd, Mn, Pb, Tl and Zn, but much lower for Co, Cr and Ni (37–59%), and increased to 127% for Sb, possibly due to contamination from the plastic storage container. The authors noted that instrumental advances may have affected results at lower concentrations and those for elements, such as Co, Cr and Ni, most likely to suffer from unresolved polyatomic interferences. The conclusions of this study highlight the possible faults and benefits of the re-use of long-term stored urine samples in retrospective studies as well as the need for a careful planning of such storages, taking into account, among others, the type and quality of containers, the reliability of the analytical measurements and the support of other analyses.

Nowadays, dietary intake and life habits, such as smoking, are likely to be the main source of exposure to chemical elements. Therefore, international and national organisations promote the assessment of the content of chemical elements in foods as well as control plans where maximum limits have been set. Several analytical techniques, as well as methods of sample preparation, can be applied to the determination of trace elements in foods. The review by El Hosry et al.⁷ includes an overview of commonly used techniques for sample preparation prior to the determination of trace elements in food. After covering the initial steps of sample preparation, such as cleaning, cutting and grinding, to produce a representative laboratory sample suitable for further processing, they considered the advantages, drawbacks and criticalities of subsequent treatments. The discussion includes dry or wet ashing and digestion, with or without the application of microwaves, as well as separation and pre-concentration techniques, aimed to reduce interferences and improve LODs.

3.2 Digestion, extraction and pre-concentration

There has been a significant increase in the numbers of publications relevant to this section compared with those for last year. As in the past, reports of pre-concentration procedures by liquid-phase or SPE have been summarised in Table 1, which is described in more detail at the end of this section. Categorising the remaining contributions has been somewhat arbitrary, but has been done largely on the basis of the procedures involved, rather than on the sample type. A number of studies have been motivated by the goal of finding alternatives to conventional HNO₃ and H₂O₂ MAD, both in terms of the reagents used and the way in which energy was coupled. Several researchers pointed out that ultrasonic baths and probes are much cheaper than microwave ovens. There has also been an interest in employing greener and safer reagents, and so the (relatively few) reports of the use of HClO₄ are steps in the wrong direction. In many articles, even those in primarily analytical journals, it is difficult to find the details of the optimised procedure developed. Often these are dispersed throughout the article or are partially (or even wholly) located in the supplementary information thereby making it challenging for readers to ascertain just exactly what the researchers have done. The determination of chemical species of potentially toxic elements in biological matrices (especially seafood) continues to attract attention with regard to extraction of the analytes. The results of a major international collaborative study involving expert laboratories⁴⁰ has shed some light on the sources of inconsistencies in the published results for As species in a number of CRMs and is discussed in more detail later in this section. Even analyses that are considered well characterized can have some surprises for the relevant community: it has been shown that during acid extraction from rice and seafood, MMA can be transformed into monomethylmonothioarsonic acid,⁴¹ a finding that is also discussed later in this section.

Several reviews of sample preparation procedures for the elemental analysis of food have appeared. In a review (224 references) that covers both instrumental techniques as well as sample preparation, Ibourki et al. discussed a number of procedures including dry- and wet-ashing, paying particular attention to MAD and UAE procedures.⁹ Slurry sampling was mentioned briefly as were dispersive SPE, dilution in organic solvents and emulsification. There was a short section on separation and pre-concentration in which only 11 papers were cited. Speciation was not mentioned. Much of the same ground was covered by El Hosry et al. in a review⁷ of 177 publications that included a table of concentration ranges for As, Cd, Cr, Hg and Pb in selected foods in different countries of the world, as well as sections devoted to the bio- and environmental chemistry of each element. They noted that MAD was widely used. Pre-concentration was afforded scant coverage (only 4 references). Interestingly, the sections on the individual atomic spectrometry techniques started with the history of the development of the technique. They identified speciation of trace elements in foods as the current challenge. In a review (182 references) of the miniaturization of conventional extraction methods for elemental analysis in different real samples, Ullah and Tuzen¹³ confined their attention to LPME and SPME in methods for the determination of potentially toxic elements published from 2010 to 2021. The vast majority of the “real samples” were environmental waters; only a few papers describing the analyses of foods (or clinical materials) were included. Unfortunately, the bibliometric analysis of the literature based on Web of Science searches was erroneous, with the results of “topic” searches reported as the results of “title” searches. It is not possible to reproduce their claims for the numbers of papers devoted to inorganic analytes and to organic analytes. The reviewers appeared to assume that the inclusion of the word “inorganic” or “organic” in the title of an article referred to the analyte(s) in question, which is not necessarily the case. Curiously, pre-concentration was not discussed. On the other hand, the paper included a number of colourful, eye-catching illustrations. In a review (106 references) of sample preparation and spectrometric methods for the elemental analyses of milk and dairy products, Soares et al.⁴² also made use of the Web of Science in surveying the status of the topics in question. However, again, presenting the numbers from “topic” searches, on “milk analysis” and “dairy analysis”, gives a grossly distorted picture, as many of the hundreds of papers retrieved each year for the period in question (2000–2021) have nothing to do with chemical analysis. The reviewers conc. on sample digestion and dissolution procedures, summarising the contents of relevant papers in three tables dealing with (1) procedures based on solubilisation, slurry preparation, dilution, extraction, micro-extraction and precipitation; (2) matrix decomposition using dry ashing, combustion, Fenton digestion and conventional wet digestion; and (3) microwave-induced combustion, MAE and MAD. They concluded that MAD was the most widely used. They also examined the use of various instrumental techniques and concluded that plasma-based techniques for the determination of macro and micronutrients (ICP-OES) and of toxic species (ICP-MS) predominated. Pre-concentration was mentioned on several occasions, but did not rise to the status of a separate section.

Several reviews of particular pre-concentration/separation procedures have appeared. Morales-Benítez et al.¹⁴ reviewed (147 references) the use of GO for the speciation of Ag, Al, As, Cd, Cr, Hg, Pb, Se and Tl in various environmental and food samples. Pre-concentration was frequently mentioned. The reviewers highlighted the advantages of the GO–MNP coupling: the agglomeration of NPs is avoided due to their dispersion on GO, which also prevents oxidation of the MNPs with atmospheric oxygen; the extraction process takes much less time than using only GO, because the magnetic separation means that neither filtering nor centrifuging is necessary. They considered that the development of fully automated procedures coupled directly with the detection instrumentation is a priority for future research. Lanjwani et al.¹¹ surveyed (205 references) the application of DESs for the extraction and pre-concentration of both organic and inorganic species in water and food samples. The reviewers focused on micro-extraction procedures. Applications for the extraction of potentially toxic elements from waters (including drinking water) and foods were summarised in tables. A substantial part of the review was devoted to the extraction of organic compounds. Curiously, although pre-concentration appeared in the title of the review, the topic was not mentioned at all in the text, though values of enrichment factors and/or pre-concentration factors were given in the tables, and the word appeared in 21 of the cited articles' titles. Speciation was not mentioned either; however, only five of the articles cited contained the word in their titles. In a review (312 references with titles) of the LLME application of DESs in analytical sample pre-treatment, Andruch et al.¹⁰ surveyed the entire field of DES from 2017 onwards. They indicated a rapid rise in the number of publications, up to 1842 in 2022, of which only about 130 were concerned with analytical applications and, of these, about 50 dealt with LLME. The remainder, supposedly dealing with the use of sorbents, will be dealt with in a future review article. Combining this subdivision of the literature for all of the years concerned, the reviewers concluded that only about 23% of the papers deal with the determination of inorganic analytes, the other 77% being concerned with organics. The work for both types of analytes was summarised in two tables; that for the inorganic analytes containing about 70 entries, the majority of which concerned elemental analysis by atomic spectrometry. Within this category most of the applications were for pre-concentration, a number of DES-based extractions and DES-based speciation procedures were highlighted. The reviewers concluded by noting that routine use is hampered by lack of wide availability of DESs and the challenges of automating such methods. A review (67 references) of the elemental analysis of environmental samples, biological materials and foods by methods involving CPE combined with ETAAS contains a summary table with 10 entries.⁴³ As the review is in Czech, little further information can be provided.

Two reviews of relevant aspects of the analytical chemistry of single elements have appeared. Hagarová and Nemček¹² examined the contents of 79 articles, published from 2006 to 2020, describing CPE in methods for the determination of Se in food, beverages, and water with quantification by spectroscopic techniques. Sample types included drinking water, wine, juices, cereals, legumes, fresh fruits, fresh vegetables, tea, mushrooms, and nuts as well as river and lake waters. Both pre-concentration and speciation were discussed. The reviewers noted that time-consuming optimization of operating parameters was needed, and they recommended the use of multivariate strategies. They also noted that for solid samples, digestion with oxidizing media would convert Se^IV to Se^VI. In a comprehensive review (205 references) of the determination of Tl in biological and environmental samples, Shi et al. devoted a considerable fraction of the review to sample preparation.⁸ Topics covered included, SPE, CPE, and DLLME. On the instrumental side, contributions to the literature were summarised in three tables devoted to detection by (1) AAS, (2) AES and (3) ICP-MS. A fourth table (18 entries) covered molecular spectrometry techniques (colorimetry, fluorimetry and chemiluminescence). Pre-concentration was mentioned frequently throughout the text, whereas speciation was dealt with in a separate section. The reviewers concluded that future developments in methodology should include new sample pre-treatment techniques for efficient enrichment and matrix separation; nanomaterials, such as ion-imprinted polymers and metal organic frameworks, were considered promising.

There have been a number of studies in which the comparison of different digestion or extraction procedures has been the major focus. In the clinical area, for the determination of Cd and Pb in whole blood by ICP-OES, researchers⁴⁴ compared MAD with HNO₃ and H₂O₂ with a simple “dilution digestion” with a HNO₃–n-butanol—Triton X-100 mixture. Recoveries ranged from 85% to 88% for the latter, compared with 96% for the former. They concluded that, despite taking longer and being less cost-effective, the MAD method was preferred. A similar comparison was made by Brouziotis et al.⁴⁵ for the determination of REEs in human urine by ICP-MS: urine was either diluted (1 + 4) with 2% (v/v) HNO₃ or subjected to MAD in a sealed vessel (2 mL urine, 0.5 mL conc. HNO₃, final volume 10 mL). The researchers found no significant differences between the results for a variety of performance parameters relating to accuracy and precision as well as for LODs and concluded that both methods were equally effective. Liu et al.⁴⁶ investigated five digestion procedures featuring conc. HNO₃ in the determination of up to 52 inorganic elements in human hair by ICP-MS. In addition to digestion (a) at room temperature for 24 h, (b) at 90 °C for 4 h, (c) with ultrasound assistance, two microwave-aided procedures were evaluated: (d) a programmed MAD procedure and (e) MAD in a domestic microwave oven (100% power for 30 min). They found that the results for heating at 90 °C for 4 h were inferior to those of the other four methods and that the domestic oven method allowed the detection of multiple elements in a small quantity of hair (3 mg), with a minimum amount of HNO₃ (200 μL) and a short digestion time (30 min). The only elements that could not be reliably determined with this procedure were Cd, Cs, Eu, and Th. Aydinoglu⁴⁷ evaluated three methods for extracting Fe and Zn from dietary supplements prior to determination by FAAS: direct acid dissolution (in conc. HCl), wet digestion (with 1 + 2 HClO₄ and HNO₃), and MAD (with HNO₃ and H₂O₂). Each procedure involved final steps of sonication for 1 h and filtering, so presumably none of the procedures effected a complete dissolution. It was found that the best performance was obtained with the MAD method, and so a method involving HClO₄ was not recommended. This is in contrast to the findings of Bankaji et al.⁴⁸ who evaluated no fewer than eight digestion procedures in the FAAS determination of Al, Cd, Mg and Mn in plants and found that in the case of Mn, the best results were obtained with a 1 + 3 (v/v) mixture of hot HClO₄ and HNO₃. Two of the methods involved cold (1% or 10%) HNO₃, the other six procedures featured two or three acids selected from HNO₃, HCl, H₂SO₄ and HClO₄. All these procedures involved step-wise heating with final incineration at 350 °C for 1 h and subsequent dissolution of the ash in 0.05% HNO₃. This means that a perchloric-acid-grade fume hood is needed to implement some of the procedures, a facility to which many laboratories no longer have access. The samples analysed were Atriplex portulacoides, Arthrocnemum indicum, and Ulva lactuca; in addition two CRM: BCR-62 (Olea europaea) and BCR-279 (Ulva lactuca) were also analysed. Concordant results were obtained for Al and Cd by all six of the hot acid digestion methods, including, therefore, those not featuring HClO₄. But they found that the results for Mg depended on the plant species. The reasons for the discrepancies were not discussed, nor was there any discussion of the wildly inaccurate results reported for the analyses of the reference materials. The occasional inadvertent substitution of a comma for a decimal point did not help: there's a big difference between 235,996 and 235.996, not to mention the inappropriateness of quoting results to six apparently significant digits (the certificate value for the element in question, Cd, was given as 0.274 with units μg g⁻¹ for all values). Chen et al.⁴⁹ also chose a digestion procedure involving HClO₄ (and HNO₃) in the determination of Se in tea leaves by HR-ICP-MS. Their rather complicated procedure involved a five-stage digestion. A sample aliquot (0.5 g) was first heated (in a semi-closed vessel—either a conical flask or a beaker) at 80 °C for 30 min with a 4 + 1 mixture of HNO₃ and HClO₄, then heated at 130 °C until all solids had dissolved. Another 3 mL of the acid mixture were added and heating continued for 2.5 h. Then a few drops of conc. H₂O₂ were slowly added, the solution evaporated to 1–2 mL and then diluted to 20 mL. This was not the end of the sample preparation: potentially interfering matrix components were removed by loading the digest (adjusted to pH 9) into an anion-exchange column at 1.0 mL min⁻¹, followed by washing with 20 mL of 6 mol L⁻¹ HNO₃ at the same flow rate. The collected eluate was evaporated to near dryness (at 130 °C) and diluted to 10 mL with 3% HNO₃. Finally, a subsample was injected onto a RP HPLC column (for separation by ion-pairing—details in an earlier publication) connected directly to the HR-ICP-MS instrument for measurement at m/z 82. Not surprisingly, the only Se species detected was Se^VI. Recovery was monitored by spiking with ⁷⁵Se and measurement by gamma spectrometry. The method was also validated by spike recovery measurements and the analysis of a CRM (MRM0822). The method was applied to the analysis of one sample, in which Se was detected at 0.25 μg g⁻¹ w/w. The LOD was not given. It is difficult to see how such a method could be described as “rapid” (the second word in the title of the paper). In the development of a method to determine major (Ca, K, Mg, Na and P) and trace (Cu, Fe, Mn and Zn) elements in cheese by ICP-OES, Deshwal et al.⁵⁰ compared MAD with both dry and wet digestion. They analysed nine different cheese samples, whose moisture content varied from 32% to 81%, and a CRM ERM-BD151 (skim milk powder). Dry digestion involved ashing 1 g at 550 °C with dissolution of the residue in 5 mL of conc. HNO₃ and dilution to 100 mL. Wet digestion involved heating 1 g with 15 mL HNO₃ until no more fumes were evolved and dilution to 100 mL. In the MAD, 1 g was heated in a sealed vessel with 5 mL conc. HNO₃ followed by dilution to 100 mL. The researchers observed strong correlation between all three digestion methods and concluded that the procedures were comparable; however, for the trace elements they reported that no concrete conclusion about the preferred digestion technique could be drawn. The results for the analysis of the CRM were not discussed in detail, but were deemed to be in “good agreement with the certified values.” The table of the results of this analysis contains some errors in the calculations of the ratio of the measured value to the certified value (or maybe some entries for the measured values were transposed). In a somewhat similar study, Mehrasebi et al.⁵¹ compared microwave, wet and dry ashing digestion methods in the determination of Cd and Pb in four dairy products (milk, dough, yogurt, and cream) by ETAAS. Evaluation was by spike recovery; no CRM was analysed. They concluded that, overall, the MAD method was best for Pb-spiked samples, and the dry digestion method was the best for Cd-spiked samples. The paper contained the details of which of the 24 possible combinations of analyte, sample and method gave unsatisfactory results.

Several research groups have devised procedures that do not involve conc. acids. For the ICP-OES determination of multiple elements (Al, Ba, Ca, Cd, Cr, Cu, Fe, K, Mg, Mn, Na, Ni, Pb, Sr, and Zn) in caffeinated yerba mate drinks, Welna et al.⁵² investigated various “green” procedures (acidification or dilution with HNO₃ and direct analysis of untreated yerba mate with or without sonication) as an alternative to total sample decomposition. Results were compared with those of a procedure involving MAD with HNO₃, whose reliability was confirmed by the accurate analysis of three CRMs: Institute of Nuclear Chemistry and Technology INCT-MPH-2 (mixed Polish herbs) and INCT-TL-1 (tea leaves), and NIST SRM 1515 (apple leaves). They found that acidification with conc. HNO₃ to 5%, coupled with ultrasonication (10 min, at room temperature) gave the best results. For the determination of 19 elements (Al, As, Ba, Ca, Cd, Co, Cr, Cu, Fe, K, Mg, Mn, Ni, P, Pb, S, Sb, Se, and Zn) in both white and brown rice by ICP-MS, Lee et al.⁵³ optimised MAD with HNO₃ and H₂O₂, with a view to minimising the reagent consumption. The accuracy of the procedure was evaluated by the analysis of NIST SRM 1648b (rice flour). They found that for the MAD of a 0.25 g sample in a single reaction chamber, at 280 °C and 160 bar, the optimal conditions were a mixture of 3 mL of 0.1 mol L⁻¹ HNO₃–5 mL 30% H₂O₂. The optimised procedure had a residual acidity of 0.017 mol L⁻¹ and a low residual carbon content (355 mg L⁻¹). The final volume was 25 mL. The LODs, which ranged from 0.16 μg kg⁻¹ (Sb) to 9.9 μg kg⁻¹ (Fe), were adequate to allow detection of 15 elements in the CRM and all 19 elements in most of the 30 samples of white and brown rice. For some samples, the concentrations of Cd, Pb or Sb were below the LODs. They calculated that the method scored 85 (out of 100) on the analytical Eco-Scale. Lemos and Dantas⁵⁴ devised a method in which samples (crab and shrimp) were digested with formic acid followed by determination of Al, Cr, K, Mg, Mn, and Zn by MIP-OES. Samples (0.25 g) were heated with 5 mL of 50% (v/v) formic acid in a thermostated bath at 90 °C for 1 h, followed by dilution to 50 mL final volume. The method was validated by the analysis of a CRM (NRCC DORM-4 fish protein) and results for one sample were compared with those obtained by a conventional MAD with acid and peroxide procedure. The LODs ranged from 0.1 mg kg⁻¹ to 2 mg kg⁻¹. In the 11 samples all analytes were detected, except for Mn, whose concentration was below the LOD for three samples. To trace the geographical origin of three medicine food homology species (Atractylodis macrocephalae Rhizoma, Lilii Bulbus and adlay) based on the measurement of 46 mineral elements by ICP-MS, Wang et al.⁵⁵ developed a simple acid UAE procedure in which 80 mg of the dried ground sample was sonicated with 8 mL of 2% HNO₃ for 30 min followed by centrifuging and filtering. No details of the ultrasonication device or of temperature or of power were given. A typical HNO₃ MAD procedure was described in the supplementary information but was not discussed in the text, so it is not clear what, if any, comparisons of results were made. Fang et al.⁵⁶ extracted minor elements from eggs by a method in which a 2 g sample was pre-digested with 5 mL of 30% (v/v) HNO₃ at 100 °C for 1 h followed by addition of 1 mL of 30% H₂O₂ prior to the MAD. The final volume was 50 mL. Curiously, the operating parameters of the microwave oven device were not part of the optimisation strategy and were not given. Both the residual carbon content and the residual acidity were measured. The ICP-MS LODs ranged from 0.8 ng g⁻¹ to 10 ng g⁻¹ for 19 elements, and the method was validated by the analysis of a CRM (GBW10213 milk powder). In the analysis of real samples, eight elements (Ba, Cu, Li, Mn, Mo, Se, Sr, Ti) were quantified at concentrations above the LOQs, five element (As, Co, Cr, Ni, V) concentrations were below the LOQs and for six elements (Be, Cd, Hg, Pb, Sb, Tl) the concentrations were below the LODs. For the extraction of As species from DBS prior to analysis by HPLC-ICP-MS, Zhang et al. constructed a miniature mixer–spinner device from LEGO® components.⁵⁷ An 80 μL DBS was first mixed with 160 μL of a solution of Triton X-100 and N-ethylmaleimide to disrupt the cells and release the As bound to proteins. Then 12 μL of DCM and 68 μL of 5% HClO₄ were added and the components mixed by the action of a magnetic stirrer in an external magnetic field. Finally the device was switched to centrifuge mode and the As species obtained in the supernatant. The whole process took about 25 min. The method was validated by comparing the results with those obtained for venous whole blood. The researchers also investigated sample stability and concluded that the As species in the DBS are preserved for at least 30 days at 4 °C. To analyse small biological sample masses (10–20 mg) by ICP-MS, Kolotov et al.⁵⁸ constructed miniaturised reaction vessels (internal volume 3 mL, external diameter 1.2 cm with caps). The researchers reported that “three assemblies of these minivessels were placed into a standard EasyPrep (100 cm³) MARS-5 autoclave” and that a “portion of nitric acid (12.5 mL) of analytical grade was added to the autoclave to provide vapour pressure equal to the pressure of acids in the minivessels. The autoclaves with minivessels were then placed on the microwave system rotor.” The digestion was effected with 1.4 mL of high purity HNO₃ and the temperature was ramped to 150 °C over 10 min and then held for 25 min. After cooling, the solutions were diluted to 10 mL with 2% HNO₃, then prior to measurement, further diluted two-fold with 2% HNO₃. The method was validated by the analysis of several CRMs (IAEA-153 powdered milk, INCT-TL-1 Argentine black tea leaves, CTA-OTL-1 oriental tobacco leaves and CTA-VTL-2 Virginia tobacco leaves) for which the results for 41 elements were given. The researchers claimed that as the volume of the sample after the decomposition is minimised, lower LODs are achieved. For a 10 mg sample, the dilution factor is 2000, so it is difficult to see how lower LODs can be obtained, as claimed by the authors. Homogeneity at masses of 10–20 mg was discussed only in relation to CRM.

Considerable efforts have been devoted to finding suitable alternatives to mineral acids for sample preparation prior to quantitative analysis. Two reports from the same research group described methods for the determination of metals by HR-CS-ETAAS in which samples were prepared by emulsion breaking. In the first report,⁵⁹ the researchers described the analyses of several oily dietary supplements for Cd and Pb. To 500 mg of sample 1.5 mL of toluene were added followed, after 20 s agitation, by 2 mL of 5% w/v Triton X-100 solution in 15% v/v HCl. The mixture was stirred for 30 s to obtain the emulsion, which was then broken by centrifugation (5 min at 4000 rpm), whereupon two phases were obtained: an organic phase and an acid aqueous phase containing the extracted metals. Results were compared with those obtained by acid decomposition in an open digester block heating system. To 1 g of sample both HNO₃ and H₂SO₄ were added and after the carbonized material obtained at 200 °C was dissolved in H₂O₂, the heating was continued for a further 2 h. The final volume was 10 mL. The LODs for the method involving the emulsion breaking procedure were 1.5 and 3.3 μg kg⁻¹ for Cd and Pb, respectively. The method was validated by spike additions and applied to the analysis of nine different samples, in only one of which (safflower oil) were Cd and Pb found. In the second report⁶⁰ they described the determination of Al, Cd, Cu, Mo, and Pb in foods for enteral nutrition of chronic renal patients by HR-CS-ETAAS. To 0.1 g of sample 2 mL of a 20% (m/v) Triton X-100–10% (v/v) HNO₃ solution were added. The emulsion, which was formed by vortexing for 2 min, was broken by centrifugation (5000 rpm for 15 min) into three phases: an oily organic phase (upper phase), an aqueous phase containing the extracted metals (intermediate phase), and a surfactant rich phase (lower phase). The intermediate aqueous phase was removed with a micropipette for analysis. The results were compared with those obtained by conventional acid/peroxide MAD in which the analytes in 1 g of sample ended up in a final volume of 15 mL. The authors stated that the accuracy of the method was evaluated through recovery tests, but no details were given. The LODs were 0.022 mg kg⁻¹, 0.0003 mg kg⁻¹, 0.001 mg kg⁻¹, 0.009 mg kg⁻¹ and 0.002 mg kg⁻¹ for Al, Cd, Cu, Mo, and Pb, respectively and all the elements, except Cd, were detected in all 5 samples. The authors indicated that the emulsion-breaking method took 17 min, whereas the MAD took 53 min, but did not indicate how many samples could be processed in parallel by each method. To avoid the spectral interferences encountered in the ICP-OES determination of As in the NMR contrast agent gadobutrol, Bunina et al.⁶¹ devised a separation scheme based on the CPE of As^V, as the arsenomolybdate derivative, with salt-induced phase separation. First, As^III was oxidized to As^V by heating with an oxidizing agent (either OXONE or Chloramine T), then the As^V was extracted. To 10 mL of sample solution at pH 2.5 were added 100 μL of 5 g L⁻¹ ammonium molybdate solution and 50 g L⁻¹ Triton X-100 solution (volume not given) and the solution mixed (no details given). Then 2 g of (NH₄)₂SO₄ was added with “intensive mixing” (again no details given) and after centrifuging (3000 rpm for 30 min) the bottom aqueous phase was separated, the remaining micellar phase acidified with 100 μL of conc. HNO₃, diluted to 10 mL, and the As content determined. The researchers indicated that sample volumes of up to 50 mL could be used, and thus the procedure had the potential for pre-concentration, for which the LOD was 0.6 μg L⁻¹. The method was validated by spike recoveries from one gadobutrol sample, though it was not clear whether the procedure was able to detect any As in the sample. The authors noted the possible application to iAs speciation analysis. Guimarães et al.⁶² determined only one analyte (As) in medicinal herbs by ICP-MS but investigated two possible extraction procedures with DESs based on β-alanine, malic acid, citric acid, xylitol, and water: an UAE matrix solid-phase dispersion procedure and a MAE procedure. It is impossible to decipher what the optimum conditions were for each procedure from the text; however, the abstract contains some relevant information. The general outline of the UAE procedure is that sample (50–150 mg) was macerated with the powdered DES precursors (β-alanine and citric acid, masses not given) in a mortar and the mixture transferred to a 50 mL conical tube and water added (volume not given) then sonicated in a bath at 40 kHz (50 °C and 60 min) and finally made up to volume in 10 mL. The optimum sample to solvent ratio was given as 10 : 1 mg mL⁻¹. For the MAE procedure, sample (200–400 mg) and 5 mL of prepared DES were mixed and subjected to a microwave program (100 °C and 40 min) and finally made up to 10 mL. The optimum sample to solvent ratio was given as 40 : 1 mg mL⁻¹. Results were compared with those obtained from a conventional acid/peroxide MAD, in which the As in 100 mg of sample ended up in 25 mL. The methods were validated by the analysis of CRMs NIST SRM 1547 (peach leaves) and Brazilian Agricultural Research Corporation (EMBRAPA) E1001a (forage grass) and applied to four real samples: (Foeniculum vulgare Mill.), yerba mate (Ilex paraguariensis), lemongrass (Cymbopogon citratus Stapf.) and chamomile (Matricaria recutita L.), all of which contained As at a concentration above the LOD (0.01–0.04 mg kg⁻¹). Several research groups have devised methods in which analytes were extracted with DESs. For the determination of Ag, Al, Ba, Cd, Cr, Cu, Fe, K, Li, Mg, Mn, Ni and Pb in used cooking oils, olive oils and animal fat by ICP-OES or ICP-MS/MS, Torres et al.⁶³ developed a procedure involving a dispersive liquid–liquid aerosol phase. To 1 g of sample in a 5 mL polypropylene tube was added 1 mL of hexane, and an aerosol, generated by a pneumatic concentric nebuliser, from the DES (ChCl + ethylene glycol) directed against the surface of the solution contained in the plastic tube. The mass of extractant was determined by weighing. On standing for 30 min, the DES formed the lower layer and was removed for analysis. The results were compared with those obtained by either (a) dilution with xylene and microsample introduction into an ICP-OES instrument or (b) vortexing of the sample-DES mixture and introduction into an ICP-MS/MS instrument. For the aerosol extraction method with ICP-OES detection, the LOQs ranged from 0.013 mg kg⁻¹ (Cd, Mn) to 0.53 (Pb) mg kg⁻¹. Accuracy was also assessed by spike recoveries of four elements in one sample. Results were reported for all elements except Ag, Cd, K, Li and Ni, in the 10 or 12 samples. Yang et al.⁶⁴ determined Co, Cr, Cu, Ni, and Zn in the traditional medicinal plant Gentiana rigescens by ICP-OES. Sample (100 mg) was added to 2 mL DES (ChCl + oxalic acid) and the mixture ultrasonicated at 40 °C for 20 min, after which 5 mL of 2 mol L⁻¹ HNO₃ was added. After stirring (5 min) and centrifuging (4000 rpm 5 min), the filtered supernatant was diluted to 10 mL. The LODs ranged from 0.4 (Co) to 2.5 (Cu) μg L⁻¹. The results were compared with those of a MAD procedure, no details of which were provided, and were also validated by spike recoveries. All analytes were detected in one sample. They showed that extraction with boiling water (decoction) was incomplete for all elements except Zn, and that the Cr extracted was <LOD (0.9 μg L⁻¹). The decoction procedure started with 25 g of powder, and the final volume, which was not specified exactly, could have been 300 mL.

Two applications of SPE for sample clean-up have been reported for the determination of Pb in high-salt matrices. For the ETAAS determination of Pb in salt, Ding et al. retained both sodium and Pb on a column of chitosan-coated polystyrene microspheres.⁶⁵ A volume of 5 mL of sample containing not more than 0.3% (m/v) of salt was left in the column overnight and the retained sodium eluted with 10 mL of 0.1 mol L⁻¹ HCl, followed by the elution of Pb with 10 mL of 2 mmol L⁻¹ Na₂EDTA. Other than reporting a LOD of 0.087 μg L⁻¹ and some column recoveries, the researchers provided almost no information about the application of the method to the analysis of real samples. In contrast, Xiao et al.⁶⁶ not only validated their ETAAS method, through interlaboratory testing in six laboratories of QC samples of soy sauce (CFAPA-QC1728-4) and spike recoveries, but also determined Pb in 20 real samples of foods, food additives and dietary supplements. Solid samples were subjected to a conventional acid/peroxide MAD, after which the acid was evaporated, the residue taken up in water and, after pH adjustment, diluted to 25 mL. This was loaded onto a Chelex-100 column (dimensions not given, but possibly containing 0.5 g of resin) at 1 mL min⁻¹ and, after washing, was eluted with 10 mL of 0.16 mol L⁻¹ HNO₃ and finally diluted to 50 mL. All of the loading and elution steps were automated. The LOD was 0.0032 mg kg⁻¹, and Pb was detected in all samples.

There are two reports of pre-concentration by co-precipitation from the same research group. In their first paper, Ozdes et al.⁶⁷ described a method for the FAAS determination of Cu in fruit (sour cherry, mulberry, apple and peach) and waters (stream and sea). The solid samples were first subjected to the usual acid/peroxide MAD (0.75 g of sample was diluted to 50 mL). To a 15 mL subsample (adjusted to pH 8) were added 2 mL of a 0.2% (in EtOH + DMSO, 1 + 1) solution of 2-[5,6-dichloro-2-(2-bromobenzyl)-1H-benzimidazole-1-yl]acetohydrazide (DIBBA) and after waiting for 5 min, the precipitate was collected by centrifugation and dissolved in 1 mL of conc. HNO₃ then further diluted to 2 mL prior to analysis. The LOD was 0.44 μg L⁻¹ and the method was validated by spike recoveries. The analyte was detected in all fruit samples, and in both water samples, provided 100 mL of sample was taken. The researchers described the method as “carrier element free”, which is true in the sense that no additional elements were added; however, the researchers did not discuss the nature of the precipitate, which could have been the product of reaction between DIBBA and other elements in the sample as well as excess of the reagent. It should be noted that the reagent is not commercially available. In the second paper, Duran et al.⁶⁸ reported a very similar procedure for the FAAS determination of Cu and Cd in vegetables (lettuce, parsley, carrot and tomato) and waters (stream and sea), in which the precipitant (also not commercially available) was 2-(2-(2-(4-bromobenzyl)-1H-benzo[d]imidazole-1-yl)acetyl)-N-ethylhydrazine-1-carbothioamide (BIMANEC), added as a 0.2% solution in EtOH and DMSO (1 + 1). The LODs were 2.3 and 0.51 μg L⁻¹ for Cd and Cu, respectively. As before, validation was by spike recoveries and both analytes were found in all solid samples, but in the waters only when 100 mL was taken. There was no discussion of the relative merits of the two methods.

Several procedures involving ultrasound-assisted extraction have been developed. For the ICP-OES determination of Ca, Cu, Fe, K Mg, Mn, P, S, in guarana, Farias et al.⁶⁹ devised an ultrasonic bath procedure in which 300 mg of sample was extracted with 10 mL of 1.4 mol L⁻¹ HNO₃ at 60 °C, with 100% power applied for 10 min. After centrifuging, a 1 mL sub-sample was diluted to 10 mL (for the determination of K, Mg, P and S) or to 2.5 mL (for the determination of Cu, Fe and Mn). Results were compared with those obtained by MAD of 300 mg with 5.0 mL of HNO₃ and 2 mL H₂O₂, with dilution to a final volume of 25 mL. Mayotha et al.⁷⁰ developed a similar procedure for the FAAS determination of Cd and Pb in sea grapes and some edible seaweed products, but with three steps. After an initial 5 min pre-sonication step in which 2 mL of a HNO₃–H₂O₂ mixture was added to 100–150 mg of sample, the vessel was heated to 60 °C in an ultrasonic bath for 10 min, then a further 1 mL of the acid–peroxide mixture was added, followed by a further 10 min of irradiation. This step was repeated and the solution finally diluted to 10 mL. Results were compared with those obtained by a three-step digestion on a hot plate: 500 mg sample and 10 mL of conc. HNO₃, overnight stand followed by heating (temperature not specified) for 1 h, then 1 mL of conc. H₂O₂ was added and the heating continued for 20 min, whereupon this step was repeated. The final volume was 50 mL. The LODs were 0.04 mg kg⁻¹ and 0.59 mg kg⁻¹ for Cd and Pb, respectively, low enough so that Cd was found in all 24 samples, whereas Pb could not be detected in 9 samples. The researchers pointed out that the sequential ultrasound-assisted digestion has the advantages of ease of operation, short operation time, and high sample throughput as more than a dozen samples could be treated simultaneously with a simple, inexpensive, temperature-controlled ultrasonic bath. For the FAAS determination of Cu in bovine and ovine liver, Iaquinta et al.⁷¹ compared four extraction procedures: extraction with dilute HNO₃ or dilute TMAH, assisted by an ultrasonic bath or an ultrasonic probe. All procedures were considered satisfactory; however, the two methods involving the probe were more efficient and faster, and were selected for subsequent validation, which included the analysis of a CRM (NIST SRM 1577c bovine liver). The researchers measured particle size distributions and found a notable decrease in particle size with increasing sonication time when the probe was used: the 250 μm in the control sample decreased to 90 μm in 0.36 mol L⁻¹ TMAH after 10 min, which resulted in complete dissolution in 20 min (about half the time needed when the bath was used).

Hao et al.⁷² used a procedure described as “graphite digestion” for sample preparation in the determination of As, Cd, Co, Cu, Li, Hg, Ni, Pb, Sb, and V in parecoxib sodium (a widely used parenteral cyclooxygenase-2 selective inhibitor to relieve acute postoperative pain following gynecologic laparotomy surgery) by ICP-MS. The procedure turns out to involve heating a “digestion tank” containing 0.05 g of sample, 3 mL HNO₃ and 3 mL of H₂O₂ “on a graphite furnace” to 100 °C for 20 min followed by dilution to 25 mL. To prevent the loss of Hg, 20 μL of a 1000 mg L⁻¹ gold standard solution was added before heating. The LOQs ranged from 0.1 μg L⁻¹ (Cd) to 15 μg L⁻¹ (Cu) and the method was validated by spike recoveries. The results for real samples are ambiguous; it quite possible that no analytes were detected in any of the samples.

A method involving pulsed electric field-assisted extraction has been developed⁷³ for the determination of carbohydrates, proteins, anti-oxidants and Ca, Fe, Mg, P, Se and Zn in mushrooms (A. bisporus) by ICP-MS. The authors described the procedure as the application of electrical pulses between two electrodes inside the treatment chamber, creating micropores in eukaryotic cell membranes that increased cell permeability allowing for the selective extraction of intracellular compounds. A commercially available device was used with a 900 mL extraction chamber to which were added 20 g of fresh sliced sample and 200 mL of water. The distance between the electrodes was 10 cm and 100 ms pulses at 2 Hz were applied. The optimum conditions were an electric field strength of 2.5 kV cm⁻¹, specific energy of 50 kJ kg⁻¹ for 6 h. Then for the determination of the elements, 1 mL of the extract was subjected to a conventional acid/peroxide MAD and diluted to 5 mL, of which 100 μL was further diluted to 100 mL. To examine the efficiency of the electric field method, 10 mg of the original sample was digested in the same way and diluted to 10 mL with subsequent dilution of 100 μL to 9.1 mL. They concluded that the electric field method was not suitable for the extraction of Ca and Zn.

A number of research groups have reported non-chromatographic speciation methods for which sample preparation procedures were an integral part. Chirita et al. reported the results of a comprehensive evaluation of possible procedures for the determination of total As and iAs by HG-CS-QT-AAS in various foodstuffs.⁷⁴ All of the methods investigated had been previously published, including a method for iAs that included LLE of the trichloride into toluene (a modification to the published method that had used chloroform) followed by back extraction into dilute (1 mol L⁻¹) HCl. Total As was determined following a conventional acid/peroxide MAD in which 0.2–0.5 g of sample was eventually diluted to 25 mL. Heating with L-cysteine under carefully controlled pH conditions reduced As^V to As^III. The researchers also optimized the generation and hydride transport processes, finding that purging the reaction cell with Ar before the addition of borohydride, to remove oxygen, and drying of the Ar–arsine stream produced substantial improvements in both sensitivity and repeatability. For the determination of iAs, three extraction procedures were evaluated: extraction with 10 mol L⁻¹ HCl (with vortexing), extraction with 0.28 mol L⁻¹ HNO₃ (95 °C in a water bath) and extraction with 0.01 mol L⁻¹ HCl (90 °C in a water bath). Two variations of the 10 mol L⁻¹ HCl extraction were examined: both involved the LLE clean-up, but involved different reductants—either hydrazine hydrochloride or L-cysteine to convert As^V to As^III. The researchers concluded that the additional complexity of the LLE step was not necessary, and that satisfactory results were obtained with both of the other extraction steps. The methods were validated by the analysis of eight CRMs, IRMM ERM-BB422 (fish muscle), BCR-627 (tuna fish tissue), ERM-CE278k (mussel tissue), ERM-BC211 (rice); NIST SRM 2976 (mussel tissue), Institute of Nuclear Chemistry and Technology CS-M-3 (mushroom powder), NRCC Tort-2 (lobster hepatopancreas) and IAEA-359 (cabbage), and applied to real samples of fish muscle, pork and chicken meat and liver, rice and infant food preparations with rice, meat and vegetables. Both species were found in all samples at concentrations above the LODs, which were 0.0044 mg kg⁻¹ and 0.0022 mg kg⁻¹, for total As and iAs, respectively. The fates of methylated forms of As (and of AB) were not directly investigated, the researchers assumed that these (including DMA, known to be present in rice) did not react in the derivatization step. For the determination of Sb species in water, tea and honey by ICP-MS, Oviedo et al.⁷⁵ devised a DSPME procedure that was selective for Sb^III. The Sb^III DDTC complex in 8 mL of pretreated sample was retained on 3 mg of GO, followed by the addition of 40 μL of the MIL trihexyl(tetradecyl)phosphonium hexachlorodysprosiate(III), instantly magnetizing the sorbent surface allowing its separation with a magnetic rod. The aqueous phase was then analysed for the remaining Sb^V; total Sb was determined by direct introduction of the sample into the ICP-MS instrument, and hence the Sb^III concentration was calculated by difference. Honey was diluted 1 + 99 with water, and tea infusions were prepared by extraction of 2 g with 200 mL of hot water. The method was validated by the analysis of NIST SRM 1643e (trace elements in water) and by spike recoveries. The LODs were 5 ng L⁻¹ and 3 ng L⁻¹ for Sb^III and Sb^V, respectively and both species were found in all eight samples. The concentrations in the original tea samples were incorrectly calculated, based on the extraction of 2 g with 200 mL. For the determination of inorganic species of As, Cr and Se in high-salt matrices by ICP-MS, Chen et al. developed a SPE procedure in which selectivity was obtained by pH control.⁷⁶ A 3-D printed monolithic column of polyamide was modified with TiO₂ NPs, and samples were loaded at pH 2 (to selectively retain As^V and Cr^VI, but both Se^IV and Se^VI), pH 6 (to selectively retain Se^IV) or pH 12 (to selectively retain As^III and Cr^III). The loading, rinsing, elution and washing steps were automated in a two-column system with appropriate switching valves. The method was validated by spike recoveries from real samples (sea, river and, agricultural waste waters, and human urine) and by analyses of CRM NTCC CASS-4 (nearshore seawater), SLRS-5 (river water), NIST SRM 1643f (fresh water), and Seronorm Trace Elements Urine L-2 (human urine). The LODs ranged from 0.7 ng L⁻¹ (As^V) to 32 ng L⁻¹ (Se^IV). No Cr^VI or Se^VI were found in any of the CRMs, Cr^VI was not found in any of the real samples, and Se^VI was found only in the agricultural waste water. To determine non-ceruloplasmin-bound Cu (NCBCu) and total Cu in biological fluids, Lu et al.⁷⁷ devised a batch SPE procedure in which the NCBCu in a 20 μL serum sample (diluted with 170 μL of buffer), was retained on dendritic spherical silica particles functionalized with EDTA, and then eluted with 100 μL of 2 mol L⁻¹ HNO₃, 30 μL of which was diluted with 270 μL prior to ICP-MS detection. For total Cu, the ceruloplasmin-bound Cu was released by reaction with 200 mmol L⁻¹ H₂O₂ for distribution among other ligands, which was then efficiently retained by the adsorbent. The LODs were 0.064 μmol L⁻¹ (NCBCu) and 0.12 μmol L⁻¹ (total), i.e. 4.1 μg L⁻¹ and 7.6 μg L⁻¹, respectively. The method was applied to the analysis of serum samples from several healthy subjects and several cancer patients. An approach to the same analysis by Tuchtenhagen et al. involved precipitation of the “exchangeable Cu” (i.e. the labile fraction not bound to ceruloplasmin).⁷⁸ To 50 μL of serum were added 50 μL of 5 g L⁻¹ EDTA solution then, after standing for 1 h, 100 μL of ACN at −20 °C was carefully layered on top. After vortexing for 10 s, the suspension was centrifuged for 10 min (20 000g and 4 °C). To avoid unintentional carryover of sedimented proteins, half of the supernatant (100 μL) was carefully removed, the solvent removed via vacuum centrifugation and the residue taken up in 50 μL of carrier solution (1 mmol L⁻¹ EDTA at pH 7) for measurement by FI-ICP-MS/MS, for which the sample injection volume was 10 μL. The LOD was 5 μg L⁻¹ and the method was applied to the analysis of the serum of five healthy volunteers. For the determination of Se species in a Se-enriched edible fungus (Ganoderma lucidum), Shi et al. investigated 12 extraction methods.⁷⁹ The material was first ground with water in a ball mill, the sediment collected and subjected to further extraction either with buffers alone (5 were investigated) or with buffers and enzymes (trypsin, protease K and XIV, pepsin); both two- and three-step procedures were investigated. Total Se was determined by HG-AFS and Se species were determined by anion-exchange HPLC-AFS (both methods were previously published). They found that the highest extraction efficiency of 65% was obtained with a three-step procedure: deionised water extraction, pepsin extraction in 0.05 mol L⁻¹ glycine buffer (pH 2.1), trypsin extraction in 0.05 mol L⁻¹ Tris–HCl buffer (pH 7.5). The total Se method was validated by the analysis of a CRM (GSV-1 shrub leaves) and the LODs were 2 μg L⁻¹, 2 μg L⁻¹, 10 μg L⁻¹ and 5 μg L⁻¹ for Se^IV, SeCys₂, SeMet, and MeSeCys, respectively. All species were found in the sample. To assess the potential bioavailability of Mn in hot infusions of green and roasted mate (Ilex paraguariensis), Leoncio and Garcia measured (FAAS) the extent to which Mn was removed from the infusion by a batch SPE procedure with Chelex-100.⁸⁰ The infusions were prepared by heating 0.5 g of samples with 40.0 mL of hot water, (80 °C) for 3 min on a water bath. Total Mn was determined following a hot plate digestion with HNO₃ and H₂O₂ in which 0.5 g of sample was diluted to 50 mL. For the fractionation study, 40 mL of the infusion and 0.45 g of Chelex-100 in the ammonium form were stirred for 5 min and the Mn remaining in the solution determined by FAAS. The methods were validated by spike recoveries and the LODs were 0.04 mg L⁻¹ and 0.07 mg L⁻¹ for the acid digests and hot infusions, respectively. They concluded, albeit somewhat tentatively, that almost all (98.4–99.7%) of the Mn in these mate infusions was potentially bioavailable.

Several of the entries in Table 1 (pre-concentration by LLE or SPE) are summaries of papers describing speciation analysis, including that of As in rice,^81,82 Cr in lettuce and water,^83,84 Fe in beverages,^85,86 Hg in food, water and blood,⁸⁷ Mn in food and water,⁸⁸ Te in beverages,⁸⁹ Tl in beverages,⁹⁰ and V in beverages.⁹¹

Several sample preparation procedures for solid samples have been reported. For the determination of Cd, Cr and Pb in waters by LIBS, You et al. prepared dried films from the samples by first preparing semi-solid agarose hydrogels.⁹² To 5 mL of sample (in this case lake water) in a screw cap bottle were added 0.125 g of agarose powder followed by heating in a water bath for 15 min at 100 °C. Then 3 g of the solution were transferred to a 3.5 cm Petri dish and allowed to cool, when a hydrogel with a smooth surface formed; finally, the water was allowed to evaporate by storage in a well-ventilated environment. The LODs were 0.011 mg L⁻¹, 0.12 mg L⁻¹ and 0.12 mg L⁻¹ for Cd, Cr and Pb, respectively, none of which were found in the one sample analysed (though spike recoveries were satisfactory). The researchers found that it was not possible to make measurements by irradiation of the hydrogel surfaces because of the interference from water. Zhu et al.⁹³ constructed an automated sample preparation and handling device for the LIBS determination of Ca and Mg in water samples (plant and breeding water) in which the analytes were extracted by a cation-exchange membrane (CMI-7000 S from Membrane International Inc. with a thickness of 0.42 mm, an exchange group of SO₃–Na, exchange capacity 1.6 meq g⁻¹, and applicable over the pH range 1–14). A 3 cm² disc was placed in 20 mL of sample solution, which was stirred by a magnetic follower at 600 rpm for 3 min. After rinsing with water, the discs were dried with a stream of warm air before moving into the focal lane of the LIBS optical system. The LODs were 3.6 mg L⁻¹ and 1.6 mg L⁻¹ for Ca and Mg, respectively. Again spike recoveries were satisfactory, but it was not clear whether the analytes were detected in any real samples. Danilov et al.⁹⁴ developed a sample preparation procedure for the TXRF determination of elements in vitamin-mineral supplement tablets, in which 10–20 mg of the powdered sample was uniformly dispersed with the help of ultrasound in 1 mL of ethylene glycol together with Rb as IS. Then 5 μL was placed onto a quartz disc and dried on an electric hotplate at 80 °C for 3 min followed by cooling in air and TXRF analysis. Results were compared with those obtained by ICP-OES after a conventional acid/peroxide MAD in which 0.2 g of sample was diluted to 50 mL. Calibration solutions containing (in total) about 50 elements were prepared, though one of the two samples was examined for only seven and the other for only 17 elements. Elements found by ICP-OES but not detected by TXRF were Mg, Mo and Ni; whereas Cl, P and Se were found by TXRF but not by ICP-OES. LODs were not given, other than a value for Ni of 0.002 mg per tablet by TXRF. The researchers pointed out that compared with MAD, their suspension-and-drying method was simpler and faster. Janeda and Slachcinski devised a microvolume slurry nebulisation procedure for the determination of Ca, Cu, Fe, K, Mg, Mn, Na, Pb, Sr and Zn in teeth and bones by MIP-OES.⁹⁵ After removing any residual soft tissue with H₂O₂, teeth were ground in a ball mill with a tungsten carbide bowl for 2 min so that the material passed through a 32 μm nylon sieve. Slurries (3% m/v) were prepared from 150 mg of sample and 5 mL of water using an ultrasonic homogeniser with a microtip, and 15 μL of the slurry was injected into the microsampling device. The results were compared with those obtained by ICP-OES following complete sample decomposition by acid/peroxide MAD in which 250–300 mg of sample were digested with 1 mL of 30% H₂O₂ and 5 mL of conc. HNO₃ (10 min at 25–40 bar). The method was also validated by the analysis of a CRM (NIST SRM 1486 bone meal). The LODs ranged from 2 μg kg⁻¹ (Na) to 1500 μg kg⁻¹ (Zn). All the analytes were found in both samples, which were teeth from a dog (Jack Russell terrier) and human deciduous teeth.

There has been considerable attention paid to the problems of extracting analytes from seafood. For the speciation of Hg in fish, Cai et al. devised a “one-pot” system that integrated the extraction, derivatization, separation and enrichment of the Hg prior to determination by OES with a DBD microplasma.⁹⁶ The apparatus was able to process 12 samples in parallel, completing the analysis in 50 min. To 100 mg of freeze-dried fish powder or 1 g of minced fresh fish in a quartz reactor (15 mm od, 14 mm id, 65 mm height), 1.5 mL of 25% (m/v) TMAH, 0.5 mL of 0.8% (m/v) SnCl₂ dissolved in 10% (v/v) HCl, and 2 mL of water were added. The vessels were sonicated for 30 min and the Hg⁰ generated from Hg^II trapped on an activated carbon rod in the needle of a syringe inserted through the septum with which the vessel was sealed. The needle was replaced and the MeHg converted to Hg⁰ by UV irradiation and trapped on a second activated carbon rod. The Hg was thermally desorbed by inserting the rod into the base of the microplasma. The method was validated by the analysis of CRM Beijng Wanjia ZKQC8236 (squid powder) and by spike recoveries. Total Hg was determined by ICP-MS after a conventional acid/peroxide MAD, and some measurements were made by HPLC-ICP-MS, but these were mentioned only very briefly in the supplementary information and there was no further information. The LODs were 2 μg kg⁻¹ and 1.2 μg kg⁻¹ for Hg^II and MeHg, respectively in freeze-dried fish. The researchers calculated that the extraction efficiencies of their method were 90% for freeze-dried fish and 86% for fresh fish, and both species were determined in five real fish samples, except for Hg^II in yellow croaker. The potential for making measurements in the field was mentioned repeatedly. For the determination of arsenosugars extracted from algae, Morales-Rodriguez et al.⁹⁷ investigated the role of SPE by anion-exchange. They investigated several options for the loading and elution conditions as well as the extractant itself and concluded that the best performance was obtained with DSC-SAX cartridges, loaded at pH 8 and eluted with 4 mL of 0.5% ammonium formate. Further pre-concentration was achieved by lyophilisation. The As species were then determined by HPLC-ICP-MS. The procedure was applied to the analysis of commercial Fucus vesiculosus (bladder wrack) dietary supplement tablets, which were finely powdered in an agate mortar, 0.25 g of which were then mixed with 10 mL of water in an end-over-end shaker at 30 rpm for 16 h at room temperature. The suspensions were centrifuged at 3000 rpm for 20 min and supernatant extract filtered through a 0.45 μm nylon filter and kept at 4 °C until analysis. Three anionic arsenosugars were the predominant As compounds in the sample extracts, accounting for 85% of the extracted As. LODs were not reported. For the determination of I species in seaweed, Jerse et al.⁹⁸ devised an HPLC-ICP-QQQ-MS method in which species were extracted in a two-step procedure, chosen after a comprehensive evaluation of several possible procedures, in which 0.3 g of dry sample was first mixed with 40 mg of pancreatin and 5 mL of water in a 50 mL tube in a shaking water bath at 37 °C and 200 rpm for 15 h. Then, 100 mL of 25% TMAH was added and the tube heated (conventional oven) at 90 °C for 3 h. Finally the contents were diluted to 50 mL, centrifuged and filtered prior to injection (5 μL) into the HPLC system. They showed that in this procedure iodate was converted to iodide, and thus all inorganic I was determined as iodide. The LODs ranged from 0.01 ng mL⁻¹ to 0.04 ng mL⁻¹. Total I was determined following TMAH extraction at 90 °C and direct introduction to the spectrometer. The procedures were applied to a range of brown (6 samples), red (6 samples) and green (3 samples) seaweeds. Iodide was the predominant species in all seaweeds, with monoiodotyrosine and diiodotyrosine as minor constituents of most seaweeds. Comparing the sum of species with the totals for each sample revealed extraction efficiencies that were a function of sample type, varying from 66–100% for the brown samples to 12–45% for the red samples. Peaks for six unknown iodine-containing species were observed. In a study aimed at providing insight into the As metabolism of the seaweed oarwrack (L. digitata), and how the As speciation may fluctuate temporally and spatially within the thallus, Sim et al.⁹⁹ compared the water-soluble speciation before and after drying to determine whether freeze-drying had an impact on the water-soluble speciation. They noted that there are only a few reports of extractions using fresh, undried sample materials. Arsenosugars were extracted with 10 mL of water by shaking (0.2 g of dried sample, 1 g of fresh sample) in a 50 mL tube for 2 h, followed by centrifugation (4000 rpm for 10 min). Finally a 1 mL subsample was further centrifuged (15 000 rpm for 5 min) and analysed directly by HPLC-ICP-MS. A sequential extraction of arsenolipids (polar and non-polar) was applied in which 0.2 g of sample was extracted with two 5 mL portions of heptane that were combined and evaporated. The residue was dried and then extracted with two 5 mL portions of DCM–MeOH. The dried extracts were redissolved in the respective solvents and transferred to digest vessels for the final stage of the sample preparation. The dried residual sample material was extracted with 10 mL of water for 2 h and the resultant slurry centrifuged at 4000 rpm for 15 min. A 1 mL subsample of the water extract was digested, the residue was washed twice and a portion was digested to determine the total As content. Total As was determined following a conventional acid/peroxide MAD in which 0.2 g of sample was diluted to 50 mL. In addition to real samples, a CRM NMIJ 7405-b (hijiki) was also analysed. They found that the extraction efficiency was lower in fresh samples (64–77%) than in freeze-dried (95–116%) from the same month, but that water-soluble, polar As–lipids, and residual As concentrations, were generally higher in February than in May. They concluded that the arsenosugars are not a product of As detoxification, but a by-product of normal biological activity. For the speciation of Hg in seafood (seaweeds, fish and shellfish) Li et al.¹⁰⁰ developed a MAE method in which 0.10 g of dried sample was mixed with 7.0 mL of 100 mmol L⁻¹ HNO₃ + 6.0 mmol L⁻¹ cysteine solution and allowed to stand for 6 h at room temperature. Then, the closed vessel was microwave heated to 120 °C within 10 min and held for 30 min at 120 °C. After cooling, the extract was separated by centrifugation. The residue was immediately extracted again with 5.0 mL of the same solution by the same procedure. The extracts were combined, and filtered (0.45 μm) prior to analysis by HPLC-ICP-MS. The LODs were 2.4 ng g⁻¹, 5.0 ng g⁻¹, 50 ng g⁻¹ in the dried solid samples for Hg^II, MeHg and EtHg, respectively. The method was validated by the analysis of a CRM (Puen Scientific Instrument Company P40219 fish). Total Hg was determined by ICP-MS following MAD in which 0.1 g was heated with 5 mL of 7 mol L⁻¹ HNO₃ at 120 °C for 3 h, and then evaporated to near dryness and the residue re-dissolved in an appropriate volume of 1.0% HNO₃ solution. The researchers also devoted considerable effort to the optimisation of the cation-exchange chromatographic separation, which featured derivatization of the analytes with L-cysteine. The pre-soak procedure ensured that all Hg^II was converted to the L-cysteine complex, and only one peak was observed in the chromatogram. Both Hg^II and MeHg were found in all eight samples (except that MeHg was not detected in the Nori seaweed Porphyra); whereas EtHg was only found in four samples. The extraction efficiencies were >90% for all three species. As a further complication for studies of arsenic speciation in rice and fish for which dilute acid extractions are used, Matsumoto et al.⁴¹ showed that MMA spiked into the extracts was partially converted to monomethylmonothioarsonic acid, which they considered more cytotoxic than MMA. They were also able to prepare the compound by bubbling H₂S through a solution of MMA, and suggested that MMA reacts with sulfur in rice during the extraction. They also reported a similar reaction with DMA, though the rate was slower. Overall, they noted that the extent of transformation during extraction was dependent on parameters such as heating, and that it is very difficult to avoid such transformations. They make no suggestions for how such transformations can be avoided, although they describe the development of error-free procedures as imperative. An international intercomparison study involving eight laboratories⁴⁰ was organised to identify the possible sources of discrepancies in the quantification of several As species in different matrices of biological RMs, including one plant tissue, three marine and threeterrestrial biological tissues. The RMs in question, all supplied by NRCC, were certified for only total As, but results for speciation analyses that are available in the literature, show wide variations. The organisers provided a set of calibration standard solutions to each participating laboratory, who analysed the samples following a common group extraction method (selected by the participants) as well as by following an in-house protocol. In the group method, 0.25 g was extracted in 50 mL vessels (an important consideration) with 10 mL of 1% v/v H₂O₂ (prepared by diluting 16.6 mL of 30% H₂O₂ to a final volume of 500 mL in deionised water) using a pre-heated hot block at 95 °C for 60 min. The supernatant after centrifugation was taken for analysis by HPLC with ICP-MS (or MS/MS) detection. In addition to direct couplings, interfaces involving HG and cryotrapping were used. The in-house extraction protocols all involved heating (to 90 °C or 95 °C) with dilute reagents that were combinations of H₂O, HNO₃, H₂O₂ or trifluoroacetic acid. One procedure involved elevated pressure (50 bar). The organisers found the choice of extraction method had little impact on the speciation results for the plant and terrestrial biological tissues; however, the prescribed extraction method led to a significant decrease in the uncertainties for the more complex marine animal tissues. They also found that significant biases were introduced by insufficient verification of the analyte mass fractions in the calibration standard solutions. The report contains detailed discussions of the results broken down in terms of the sample type, extraction method, and instrumental technique, and is recommended reading for laboratories with an interest in As speciation in biological materials. Researchers should note that the study identified extractant-dependent behaviour of As species in TORT-3 (lobster hepatopancreas) and possibly other similar matrices, whose origin and implications were not clear and which will require further investigation. None-the-less, the plan is to publish combined consensus values as certified reference values of As species in the materials studied.

Costa et al.¹⁰¹ investigated the tricky concept of bioaccessibility of Hg species from seafood in a detailed study of three extraction schemes applied to eight fish or shellfish, including swordfish, blue shark, tuna, black scabbardfish, Atlantic salmon, grey mullet, scallops and clams. They also investigated the effects of cooking (frying, grilling, boiling) on the bioaccessibility from black scabbard fish. The three methods of extraction were (a) Simplified Bioaccessibility Extraction Test (SBET) (Method EPA 9200.1-86), also called the relative bioaccessibility leaching procedure (RBALP), (b) the Relative In vitro Model (RIVM), and (c) Unified Bioaccessibility Method (UBM), which resulted from an effort by the Bioaccessibility Research Group of Europe (BARGE) to harmonise the use of bioaccessibility in human health risk assessment of contaminants in soils. Each of these procedures uses mixtures of chemicals to mimic biological fluids at different stages of consumption and digestion (saliva, gastric, bile and duodenal). Total Hg was quantified by a direct mercury analyser (LECO AMA 254) in which dried samples were thermally decomposed, the evolved Hg⁰ trapped by amalgamation and then released for measurement by AAS. The same instrument was used to determine organic Hg after selective extraction of species, following the addition of KBr and CuSO₄, into toluene and back extraction into aqueous sodium thiosulfate. The methods were validated by the analysis of CRM NRCC DORM-4 (fish protein). IAEA-436 (fish muscle) was digested in parallel with real samples as part of the QC/QA measures. They found that bioaccessible Hg fractions ranged from 10 to nearly 90% of total Hg and increased in predator species (swordfish, blue shark and tuna), and that the Unified Bioaccessibility Method (UBM) provided the highest estimation of Hg bioaccessibility for consumers. The cooking procedures (frying, grilling and boiling) considerably decreased the bioaccessible fraction. A study of the contents and the bioaccessibilities of seven trace elements (As, Cd, Hg, I, Ni, Pb and Se) in 231 seaweed product samples available in Belgium has been made.¹⁰² In addition, As and Hg speciation were measured and the effects of washing investigated. For total element determination, 0.2 g of the dried, ground samples were subjected to MAD with 65% HNO₃ and diluted to 50 mL. To stabilise Hg, gold was added at a concentration of 1 mg L⁻¹. Analyses were made by ICP-MS for which the LODs ranged from 0.98 μg L⁻¹ (I) to 0.007 μg L⁻¹ (Cd), and the accuracy was verified by the analysis of several CRMs, including BCR-279 (sea lettuce), SRM 3232 (kelp powder), CRM-7405a (hijiki) and ERM-CD200 (seaweed). To speciate Hg, MeHg was selectively extracted into toluene with back extraction into L-cysteine solution. For arsenic speciation, samples were subjected to UAE with 50% MeOH in water and the extracted As compounds determined by HPLC-ICP-MS. The bioaccessibilities of the trace elements for 13 selected seaweed products were estimated by incubation with enzymatic mixtures under conditions resembling the gastric and small intestinal phases of human digestion. The samples were first prepared as indicated on the package, e.g. by wetting or boiling, without discarding the liquid. Three simulated fluids were prepared (saliva, gastric and intestinal) and the extractions with simulated saliva and gastric juice were performed sequentially on the same sample, thus two fractions were obtained after 2 h of incubation: a gastric phase and a small intestinal phase. In terms of total concentrations, values for As ranged from 2 μg kg⁻¹ to 186 000 μg kg⁻¹. One protein food supplement, prepared from dried seaweed (Ascophyllum nodosum), contained 32 700 μg kg⁻¹, a result the researchers described as remarkable. They also applied this term to the decrease in concentrations of all elements caused by washing (described in detail in the supplementary information). The results for bioaccessibility were discussed in detail for all elements, and in the small intestine was summarised as As (52%) > Ni (35%) > Pb (27%) > Cd (20%). For As, on average for all 13 samples assessed, 37 ± 16% was released in the stomach and 52 ± 18% in the small intestine. They observed an increase in bioaccessibility on cooking, to such an extent that for one cooked seaweed (S. fusiforme), the iAs in the bioaccessible fraction was 26 μg g⁻¹, a concentration that was considered to be a serious health risks for consumers.

As in previous Updates, papers describing pre-concentration by LLE and SPE are summarised in Table 1. As the number of studies featuring extraction onto nanoparticulate phases has increased substantially this year, these have been collected together as the first 32 entries in the table. To make the tables more reader-friendly, the level of detail provided about the methodology has been decreased compared with that given in last year's Update.¹ Details of the sample amount and the final volume in which the analytes are dissolved are given, so that in conjunction with the values of the LODs given by the researchers, which are always for the solution after pre-concentration, readers can calculate the LODs in the samples taken for analysis. Also provided is information (a) as to whether analytes were detected in the samples and (b) concerning any sample dissolution or digestion. Not given are details of standard procedures such as drying, grinding, filtering, pH adjustment, heating, cooling, or of values of operating parameters relating to MAD, UAE and centrifugation.

Table 1 Pre-concentration by liquid- or solid-phase (micro) extraction

Element	Matrix	Technique	Extraction mode/reagents	Procedure/comments	LOD in μg L^−1 (unless stated otherwise)	Validation	Ref.
Nanomaterial-based methods (N = 32)
Ag, Ag NPs, Au, Au NPs	Water (bottled, tap, river, sea)	ETAAS	SPME on non-functionalised ferrite NPs. Magnetic separation. Elution with 0.1 mol L^−1 thiourea	Analytes separated by pH control. Sample 10 mL, eluent 100 μL. No Au found in any samples or standards, Ag found is five out of seven samples, Ag NPs found in tap and river waters, but not in bottled or sea waters	4 (Ag), 12 (Au) ng L^−1	Environment and Climate Change Canada TM-23.4 (trace element fortified water), TM-25.4 (low level fortified water), and TMDA-62.2 (Lake Ontario fortified water), NIST SRM 1640a (spring water), spike recoveries	240
As, Cd, Cu, Hg, Pb	Seafood (mussels, shellfish, shrimp)	ICP-MS	SPE on functionalized fibrous silica (KCC-1) NPs (thiol or amine groups) coated on a stir bar. Elution with 1 mol L^−1 HNO3 and 2% thiourea	Glass stir bar (10 mm × 0.5 mm). The thiol functionalized bar was selected. Sample (0.5 g). MAD but final volume not given. Method difficult to follow, 5.0 mL of digest placed in 20 mL vial, but then “sample volume” was varied from 5 to 25 mL, with 20 mL reported as optimum. There's ambiguity over whether the eluent (10 mL) contained thiourea. All analytes found in all samples. Extraction efficiency not measured	0.03 (As), 0.08 (Cd), 0.07 (Cu), 0.02 (Hg), 0.04 (Pb) μg kg^−1	GBW10024/GSB-15 (shellfish) GBW10050/GSB-28 (shrimp)	241
iAs	Rice	HGAAS	SPE on an anion-exchange nanofibre membrane in pipette tip. A covalent organic framework was prepared from reaction of 1,3,5-trialdehyde phloroglucinol and 3,3′-dinitrobenzidine. This was derivatized with polyacrylonitrile by electrospinning	Sample mass 0.25 g. MAD with 0.06 mol L^−1 HNO3 + 3% H2O2, which apparently did not convert organic species to inorganic, but final volume not given. 5 mL loaded onto column at pH 6 and eluted with 5 mL of 1.6 mol L^−1 NaOH, which diluted to 10 mL. iAs found in all four samples, but results were not given. Spike recoveries from SRM are calculated incorrectly. There is confusion as to whether the LOD is 0.015 μg L^−1 (abstract) or μg kg^−1 (text) and hence whether the value applies to the digest or to the rice	0.015	NIST SRM 1568b (rice flour) and spike recoveries	81
As (inorganic and total)	Rice	ICP-OES	Magnetic SPME on Fe3O4 NP modified with dimethyl triamine-pentamethylene phosphonic acid. Elution with 10% (v/v) HNO3	iAs was extracted from 0.500 g sample by 10 mL of 2% (v/v) HNO3 by heating at 90 °C for 1 h. Final volume was 40 mL. The residue after filtering was digested as for total As (0.250 g sample plus 5 mL of 65% (w/w) HNO3 by heating at 120 °C for 3 h, final volume 25 mL). After extraction, the iAs retained from the 40 mL digest was eluted with 2 mL of 10% (v/v) HNO3. iAs was found in all five samples	1 μg kg^−1	NIST SRM 1568b (rice flour), USP-EMBRAPA CRM-Agro Ar_01/2015 (rice) and spike recoveries	82
Cd	Shalgam juice	SQT-FAAS	DSPME on copper-based nanoflowers with elution by 3 mol L^−1 HNO3	Sample volume 30 mL, elution volume 100 μL. Only one sample analysed and result not given	0.14	Spike recoveries	242
Cd	Food (spices, milk powder) and water (tap, waste)	FAAS	Magnetic DSPME on Fe3O4 NP modified with tetraethyl orthosilicate and 3-aminopropyl silane. Elution with 0.1 mol L^−1 HCl	Preparation details not given. Sample volume 20 mL, elution volume 10 mL, but an enhancement factor of 84 was claimed. Analyte only found in waste water	0.17	NIST SRM 1643e (trace elements in water), ERM-CA011 (hard drinking water), NWTMRAIN-04 (simulated rain water), NWTM-15.2 (fortified water) and spike recoveries	243
Cd	Food (tea, coffee, bread, tobacco, radish, spinach), water (tap, waste)	FAAS	Magnetic DSPME on multi-walled CNTs modified with Fe3O4 and an ion-imprinted polymer. Elution with 1 mol L^−1 HNO3	Samples wet digested (references given). Details of analytical procedure impossible to find. Sample volume 250 mL, elution volume 5 mL, but sample mass not given. Analyte found in all samples	1	CRM, NIES 10c (rice flour) and spike recoveries	244
Cd, Pb	Milk	FAAS	Magnetic SPME on restricted-access CNTs. Elution with HNO3	Proteins were excluded by coating the CNTs with a chemically crosslinked bovine serum albumin external layer. Sample volume 45 mL, elution volume 0.4 mL (1.00 mol L^−1 for Cd and 1.26 mol L^−1 for Pb). Not clear whether analytes were found in the sample	3 (Cd), 2 (Pb)	Spike recoveries	245
Cd, Cu, Pb	Water (tap, river, ground) and industrial effluent	ICP-OES	SPE (column) on cellulose NP functionalised with EDTA-linked polyethyleneimine. Elution with 1 M HCl	Sample volume 100 mL (loaded at 8 mL min^−1), elution volume (flow 1 mL min^−1) 5 mL. All analytes found in all samples, except Cd and Cu in river water	0.4	CRM NIES 10C (rice), NIST SRM 1572b (citrus leaves) and spike recoveries	246
Cd, Cr, Cu, Mn, Ni	Rice and water (mineral, river, agricultural waste)	ICP-OES	Thin film (batch) SPME on electrospun nanofibers consisting of a composite of polyacrylonitrile agar/silver NPs. Elution with 2 mol L^−1 HNO3	Rice (1 g of ground material) extracted with 20 mL ACN, 20 mL of liquid samples also taken. Elution volume 500 μL. Not all analytes were found in all samples (of 35 possible results, 17 were <LOD)	0.2 (Cd, Ni), 0.3 (Cu, Mn), 0.5 (Cr)	NIST SRM 1643e (trace elements in water) and spike recovery	247
Cu	Water (tap, river, lake)	FAAS	Magnetic SPME 2,2′-((1E,1′E)-hydrazine-1,2-diylidenebis(methanylylidene)) diphenol coated magnetite NPs. Elution with 1.5 mol L^−1 HNO3	Few details of analytical procedure given apart from sample volume 20 mL, eluent volume 2 mL (and pH and mass of sorbent). Not clear whether any analyte detected in samples as no results were given	1.6 μg mL^−1	Spike recoveries	248
Cu	Food (green tea, black tea, rice powder)	FAAS	SPE on CeO2/SiO2 nanocomposite based on nanosilica rice husk. Elution with 2 mol L^−1 HNO3	No information about sample preparation provided other than MAD. Sample mass, digest volume, volume taken for extraction, and eluent volume not given. Analyte found in all three samples	Not given	Comparison with results obtained by ICP-MS	249
Cu, Pb	Milk	FAAS	Magnetic DSPME (batch) on Fe3O4 NPs modified with Tanacetum leaves extract. Elution with methanol. Then DLLE into 1,1,2,2-tetrachloroethane	Proteins in 40 mL sample precipitated with trichloroacetic acid. Eluted into 1.5 mL of MeOH, to which were added 250 μL of 1,1,2,2-tetrachloroethane followed by injection into 5 mL of H2O with separation by centrifugation. Both analytes were <LOD in all five samples	0.16 (Cu), 0.22 (Pb)	Spike recoveries	250
Cd, Cu, Mn, Pb, Zn	Food (rice, wheat, corn, beans), feed (sheep, cattle, chicken) and the feed raw materials (distiller's grains)	ICP-MS	Magnetic SPE on carboxyl-functionalized Fe3O4@SiO2 elution with 5.0% (v/v) HNO3	Sample (1.00 g) was extracted into 5 mL of 5% (v/v) HNO3, and 1 mL taken for extraction in an automated system with elution into 5% HNO3 (volume not given). Difficult to see how this procedure can be described as preconcentration. Analytes found in all samples except Pb in a rice and a bean sample	0.16 (Cd), 12 (Cu), 3.8 (Mn), 6 (Pb), 2.7 (Zn) ng L^−1	CRM GBW(E) 100377 (brown rice flour). Comparison of results with those obtained with MAD	237
Co, Ni	Food (green tea, black tea, cabbage, mint, potato, parsley, spinach, hazelnut and tomato), juice (apple, grape, orange and peach), water (well, tap, mineral and sea)	FAAS	DSPME on oxidised multiwalled CNTs modified with 3-(2,4-dihydroxyphen-1-ylazo)-1,2,4-triazole. Elution with 1 mol L^−1 HNO3	Solids (1 g wet weight) MAD and diluted to 50 mL. Sample volume 50 mL, elution volume 2 mL. Analytes found in most samples, except tap and mineral waters. For most juices, Ni <LOD	0.3 (Co), 0.6 (Ni)	NIST SRM 1570a (spinach leaves), TMDA-52.3 (fortified water) and spike recoveries	251
Co, Pb	Natural medicine P. polyphylla var. yunnanensis	ICP-OES	DSPME on GO–TiO2–DES (choline chloride and urea) nanocomposite. Elution with 1 mol L^−1 HNO3	Sample mass 0.5 g (dry weight), MAD and made up to 10 mL. Sample volume 50 mL, elution volume 1.5 mL	0.11 (Co), 0.24 (Pb)	Comparison of results with those obtained by ICP-MS and spike recoveries	252
Cu, Hg, Ni, Pb, Zn	Fruit juices (orange, barberry, mango, pomegranate, sour cherry, grape and pineapple), water (tap, waste)	ICP-OES	DSPME on cobalt ferrite NPs functionalized with a DES prepared from choline chloride and p-aminophenol. Elution with 5% (v/v) NH4OH	Sample volume 5 mL. Elution volume 100 μL. Of 70 results, 40 <LOD. Hg, Ni, Pb <LOD in fruit juices, Pb, Zn <LOD in all waste waters and tap water. Cu found in all samples except municipal waste and tap water	0.54 (Cu), 1.3 (Hg), 1.1 (Ni), 0.62 (Pb), 0.96 (Zn)	SPS-WW2 batch 108 (water) and spike recoveries	253
Cu	Milk (fresh, pasteurized, curd, powder, infant formula), water (waste)	FAAS	Magnetic SPE (batch) on Fe3O4@GO–vinyl pyridine ion-imprinted polymer. Elution with 1 mol L^−1 HNO3	Samples MAD, but no details of sample masses or final volume given. For SPE, 300 mL taken and the extracted Cu desorbed into 5 mL. Analyte found in all samples	0.9	SRMs JSS 800-3 (Rompin hematite) and NIES 7 (tea leaves) and spike recoveries	254
Cu	Eggs (preserved)	FAAS	SPE on magnetic (Fe3O4) mesoporous silica microspheres modified with polyethyleneimine. Elution with 1 mol L^−1 HNO3	Samples (1 g of either egg white or yolk or outer coating mixture) ashed and dissolved in HNO3 and diluted to 50 mL, of which 20 mL was taken for extraction. After separation, Cu dissolved in 5 mL of eluent. Analyte found in all samples	0.14 mg kg^−1	Spike recoveries	255
Hg	Food (fish, shrimp, canned tuna), seawater	CV-AAS	Magnetic SPE on MOF composite functionalised with dithizone. Elution with 1.5 mol L^−1 HNO3	Muscle tissue separated and dried and 1.0 g taken for MAD with dilution to 250 mL. Sample volume up to 500 mL, elution volume 1.23 mL. Analyte found in all 10 samples	0.015 μg kg^−1	CRM (details not provided) and spike recoveries	256
Hg	Food (fish, canned tuna fish, black tea, canned beans and edible oil)	CV-AAS	Magnetic SPE on magnetite@Co-MOF-71 modified with 1,3,4-thiadiazole-2,5-dithiol. Elution with 0.7 mol L^−1 HCl	Solid samples (0.5 g) acid digested on hot plate and diluted to 50 mL, all taken for preconcentration, eluent volume 0.8 mL. Analyte found in all 10 samples	0.004	CRM IAEA-436 (tuna fish homogenate) and spike recoveries	257
Hg species	Food (fish, lettuce, spinach, eggplant, tomato, cucumber, pepper and cowpea), water, human blood	CV-AAS	DSPME on pyrrolic/pyridinic nitrogen-doped porous graphene nanostructure dispersed in an IL ([OMIM][PF6]), with 500 μL acetone as dispersant. Elution with 0.2 mol L^−1 HNO3	Speciation was by selective extraction of iHg. For solids, sample mass 1 g MAD and diluted to 20 mL, which was then extracted, and the iHg eluted into 0.5 mL eluent. Mercury species found in all 22 samples, except organic Hg in one serum sample	2.6 (blood), 1.2 (food, water) ng L^−1	CRM NIST, SRM 1946 (Lake Superior fish tissue), CRM ERM CE464 (mercury in tuna fish) SRM 1570a (spinach leaves), 1573a (tomato leaves), SRM 955c, (toxic elements in caprine blood)	87
Mn	Tea (jasmine)	SQT-FAAS	Magnetic DSPE on a colloidal gel (MCG), formed by mixing cobalt magnetic NPs and a DES (choline chloride and urea). Elution with 0.2 mol L^−1 HNO3	Hot water extract (volume not given) of 5.5 g of tea. Sample volume 8 mL, elution volume 100 μL. Mixing was by passing sample and extractant from one syringe to another via a custom 3D printed chamber. Analyte found in extracts of all three samples	4	None, other than use of standards addition calibration	258
Mn (total, Mn^II, Mn^VII)	Food (almond, white rice, brown rice, spinach, apple leaves, beans, pineapple, carrot, broccoli, bran, red pepper, beetroot, radish, celery, mutton, fish, dark chocolate) and water (drinking, well and waste)	QT-AAS	DSPME on chloro-functionalized multi-walled CNTs derivatized with N-acetylcysteine. Elution with 0.5 mol L^−1 HNO3 followed by 0.5 mol L^−1 NaOH	Total Mn determined in solids (mass 1 g) after microwave-induced oxygen combustion and UV digestion, dilution to 50 mL, with eluent volume 250 μL. Speciation only for water samples based on pH-selective extraction. Analyte found in all food samples. Both species found in all water samples except for Mn^VII in drinking water	0.12 (Mn^II), 0.14 (Mn^VII)	CRM NIST SRM 1640a (spring water) 1570a (spinach leaves) 1548a (typical diet) 1568b (rice flour) 1515 (apple leaves) and spike recoveries	88
Pb	Food (green tea, sage tea, kefir, oat, olive leaves, canned tuna fish, garlic, table salt), water (tap, sea), tobacco, multi-vitamin tablet	CS-FAAS	Magnetic DSPE on multi-walled CMTs modified with MgAl2O4 and TiO2. Elution with 3 mol L^−1 HNO3	Solid samples (1 g) dissolved in acid on hot plate, diluted to 50 mL. For tap and seawaters 5 mL were taken, for salt 0.5 g was dissolved in 10 mL water. Sample volume 50 mL (or 5 or 10 mL), eluent in all cases 0.5 mL. Analyte found in all samples except, tap water, sea water and table salt	0.42	CRM TMDA 64.3 (fortified lake water) and NCS DC 73349 (bush, branches and leaves) and spike recoveries	259
Pb	Food (biscuit, infant formula, sumac, powdered chili pepper, oat flour, dried chili pepper, garlic), tobacco, water (tap, waste)	CS-FAAS	DSPME on Ag-modified ZnO nanoflowers. Elution with 0.25 mol L^−1 HNO3	Solid samples 0.5 g acid digested in beaker on a hotplate with final volume 5 mL, liquid samples (40 mL), eluent volume 0.4 mL. Analyte found in all samples except, tap water and biscuit	8.5	CRM GBW07424 (GSS-10) (soil),G BW07425 (GSS-11) (soil) and spike recoveries	260
Pb	Food (spinach, tomato, eggplant, onion, honey (four varieties), grilled meat, chicken and fish), water (mineral, bottled, well, waste)	FAAS	LLME of sulfadiazine complex into nano-structured supramolecular solvent (375 μL 1-octanol and 225 μL THF)	Water samples (100 mL) evaporated to 10 mL. Solid samples (1 g) MAD and diluted to 10 mL. For extraction 10 mL were taken and extracted with 600 μL of the supramolecular solvent. Analyte found in eight of the 14 samples (not found in mineral water, bottled water, two honeys, onion and eggplant)	0.15	CRM NIST SRM 1548a (typical diet), SRM 1643e (trace elements in water) and spike recoveries	261
Pb	Tea extract, tap water	SQT-FAAS	SPME on cobalt nanoleaves. Elution with 5.0 mol L^−1 HNO3	Blue butterfly tea (1 g) extracted with 1 L drinking water. The extract diluted 10-times. Sample volume 30 mL, eluent 75 μL. Analyte found in two tea samples, but not clear if detected in tap water	2.4 μg kg^−1	Spike recoveries from one tap water sample and standard additions calibration for the tea	262
Pb	Water (tap, river, packaged drinking)	ICP-OES	SPE on a column of a GO/Fe3O4@beta-lactoglobulin nanocomposite. Elution with 1.0 mol L^−1 HCl	Sample volume 100 mL, eluent volume 3 mL. Analyte only found in river water. No details of sample preparation for CRMs given	0.3	CRM NIST SRM 1572 (citrus leaves), NIES 8 (vehicle exhaust particulates) and spike recoveries	263
Pb	Food (lettuce), water (tap, mineral)	ETAAS	SPE on magnetic MOF (Fe3O4, copper(II), and benzene-1,3,5-tricarboxylic acid). Elution with 0.5 mol L^−1 HCl	Lettuce (0.5 g) acid digested and diluted to 50 mL, 2 mL of which was taken for preconcentration. Elution volume 500 μL. Standard additions calibration and analyte found in all samples	0.026	Spike recoveries	264
Pb	Juice (soy, whole grape, reconstituted grape, orange nectar)	FAAS with magnetic probe insertion into flame	DSPME (magnetic) with orange peel powder	Samples also analysed by FAAS after mineralization (5 mL wet ashed and made up to 2 mL). For extraction 1 mg of adsorbent was added to 40 mL of sample, collected by an external magnet then removed on the end of a magnetic probe and inserted directly into the flame, producing a rapid transient signal. Not clear if any analyte detected in any samples	3–5	Comparison with results of the mineralization method, and spike recoveries	265
Pb	Water (tap, well, lake, waste)	ETAAS	SPE of DDTC complexes on chitosan-functionalized magnetic (Fe3O4) graphene oxide composite. Elution with 3 mol L^−1 HNO3	To sample (200 mL) were added 90 mg of adsorbent, and after extraction and separation, the Pb was dissolved in 5 mL of 3 mol L^−1 HNO3. Analyte not detected in any samples	0.21	Spike recoveries	266
Methods unrelated to nanomaterials (N = 48)
Cd	Water (tap, mineral, lake, physiological solution)	Thermospray FAAS	Column (1.5 cm × 0.5 cm) SPE with MOF UiO-66. Elution with 1.0 mol L^−1 HCl	Sample volume 10 mL, elution volume 1 mL. Analyte <LOD in all samples except mineral water	0.03	Spike recoveries	267
Cd	Water (tap, spring, lake), mushroom, tobacco, sediment	FAAS	LLE with a DES consisting of tetrabutyl ammonium iodide, tetrabutyl ammonium fluoride, n-pentanol (1 + 1 + 10) and dithizone	Soil CRM 0.2607 g acid digested made up to 8 mL, sediment 0.25 g acid digested made up to 6 mL, tobacco 1.0 g to 8 mL, mushrooms 1 g to 7 mL. Extractant solution 400 μL, after phase separation, mixed with 500 μL of 1 mol L^−1 HNO3 in methanol. Analyte not found in any water samples, but in all solid samples	2.1	CRM UME EnvCRM 03 (elements in soil) and spike recoveries	268
Cd	Edible oils (sunflower, olive, hazelnut)	FAAS	LLME (spray-based) into 2% HNO3	Sample volume 7 mL, extractant volume not given but the mass was 230 mg. No details of spray system given other than a schematic diagram and the country of supplier (Turkey). No analyte detected in any samples	2.9 μg kg^−1	Spike recoveries	269
Cd, Cu, Pb	Beer, urine, water (sea, lake)	ICP-OES	SPE of complexes with DDTP on sol–gel poly(caprolactone)–poly(dimethylsiloxane)–poly(caprolactone)-coated polyester fabric disks. Elution with diisopropyl ketone and back-extraction into 20.0 mmol L^−1 KIO3 in 1.0 mol L^−1 HNO3	Sample digestion (heating block) not described. Sample volume 9 mL, elution volume 400 μL, of back-extractant volume 300 μL. Analytes not found in samples, except for Cu (all samples) and Pb in lake water and beer. CRM IAEA 433 incorrectly described as mussel tissue. Maybe the authors meant IAEA 432	0.25 (Cd), 0.13 (Cu), 0.37 (Pb)	CRM NIST SRM 1643e, (water) IAEA-433 (marine sediment), and Seronorm (trace elements in urine level-1) and spike recoveries	270
Cd, Cu, Pb	Distilled spirits (gin, rum, vodka, tsipouro)	ETAAS	SPE on sol–gel graphene oxide-coated polyester fabric. Elution with MIBK	Samples diluted 1 + 1 with water. Volumes of sample and eluent not given. Of the 24 analytical results, 7 were <LOD	1.9 (Cd), 7.1 (Cu), 17 (Pb)	CRM NIST SRM 1643e, (water) IAEA-433 (marine sediment) BCR 278-R (mussel tissue) and spike recoveries	271
Cd, Hg, Pb	Fish (Tigris river snapper and croaker; Euphrates river perch and carp, and two samples of farmed trout)	ETAAS	LLME with DES of l-menthol and ethylene glycol containing DDTP. Separation by freezing	Sample (0.1 g) MAD and made up to 10 mL, then extracted with 50.0 μL of DES containing 15.0 μL DDTP. Analytes found in all samples	0.05 (Cd), 0.1 (Hg), 0.05 (Pb) μg kg^−1	CRM NRC DORM-2 (dogfish muscle), NIST SRM 2976 (mussel tissue) and spike recoveries	272
Cd, Hg, Zn	Waters (tap, river)	AAS used in method development, final analysis by voltammetry	SPE (column) on an ARA-8p highly basic anion-exchange resin modified with 5-(4-carboxyphenyl-azo) rhodamine. Elution with HCl or HNO3	Sample volume 1 L, eluent 10 mL of 4 mol L^−1 HCl for Cd and Zn, then Hg 25 mL of 3 mol L^−1 HNO3. Analytes found in all samples	Authors give LODs as n × 10^−3 mg L^−1 (Cd, Zn) and n × 10^−5 mg L^−1 (Pb)	Spike recoveries	273
Cd, Pb	Food (rice, apple, onion), water (tap, river)	ETAAS	DSPME on amine-functionalized mesoporous silica KCC-1. Elution with 1.57 mol L^−1 HNO3	Solid samples 1.0 g acid digested and diluted to 10 mL. 10 mL taken for extraction with elution into 1.6 mL. Analytes found in all samples, except Cd in tap water	0.02 (Cd), 0.18 (Pb)	CRM046 (Taiwan clay soil) and spike recoveries	274
Cd, Pb	Food (leek, onion, tomato, eggplant, rice, tea, honey), water (tap)	FAAS	DLLME of methyl violet chelates into a DES (methyltriphenylphosphonium bromide and ethylene glycol)	Solid samples (0.5 g) acid MAD, final volume 100 mL to which were added 700 μL of DES and 300 μL ethanol (as disperser). Analytes found in two tap water samples, but <LOD for others. Analytes found in all vegetables	0.33 (Cd), 1.3 (Pb)	CRM NIST SRM 1547 (peach leaves), SRM 1643e (simulated fresh water) and spike recoveries	275
Cd, Pb, Zn	Vegetables (cucumber, tomato, carrot, lettuce, broccoli), water (tap and mineral)	FAAS	Magnetic DSPME as derivatives with 4-(2-thiazolylazo) resorcinol on toner powder. Elution with 1 mol L^− 1 ethanolic HNO3	Solid samples (250 or 400 mg) acid digested and diluted to 25 mL, all taken for extraction. Eluent volume 1.5 mL. No Cd found in any sample, Pb found in broccoli and tomato, and Zn found in all food samples	0.3 (Cd), 3.6 (Pb), 0.11 (Zn)	Spike recoveries	276
Co	Tea (chamomile)	FAAS	Magnetic SPE on PVA hydrogels with elution into 1 mol L^−1 HNO3	Hot water extraction of 1.0 g with 200 mL. Sample volume 40 mL, after magnetic separation hydrogels dried at 50 °C, then extracted with 150 μL eluent. Results of sample analysis not given. Acceptable recoveries only obtained with standard additions calibration	4.2	Spike recoveries	277
Co	Food (white cabbage, cucumber), water (tap, waste), drug (vitamin B-12 ampoule)	FAAS	Column SPE with polyvinyl chloride modified by 3-(2-thiazolylazo)-2,6-diaminopyridine. Elution with 1.0 mol L^−1 HCl	Solid samples (1.5 g) acid digested on hot plate and made up to 250 mL, 10 mL taken for extraction. Elution volume not specified as eluent delivered directly to nebuliser of spectrometer. Analyte found in both samples	1.3	Spike recoveries	278
Co	Water (bottled)	FAAS	Spray-assisted LLME of derivatives with a new Schiff base into CHCl3	Sample volume 8 mL, eluent mass 0.62 g. After separation and evaporation, the residue was dissolved in 50 μL of conc. HNO3. Analyte not found in samples	2.2	Spike recoveries	279
Cd, Hg, Pb	Medicinal plant infusions (arnica and passion flower)	ICP-MS	DLLME of DDTP complexes into an IL [C8C1Im][NTf2]. Back extraction with conc. HNO3	Samples (3 g) extracted with 150 mL of water. Extractant was 25 μL of IL in 200 μL of diethylcarbonate as dispersant. After phase separation, 750 μL of conc. HNO3 added, diluted after 24 h to 15 mL with water. No results given	0.20 (Cd), 0.078 (Hg), 0.27 (Pb)	Spike recoveries	280
Co, Cu, Ni, Pb, Tl, Zn	Honey	ICP-OES	LLE into a ternary DES (n-butanol and choline chloride : menthol : p-aminophenol) followed by DLLME	Honey (50 g) was mixed with 150 mL deionized water, 5 mL taken and 0.75 mL of n-butanol (as the extraction solvent in LLE and disperser in DLLME) was mixed with 50 μL of ChCl : menthol : p-aminophenol DES (as the complexing agent in LLE and extraction solvent in DLLME). Then 0.5 g NaCl (as the phase separation agent) was added and the organic phase (consisting of n-butanol and DES and the extracted analytes) was injected into 5 mL deionized water when the n-butanol dissolved, leaving the analytes in the DES. In 10 samples, Cu and Zn were found in all, and Pb was found in three	0.24 (Co), 0.49 (Cu), 0.69 (Ni), 0.42 (Pb), 0.64 (Tl), 0.74 (Zn) ng g^−1	Spike recoveries	281
Co, Ni	Juice (pineapple, cherry, fruit cocktail), water (dam, well, river)	FAAS	DLLME of complexes with DDTC into DES (ChCl and phenylacetic acid) with DES dispersive solvent (ChCl and butyric acid)	Sample (5 mL), 750 μL of the dispersive DES + 280 μL of the extraction DES rapidly injected after phase separation approximately 208 μL obtained of which 100 μL was taken for analysis. Co found only in dam water and pineapple juice, Ni found in well water, dam water and cherry juice	0.10 (Co), 0.17 (Ni)	CRM SPS-WW2 (surface water) and spike recoveries	282
Co, Ni, Pb	Food (dill, coriander, spinach), water (tap, mineral, river, sea)	FAAS	SPE (stir bar made from face mask) of complexes with 2-(5-bromo-2-pyridylazo)-5-diethylaminophenol on the meltblown layer modified by DES. Elution with 1 mol L^−1 HNO3 in THF	Solid samples (2 g) hot-plate digested and made up to 100 mL, all of which was taken for extraction. After extraction, the solvent bar was removed and analytes dissolved in 4 mL of eluent. In foods, Ni was found in all samples, Co in two out of three, but lead was <LOD. In waters, no analytes found in tap or mineral waters, and Pb <LOD in river and seawater	0.91 (Co), 0.61 (Ni), 2.3 (Pb)	CRM SPS-SW2 (surface water), QC-1187 (trace metals ICP-MS sample 1) and spike recoveries	283
Cr (Cr^III and total Cr)	Lettuce, water (tap, waste, lake)	FAAS	DSPME on chlorosulfonated polystyrene resin modified with 5-amino-1,3,4-thiadiazole-2-thiol. Elution with 2 mol L^−1 HNO3	Solid sample (0.3 g) extracted with 4.5 mL of the artificial saliva solution (prepared according to the unified bioaccessibility method of the BARGE) by shaking at 37 °C for 15 min. Sample volume 20 mL, elution with 5 mL. Reduction of Cr^VI by 2 mL of 1% (v/v) NH2OH/HCl and 500 μL of 1 mol L^−1 HNO3 (30 min at room temperature). Neither species found in tap and waste waters or in lettuce	2.9	CRM TMDA-70 (fortified lake water) and spike recoveries	83
Cr^VI	Water (tap, surface, ground, waste)	FAAS	LLME into IL (1-(2-(quinolin-8-yloxy)ethyl)pyrrolidinium chloride) with NH4PF6 as counter ion	Sample volume 5 mL, 150 mg of extractant (volume not given). Analyte not found in any samples. Authors given in Web of Science database do not appear on article at journal website	5.5 ng L^−1	Spike recoveries	84
Cu	Water (bottled)	FAAS	SPE (column) on a chitosan:alginate:carbon composite. Elution with 1 mol L^−1 Na2EDTA	Sample volume 25 mL, eluent volume 5–10 mL. Analyte detected only in bottled water	LOQ 19	Spike recoveries	284
Cu	Vegetable oil (olive, sunflower, rapeseed)	ETAAS	RP LLME of lactic acid complex into a DES (ChCl + lactic acid + water)	Sample volume 4 mL, extractant volume 100 μL, diluted to 500 μL with water prior to analysis. For MAD, 0.5 g sample made up to 25 mL. Analyte <LOD in all samples by both methods	0.1 μg kg^−1	CRM Petroanalytica, (St. Petersburg, Russia) 2046:2016 (oil), comparison of results with those obtained by MAD and ICP-OES. Spike recoveries	285
Cu	Milk, water (lake)	FAAS	SPE (mini-column) on a restricted-access poly(protoporphyrin-co-vinyl pyridine) adsorbent. Elution with 0.4 mol L^−1 HCl	Sample (30 mL) loaded at 14 mL min^−1 and eluted at same flow rate directly into the spectrometer. The exclusion of macromolecules by the polymer was verified by the addition of bovine serum albumin or humic acid. Analyte <LOQ in all samples	0.9, 2.9 (LOQ)	Comparison of results (for milk) with those obtained by MAD. Spike recoveries	286
Cu	Beer	FAAS	DLLME of DDTC complexes into CCl4 (no disperser)	Sample (5 mL), extractant volume 100 μL (ethanol in sample acts as disperser). Analyte found in all 7 samples. For the MAD procedures 2 mL of sample were diluted to 15 mL	3.2 (DMLLE procedure), 8.8 (ICP-MS), 12 (ICP-OES)	Comparison of results with those obtained by MAD and ICP-MS and ICP-OES. Spike recoveries	287
Cu, Mg, Zn	Water (bottled, mineral, river)	FAAS	LLE into three-phase DES (ChCl + urea) with di(2-ethylhexyl)phosphoric acid and salt (K2HPO4)	It was found that 97% of Zn was conc. in the D2EHPA-rich top phase, 98% of Cu was in the DES-rich middle phase, and 98% of Mg was in salt-rich bottom phase. The Zn in the D2EHPA could not be determined directly by FAAS, so was back-extracted into 1 mol L^−1 H2SO4. Sample volume was 8.5 mL, but no details of volumes of extracts or of back-extraction were given. Analytes found in all three samples	1.6 (Cu), 1.6 (Mg), 2.4 (Zn)	CRM NIST SRM 1640a (tea), GBW 07605 (water) and spike recoveries	288
Cu, Pb	Fruit juice (sour cherry, pineapple and peach), water (well, dam, rain)	FAAS	DLLME, of DDTC complexes, after coprecipitation with folic acid, into CCl4 with DMSO as disperser	To 5 mL sample, DDTC was added at pH 8, folic acid was added and the pH decreased to 3. The precipitate was separated after centrifugation and dissolved in 1.5 mL DMSO. 250 μL of CCl4 added and mixture dispersed in water. After centrifugation the CCl4 layer (approx. 205 μL) taken for analysis. Cu found only in well and dam water, Pb found in all three water samples. Both analytes <LOD in fruit juices	0.08 (Cu), 0.07 (Pb)	CRM SPS-WW2 batch 108 (surface water) and spike recoveries	289
Cu, Pb, Zn	Food (spinach, chicken, fish), water (river, sea)	FAAS	SPE (batch) on a composite of sodium aluminium silicate hydrate and (3-aminopropyl)trimethoxysilane modified by thioglycolic acid. Elution with 0.5 mol L^−1 HNO3	The sodium aluminium silicate hydrate was synthesized from rice husk, as the silicon source, and scrap aluminium cans as the aluminium source. Solid samples (0.5 g) MAD and diluted to 50 mL, all taken for extraction. Water (50 mL) taken. After extraction and filtration, analytes desorbed into 5 mL eluent. All analytes found in all five samples	LODs not given	Spike recoveries	290
Fe^II, Fe^III	Beverages (bottled water, green tea, sprite drink and orange juice)	ETAAS	Direct immersion single-drop (CHCl3) microextraction of complex with N-benzoyl-N-phenylhydroxylamine	Speciation based on pH-controlled selective extraction. Sample volume 20 mL, drop volume 15 μL diluted to 100 μL with ethanol. Both species found in all samples, except for Fe^II in bottled water and Fe^III in sprite	0.06	CRM GBW 10017 (milk powder) and spike recoveries	85
Fe^II, Fe^III	Beverage (Chinese yellow rice wine)	ETAAS	Dual direct immersion single-drop (CHCl3) microextraction of selective reaction complexes of 1-(2-pyridylazo)-2-naphthol (PAN) and N-benzoyl-N-phenylhydroxylamine (BPHA) with Fe^II and Fe^III, respectively	Sample (5 mL) heated to 80 °C in ultrasonic bath with a reduced-pressure evaporator to remove ethanol, and diluted ten-fold. Then 20 mL placed in a beaker and two 15 μL chloroform drops containing PAN or BPHA were suspended in the solution from chromatography microsyringes. Following extraction, the drops were diluted to 100 μL and 10 μL taken for analysis. Both analyte species found in four samples. No Fe^II found in CRM due to oxidative digestion conditions	0.067 (Fe^II), 0.054 (Fe^III)	CRM GBW 10010 (rice) and spike recoveries	86
Gd, REEs	Water (tap)	ICP-MS	SPE on a seaFAST-pico column (mixture of EDTA and imino-diacetate)	Samples (60 mL) were acidified with 3 mL of 20% HNO3 (v/v) and were analysed at least 12 days after sampling date to ensure the degradation of Gd-based contrast agents before the preconcentration step. Details of the (previously published) SPE procedure were not given. Analyte found in all three samples from major Polish cities	0.01–0.1 ng L^−1	CRM NRCC SLRS-6 (river water)	291
Hg	Water (tap, surface)	HGAFS	LLE (CPE) of thio-Michler's ketone complex with Triton X-114	Sample volume 50 mL, surfactant-rich phase diluted to 5 mL with 5% (v/v) HCl. Before AFS analysis, tributyl phosphate antifoam agent was added. Analyte detected in all 4 samples	0.003	Spike recoveries	292
Ni	Food (fish, black tea)	ETAAS	DLLME of complex with 1-[4-[(2-hydroxynaphthalen-1-yl)methylideneamino]phenyl]ethanone into orange peel oil	Samples (5 g) acid digested on hot-plate and diluted to 100 mL. Wheat CRM 0.5 g MAD and diluted to 50 mL. Semi-automated system in which 3 mL of sample was pumped into an extraction vessel together with 300 μL of complexing agent dissolved in orange peel oil. Following dispersion aided by N2 gas, and separation by centrifugation, 20 μL taken for analysis. Analyte found in both samples	0.87 ng L^−1	CRM NCS ZC11018 (wheat), comparison of results with those obtained by MAD and ICP-MS. Spike recoveries	293
Ni	Food (cornflakes, sage, green tea, yogurt), water	CS-FAAS	Magnetic SPE of complex with 2-(5-bromo-2-pyridylazo)-5-(diethylamino)phenol on Fe3O4@Diaion HP-2MG resin. Elution with 3 mol L^−1 HNO3	Samples (0.15 g) acid digested on hot-plate and diluted to 20 mL, all of which was taken for extraction, eluent volume 0.4 mL. Analyte <LOD in green tea and yogurt	15	CRM NIST 1573a (tomato leaves) and TMDA-64.3 (fortified water) and spike recoveries	294
Zn	Serum	FAAS	CPE (air-assisted) of the complex with 3,5,7-trihydroxy-2-(4-hydroxyphenyl)-6,8-bis(3-methylbut-2-enyl) chromen-4-one in the presence of Triton X-114	Serum samples (0.2 mL) were wet oxidized (room temp and MAD) and diluted to 5 mL complexing agent, salt and 1 mL of 0.5% (v/v) Triton X-114 were added. After extraction and separation, the surfactant-rich layer was diluted with 0.5 mL of 2.2 mol L^−1 HNO3, the air compressor was used to transform the solution back into two phases, and the top layer taken for analysis. Analyte found in all samples	0.09	CRM (serum) details not given	295
Pb	Food (canned green lettuce, mustard green, corn, cucumber; fresh chili pepper, tomato paste, garlic, onion), water (tap, drinking)	ETAAS	LLME (electromembrane hollow fibre) into DES (ChCl + phenol) diluted with buffer	Samples (0.5 g) MAD and diluted to 50 mL, 30 mL taken for extraction. Extract volume not given, but is calculated from the preconcentration factor of 750 to be 40 μL. Analyte found in all samples	0.011	CRM NRCC SLRS-6 (river water), NIST SRM 1515 (apple leaves) and spike recoveries	296
Pb	Water (tap, canal), beverages (tea, coffee)	LIBS-LIF	SPE on Lewatit MonoPlus TP 207 (macroporous cation-exchange resin with weakly acidic iminodiacetic acid groups)	To 5 mL of sample, 100 mg of resin beads added. After extraction 10 resin pellets were removed dried, and attached to a glass slide. Analyte not detected in any samples	0.88 ng L^−1	Spike recoveries	149
Pb	Medicinal plant (Gentiana rigescens)	ICP-OES	CPE of dithizone complexes in N,N-dimethylcyclohexylamine (DMCHA) and Triton X-100	Sample (0.5 g) MAD and diluted to 10 mL. Then 0.2 mmol L^−1 dithizone added at pH 2 followed by 2 mL of DMCHA and Triton X-100. After separation, the upper organic phase was heated to 120 °C to remove the organic reagent, then diluted to 1.5 mL with water. Analyte found in all three samples	0.11	Comparison of results with those obtained by ICP-MS and spike recoveries	297
Pb	Water (tap, lake)	ICP-OES	CPE of complex with sulfonate-functionalized C-iso-butyl-calix[4]resorcinarene and Triton X-114	To sample (50 mL) was added an unspecified volume of 1.0 g L^−1 C4RS solution, 5% (v/v) TX114 solution and 2 mL Britton–Robinson buffer solution. After extraction and separation, the surfactant-rich phase was diluted to 5 mL with 5% (v/v) HNO3 after the addition of 0.4 mL tributyl phosphate. Analyte not found in any samples	0.6	Spike recoveries	298
Pb	Fruit juice (orange, pear, multi-fruit and herbal)	ETAAS	SPE (online) with biodegradable hybrid material (BHM) composed of Rhodococcus erythropolis AW3 bacteria and Brassica napus hairy roots. Elution with 0.05 mol L^−1 HNO3	Samples were diluted (1 + 1) with 0.5% (v/v) HNO3 and filtered (0.22 mm) pore size PTFE membrane and 5 mL were pumped through a microcolumn containing 0.1 g of the BHM and the retained Pb dissolved in 80 μL of eluent in an air carrier stream and delivered directly to the spectrometer. Analyte found in all eight samples	5 ng L^−1	Spike recoveries	299
Pb	Blood plasma (45 opium users, 45 control subjects)	ETAAS	DLLME of DDTP complexes into DES (l-menthol and (1S)-(+)-camphor-10-sulfonic acid	To 200 μL of plasma were added 500 μL of a mixture of HNO3 (65%) and H2O2 (30%) mixture (3 + 1 v/v), 100 μL of 15% (w/v) zinc sulfate solution, and 70 μL of ACN. After mixing, cooling and centrifuging, the supernatant was transferred and diluted to10 mL with distilled water. For extraction, 0.50 mL of acetone containing 50.0 μL of DES and 80.0 μL of DDTP were then added. Following extraction and phase separation by centrifuging and freezing, 20 μL of the DES was taken for analysis. Analyte found in all 90 samples	0.1	Spike recoveries	300
Pu	Water (drinking, glacier)	ICP-MS	SPE on TK200 resin column. Elution with 0.5 mol L^−1 HCl + 0.1 mol L^−1 HF	Sample (1000 mL) loaded at 15 mL min^−1, elution with 2 mL eluent at 0.5 mL min^−1. O2 reaction gas used and measurements made at m/z 271 (^239PuO2^+) and 272 (^240PuO2^+). Analytes not found in drinking water sample	0.32 (^239Pu), 2.00 (^240Pu) μBq L^−1	CRM IAEA-443 (seawater) and spike recoveries	301
Sb	Beverages (water, coke, orange juice, apple juice, vinegar, lemonade, cherry juice, iced tea and energy drink)	HGAAS	DSPME on polyoleic acid–polystyrene (PoleS) block/graft copolymer. Elution with ACN	Sample preparation not fully described; 5 mL digested with HNO3 and H2O2, but final volume not specified, but could be 50 mL. Volume taken for SPE not clearly specified either. Text says 5 mL, but a figure in supplementary information suggests 100 mL, 105 mg adsorbent, and eluent volume 1 mL. The researchers claim a preconcentration factor of 90. Out of 12 samples, Sb was <LOD in five of them (two bottled waters, bottled orange juice, bottled lemonade and bottled iced tea)	1.5 ng L^−1	CRM SRM-1643e (trace elements in water) and spike recoveries	302
Te^IV, total Te	Beverages (well water, bottled water, green tea, sprite drink and carbonated drink)	ETAAS	DLLME of APDC complex into DES (ChCl + phenol)	To sample (20 mL) adjusted to pH 6.0, 0.1 mL of 0.8 mol L^−1 APDC, 0.3 mL DES, and 0.5 mL of THF were added. After extraction and phase separation, the DES phase was diluted to 0.5 mL with ethanol and 20 μL taken for analysis. To reduce Te^VI to Te^IV, L-cysteine (0.5% w/v) was added and the mixture heated. Te^IV found only in well water, Te^VI <LOD for all samples	3.5 ng L^−1	CRM GBW(E)080548 (water) and spike recoveries	89
Tl^I, Tl^III	Beverages (orange juice, soda, cola and sprite drink)	ETAAS	LLE (single-drops) Tl^I as the dicyclohexano-18-crown-6 (DCH-18-C-6) complex into nitrobenzene; Tl^III as the 1-(2-pyridylazo)-2-naphthol (PAN) complex into 1-dodecanol	Samples were filtered, degassed, diluted five-fold with water and pH adjusted to 6. Tl^III from a 20 mL sample was extracted into a 15 μL drop of 1-dodecanol containing PAN and then, after adding picrate as a counter ion, Tl^I was extracted into a 15 μL drop of nitrobenzene containing DCH-18-C-6. Each drop was diluted to 100 μL with ethanol and 20 μL taken for analysis. Tl^I, was detected in all samples, but Tl^III was detected only in orange juice	8.3 (Tl^I), 6.3 (Tl^III) ng L^−1	CRM GBW 10017 (milk powder) and spike recoveries	90
U	Water (tap, surface, ground, river, lake, sea), wine (red, rose and white), honey (monofloral lime, rapeseed and sunflower)	ICP-OES	SPE (batch) on ion-imprinted polymer for which the PAR complex with U^VI was the template; the monomer was methacrylic acid (MAA), and the crosslinking agent was trimethylolpropane trimethacrylate (TMPTMA). Elution with 2 mol L^−1 HCl	Waters: to 30 mL sample, adjusted to pH 7, 100 mg polymer gel particles were added. After mixing, centrifuging and washing, U was eluted in 2 mL. Wines: same procedure but with 20 mL sample. Honey: the same procedure applied to a 5% diluted solution. Analyte was reported in two wine and two honey samples. To detect U in tap water, ICP-MS was used	0.05 (waters), 0.07 (wines) and 1.0 μg kg^−1 (honey)	Comparison of results with methods based on alpha spectrometry or MAD and ICP-MS	303
V^IV, V^V	Beverages (apple juice, green tea, cola, lemon-lime beverage and soda)	ETAAS	LLE (single-drops) theonyltrifluoroacetone (HTTA) and chloroform, V^IV at pH 4.5, V^V at pH 2.5	To 20 mL sample, pH adjusted to 2.5, a 15 μL drop of HTTA–chloroform was introduced with a chromatography microsyringe. Following the extraction of V^V, the organic phase was diluted to 100 μL (diluent not specified), and 20 μL taken for analysis. For the determination of V^IV, the pH was adjusted to 4.5 and the procedure repeated with 20 mL of the resulting solution. V^IV was found only in apple juice and the CRM; V^V was found in all samples (including the CRM)	3.1 (V^IV), 2.6 (V^V) ng L^−1	CRM GSBZ 50029-94 (water) and spike recoveries	91
V	Food (whole milk, fungi and shitake mushrooms, brown rice, black tea, parsley and oats)	FAAS	LLE of the 2-(5-bromo-2-pyridylazo)-5-(diethylamino) phenol (Br-PADAP) complex into an aqueous two-phase system formed from (NH4)2SO4 + PEG400 + H2O	Samples (0.5 g), digested with HNO3 and H2O2, and diluted to 10 mL. The extraction procedure is impossible to understand. Analyte <LOD in all samples	0.022 μg kg^−1	CRM NIST SRM 1515 (apple leaves) and spike recoveries	304
Zn	Serum	FAAS	LLE (air-assisted CPE) of complex with 3,5,7-trihydroxy-2-(4-hydroxyphenyl)-6,8-bis(3-methylbut-2-enyl) chromen-4-one with Triton X-114	To 0.2 mL of serum was added conc. HNO3 + H2O2. After 10 min at room temperature, samples were heated in a microwave oven for 3 min and diluted to 5 mL. About 1 mL of 1.0 × 10^−4 mol L^−1 complexing agent, 1 mL of 0.5 mol L^−1 salting-out, and 1 mL of 0.5% (v/v) Triton X-114 were added, and following 1 min air-assisted dispersion, the phases were separated by centrifugation. The aqueous layer was removed after cooling in an ice bath. The surfactant-rich layer containing the zinc complex was diluted with 0.5 mL of 2.2 mol L^−1 HNO3, and the compressor used for 1 min. The supernatant (top layer) was taken for analysis. Analyte found in all 45 samples	0.09	CRM details not clear (blood or maybe lyophilized human serum)	295

---

4. Progress with analytical techniques

4.1 Mass spectrometry

The majority of published advances in ICP-MS this year involved applications in biological materials, reflecting the ongoing interest in use of this methodology in the biomedical field. This was captured in a comprehensive two-part review, where the first part¹⁵ provided an overview of different aspects of ICP-MS and its current biomedical applications, while the second part¹⁶ focussed on newer and potential directions, such as characterisation of novel nanomaterials and combination with immunochemical techniques to expand application to a wider range of biomolecules. Developments in and applications of scICP-MS in particular, received a large amount of attention over the review period with less focus on sample introduction compared with previous years and more emphasis on multi-element/isotope detection along with an interest in applying scICP-MS to real cells rather than experimentally produced cell-lines. A review of scICP-MS was also published,¹⁷ presenting a comprehensive account of instrumentation and sample introduction techniques, hyphenation with other techniques, data manipulation and examples of biomedical applications.

Conventional scICP-MS has the capability to measure only one element concentration at a time in an individual cell, meaning that it is not possible to glean information about multiple elements or isotopes within the same cell. In order to explore multi-isotope/element determination with sc/spICP-MS, Tian et al.¹⁰³ described a systematic comparison of the performance of three techniques, ICP-QMS, ICP-TOF-MS and MC-ICP-MS, to simultaneously detect two isotopes (¹⁰⁷Ag and ¹⁰⁹Ag) in both AgNPs using sp-mode and individual Ag-exposed cyanobacteria cells with sc-mode. The ICP-QMS platforms were used in “peak hopping” mode, which provides an approximation of simultaneous monitoring by alternatively scanning two m/z ratios at a very high frequency. Using optimised dwell and settling times of 100 μs, both isotopes could be detected within a NP or cell up to 85% of the time where the event duration was relatively long (i.e., larger sized NPs or higher intracellular content of cells), and was deemed suitable for qualitative detection of paired events. Perhaps unsurprisingly, the accuracy of the measured isotope ratios was poor (<20% of events had a ratio deviation of <10%), rendering the method unsuitable for quantitative isotopic sp or scICP-MS applications. The MC-ICP-MS platform, with settings of 30 000 cycles and an 8 ms dwell time, achieved excellent isotopic ratio accuracy with a ratio deviation of ≤5% for all events. The main limitations of this type of ICP-MS were low signal intensity, meaning that it was unable to reliably detect events of NPs/cells with only a low elemental mass and that MC-ICP-MS is limited to monitoring isotopes with similar m/z ratios. The most suitable method overall for this purpose was found to be ICP-TOF-MS, using a dwell time of 5 ms, owing to its ability to detect almost all paired events, reasonable accuracy in measuring isotope ratios and capability to acquire a full mass spectrum. Signal intensity was still a limiting factor although this can be expected to improve as more sensitive instruments become available. Tian et al.¹⁰⁴ also utilised the scICP-TOF-MS approach to achieve simultaneous determination of multiple elements in individual human sperm cells. A full mass spectrum was acquired for each cell, however, data analysis was only performed for selected essential and non-essential elements, believed to be important in sperm cells. Cells were identified using P signals and this enabled false positive signals from cell debris to be identified and accounted for. The approach allowed correlations between different elements within individual cells to be explored. Again the disadvantage of using ICP-TOF-MS was inadequate sensitivity to detect some expected elements (e.g., Fe, K and Se).

Application of scICP-MS to cell suspensions from solid tissues was the subject of another study, reported by Garcia et al.¹⁰⁵ This represents a significant step forward in extending the application of scICP-MS to real-life medical applications. The one-step sample preparation protocol, involving an optimised mixture of proteolytic and collagenolytic enzymes, was validated for disaggregation of mouse liver and spleen tissue samples into single intact cells. Cell suspensions were pumped into the ICP-MS using pneumatic nebulisation at a flow rate of 10 μL min⁻¹. Intracellular elemental concentrations of Cu, Fe, P and S were determined sequentially using ICP-QQQ-MS (P and S in O₂ mode (detected as oxides), Cu in He mode and Fe in H₂ mode). The mass of element per cell in the disaggregated tissue cells was compared with that of the in vitro produced cell line, HepG2. Moreover, to ascertain that cell surface receptors were not perturbed by the sample preparation procedure, transmembrane protein transferrin receptor 1 expression levels were assessed in both the cultured and disaggregated tissue cells, through detection of a Nd-labelled antibody marker.

Indirect measurement of biomolecules through immunoassay selection with ICP-MS detection of either tagged or endogenous elements has been discussed in these Updates for a number of years. However, the use of scICP-MS detection for this purpose is a more recent development. One report employed scICP-MS and LA-ICP-MS as complementary techniques to study intracellular content and distribution of two proteins, Apo E and claudin-1, in individual cells from an in vitro cell model, ARPE-19.¹⁰⁶ The highlights of the work were the achievement of a high transport efficiency for the ARPE-19 cells of 85 ± 9%, by way of a microFAST single cell sample introduction system, and use of Ir-nanoclusters, which contain a large number of Ir atoms and therefore act as an amplification system to increase sensitivity for both the scICP-MS and LA-ICP-MS analysis. The LODs achieved were 0.02 fg per cell for Apo E and 3 ag per cell for claudin 1.

Two papers described new methods for isotope ratio determination this year. A high precision double spike-standard addition (DSSA) MC-ICP-MS method was evaluated for the determination of Cd isotopes in animal organs with very low Cd content down to 0.004 μg g⁻¹.¹⁰⁷ The DSSA technique involves adding a certified reference double spike (¹¹¹Cd–¹¹³Cd) standard solution with a known isotope composition to the sample to increase the concentration of analyte. Advantages afforded by the approach include a significantly reduced sample mass requirement to achieve acceptable precision and a simplified single column purification process. A sample fraction of 20% to 50% was found to be optimal and yielded precision for δ^114/110Cd of better than 0.040‰ in ovine and pig kidneys, livers and lungs. A less precise but pragmatic method involved ICP-QMS measurement of ¹³C : ¹²C ratios to quantify [¹³C]glucose that had been extracted from bovine liver tissue using probe-assisted ultrasonication.¹⁰⁸ A substantial part of the paper was devoted to establishing the ICP-QMS instrumental conditions to achieve a stable ¹³C : ¹²C signal that was reliably distinguishable from the background of the natural abundance C. It was necessary to measure a daily ¹³C : ¹²C ratio in an un-treated or un-spiked sample to overcome the inherent mass bias of the ICP-MS instrumental components. The optimised stability of the ¹³C : ¹²C ratio was <3%, maintained over a 6 h run, which although not comparable to performance of other more accurate isotope measurement techniques, was deemed to be fit for the intended purpose in this study. A much improved LOD for ¹³C of 80 μg L⁻¹ was achieved, which was equivalent to 60 μmol L⁻¹ to 170 μmol L⁻¹ [¹³C]glucose in bovine liver tissue the potential of the method was demonstrated by determination of [¹³C]glucose as well as Rb in contracting skeletal mouse muscles.

Determination of radionuclides by MS has been a topic featuring in this section of the Update in recent years with the focus being on development of more rapid methods that require smaller sample volumes than traditional radiometric techniques. Aoki et al.¹⁰⁹ report a significant advance in the measurement of a radionuclide, ⁹⁰Sr, in microvolume (1 μL) biological samples by way of ID-TIMS-MS. Quadrupole energy filtering was employed to reduce background noise, associated with the presence of natural Sr, by a factor of 6.7. The reported LODs were dependent on the natural Sr background but were impressively low considering the small sample volume at 0.0615 ag to 0.39 ag for a 1 μL sample. Isobaric ⁹⁰Zr did not interfere in the analysis when present in sub-mg L⁻¹ concentrations, however, removal using an Sr resin was required prior to ICP-MS detection, in samples containing higher concentrations. Recoveries of ⁹⁰Sr from addition and recovery experiments in tears, saliva, eyelashes and teeth were between 96% and 104%. Results were also in good agreement with those obtained by radiometric analysis.

Improvements in sample introduction systems for ICP-MS to facilitate the analysis of microvolume biological samples remained an area of interest over the review period. Dong et al.¹¹⁰ published a comprehensive evaluation of a novel high-efficiency sample introduction system, consisting of a miniaturised ultrasonic nebulisation (MUN) unit and a bespoke no-waste spray chamber. The authors also designed a power-adjustable MUN circuit board that allowed continuous adjustment of the nebulisation rate between 8 and 300 μL min⁻¹. An almost 100% sample introduction efficiency with a nebulisation rate of 10 to 30 μL min⁻¹ was achieved. Furthermore, sensitivity was increased 10-fold with respect to conventional pneumatic nebulisation and sample consumption was as low as 7 μL even for multi-elemental analysis. Oxide and doubly charged ratios were very low at 0.25% and 0.5% respectively. The LODs achieved for 26 elements with the optimised system were between 0.0005 ng mL⁻¹ (U) and 3.0 ng mL⁻¹ (Mg) and they represented a significant improvement, of one to two orders of magnitude, in absolute LOD vs. pneumatic nebulisation. Satisfactory RSDs of 0.8% to 2.7% on multi-element standards were reported. Analysis of serum, urine and food-related CRMs demonstrated excellent agreement with certified values. In other work, a commercially available micro-flow injection system coupled to a high efficiency total consumption spray chamber was evaluated.¹¹¹ Once optimised in terms of sample volume (25 μL) and uptake rate (25 μL min⁻¹), the system afforded a four-fold enhancement in sensitivity compared with a conventional high volume set up. Upon combination with ICP-QQQ-MS, employing NH₃/He reaction gas to overcome polyatomic interferences, instrumental LODs for serum Co, Cr, Mn, Ni, Ti and V ranged from 0.4 ng L⁻¹ (Co) to 60 ng L⁻¹ (Cr and Ni). The RSDs were <1% under almost all measurement conditions and matrix effects were noted to be reduced. Use of Y as an IS was shown to effectively correct for small differences in sample volume and sample flow rate as well as instrumental variation and the authors highlighted the potential use of the IS correction to enable analysis of calibrators and samples on the same run using different sampling parameters. The largest sample dilution investigated (1 + 24) only required 0.4 μL of neat sample but didn't have adequate sensitivity for determination of low concentrations of Cr in serum and Ni in urine. However, with a 1 + 4 dilution (neat sample volume, 5 μL) satisfactory accuracy for Seronorm serum and urine samples was achieved.

Two different novel approaches were explored to achieve high efficiency sample introduction of biological and food samples. For biological samples, sensitive determination of Cr was performed by way of a low cost, low power, chelate-enhanced nebulised film dielectric barrier (NFDBD) VG sample introduction system coupled to ICP-MS.¹¹² Generation of Cr species using DBD plasma offers advantages of high efficiency sampling and low matrix interference compared with pneumatic nebulisation or CVG. The sample introduction system consisted of an FI valve and the NFDBD vapour generator, followed by two gas–liquid separators for solvent removal. Sampling efficiency of Cr^VI was optimised using the chelate with DDTC, resulting in a 10.5-fold increase in sensitivity with respect to FI pneumatic nebulisation. Matrix effects were sufficiently minimised with use of a blank DDTC subtraction step to allow a standard calibration curve to be employed. An RSD of 1.4% at a Cr^VI concentration of 5 μg L⁻¹ was achieved and the LOD was 0.023 μg L⁻¹; the latter being very comparable or better than other techniques for Cr determination. The method was successfully applied to the determination of total Cr, following oxidation of all Cr to Cr^IV in a biological tissue CRM, Nibea albiflora, as well as number of environmental CRMs. On the other hand, Lan et al.¹¹³ described a direct slurry sampling GF system for food samples, which comprised two novel features. First, a gas turbulator was used to reduce deposition of the analyte on the GF outlet and improve analyte transportation and second, a signal delay device was incorporated between the GF and ICP-MS to overcome the production of transient signals, which is the main hindrance to multi-elemental analysis using ETV-ICP-MS. Furthermore, the modification had a marked effect on precision with an improvement in RSDs in a slurry rice sample from ≤16.2% to ≤2.5%. A non-detrimental reduction in analytical sensitivity was also observed. Following optimisation of the ICP-MS and sample introduction parameters, LODs (0.3 ng g⁻¹ for As and Cd; 0.6 ng g⁻¹ for Pb and 0.5 ng g⁻¹ for Se) were lower than or comparable to previously described ETV-ICP-MS methods. The accuracy of the developed method was validated by analysis of six food CRMs (spinach, celery, carrot, soybean, tea powder and rice powder) and recoveries were between 86% and 118%. The obvious advantage of the method was a very rapid total analysis time of eight minutes, inclusive of sample preparation.

Direct sampling by way of LA-ICP-MS to achieve rapid analysis times and avoid acid digestion, was the focus of two papers over this Update period. The first considered sampling of whole blood using LA-ICP-MS,¹¹⁴ an approach that has a very small sample volume requirement and high tolerance to high salt samples but can suffer from poor precision due to splatter of liquid droplets. To overcome this, Li et al. incorporated a cryogenic ablation cell into their fsLA-ICP-MS set-up, which pre-froze the blood, thus allowing ablation of a solid sample and greatly improving measurement precision (RSDs 1.9 to 6.3% compared with 14.2% to 36.4% with ablation at room temperature). The consequential attenuation of sample signal with ablation at low temperatures was counteracted by use of an ablation cell with a reduced inner volume. A matrix-matched blood CRM was used for calibration to overcome matrix effects. Sampling from a microwell plate facilitated high sample throughput and the total analysis time was less than 1 min. Validation of the method for nine elements (Al, Ba, Cr, Cu, Li, Mg, Mn, Pb and Zn) was undertaken. The LODs were between 0.17 μg L⁻¹ (Mn) and 0.94 μg L⁻¹ (Cr) with the exception of Mg (21.39 μg L⁻¹). Recoveries from standard addition experiments ranged from 97% to 106% and the measured concentration of a whole blood CRM was within 7.3% of the certified value. Acceptable comparison data were obtained for 57 blood samples also analysed by standard ICP-MS. In the second report, Kim et al.¹¹⁵ investigated direct sampling of pressed rice powder samples using fsLA-ICP-MS. For this method, the authors used a calibration curve, generated from various rice CRMs, for the quantification of As, Cd and Pb elemental content in rice. The approach was then validated using rice samples (n = 18) and CRMs. Figures of merit, following optimisation of the ablation conditions, were: LODs: 0.01, 0.02 and 0.01 mg kg⁻¹ for As, Cd and Pb, respectively; recoveries for a rice flour CRM (IRMM-804) between 87.1% and 118.0% and recoveries vs. the ICP-MS determined values in the real rice samples, 86.1% to 116.7%. The performance was notably poorer than that observed for conventional MAD-ICP-MS (although deemed to be acceptable by the authors), however, this represented a trade-off in terms of the much smaller mass of sample required (2 g vs. 200 g).

Capillary electrophoresis is a much less common coupling to ICP-MS than chromatographic separation techniques. However, it offers advantages of high resolution and separation efficiency in small volume samples and the ability to employ milder physiological separation conditions. Two papers this year reported developments in application of CE-ICP-MS to biological systems. Men et al.¹¹⁶ utilised scICP-MS, with a homemade single cell focussing serpentine microfluid device, and CE-ICP-MS, together with an interface to allow flexible switching between the two, to investigate the cellular transformation of As^III in HepG2 cells. While the scICP-MS facilitated the study of cellular uptake of As and its distribution among, and elimination from, individual cells, the CE-ICP-MS method was utilised for As speciation (AB, As^III, As^V, DMA and MMA) to examine the fate of As within cells. Arsenobetaine was identified as the main metabolite excreted by HepG2 cells, however, MMA was also detected within As^III-exposed cells, suggesting MMA as an intermediate metabolite. Zajda et al.¹¹⁷ in a proof of principle study, employed CE-ICP-QQQ-MS to study the formation of Pt–DNA adducts under simulated physiological conditions; an area of interest in the development of chemotherapy prodrugs containing Pt^IV. Elements, P, Pt and S were measured as markers of DNA, Pt complex and proteins respectively and the adducts formed between the components detected as co-migrating signals. To avoid polyatomic interferences, ³¹P and ³²S were detected off mass as their respective oxides, ³¹P¹⁶O⁺ and ³²S¹⁶O⁺, whereas ¹⁹⁵Pt could be detected on mass. Following optimisation for sensitivity, the CE-ICP-MS method showed promise and four co-migrating ¹⁹⁵Pt⁺ and ³¹P¹⁶O⁺ signals were detected for the Pt-prodrug and DNA oligonucleotide model system used but not for the Pt-prodrug alone suggesting the formation of Pt–DNA adducts. These signals only accounted though for 12% of the total Pt content, which was postulated to be due to limitations of the model system used.

A method was developed and validated for the accurate quantitation of total haemoglobin, traceable to SI units, employing HPLC-HR-ICP-MS.¹¹⁸ Following chromatographic separation of blood proteins, Hb concentration was determined through measurement of Fe and S, which are present within the haem moiety. Diluted blood was vortex mixed to lyse the cells before centrifugation and filtration (0.22 μm) and then injection onto the size exclusion column. The enriched isotope solutions, ⁵⁴Fe and ³⁴S, were continuously added on-line post column and the isotope ratios of ⁵⁶Fe : ⁵⁴Fe and ³²S : ³⁴S were measured by HR-ICP-MS in medium resolution mode, which effectively removed polyatomic interferences. It was also possible to detect transferrin via Fe detection using the same method with an adjusted flow rate of ⁵⁴Fe. The method was validated using a purified Hb CRM and figures of merit were a measuring range for Hb of 10 to 240 g L⁻¹, LODs of 0.01 g L⁻¹ and 0.07 g L⁻¹ for measurement of Hb via Fe and S respectively and RSDs of 1.2–2.2% and 1.6–2.6% for measurement via Fe and S respectively. Good agreement was observed for 21 samples with respect to the HiCN (photometric determination of cyanmet Hb) reference procedure.

Redeker et al.¹¹⁹ reported an entirely different approach to elemental mass spectrometry using LC-ICP-HR-molecular MS (nanospray-Orbitrap) for the simultaneous and species-independent determination of fluorinated and chlorinated drug metabolites via BaF⁺ and BaCl⁺ respectively. The approach, involving plasma-assisted reaction chemical ionisation using Ba, offers a significant way forward in determination of these elements, which are problematic to measure using ICP-MS owing to high ionisation potentials and isobaric interferences. Elemental F and Cl were first converted to HF and HCl by an ICP with post-plasma recombination followed by reaction with Ba-containing ions originating from a nanospray. Isobaric interferences were eliminated through use of the HR-Orbitrap at resolving powers of 35 000. A rapid throughput workflow for determination of the drug metabolites was described with initial separation by RP-LC followed by quantitation of the Ba reporter ions using a single point compound-independent calibration. There was good agreement between independent measurements of the compounds via F and Cl, on-column recoveries >90% and on-column LODs of 8–12 pmol (Cl) and 5–12 pmol (F). The LODs were considered by the authors to be comparable or better than those reported for gradient LC-ICP-QQQ-MS with the added advantage of being able to detect both elements simultaneously.

4.2 Atomic absorption and atomic emission spectrometry

Although the majority of applications for the analysis of clinical specimens or food and beverage samples refer to ICP-MS, in several situations other atomic spectrometry techniques may be preferred. The choice among different techniques is driven by the LOD to be achieved, that may, in turn, depend on available sample size, element expected concentration, potential matrix interferences, related analysis costs and local regulations. Not surprisingly, this year's review also includes a number of papers describing analytical procedures based on AAS as well as GD-ICP- or MIP-OES. Applications mainly devoted to the development of sample pre-concentration and/or extraction procedures to improve the sensitivity of methods based on these atomic spectrometry techniques are covered in Section 3.2 and Table 1, whereas work focusing on coupling of VG techniques with OES or AFS are described in Section 4.4.

Atasoy,¹²⁰ aiming to improve the LOD for the determination of Pb in drinking water and fish samples using AAS, developed a procedure based on HGAAS with an externally heated platinum-coated tungsten-coil atom trap for in situ Pb pre-concentration and compared its analytical performances with those of a conventional method, based on Zeeman corrected GFAAS. The instrumental apparatus involved a quartz T-tube atomiser, externally heated with an air–acetylene flame, and connected to a smaller QT where the tungsten coil was placed. Drinking water samples were acidified to contain 1.0 mol L⁻¹ HNO₃, stored in a refrigerator and spiked with 1.0% (w/v) K₃[Fe(CN)₆] before analysis. Aliquots of 0.1–0.5 g of the fish samples (liver, muscle, and gill tissues of red mullet and common pandora) were digested with 10 mL of 70% (w/w) HNO₃ in a microwave system, then the digests were diluted with UP water to a final volume of 50 mL. The HGAAS method achieved an LOQ of 11.0 ng L⁻¹ compared to 724.4 ng L⁻¹ for GFAAS and better RSD% (2.3% vs. 3.5%), but required a larger sample volume (6.83 mL vs. 0.01 mL). The linear range was 0.01–10.0 μg L⁻¹vs. 5.0–200.0 μg L⁻¹. Analysis of CRMs (NIST SRM 1640a elements in natural water; DOLT5 dogfish liver) gave recoveries of 100.3% and 109.9%, respectively. Paired measurements carried out on drinking water samples did not show significant differences by the t-test. The author also provided a comparison of the LOD and RSD% of the proposed method with those of ten other plasma-based and trap methods for Pb determination. Notwithstanding the rather complex equipment necessary, the proposed method may offer a suitable alternative when other sensitive techniques are not available.

Furthermore, the group led by Atasoy¹²¹ managed to achieve an LOD of 0.075 ng mL⁻¹ and an LOQ of 0.25 ng mL⁻¹ for the determination of Cd in drinking water and fish samples, using a SQT with in situ atom trapping coupled with FAAS. These values outperformed those achieved, in the same study, for FAAS (LOD 110 ng mL⁻¹; LOQ 366 ng mL⁻¹) and SQT-FAAS (LOD 5 ng mL⁻¹; LOQ 16 ng mL⁻¹). As it may be expected, the linear range with this approach did not extend to the same concentration levels as other techniques (0.25–2.0 μg L⁻¹ as compared to 25–1000 μg L⁻¹ for SQT-FAAS and 250–5000 μg L⁻¹ for FAAS), but allowed reliable measurements at lower concentrations, using less expensive equipment. The in situ atom trapping approach (only applicable to volatile elements) aims to pre-concentrate the analyte on the inner surface of the SQT, followed by its re-volatilisation by the introduction of an organic solvent (e.g. MIBK) and subsequent analysis by FAAS. Sample preparation for drinking water consisted of addition of HNO₃ to a final concentration of 1.0 mol L⁻¹. Approximately 0.1–0.5 g of each fish (common pandora and red mullet) tissue (liver, muscle, and gill) sample underwent MAD with 10 mL 65% HNO₃ in Teflon vessels. The digests were further diluted, presumably with UP water as reported by Atasoy,¹²⁰ to achieve the volume of at least 29.6 mL necessary for the in situ atom trapping approach. The method was validated by the analysis of different matrix CRMs (NIST SRM 1640a trace elements in natural water; BCR 146R Sewage Sludge-industrial origin and DOLT5 dogfish liver) that underwent the same sample pre-treatment as the real samples. Recoveries ranged from 94.5% to 99.6%. In addition, given the high Cd content in NIST SRM 1640a (3.992 ± 0.074 μg L⁻¹), recovery experiments were carried out on four drinking water samples, spiked (two each) with 0.5 ng mL⁻¹ or 1.0 ng mL⁻¹ Cd, yielding recoveries between 102% and 105%. The authors highlighted the reduced matrix effects of the proposed method, as well as its simplicity and low cost, as compared to other techniques.

Liu and co-workers¹²² were concerned with the potential to achieve on-site measurements of Cd and Hg in grain samples, to fulfil the need for rapid analysis of these staple foods and warrant their safety for human consumption. To this aim they designed a portable instrument, based on a metal–ceramic heater for ETV, an on-line catalytic pyrolysis furnace with aluminum oxide fillers, a composite Pt/Ni trap, and a miniature AA spectrometer, including a charge coupled device to achieve the simultaneous determination of the two elements. The instrument weighed <14 kg and required only 270 W to operate. Matrix interferences were successfully eliminated, thus allowing for calibration with aqueous standard solutions (R² > 0.995). The method did not require sample digestion and could be applied to 20 mg grain samples, achieving LODs of 0.3 μg kg⁻¹ for Cd and 1.1 μg kg⁻¹ for Hg. The analysis of two rice CRMs (GBW(E)100351 and GBW(E)100360) yielded recoveries of the certified values of 99.7 and 103.7% for Cd and of 105.6% and 120% for Hg, respectively. In addition the results obtained on seven grain samples, including corn, rice and wheat, compared well with those obtained by either ICP-MS (Cd) or commercial Hg analyser (p > 0.05, Student's t-test). Recoveries of spiked amounts of the analytes to these samples ranged from 96% to 111% for Cd and from 97% to 103% for Hg, respectively.

Two groups of researchers applied GD-OES to the analysis of drinking water¹²³ and other beverages.¹²⁴ To achieve improved sensitivity for the determination of As in water, Liu et al.¹²³introduced a hydrogen-doped solution anode to perform GD-OES (SAGD-OES). This arrangement allowed simultaneous efficient volatilisation and excitation of As, as well as elimination of the need for an HG sampling unit and related chemicals. Since cations present in water samples suppressed the As signal, samples were preliminary run through a cation exchange resin. At a wavelength of 193.7 nm, the LOD was 1.4 μg L⁻¹ and the linearity ranged from 5 to 500 μg L⁻¹. The analysis of the CRM GSB 07-3171-2014 (simulated water sample) with a certified As value of 55 μg L⁻¹ gave an average result of 53.7 ± 1.8 μg L⁻¹. The As values determined on seven real water samples compared well with those obtained by ICP-MS.

Gorska et al.¹²⁴ proposed a simplified procedure for the rapid and simultaneous determination of Ca, K, Mg, Na and Zn in beverages widely consumed by athletes (Coca-Cola Zero, energy and vitamin drinks, pre-workout, branched-chain amino acids, almond drink, and whey protein). After dilution and acidification with HNO₃ to a concentration of 0.2 mol L⁻¹, the samples were analysed by APGD-OES, with a flowing liquid cathode and a gas (He) jet anode. Instrumental LODs (μg L⁻¹) were 20 (Ca), 0.14 (K), 0.91 (Mg), 0.062 (Na) and 21 (Zn), respectively, thus allowing for large dilution and external calibration for all but three samples. A comparison with results obtained on the same samples by ICP-OES yielded recoveries between 88% and 111%, that were deemed fit for the purpose.

Microwave-induced plasma AES also continues to attract attention, since the use of N₂, that can be produced in situ, to generate the plasma, makes it a more sustainable and less expensive multi-element analytical technique than ICP-OES or ICP-MS. In this year's Update, we report of several applications of this technique, for the analysis of biological, food, beverage or supplement samples. Martins and co-workers¹²⁵ compared the performances of ICP-OES and MIP-OES coupled with a Marin-5 nebulisation system, for the determination of 10 trace elements (Al, Cd, Co, Cr, Cu, Fe, Mn, Mo, Ni, and Pb) in mineral water. They described this equipment as a pneumatic nebuliser, operating at a maximum flow of up to 0.5 mL min⁻¹, coupled to a heated spray chamber, at 120 °C, and a desolvation/condensation unit, at 3 °C, for solvent reduction. This sample introduction method improved the LODs for both techniques. The corresponding LOQs ranged from 0.49 μg L⁻¹ to 19.73 μg L⁻¹ when the Marin-5 nebulisation system was used in conjunction with ICP-OES, in comparison with a range between 0.90 μg L⁻¹ and 38.68 μg L⁻¹ achieved with a standard Meinhard concentric glass nebuliser. When MIP-OES was used for the measurements, LOQs ranging from 0.43 μg L⁻¹ to 19.73 μg L⁻¹ were achieved, in comparison with those obtained with a standard nebuliser (1.03–50.91 μg L⁻¹). Recoveries of the certified values for a CRM (NIST SRM 1643e trace elements in fresh water) varied between 83% and 117%.

Matrix effect is one of the main problems affecting the determination of trace elements. The work of Fontoura et al.¹²⁶ investigated the effect of a complex matrix, such as urine, on the optical emission signals of 14 elements (Al, As, Ba, Be, Bi, Ca, Cd, Co, Cr, Cu, Li, Mg, Pb and Sb). Since these effects may involve enhancement or depression of the signals, as well as being absent at all, depending on the analyte, the authors studied the plasma operational conditions for MIP-OES, as well as external calibration vs. internal standardisation, and three levels of dilution (2-, 20-, and 200-fold). After investigating several elements and four molecular species present in the nitrogen plasma (CN, N₂, N₂⁺, and OH), they chose Ge, Rh and Sc as ISs for the 20-fold dilution and N₂⁺ for the 200-fold dilution. They reckoned that the proposed strategy overcame the effects of the matrix in a relatively colder plasma (5000 K). Although the LODs, ranging from 0.33 μg L⁻¹ to 329 μg L⁻¹ under the different dilutions and whether internal standardisation was applied, were higher than those obtained by means of ICP-OES, they were deemed acceptable for the determination of these elements in urine. However, the recoveries of spikes of 1 mg L⁻¹ in a urine matrix remain highly variable depending on the dilution and the element (between 80% and 120% for most elements only at a 200-fold dilution). This study provides an important insight on the spectral effects during urine analysis using ICP-OES. However, the simultaneous determination of several elements in urine by this technique appears to remain a challenge.

Another group of researchers¹²⁷ investigated the determination of toxicological relevant As species (As^III, As^V, DMA and MMA) in urine, as the sum of the so-called “hydride-reactive” species, by means of HG-MIP-AES, as a simpler and economic alternative to HPLC-ICP-MS for biomonitoring studies. Following previous literature reports, the authors adopted a pre-reduction step, diluting urine samples (1 + 1) with 2% (w/v) L-cysteine–2% (w/v) HCl. This allowed to generate As^III–L-cysteine complexes from these four As species, considering that better yield is expected for hydrides formed from As^III than As^V. Non-toxic forms of As, such as AB and AC, do not form hydrides on treatment with NaBH₄ and therefore are not determined by this procedure. Subsequently, AsH₃ was generated with 2% (w/v) NaBH₄ – in 0.5% (w/v) NaOH and 0.1 mol L⁻¹ HCl, prior to determination of As by MIP-AES. A matrix-matched calibration approach was adopted, spiking a pooled urine matrix with As^V, at eight levels between 0 and 200 μg L⁻¹. For the validation of this method, the authors determined the LOD (1.8 μg L⁻¹) and the LOQ (5.4 μg L⁻¹), accordingly to the Eurachem guidelines,¹²⁸ as 3.3- and 10-fold the SD of results obtained on urine blank samples. Repeatability (1.8–6.2%), intermediate precision (6.9–5%) and recoveries of spiked amounts (95–97%) were assessed on in-house QC samples at three concentration levels (10, 35, and 100 μg L⁻¹). Unfortunately the QC samples were prepared at the same time and using the same matrix as that for calibration, therefore making the bias assessment results self-reliant. Although the authors also reported independent assessments of this method, by participation in an EQA scheme (GEQUAS), achieving results within the tolerance intervals established by the provider, it is noted that, from the reported data, such tolerance intervals allow for wide acceptance limits, e.g. from 20% to 30% of the assigned value.

Another paper, by Vazquez-Quintal et al.,¹²⁹ reported the application of MIP-AES to the determination of several trace elements (including Al, Cd, Co, Cr, Cu, Fe, K, Mg, Mn, Na, Ni, Pb and Zn) in ethanolic extracts of propolis (propolis tinctures), to support future recommendations by regulatory authorities in terms of element content or EtOH percentage. The authors warned that the EtOH content of the products influenced the LODs for some elements. This effect needs to be considered as the median concentrations of Cd and Pb in 23 commercial propolis tinctures were low (<0.28 μg L⁻¹ and <2.8 μg L⁻¹, respectively).

Inductively coupled plasma optical emission spectrometry, as other atomic spectrometry techniques, relies on appropriate calibration to provide reliable and traceable measurement results. Owing to interferences arising from complex matrices, over the years various calibration methods have been explored, including external calibration, internal standardisation, matrix-matched calibration and the standard addition methods. The determination of nutrient elements in foods by ICP-OES was the case of study for Carter and co-workers,¹³⁰ who wanted to overcome the difficulties arising from variable and complex matrices as well as the time-consuming approach of standard additions. To this aim, they investigated the potential of automated online standard dilution analysis, an approach that combines internal standardisation and matrix matched calibration, for the determination of 10 nutrient elements (Ca, Cu, Fe, K, Mg, Mn, Na, P, S, and Zn) in ten food CRMs and four fortified real samples (breaded chicken, corn flakes, Italian dressing and vegetable oil). Samples (0.5 g, but for vegetable oil: 0.25 g) underwent MAD with 8 mL HNO₃ and 1 mL H₂O₂. Digests were diluted to 100 mL in 5% HNO₃–0.5% HCl. The samples were presented to the autosampler and known amounts of the two solutions containing the analytes for standard addition and either Lu or In as ISs (1 mg kg⁻¹ in 5% HNO₃–0.5% HCl) were sequentially added via a y-joint in the pump tubing, diluting the sample 1 + 1 with increasing and decreasing amounts of standard and blank. The LODs for the ten elements, ranging from 0.39 mg L⁻¹ to 110 mg L⁻¹ and the LOQs, between 1.3 mg L⁻¹ and 370 mg L⁻¹, were considered adequate for the analysis of food samples. Recoveries within the acceptable range for food analysis (80–120%) were achieved for all studied materials and elements, when the ratio between the concentration of the analyte in the processed sample and the calibration solution was <10. This study, covering a range of elements and food matrices, contributes to work exploring a viable calibration alternative to obtain fit-for-purpose analytical results.

Inductively coupled plasma optical emission spectrometry also offers the potential for the determination of elements which other atomic spectrometry techniques do not provide. In this year's Update we report of two papers, addressing the determination of Cl¹³¹ and phosphine.¹³² Chloride was determined indirectly, after precipitation of AgCl. Liquid samples (10 mL) were mixed with 2.5 mL of a solution containing 300 mg L⁻¹ Ag⁺, as the precipitation agent, and 60 mg L⁻¹ Y³⁺as the IS, followed by dilution to 15 mL with UP water, thus providing a baseline concentration of 50 mg L⁻¹ Ag⁺. After standing in the dark for about 30 min, the resulting solution was filtered (0.22 mm) and presented to the ICP-OE spectrometer to determine the changes in the Ag concentration. These values were then used to determine the Cl concentration on a negative slope calibration curve, given that the Ag concentration decreases as the amount of Cl in the sample increases. Taking into account the instrumental LOQ for Ag determination and a chemical LOQ, defined as the amount of unreacted Cl remaining in the solution after AgCl precipitation, the method working range was defined as between an LOQ of 0.2 mg L⁻¹ and 15 mg L⁻¹ Cl⁻. The authors explored the effect of various parameters, such as temperature, pH and sample filtration time, on the performance of the method and compared the results obtained on four samples (spiked-purified water, seawater, wine, and urine) with those obtained by means of IC performed on a Metrosep A column (250 × 4 mm), with no observed statistical difference. The determination of PH₃ residues in cereals and tissues was achieved¹³² by means of dynamic headspace ICP-AES. Reaction of the sample with H₂SO₄ in a headspace vial produced PH₃ vapours, that were introduced into the plasma for P measurement at 213.6 nm. The authors reported a working range between 0.10 and 0.50 μg g⁻¹ PH₃ (R² = 0.990) and an RSD% <12% (n = 10), as well as comparable results with PH₃ determinations by GC with nitrogen phosphorous detector. These procedures appear to extend the range of applicability of ICP-OES, offering simple and rapid alternatives to other analytical techniques.

Finally, an interesting development was reported on point-of-care testing, using miniaturised equipment. Out of concern for the detrimental effects of exposure to Pb for children intellectual development, He et al.¹³³ designed a portable point-discharge microplasma OES device integrated in a ballpoint pen format. They applied this innovative equipment to determine blood Pb, using a sample size of only 100 μL blood and reported an LOD of 0.73 μg L⁻¹ Pb, thus fulfilling the need for rapid on-site screenings.

4.3 Laser induced breakdown spectroscopy

Within this Update period, the number of papers covering LIBS has significantly increased, perhaps reflecting the development and maturity of the technique. There has been continued interest in utilising LIBS for species identification and authenticity of food products. Ma and co-workers¹³⁴ presented a new approach for LIBS data collection for the identification of adulterants in brown rice flour. The researchers termed this approach ‘time resolved’ LIBS, where spectra were collected from multiple gate delay times and spliced together to increase the amount of data per sample. To develop the model, samples of brown rice flour were gravimetrically mixed with common additives, namely buckwheat flour, corn flour, gypsum powder, sorghum flour and talc powder, and pressed into pellets. The spectral lines of C, Ca, Cu, Fe, H, K, Mn, N, Na, O, P and Zn were selected and eight time resolved spectra were spliced together to create a single input. The model was designed using a one-dimensional convolutional neural network and was compared with traditional LIBS and statistical approaches such as LDA and SVM. Overall, the time resolved LIBS library provided higher accuracy, significantly improved discrimination power and lower LODs, highlighting a considerable development in this field. Wei and co-workers¹³⁵ investigated the adulteration of the Chinese medicine Fritillaria cirrhosa with the lower cost species Fritillaria thunbergia. Blends of the two herbs were gravimetrically prepared over 21 levels. The samples were pressed into pellets and 100 LIBS spectra were acquired from each pellet. The spectral lines from C₂, Ca, CN, H, K, N, Na and O were used in the statistical models. Several standardisation methods were applied to normalise the data before ‘least absolute shrinkage and selection operator’ was utilised for dimension reduction. Then PLS regression and SVR algorithms were implemented for the final accuracy assessment. The complex statistical tools were required to achieve the discrimination level necessary where the difference between adulterants is very small. Another group of researchers¹³⁶ analysed four different spices (clove, common wormwood, sweet wormwood and lemongrass), implementing PCA for dimension reduction followed by error back propagation artificial neural network modelling. After cross verifications, the recognition accuracy was 85.25%. Jiao and co-workers¹³⁷ focused on the popular Chinese medical herb Salvia miltiorrhiza, or red sage, used to treat cardiovascular disease. Samples were collected from six different geological regions and analysed by LIBS. A classification model was built using convolutional neural network which achieved an accuracy of 97% for location discrimination.

Sun and co-workers¹³⁸ demonstrated the power of combining LIBS and Raman spectroscopy to identify meat species, namely beef, mutton and pork. Here, a random forest (RF) model was applied to the individual data sets and the combined library to identify the most relevant variables followed by use of a back propagation neural network algorithm. The merged data set provided the highest classification accuracy at 99.42%. Ren et al.¹³⁹ focused on the identification of fish species with LIBS in combination with Raman spectroscopy 13 fish varieties. Two machine learning methods were investigated, namely SVM and convolutional neural network, using the separate data sets and the pooled library. Perhaps unsurprisingly, the combined data set again provided enhanced discrimination power with the neural network model outperforming the SVM. The classification accuracy achieved for the optimised algorithm was 98.2%. Both publications demonstrated the increased discrimination ability of fused data libraries from complementary analytical techniques. Also examining fish from closely related species, Yan et al.¹⁴⁰ utilised qualitative LIBS data for mislabelling or adulteration detection. A handheld LIBS device was used in this work and the spectral lines of C, Ca, CN, H, K, Mg, Na and O were detected. Initially, PCA easily identified fresh and saltwater fish but differentiation within the group was limited. Therefore, a nonlinear RF design was implemented and the selection parameters were optimised, achieving an overall classification accuracy of 94%. The work demonstrated the potential of a handheld device for the rapid discrimination of fish species.

Two publications have applied LIBS to support the quality and safety of food products. Currently quality labelling of rice grains is based on imaging or visual inspection and is therefore subjective. Perez-Rodriguez et al.¹⁴¹ tested the potential of LIBS to provide an objective view. Rice samples were grouped based on grain quality features and LIBS spectra were collected after simply grinding and pelleting. The data was first processed using PCA for normalisation, followed by the k-nearest neighbour algorithm to achieve a classification prediction accuracy of 94%. It demonstrated the ease of analysis with the potential to replace the current subjective assessment of rice quality. Yan et al.¹⁴² focused on the detection of Cr and Pb in oysters for food safety monitoring. The impact of drying, carbonisation and ashing of the samples was compared, finding that levels of C, H, N and O were significantly reduced after carbonisation and ashing, leading to signal enhancements of 3–6 fold for carbonisation and a further 5- to 9-fold for ashing due to the simplified matrix. The LODs obtained were 0.08 mg kg⁻¹ for Cr and 0.52 mg kg⁻¹ for Pb, which are appropriate for compliance with Chinese food safety standards. The method represents a rapid and simple approach for heavy metal contamination monitoring.

The issue of appropriate calibration materials for LIBS was tackled by Gamela et al.¹⁴³ by implementing matrix matched solid standards. Bean seed samples were used in this work and matched calibrants were prepared for Ca and Mg. Various univariate and multivariate statistical approaches were evaluated, namely PLS, principal component regression multi-energy calibration and single point gravimetric standard addition. It was found that univariate methods performed better with the gravimetric standard addition, providing recoveries of 76–112% for Ca and 87–127% for Mg, demonstrating the importance of appropriate calibration materials.

The application of LIBS for electrolyte analysis in blood samples has received particular attention. Wang et al.¹⁴⁴ described the use of ‘ordered microarray silicon substrates’ to directly add serum prior to LIBS detection. The researchers showed that the substrate prevented cracking and the ‘coffee-ring’ effect compared to silicon wafers, which improved signal stability and accuracy. The paper provided a robust assessment of the approach which achieved accurate results from 10 μL of serum with RSDs of <6% for Na and <4% for K and R² exceeding 0.98 with analysis times under 5 min. Another group¹⁴⁵ took a slightly different approach by employing surface enhancement to improve sensitivity. Blood plasma was dropped onto a metal substrate, allowed to dry and coagulate into a gel then into a solid state. Several parameters were optimised including the sample volume, standing time and substrate metal for analysis of K and Na. With 15 μL of plasma, 6 min standing time and zinc metal substrate, the figures of merit were assessed, obtaining a LOD of 0.21 mmol L⁻¹ and 0.15 mmol L⁻¹ for K and Na respectively, RSDs were 3.19–6.61% and the R² was greater than 0.98. The quantitative LIBS results were compared against ISE, finding the relative errors were <2%. Overall, the work demonstrated a rapid and straightforward approach for analysis of multiple electrolytes in a single run. Feng et al.¹⁴⁶ employed a portable LIBS device for the determination of Ca, K and Na in blood serum. Filter paper and glass slides were used as the substrate, however, both required blank correction due to their contribution to the signal. The serum was then dried for 10 min prior to analysis. Calibration was achieved with PLS regression and reference values for Ca, K and Na were obtained from a medical electrolyte analyser instrument. The prediction accuracies were 0.61–3.80% for the glass slide and 0.52–7.47% for the filter paper, demonstrating good agreement for a portable device. However, the authors did not provide any other analytical figures of merit or consider the variability of the blank contribution, leaving the paper somewhat lacking.

In an interesting application by Winnand et al.,¹⁴⁷ LIBS was presented as a potential tool for real time analysis to support surgeons during bone removal caused by oral cancer. Tumour cells in head and neck cancers can infiltrate the mandible which can lead to invasive bone tumours, resulting in surgical removal. However, a wide margin of healthy bone surrounding the tumour must also be removed to prevent reoccurrence. The levels of Ca, K and Na vary in bone, where high concentrations of Ca indicate healthy bone, but increased K and Na is observed in invasive bone cancer. Additionally, LIBS can provide rapid results. With PCA, the regions were clearly identified and variations in bone morphology did not significantly impact the detection capability. The study demonstrated the potential of LIBS as an analytical surgical tool to assess healthy and diseased bone regions in situ. Another interesting use of LIBS was published by Priyanka et al.¹⁴⁸ to assess the elemental profile in teeth with respect to age and sex. Elemental analysis of 170 tooth samples was performed at the cemento-enamel junction with LIBS. Although several trace elements such as Cu, Mn and Zn were detected, the researchers focused on Ca, Fe and P. Qualitative comparisons were made by splitting the subjects into age groups of 10 years and male/female categories. In general, levels of Ca, Fe and P decreased over time with each element displaying different rates. Additionally, the intensity levels for female teeth were relatively higher compared to the male. Whilst the data showed some potential, the variations or standard deviations within the categories were not discussed which is an important consideration.

A paper by Wen and co-workers¹⁴⁹ described the use of an enrichment procedure for the determination of Pb at the ppt level in water samples by LIBS assisted by LIF. A macroporous cation exchange resin was mixed with the water to trap Pb from solution. The resin was then dried and attached to a glass slide. Calibration standards were prepared in the same manner. The combination of LIBS and LIF provided a 7-fold signal enhancement over LIBS alone. The LIBS-LIF system was thoroughly optimised and achieved an LOD of 88 ng L⁻¹, linearity up to 10 μg L⁻¹ with RSDs of 9.73%, which is a remarkable accomplishment. The approach was then tested with tap water, canal water, tea and coffee, with Pb spiked into the matrices, obtaining recoveries in the range of 91.5% to 106.5%. This development represents a large step forward in the sensitivity of LIBS and places it on par with other atomic spectrometry techniques.

4.4 Vapour generation procedures and atomic fluorescence spectrometry

Pneumatic nebulisation is the routine sample introduction method for a number of atomic spectrometry techniques, although its efficiency is low (from 3% to 7%). Therefore, alternative sample introduction techniques, such as VG and others, aimed to increase the measurement sensitivity for one or more elements, are explored, as regularly reported in our Updates. Also in this year's Update, we report interest for VG procedures coupled with AFS or ICP-OES.

Three papers reported improved vapour generation methods for sample introduction in atomic fluorescence spectrometry. Štádlerová and co-workers¹⁵⁰ developed a method to determine ultratrace levels of Cd by means of FI-CVG-AFS, with an impressive LOD of 0.42 ng L⁻¹, that, depending on the matrix, may be better than those reported for other VG techniques coupled to AFS or ICP-MS. On the basis of their previous findings to improve the efficiency of CVG, the authors used a four-channel FI-CVG unit. This allowed separate introduction of the reductants (5% m/v NaBH₄ and 0.2 mol L⁻¹ HCl), both at 1 mL min⁻¹, and the modifiers (0.6 mmol L⁻¹ Cr³⁺, prepared from Cr(NO₃)₃·9H₂O and 0.01 mol L⁻¹ KCN) both at 0.5 mL min⁻¹. They used an in-house assembled non-dispersive AF spectrometer and compared the performances of two atomisers (a miniature diffusion flame atomiser and a flame-in-gas-shield atomiser), the latter showing a two-fold sensitivity improvement vs. the former one. The performance of the method with the latter atomiser was evaluated on water-based CRMs (NIST SRM 1643f fresh water; ERM-CA713 wastewater; NRCC CASS-4 nearshore seawater and NRCC NASS-5 seawater), diluted to a final concentration of 0.2 mol L⁻¹ HCl. For the analysis of a rice flour CRM (SRM 1568b rice flour) and three (one rice and two rice flour) commercial samples, milled samples were dried for 2 h at 95 °C. Approx. 0.25 g of sample, placed into 15 mL quartz digestion vials, with 2.5 mL of 2 mol L⁻¹ HNO₃, were left overnight at room temperature and, after the addition of 1 mL 30% H₂O₂, MAD was carried out. The authors stated that three measurements were performed on each sample, but, surprisingly, claimed to report their “median” accompanied by SDs – a translation error? Taking into account the reported values, recoveries from water-based CRMs were 101.5% and 97.3% for fresh and waste water, respectively, at concentrations of about 5 μg L⁻¹, but lower (96.2% and 87.0%) for the nearshore seawater and seawater CRMs, at about 0.02 μg L⁻¹ as well as for SRM 1568b rice flour (88.8%). Comparison with the ICP-MS results obtained for SRM 1568b rice flour and the commercial samples yielded recoveries between 91.8% and 105.7%. The identification and quantification of Se metabolites in human urine is still an open field of research, considering their analytical difficulties. Slejkovec et al.¹⁵¹ evaluated how to optimise the HG step as part of the determination of trimethylselenonium (TMSO) in human urine by HPLC-HG-AFS. They reported an evaluation of the analytical conditions and techniques applied in recent literature, then explored the mechanism for Se^IV, SeMet and TMSO volatilisation, using various mobile phases (20 mmol L⁻¹ pyridine, 15 mmol L⁻¹ KH₂PO₄ or 50 mmol L⁻¹ CH₃COONH₄) and reaction coils in terms of types of Teflon (fluorinated ethylene propylene (FEP), perfluoroalkoxy (PFA), or PTFE), id (0.5–0.8 mm) and lengths (1 or 2 m). They eventually chose a Dionex™ IonPac™ AS7 IC anion exchange column with 20 mmol L⁻¹ KH₂PO₄, pH 4.65, as the mobile phase, at a flow rate of 1 mL min⁻¹, as the chromatographic settings and a sample size of 50 μL. The eluent went through a mixing coil (any type of Teflon, id 0.5 mm, length 1 m) with 0.7% NaBH₄ in 0.1% NaOH, followed by 1 mol L⁻¹ HCl to form volatile Se hydrides that were determined by AFS. Selenomethionine, SeCys₂ and selenomethylcysteine (SeMeCys) eluted at or close to the void volume (1.95–2.20 min), TMSO at 4.4 min and Se^IV at 7.5 min, whereas Se^VI, not forming a volatile hydride, did not interfere with the determination. These analytical conditions were then applied to the analysis of urine samples collected from 16 unexposed volunteers (4 men and 12 women, aged 41.7 ± 11.0 years), over 4 consecutive weeks, for a total of 64 urine collections. Sample aliquots (0.5 mL), placed in glass tubes with 0.5 mL of conc. UP HNO₃, were treated by MAD, and the digests diluted to 5 mL with UP water. All samples were analysed in duplicate and the observed differences between paired determinations on the same sample were within 10%. An LOD of 0.2 ng mL⁻¹ was claimed. The results obtained for TMSO confirmed higher levels (2.5 ± 1.7 ng mL⁻¹) for the four subjects genetically characterised as TMSO producers (hINMT genotype GA) than for the 12 other subjects (0.2 ± 0.2 ng mL⁻¹). As a quality control measure, total Se was determined in all samples and in the control samples ClinCheck® Urine Control Level 1 (Recipe, Germany) and Seronorm™ Trace Elements Urine by HPLC-HG-AFS as well as by FI-UV-HG-AFS and ICP-QQQ-MS, the last one with H₂ as the reaction gas. Results on the QC items fell within the expected range. The authors do not provide further evidence for the reliability of their TMSO measurements, other than a repeatability precision within 10%, thus suggesting that the quantitative performance of this method may need further validation. However, given the large difference between urinary TMSO levels according to genetic features, the procedure provides a reliable basis for a qualitative assessment and can be considered fit-for-purpose. In another paper, by Liu et al.,¹⁵² a preliminary high-efficiency photo-oxidation step was proposed, based on a reaction tube passed through a low-pressure mercury lamp, for the complete degradation of Hg species (EtHg, MeHg) prior to determination of total Hg by AFS, with no need for additional reagents or heating. Under these conditions, the determination of iHg and total Hg could be achieved within 120 s, using iHg standard solutions for calibration, within a linear range (based on peak area) of 0.5–20 ng mL⁻¹ (R = 0.9999), and an LOD of 0.02 ng mL⁻¹. In addition, with appropriate connections to FI or HPLC, as well as time control, the procedure allowed for further Hg speciation on which we report in Section 6.1.

In an attempt to improve the LOD for Hg determination in biological and environmental samples by ICP-OES, Greda et al.¹⁵³ devised a microplasma-induced VG system, operating as a FI gas analysis apparatus, coupled to ICP-OES. This approach stems out of previous work on plasma-induced VG, resulting in improved sample introduction efficiencies (>50%) as well as a greener practice, owing to the reduction or elimination of chemicals. Based on its high efficiency (>80%) for Hg VG, the authors chose SAGD as the microplasma source. After delivery of the sample solutions to the SAGD system, Hg vapours were generated and accumulated in a closed discharge chamber for a chosen period, ranging from 10 to 120 s, then injected into an argon carrier gas flow entering the ICP torch. This novel arrangement, exploiting the potential of accumulating Hg vapours, led to an outstanding LOD of 0.035 μg L⁻¹vs. the same parameter values obtained by either ICP-OES with pneumatic nebulisation (5 μg L⁻¹) or CVG-ICP-OES (0.15 μg L⁻¹), as well as to better RSD% (2.5%). The researchers tested this instrumental equipment on a CRM (ERM-BB422 fish muscle) and spiked matrices with not detectable Hg levels (MODAS-2 bottom sediment and two soil samples), with recoveries ranging from 97.7% to 104%. This is a promising experimental approach for the improvement of ICP-OES LODs, that may be applied to other volatile species. However, further evidence on a larger number of samples/sample types may well be needed to support the implementation of this arrangement in routine laboratory practices.

4.5 X-ray spectrometry

A comprehensive review of recent advances in X-ray spectrometry² complements the applications with clinical and biological materials, foods and beverages covered within this Update. Imaging applications of X-ray spectrometry are described separately in Section 6.2.

This Update period features two papers covering TXRF spectrometry for clinical and food analysis. Carvalho and co-workers¹⁵⁴ compared the impact of sample preparation for biological tissues and two types of TXRF spectrometers for the determination of Ca, Cu, Fe, K, S and Zn. Suspensions of powdered sample, using only 20 mg, were compared against total acid digestion. Numerous CRMs, namely BCR-184 and NIST SRM 1577a Bovine liver, BCR-184 Bovine muscle, BCR-186 Pig kidney and NIST SRM 1566b Oyster tissue were utilised for method validation. It was found that 2 mol L⁻¹ HNO₃ was suitable as a dispersing agent and was implemented for the analysis of human tissues with different adipose content (aorta, bladder, colon, heart, intestine, liver, lung, muscle, skin, stomach and uterus). The samples were lyophilised and ground before suspending in 2 2 mol L⁻¹ HNO₃ and analysed by TXRF equipped with a tungsten anode X-ray tube and using Ga as an IS. Comparative data was obtained by completely digesting the samples with acid MAD followed by analysis with TXRF spectrometry and ICP-OES analysis. Tissues with higher adipose content performed better with complete digestion rather than suspension. Additionally, the molybdenum anode X-ray tube provided enhanced LODs compared to the tungsten anode. Therefore, total digestion and the molybdenum anode system were taken forward for comparison with ICP-OES and with direct analysis from pressed pellets by μEDXRF. Overall good agreement was obtained throughout. Although the suspension preparation was quick and simple, it was subject to impact from the fat content of the tissues, requiring complete matrix destruction for maximum accuracy. Using the molybdenum equipped TXRF spectrometer, provided better LODs and improved accuracy and precision for elements <10 mg kg⁻¹ when compared to μEDXRF. Allegretta et al.¹⁵⁵ described the application of TXRF spectrometry for food traceability and authenticity. Samples of the common bean (Phaseolus vulgaris), cultivated from two different sites, were prepared as a suspension and Br, Ca, Cl, Cu, Fe, K, Mn, P, S, Se, Rb and Zn, were detected. The approach was validated by comparative measurements by ICP-OES. Statistical models were applied to the spectra, concluding that PLS-DA correctly identified the sample origins and that quantification was not actually required due to the nature of the TXRF continuum spectra.

The application of SR-XRF spectrometry for biological investigations was the subject of a paper by Salles et al.¹⁵⁶ to determine the levels of Ca, Cu, Fe, K, Mn, P and S during tumour progression in lung cancer. The liver, lung and primary tumour tissues were extracted from mice at 0, 1, 3 and 5 weeks and analysed by SR-XRF spectrometry. Changes to the elemental levels in liver tissues were observed over time in the lung carcinoma group but not in the control group. It was postulated that the presence of the tumour cells and tumour progression may impact other organs and change the likelihood of metastasis. The authors appeared surprised by the data and suggested further research is urgently needed to understand the tumoral processes and associated metallomics. Luo et al.¹⁵⁷ described the development of new procedures to improve SR-XRF tomography analysis at the nanoscale. In order to produce 3D images, the sample was rotated on the vertical axis and multiple 2D scans collected then reconstructed, but there were limitations due to angular restrictions and even the sample itself blocking regions of interest. Whilst in theory the whole sample could be scanned point-to-point at multiple rotations, it was extremely time consuming on the nanoscale. Therefore, the authors developed a workflow to address this issue. HeLa cells incubated with Fe₃O₄/TiO₂ NPs were used as the test sample. Initially, the sample was analysed with SR-μXRF spectrometry to identify the regions of interest, followed by focused ion beam milling with dual beam SEM to remove excess material around these areas, enabling a clearer path for the subsequent SR-nanoXRF measurement. The researchers demonstrated the approach increased the angular range for sample rotation and achieved improved high resolution 3D elemental data at the sub-cellular level.

A number of publications have focused on analysis with portable XRF spectrometer devices for a range of applications. Zhou and co-workers¹⁵⁸ investigated the impact of surface thickness for portable XRF spectrometers and developed a model to counteract any matrix effects. A variety of organic samples (fungi, vegetation and animal tissues, n = 39) were selected and analysed for As, Cu, Fe and Zn. The model was established and verified by comparison with ICP-MS data, achieving relative errors of between 18% and 34%. The authors concluded that the portable instrumentation was fit for purpose as a rapid measurement technique but that the sample mass per unit area (i.e. surface density) should be considered as the major factor for optimal accuracy rather than sample thickness. Fleming et al.¹⁵⁹ assessed the calibration approaches for determining Zn in keratin-rich matrices such as nail and hair. Four RMs produced from caprine horns were used to test three methods, namely the Zn signal to total signal ratio, Zn/S signal ratio and the system output for Zn concentration. It was concluded that Zn to total signal ratio and system output were sufficiently accurate and precise when compared to the results from ICP-MS. Overall, the method performed well demonstrating the potential of portable XRF systems for non-invasive biomonitoring through hair and nails. Specht et al.¹⁶⁰ determined Pb in human bone with a portable XRF spectrometer for rapid biomarker analysis. Bone samples from cranium, distal tibia (ankle), ilium, mid-tibial shaft and proximal tibia were analysed from 22 donors. It was found that Pb did not significantly vary across the bone types compared to the mid-tibia except for the ilium, with the authors recommending the tibia as the preferred location for comparison studies. Zhu et al.¹⁶¹ described the application of a high-resolution portable EDXRF spectrometer for pharmaceutical analysis, considering the requirements of International Conference on Harmonisation (ICH) Q3D and US Pharmacopoeia 233 standards. The proposed method covered 22 elemental impurities (Ag, As, Au, Cd, Co, Cr, Cu, Hg, Ir, Mo, Ni, Os, Pb, Pd, Pt, Rh, Ru, Sb, Se, Sn, Tl and V) and was demonstrated to meet the requirements for accuracy, precision, LOD, LOQ and linearity under these standards, based on a 10 g maximum daily intake. The approach was suitable for the monitoring of elemental impurities in drugs using a rapid high throughput and non-destructive analytical method. Sultana and co-workers¹⁶² also directly determined Zn in rice using portable XRF spectrometry. The results were compared against acid MAD and ICP-OES detection. Good correlation was found between the two approaches (R² = 0.83), highlighting the suitability of portable XRF spectrometers as a rapid and simple detection technique.

5. Nanomaterials

This section covers developments for the detection of NPs in biological, as well as food and beverage samples. The use of NPs as a reagent for sample preparation/pre-concentration is covered in Section 3.2 and Table 1.

Nanoparticles are defined as materials in which at least one of the three dimensions ranges from 1 to 100 nm. They are increasingly used in a variety of applications and their widespread distribution into the environment poses a number of questions related to their accumulation and potential environmental and biological risks. Therefore, research for the development of reliable analytical methods is a novel and rapid expanding field. Characterisation of NPs requires information on their chemical composition, size, morphology, and surface chemistry. Atomic spectrometry plays an important role in this area, and ICP-MS is the most frequently used technique. However, researchers are faced with new challenges, such as reaching even lower LODs, devising appropriate calibration strategies and dealing with novel substrates, such as single cells.

In this Update, we report three papers addressing general issues in NP measurements. A review¹⁸ summarised the recent trends in analytical methodologies for the characterisation and quantification of NPs in biological matrices. Based on the analysis of 60 publications, over the period 2007–2018, the authors described the applications of techniques for the determination of the different characteristics of NPs, such as AAS, DLS, FFF, FTIR, HPLC, GC-LC-MS, ICP-MS, SEM, SIMS, TEM, UV-VIS and XRD. In addition they discussed the challenges of sample preparation and techniques for the extraction of NPs from their biological matrices (e.g. liquid–liquid extraction, centrifugation, dielectrophoresis, and FFF). Starting from the observation that NPs can interact with proteins forming a biological corona complex, Fuentes-Cervantes et al.¹⁶³ considered the potential of ICP-MS in this area of research, given the capabilities for simultaneous determination of both chemical elements, as part of NPs, and proteins, via the measurement of S. They reviewed 79 papers and concluded that the combination of several state-of-the-art imaging techniques, enabling the visualisation of biomolecules associated with the surfaces of NPs, with the power of quantitation of ICP-MS are necessary to explore the characterisation of corona complexes. In particular, separation techniques (e.g., CE, SEC, AF4, HPLC) coupled on-line with ICP-MS/MS to provide a sensitive and interference free detection of S could play a prominent role. Improvement of analytical performances in this area may require further instrumental development, that may be stimulated by systematic evaluation and comparison of the performance of the existing techniques. An impressive array of instrumental techniques, consisting of three ICP-QMS, two ICP-TOF-MS, and one MC-ICP-MS, was applied to simultaneously detect ¹⁰⁷Ag and ¹⁰⁹Ag on single Ag NPs and Ag-exposed cyanobacteria cells,¹⁰³ used as a model to compare the performances of these techniques for dual-isotope single-particle analysis and assess their advantages and limitations. To this aim, the measured event-specific ¹⁰⁹Ag : ¹⁰⁷Ag ratios were compared with the natural ratio. The authors concluded that, although ICP-QMS is the most common technique applied in NP studies, its accuracy for measuring isotope ratios within single NPs or cells was poor. They found ICP-TOF-MS to be the most promising method to investigate the bioaccumulation and transformation of NPs in single cells. This technique featured capabilities for both multielement and isotope ratio determinations as well as accurate relative abundances, due to the ability to monitor simultaneously multiple m/z. Notwithstanding its unrivalled performance for isotope ratio determination, MC-ICP-MS did not provide sufficient sensitivity for single particle or single cell analysis.

Two relevant applications of atomic spectrometry to the measurement of NPs are included in this year's Update. The determination of Se NPs in cultured cells of lactic acid bacteria, bifidobacteria, and yeast was achieved using spICP-MS.¹⁶⁴ Selenium NPs were prepared in-house and characterised using TEM. Selenised bacteria and yeasts were obtained by adding 10 mg L⁻¹ of Na₂SeO₃ to the growth media. Afterwards, the cells were centrifuged, washed twice with sterile saline solution to remove unbound Se, and freeze dried. The authors tested two sample preparation procedures. In the former, approx. 0.005 g of sample of selenised strain was dispersed in 5 mL of 1% (v/v) MeOH, sonicated for 20 min, then diluted 250-fold with 1% (v/v) MeOH. In the latter, aiming to develop a procedure suitable for both NPs determination and speciation analysis, they applied enzymatic extraction. The same amount of sample was incubated at 37 °C for 24 h with 2 mL of Tris–HCl buffer (pH = 7.5) and 25 mg of proteinase XXIII, then 200 μL of extract were diluted 250-fold with 1% (v/v) MeOH. Samples were analysed by spICP-MS, using the instrument software for data acquisition and evaluation. The authors compared the carbon-induced signal enhancement effect of MeOH on the ⁸⁰Se signal using solutions of 50 μg L⁻¹ Se and prepared SeNPs. In both cases they observed a significant increase that was optimised at a concentration of 1% (v/v) MeOH. An improved size LOD (47 nm) was achieved with 1% (v/v) MeOH, as compared to 61 nm for aqueous solutions. The NP concentration LOD (180 mL⁻¹) and LOQ (6000 mL⁻¹) were calculated from transport efficiency and sample uptake, on the assumption that for reliable NP detection and quantification, at least three and 100 NPs, respectively, should have been detected. The corresponding LOD and LOQ values in the original sample were then calculated as 4.5 × 10⁶ g⁻¹ and 1.5 × 10⁸ g⁻¹, respectively. In the absence of suitable CRMs, the authors analysed 4 strain samples spiked with NPs. The recovery of number concentration ranged from 46% to 102%, the most frequent diameter varied between 124 nm and 136 nm, toward an expected value of 133 nm, and the mean diameter values were within the range 138–146 nm vs. 141 nm. Repeatability for average diameter and number concentration, measured on samples of different strains (n = 3) were within the ranges of 0.6–15% and 4–11%, respectively. To extend the applicability of the method, the authors also tested two fermented products (sour cream and kefir), of which three aliquots were spiked each with one of the three selected strains analysed previously. After enzymatic digestion, recoveries of number concentration ranged from 54% to 98%. Several aspects of NP determination were explored in this work, supporting the science behind the reliable determinations of NPs in cells, and making the proposed procedure suitable for further development toward the routine measurement of NPs.

Owing to concern for the potential bioaccumulation of NPs, research is in place to unveil the mechanisms of association of NPs with living organisms. Fish and seafood are a primary target, as a vehicle of potential health risk for consumers. Single-cell inductively coupled plasma-mass spectrometry (scICP-MS) represents a powerful tool, since it allows the simultaneous quantification of both dissolved metals and inorganic NPs in single cells, in the range of attograms per cell. The group of Moreda-Pineiro¹⁶⁵ applied this technique to assess the association (either membrane adsorption or internalisation) of polyvinylpyrrolidone-coated Ag NPs and citrate–TiO₂ NPs in cells from sea bass (Dicentrarchus labrax) kidney, sea bream (Sparus aurata) kidney, and clam (Ruditapes philippinarum) gills, after several trials of exposure to NPs at different levels and of different size. They used PVP–Ag NPs (15 and 100 nm, nominal diameter) and citrate–TiO₂ NPs (5, 25, and 45 nm, nominal diameter) at nominal concentrations of 10 and 50 μg mL⁻¹ for citrate–TiO₂ NPs and 5.0 and 50 μg mL⁻¹ for PVP–Ag NPs. After exposure, cells were suspended with 4.0 mL of a freezing mixture and stored at −20 °C till analysis. After thawing and repeated washing, cells were re-suspended in 1.0% (w/v) phosphate buffered saline (PBS) solution. An aliquot of 100 μL, sampled via a loop, was introduced to the instrument at a flow rate of 10 μL min⁻¹. Silver was determined in standard mode, whereas Ti measurements were performed using NH₃ at a flow rate of 0.75 mL min⁻¹ as the reaction gas, to remove polyatomic interferences at 131 m/z. The authors assessed the effect of dwell time, finally set to 50 μs (Ag) and 100 μs (Ti), respectively, as well as optimal cell concentration and the number of washings necessary to remove any remaining NPs in the extracellular medium. The LODs, calculated by the instrument software, were 0.005 ± 0.001 fg per cell for Ag and 0.095 ± 0.011 fg per cell for Ti. Total Ag and Ti concentrations were measured using ICP-MS, after MAD of 1.0 mL of cell suspensions in 1.0% (w/v) PBS with 2.0 mL UP water–3.0 mL 69% HNO₃–1.0 mL 33% H₂O₂ and dilution of the digests to 25 mL with UP water. For comparison purposes, the measured values were divided by the number of cells counted by cytometry (9.24 × 10⁴). Values obtained using ICP-MS were found higher than the concentrations measured using scICP-MS, with ratios ranging from 6.4 to 37 for Ag (n = 4) and from 3.3 to 9.9 for Ti (n = 5). The authors attributed these differences to a lower number of cells involved in scICP-MS measurements and/or the possible saturation of the detector for scICP-MS since the exposure tests were conducted at high NP concentrations leading to a high number of NPs per cell. This work, while providing an interesting approach to for the assessment of the uptake of NP uptake by marine organisms, also sheds light on the difficulties of achieving reliable quantitative results in this new area of testing.

6. Applications: clinical and biological materials

6.1 Metallomics

A comprehensive review⁶ of recent advances in speciation and related applications complements the work covered in this Section, where we specifically address advances in speciation techniques related to the analysis of clinical and biological materials (Table 2), foods and beverages. Speciation based on extraction techniques is considered in Section 3.2 and Table 1 of this Update, whereas work related to NPs is included in Section 5. Arsenic, Hg and Se speciation, that are a rather regular feature in this Section, were also prominent over the period covered by this Update.

In the clinical area, two papers addressed the identification and quantification of metabolites of elements, following clinical treatments. Feng et al.¹⁶⁶ compared the performances of three analytical techniques (HPLC-ICP-MS, HPLC-MS/MS and IC) for the speciation of seven Gd-based MRI contrast agents (Gd-DTPA, Gd-BTPA, Gd-EOB-DTPA, GD-HP-DO3A, Gd-BT-DO3A, Gd-DTPA-BMA and Gd-DTPA-BMEA) in wastewater and whole blood. The first two techniques achieved identification of all seven compounds, but only three (Gd-DTPA, Gd-BTPA, and Gd-EOB-DTPA) could be determined by IC. The best LODs, ranging from 0.15 pg to 0.55 pg, were obtained using HPLC-ICP-MS. As part of a series of studies aimed to assess the potential of Se^IV as a chemotherapeutic drug, a group of researchers¹⁶⁷ applied LC-ICP-MS, with post-column ID with ⁷⁷Se for quantitation,¹⁶⁸ to identify Se metabolites in plasma samples of end-stage cancer patients, undergoing a clinical trial with intravenous infusion of high doses of Se^IV. Plasma samples (200 μL) were filtered to an M_r cut-off of 3000 Da. Species separation was achieved on a sample volume of 10 μL, using isocratic chromatography on a Gemini C18 column (250 × 2 mm), with a 200 mmol L⁻¹ CH₃COONH₄–5% MeOH mobile phase, at pH 6.7, and a flow rate of 0.2 mL min⁻¹. In addition to Se^IV and selenosugars, two unidentified compounds were eluted close to the SeMet and SeCys calibration standards, whereas TMSO was not detected. The concentrations of Se^IV and total Se were consistently quantifiable, at baseline, with a LOQ of 1.3 μg L⁻¹ and RSDs% of 0.37% (Se^IV, n = 20, level 2.34 μg L⁻¹) and 0.57% (total Se, n = 21, level 68.4 μg L⁻¹), respectively.

Not surprisingly, As speciation in food continues to attract interest. To improve the analytical precision of HPLC-ICP-MS for the determination of As species in food, Narukawa et al.¹⁶⁹ developed a procedure based on on-line internal standard correction, with rhenium as the IS. Both sample and IS were taken up by the autosampler, one after the other, then injected into the sampling loop. This approach improved the sampling repeatability (from 2.5% to 1.2%). Besides, sample throughput was also improved, since the step of manually adding the IS to each sample was no longer necessary. Arsenic species (AB, AC, As^III, As^V, DMA, TMAO and tetramethylarsonium, as well as arsenosugars) were separated on a RP C18 ODS L-column (250 mm, id 4.6 mm) using 10 mmol L⁻¹ sodium 1-butanesulfonate–4 mmol L⁻¹ malonic acid–4 mmol L⁻¹ TMAH–0.05% MeOH (pH 3.0) as the mobile phase, at a flow rate of 0.75 mL min⁻¹. The ICP-MS was operated with He as the CRC gas (3 mL min⁻¹). Two CRMs were analysed to test the procedure. Sample preparation consisted of (a) for NMIJ CRM 7533-a (brown rice flour), heat-assisted extraction of 0.5 g with 2 g of 0.15 mol L⁻¹ HNO₃ at 100 °C for 2 h, followed by dilution to 10 g with water; (b) for NMIJ CRM 7405-b (hijiki seaweed) ultrasonication of 0.5 g with 20 g water for 1 h, followed by centrifugation, filtration and dilution of the filtered solution 1 + 9 with water. Remarkably, recoveries were 100% for all certified values (As^V in NMIJ CRM 7405-b and iAs and DMA in NMIJ CRM 7533-a). Water-soluble As species (As^III, As^V, AB, DMA and MMA) from oil and brine canned solid tuna samples were extracted in 10 mmol L⁻¹ (NH₄)₂CO₃–0.05% (w/v) EDTA, at pH 10.5, and separated using IC, with a Dionex IonPac AS7 (2 × 250 mm; 10 μm) anion exchange column and the same solution as the mobile phase.¹⁷⁰ Three mL of the mobile phase were added to 200 mg of the freeze-dried samples and the mixture sonicated for 15 min. The procedure was repeated twice, after centrifugation at 3200 rpm for 10 min. The combined supernatant from all three extractions was then diluted to a final volume of 10 mL with the same solution. With this sample preparation, there was no residual matrix effect and As species could be quantified, using external calibration, by means of ICP-MS with a KED cell, using He as the reaction gas. The LOQs were 0.59 mg kg⁻¹ (AB), 0.12 mg kg⁻¹ (As^III), 0.10 mg kg⁻¹ (As^V), 0.10 mg kg⁻¹ (DMA) and 0.07 mg kg⁻¹ (MMA). The analysis of a CRM (BCR-627 tuna fish muscle tissue) yielded recoveries of 93.8% for AB and 95.5% for DMA. After reports of high As levels found in two mushroom species (Agaricus blazei Murrill and Tricholoma matsutake), the concentrations of both total As and As species (iAs, AB, AC, DMA, MMA, TMAO and TMA⁺) were investigated by Chen et al.¹⁷¹ to assess potential health risks for the consumers. To this aim, total As was determined using ICP-MS, whereas HPLC-ICP-MS with a CRC using He to minimise polyatomic interferences was applied to quantify the As species. Anion and cation exchange chromatography provided complementary separation of the As species, so the method employed both. Anion-exchange chromatography was carried out on an IonPac AS7 column with 30 mmol L⁻¹ NH₄HCO₃ (pH 9.5) as the mobile phase, at a flow rate of 1.2 mL min⁻¹. A PRP-X200 column with a mobile phase of 10 mmol L⁻¹ pyridine (pH 3.3) at a flow rate of 1.0 mL min⁻¹ was used for the cation-exchange chromatography. Sample pre-treatment consisted of ultrasonication of 0.2 g of sample powder, with 10 mL deionised water, at 60 °C for 30 min, followed by centrifugation and filtering of the supernatant. The LOQs for the seven studied species ranged from 3.9 μg kg⁻¹ to 6.9 μg kg⁻¹ and recoveries between 89% and 99% were obtained for spiked amounts of all the analytes at 50, 200 and 1000 μg kg⁻¹.

Given the continuous interest in Se supplementation through food, a method was developed,¹⁷² based on HPLC-ICP-MS/MS, for the determination of Se species in selenium-enriched foods (spring water, salts, and tea). The chromatographic separation was accomplished using a Hamilton PRP-X100 (10 μm, 4.1 × 250 mm) anion exchange column and a mobile phase of 25 mmol L⁻¹ citric acid–2% MeOH, at pH 4.0. Samples of Se-enriched drinking water were only shaken evenly and filtered before analysis. Salt and tea leaves were dried at 50 °C overnight. A 2.0 g aliquot of salt was dissolved in 50 mL of water and the solution filtered. About 5.0 g of the tea leaves were extracted twice with 50 mL boiling water for 30 min, then the supernatant was collected for analysis. The ICP-MS/MS applied O₂ as the reaction gas, to overcome polyatomic interferences. Under these conditions, six Se species were completely separated within 20 min, with LODs of 0.04 ng mL⁻¹ (Se^IV), 0.02 ng mL⁻¹ (Se^VI), 0.02 ng mL⁻¹ (SeCys₂), 0.15 ng mL⁻¹ (selenoethionine), 0.03 ng mL⁻¹ (SeMeCys), and 0.05 ng mL⁻¹ (SeMet), respectively. Recoveries between 93.7% and 105%, with RSD% <5%, were obtained on spiked samples of selenium-enriched foods.

To achieve the speciation of iHg, EtHg and MeHg from fish (hairtail, cuttlefish and octopus) tissues and lake and tap water samples, Liu et al.¹⁵² developed a high-efficiency photo-oxidation reactor, passing through a low-pressure mercury lamp, allowing the complete degradation of the organic Hg species, within 4 s (MeHg) and 6 s (EtHg), respectively. Quantification of the Hg species, by means of AFS, could then be achieved by on-line coupling with FI (FI-UV-CV-AFS) or (HPLC-UV-CV-AFS). Water samples were filtered and acidified with diluted HCl, then stored at 4 °C. Fish tissues were freeze-dried, then a 0.2 g aliquot was incubated overnight with 2 mL of 25% (w/v) KOH in MeOH. Afterward the solution, neutralised by addition of 1.5 mL of conc. HCl dropwise, was extracted with 6 mL of CH₂Cl₂, shaking for 15 min, and the organic phase was back-extracted with 1 mL of 10 mmol L⁻¹ sodium thiosulfate, shaking for 45 min. Hg species were separated by HPLC on a ZORBAX SB-C18 column (2.1 × 50 mm, 5 μm) with 0.06 mol L⁻¹ CH₃COONH₄–0.1% (v/v) 2-mercaptoethanol as the mobile phase. The effluent was mixed with the carrier stream (0.5% HCl), entered the photo-oxidation reactor, then was mixed with SnCl₂ to form volatile Hg vapour, that was introduced into the AF spectrometer for quantification. The LODs were 0.55 ng mL⁻¹ (Hg²⁺), 0.56 ng mL⁻¹ (EtHg) and 0.84 ng mL⁻¹ (MeHg), respectively. Recovery for the certified value of MeHg (840 ± 30 ng g⁻¹) in the NRCSM GBW10029 tuna fish was 100%, but no other information is provided to validate the determination of EtHg. Only iHg and total Hg could be separately determined by FI-UV-CV-AFS, with an LOD of 0.02 ng mL⁻¹, by turning the photo-oxidation reactor on and off. Analysis of lake water and tap water samples spiked with different concentrations of iHg, EtHg and MeHg gave recoveries of 100% for both iHg and total Hg. Repeatability for both methods was <5% RSD.

6.2 Imaging with LIBS, MS and X-rays

Over this review period, the number of papers in this area has increased in comparison to our previous Update.¹ Additionally, two review papers have covered the use of spectroscopy-based imaging for biological research. Davison et al.¹⁹ provided an evaluation of single cell analysis utilising spICP-MS, LIBS and LA-ICP-MS. The advantages and limitations were assessed, concluding that more than one technique was required to fully elucidate processes and that calibration and standardisation still require further development. However, these techniques offer significant potential for bioanalytical chemical research at the single cell level. Peng et al.²⁰ reviewed multiple laser imaging techniques to examine the use of nanomaterials to improve cancer diagnosis in vivo. The authors covered fluorescence imaging, LIBS, photoacoustic imaging and surface enhanced Raman spectroscopy (SERS), citing examples of research demonstrating the potential of these techniques for cancer detection and elucidation. The current limitations and challenges were also discussed, providing a useful oversight of the current status for bioimaging with lasers.

The application of LA-ICP-TOF-MS for bioimaging has continued to demonstrate its ability as a powerful multi-elemental quasi-simultaneous technique. Strekopytov and co-workers¹⁷³ published a comprehensive study on the quantitative imaging of Cu Fe, Mg, Mn and Zn in animal tissue. A low-dispersion ablation cell was utilised alongside a dual concentric nebuliser for maximum sample transfer. Following optimisation, the method achieved a spatial resolution of 3 μm correlating to detection limits of 4–9 fg per pixel for Cu, Mn and Zn and 40 fg for per pixel for Fe and Mg. Quantification was achieved through production of doped gelatin standards. The researchers compared the impact of the CRC with H₂ gas and mass resolution of the TOF (m/Δm approximately 6000 at full width at half maximum height) for interference removal of ⁴⁰Ar¹⁶O⁺ on ⁵⁶Fe⁺. Whilst H₂ did successfully reduce the interference, there was a three-fold increase in the LOD when compared to no CRC gas with mass resolution and peak deconvolution. This impact was also reflected in the other analytes of interest due to poorer counting statistics from sensitivity loss in H₂ mode. The optimised approach was applied to rat kidney tissue samples from a trial assessing the impact of bis-choline tetrathiomolybdate to regulate Cu levels in Wilson's disease patients. Therefore, the method was extended to include the determination of Mo. High-resolution images were generated showing the difference in elemental levels and Mo/Cu ratios in the kidney samples. Metarapi et al.¹⁷⁴ developed a semi-quantitative calibration model for rapid multi-elemental bioimaging with LA-ICP-TOF-MS. Gelatin standards were prepared in three sets for HNO₃ stable elements (n = 48), HCl stable elements (n = 7) and HNO₃/HF acid stable elements (n = 17), giving a total of 72 elements. The mixtures were deposited onto glass slides using a microdroplet spotter. A response factor library was generated using the standards which was tested to show errors were <25% for the majority of elements, with several below 10% which is excellent for a solid sampling semi-quantitative approach. The method was then applied to tissue samples and compared against full quantitative data, again achieving good correlation. The use of response factors could enable researchers to gain information about other elements which may not be the focus. Additionally, the authors also provided the calculation model as an open access web-based tool. Researchers from this paper also contributed to the study by Schaier et al.¹⁷⁵ who utilised LA-ICP-TOF-MS as an imaging mass cytometry technique. The impact of sample preparation of tissues (sectioning and immunolabelling) on the elemental distribution was assessed in five mouse tissue types (kidney, liver, lung, spleen and tumour). As in the previous work, spiked gelatin standards were produced with a microdroplet spotter and employed for quantitation. By using consecutive slices, the impact of preparing the tissues by either formalin fixation paraffin embedding and cryo-sectioning was determined by analysis of the endogenous levels of Ca, Cu, Fe, K, Mg, Na, P and Zn. It was found that significant changes occurred, with losses due to washing steps or gains due contamination from reagents, therefore quantification was not possible, but qualitative distributions of Fe, Na and P were possible. A panel of 18 metal-conjugated antibodies (tagged with enriched isotopes of various REEs) covering structural, immune cell and tumour markers were used for immunolabelling along with the traditional haematoxylin and eosin stain. The subsequent quasi-simultaneous LA-ICP-TOF-MS analysis provided an incredibly insightful view into the distribution and co-localisation of relevant biomarkers in a single tissue section, highlighting the power of the technique to extract significant information at the cellular level.

Continuing on the theme of calibration approaches for LA-ICP-MS, several papers have focused on this challenge. Seiffert et al.¹⁷⁶ developed a strategy to determine the location and size of nanoparticles in the tissues. The first consideration was whether the ablation process caused disintegration of the NPs which would affect the size distribution. The researchers demonstrated this did not occur for LA-ICP-MS with Ag and CeO₂ NPs embedded in gelatin with equivalent data obtained by spICP-MS and TEM. The gelatin calibration method was then applied to spleen tissue sections from rats exposed to CeO₂ particles after 3 h, 3 days and 3 weeks. The NPs were located and sized accurately which was confirmed via TEM demonstrating the suitability of the approach. The importance of matrix matched calibration standards for LA-ICP-MS was highlighted by Westerhausen et al.¹⁷⁷ during the elemental analysis of teeth. Often the NIST glass SRM series are used as calibrants for hard materials but differences in ablation behaviour can lead to errors in quantitation. The researchers compared three preparation methods using doped co-precipitated hydroxyapatite to better match the tooth composition. The simplest method of grinding the material into a powder and spiking with Al, Ba, Cd, Cu, Mg, Ni, Pb, and Zn, achieved sufficient homogeneity and concentration accuracy. Good linearity was observed with R² > 0.99 for all elements except Cu which was 0.98. The LODs were in the range of 0.1 to 2 μg kg⁻¹ which is especially respectable for a solid sampling technique. These standards were then used to determine the element levels in two teeth, one from a remote indigenous Australian community and the other from Sri Lanka. Structural features could be identified in the teeth cross sections based on the elemental profiles as well as providing a time record of exposure to Cd and Pb. In a similar approach, Deng et al.¹⁷⁸ produced calibrants from doped calcium oxalate for analysis of elements in urinary stones. Co-precipitation was also used and was suitably homogeneous for Co, Cr, Cu, Fe, Mg, Mn, Sr and Zn. This was applied to a single sample, showing correlations with physical features of the stone which may provide useful data for urological investigations.

A number of publications have focused on quantitative imaging in food products. Pereira and co-workers¹⁷⁹ assessed the elemental distribution of As, Ba, Co, Cu, Fe, Mg, Mn, Pb, Sb, Sr and Zn in rice grains. Calibrants were prepared simply by spotting aqueous standards onto pre-cleaned filter paper. Accuracy was checked with two CRMs (NIST SRM 1515 and BCR-060) pressed into pellets. For the detectable elements, reasonable recoveries were found and the LODs were in the range of 0.03 to 0.60 μg g⁻¹. Three rice samples (white, parboiled and brown) were then quantitatively analysed. The images showed the elemental levels were decreased in the inside layer compared to the outside. Interestingly, Pb was only found in the first layer of the white and brown rice whereas Pb had penetrated the second layer in the parboiled rice. Braeuer et al.¹⁸⁰ investigated the levels of Hg and Se in edible mushrooms in view of the health risks from Hg accumulation or potential health benefits from Se. Porcini mushrooms from three different species (Boletus edulis, Boletus aereus and Boletus pinophilus) and one parasol mushroom (Macrolepiota procera) were collected and analysed. For quantitation, gelatin calibrants were produced doped with Hg and Se, and the impacts of chitosan and L-cysteine were assessed to provide better matrix-matching. An in-house reference material was prepared from an homogenised mushroom paste and characterised by acid MAD and ICP-MS analysis. The RM was used for method validation, establishing that gelatin with L-cysteine was most suited and was used for the quantitative imaging of the mushroom components. Excellent LODs were achieved at 0.006 μg g⁻¹ and 0.3 μg g⁻¹ per 20 μm² pixel size for Hg and Se respectively. The Hg concentrations were 0.07–4.45 μg g⁻¹, which thankfully means 0.3–2.5 kg of mushrooms could be safely consumed per week. The Se levels were 0.19–26.2 μg g⁻¹ which distinctly varied between the species. The results also showed Hg and Se were 2 to 6 times higher in the peripheral tissue of the cap and stalk compared to the inner tissue. The study provided a useful insight into the distribution of Hg and Se in mushrooms but also highlighted the need for well-matched calibrants for quantitative LA-ICP-MS.

The use of LA-ICP-MS as a tool for disease diagnosis and biomarker detection has received significant attention during this review period. Stiborek et al.¹⁸¹ described the development and use of a new IR LA system at 2940 nm compared to the traditional UV lasers, typically at 193–266 nm. Gold NPs were used as an elemental tag to demonstrate the benefits of the new instrument for tissue imaging. Here, the proliferation marker for human tumour cells, protein Ki-67, was selected as the biomarker in colorectal carcinoma cells. The IR LA-ICP-MS was utilised in sp mode for particle counting and compared with confocal fluorescence microscopy and UV LA-SP-ICP-MS. The analysis produced clearly defined distribution maps of the Au NPs, therefore enhancing the detection of the biomarker. With IR LA-spICP-MS, the images were more distinct due to background suppression in the regions surrounding the tissue. Although it is an in-house designed and built system, it demonstrated benefits over existing commercial UV systems and could be easily extended for multiplexed mapping of low-abundant important biomarkers. Tisza et al.¹⁸² employed LA-ICP-MS to investigate the distribution of Pt in tumours from patients with pleural mesothelioma, a rare cancer with poor prognosis, following treatment with Pt-based drugs. Large differences were found in the Pt distribution between tumorous and non-tumorous (fibrotic) areas with the latter at significantly higher levels. In combination with immunostaining, the Pt locations could be attributed to specific features. Correlations were also observed in comparison with the tissue sections and Pt concentrations in serum. Furthermore, overall survival outcome was longer in patients with low serum Pt concentration compared with those with high serum levels. The work provided useful data on the impact of Pt-drug distribution for this aggressive disease. Veverkova and co-workers¹⁸³ published an investigation on the use of magnetic particles with molecular imprinted polymers for the detection of metallothionein as a marker for melanoma. The molecular imprinted polymers are an alternative to antibody labelling and in combination with magnetic particles provided specific isolation of the analyte from the tissue matrix, in this case metallothionein. The extraction efficiency was determined as 63.3% and was suitably selective for metallothionein. The methodology was then applied to melanoma and control tissues, with the tissue homogenates spotted on to polyvinylidene fluoride (PVDF) membranes for spot analysis by LA-ICP-MS, providing rapid detection and quantification of metallothionein in melanoma and healthy skin. Additionally, Zn in the tumour samples was significantly higher than in healthy skin. In another interesting application of LA-ICP-MS, Bucker et al.¹⁸⁴ utilised enriched ⁵⁸Fe to investigate the impact of blood release caused by intracranial haemorrhaging. The haemorrhage was induced in mouse models with ⁵⁸Fe enriched whole blood. The mice were sacrificed at different time points after injection (0, 48, 96 and 144 h), with the brain tissue immediately frozen and cryosectioned. The use of isotopically enriched ⁵⁸Fe, which has a natural abundance of 0.282%, provided differentiation of the endogenous blood Fe enabling the researchers to track the fate of Fe in the brain. A visualisation model was then built to generate images of the ⁵⁸Fe delocalisation over time. The work was a clever approach to understand Fe and Hb distribution following traumatic brain injury.

The determination of fluorine is a challenge with conventional atomic spectrometric techniques, so Ma et al.¹⁸⁵ reported an approach employing cryogenic laser ionisation TOF-MS to assess the F content of teeth. The setup used liquid nitrogen to cool the ion source to −25 °C and pressurised the chamber to 300 Pa with He gas. A 532 nm laser at 8 ns duration and 5 Hz repetition rate was used to ablate the tooth sample, followed by extraction into the TOF-MS. The interferences at m/z 19 were effectively removed allowing F analysis at a resolution of 20 μm. It was also possible to detect C, Ca, Cl, Mg, N, Na, O and P, with quantification achieved using doped hydroxyapatite pressed pellets. A number of human samples were analysed with and without treatment for anti-caries topical drugs to assess the penetration levels. The work provided a unique insight into F behaviour in teeth. The use of SIMS was described by Lovric and co-workers¹⁸⁶ for sub-cellular imaging of Fe in lung macrophages. In this study, nanoSIMS and helium ion microscopy SIMS were compared against traditional TEM and backscattered electron microscopy to track the distribution of Fe. The nanoSIMS achieved a resolution of ∼250 nm whilst the helium ion microscopy SIMS obtained sub-20 nm resolution. The nanoSIMS and TEM demonstrated strong correlation and similarly with helium ion microscopy SIMS and backscattered electron microscopy for the high-spatial-resolution structural cell images. The results revealed accumulation of iron in mitochondria, lysosomes and vacuole-like organelles, with SIMS providing more in-depth chemical data, providing insights into Fe metabolism at the cellular level. Penen et al.¹⁸⁷ utilised TOF-SIMS imaging to investigate the distribution of the cancer drug trastuzumab which is used for tumours with high levels of human epidermal growth factor receptor 2 (HER2) present, such as ovarian and breast cancers. Mouse models were exposed to Zr-labelled trastuzumab and sacrificed 3 days after exposure, with tumours extracted, frozen at −80 °C and sectioned. The TOF-SIMS imaging was compared against immunohistological staining. The results demonstrated good correlation between the techniques and TOF-SIMS achieving a resolution of 1.8 μm in the focused mode. Regions of interest were clearly defined and differences in uptake were observed. The study provided useful data for the distribution of the therapeutic drug in tumours.

In a short communication by Busser et al.,¹⁸⁸LIBS was implemented for elemental imaging of brain tissue for endogenous Zn and NPs. For Zn, quantitative data was obtained using doped resin standards, achieving a LOD of 2 ppm and R² of 0.9995. Additionally, Ca, Cu, Fe, Mg, Na and P were detected simultaneously. Following exposure to LaF₃:Ce NPs, rat brain tissue was collected from various stages of embryonic development and analysed. Quantification of elemental levels and localisation of NPs demonstrated the potential of LIBS, which showed considerable promise.

Imaging applications utilising XRF spectrometry for food products are not typically a regular feature within this section of the review, however, two papers have covered this topic. Forste et al.¹⁸⁹ focused on 2D and 3D imaging of cocoa beans by multiple XRF spectrometry approaches, namely, μXRF, XRF tomography and quantitative XRF. This provided distribution maps for the cocoa bean, shell and cotyledon tissue and the impact of fermentation and roasting was also considered. Bulk quantitative analysis of pressed pellets was achieved for Ag, Ca, Cd, Cr, Cu, Fe, K, Mn, Mo, Ni, P, Rb, S, Se, Sr, Zn and Zr. Furthermore, the LODs and measurement uncertainty were calculated. The work provided a robust assessment of elemental distributions in cocoa beans. In a similar approach, Deng and co-workers¹⁹⁰ used SR-μXRF spectrometry to determine the distribution of Cu, Fe, K, P, S and Zn in oat (Avena sativa L.). Different parts of the grain were imaged showing different elemental patterns. This information could be utilised in improve nutrient fortification in this important crop.

The use of XRF spectrometry for biochemical research has received some interest in this review period. Kim and co-workers¹⁹¹ combined XRF spectrometry and computed tomography (CT) to investigate the in vivo biodistribution of Au NPs. A transmission CT detector was installed in an existing pinhole XRF instrument, enabling dual detection. The technique was tested using mice injected with 1.9 nm Au NPs, with XRF and CT scans of the kidney and tumour areas collected within 60 min. The images provided a time lapse of the Au NP distribution. The animals were sacrificed and the ex vivo concentrations of Au NPs were determined by ICP-MS, agreeing well with the in vivo data. The study provided a unique and novel insight into real time biodistribution and behaviour of NPs in vivo. In another dual technique application, Szczerbowska-Boruchowska et al.¹⁹² published a method to correct for the impact of sample thickness in tissue slices leading to errors in imaging. By using FTIR microscopy to determine the total absorbance, this was correlated to the sample surface mass and a correction factor was calculated. Tissue samples of brain, heart, kidney, liver and muscle from rats were analysed by FTIR and SR-XRF. The correction factor from FTIR was compared to that generated from the Compton scattering, with equivalent results obtained. However, the FTIR offered the benefit of providing additional molecular data and applicability when Compton scattering cannot be used. By including this correction, more accurate quantitative imaging data could be obtained.

Morrell et al.¹⁹³ proposed an approach to overcome the limitations of using consecutive tissue sections to determine the distribution of exogenous particles when employing traditional histological staining and elemental imaging. High variability was found when using adjacent slices, leading to poor comparability between techniques and potential misinterpretation. The researchers proposed using confocal SR-XRF spectrometry and lanthanide antibody tagging for simultaneous detection in a single section. Proof of concept was demonstrated with spleen and liver samples from mice exposed to Ti NPs and tissues from human subjects exposed to Ti hearing aid implant anchors. The study established the approach as a valuable method for simultaneous detection of biomarkers and exogenous particles.

6.3 Elements as tags for indirect determinations

Elemental tagging for biological detection and imaging continues to be an expanding area of interest, with the emergence of various novel platforms for multiplexing in nucleic acid-based diagnostics. In the method developed by Zhan et al.,¹⁹⁴ multiplexed HPV-DNA detection was achieved using a clustered regularly interspaced short palindromic repeats (CRISPR) platform with DNA tetrahedron (DTN) functionalised magnetic beads (MBs) and tagged metal NP probes. NP-bound linker probes identified by the CRISPR system were cleaved, preventing hybridisation with DTN capture probes on MBs, thus allowing for ICP-MS detection following magnetic separation. Multiplexing of HPV-16, HPV-18 and HPV-52 target probes was achieved via differential functionalisation of Ag, Au and Pt NPs with ICP-MS detection. The proposed assay was used to quantify HPV-DNA in a surgical swab with an LOD of 218 fmol L⁻¹. Kang et al. reported on a renewable platform for multiplexed microRNA (miRNA) quantification using lanthanide labelling with entropy-driven catalytic (EDC) amplification.¹⁹⁵ Target miRNA displaced lanthanide labelled single stranded DNA from linker DNA on MBs in a competitive type immunoassay.¹⁹⁵ Quantification of lanthanide labelled DNA was carried out by ICP-MS and fuel DNA then added to bind linker DNA, displacing miRNA and allowing for regeneration of the EDC complex on MBs. Whilst the authors state the MB platform can only be used four times, this is an important step towards renewable bioassays and reducing reagent consumption.

Various research groups have investigated the use of multiplexed elemental tagging for cancer diagnosis. In the paper by Wen et al.,¹⁹⁶ quantification of cancer related miRNAs demonstrated accurate tumour classification in clinical samples from a cohort of 42 patients (14 breast cancer, 5 lung cancer, 5 gastric cancer, 3 cervical cancer, 3 colon cancer and 12 healthy controls). Ten different NaLnF4 nanosatellites, with specific DNA modulations, were prepared and hybridised to catalytic hairpin assemblies (CHA) on a MB core. Target miRNA competes for binding to the CHA, releasing the lanthanide-based nanosatellites for ICP-MS detection following magnetic separation. Zhu et al.¹⁹⁷ successfully employed spICP-MS to quantify AFP and CEA in a solid phase immunoassay platform. The recognition AFP and CEA caused dehybridisation of surface bound aptamers initiating a RecJ_f exonuclease-assisted target recycling amplification process. Chelation of AuCl⁻ and Cu²⁺ into resultant double stranded DNA chains enabled the formation of Au and Cu nanoclusters which were detected by spICP-MS. While this technology achieved wide dynamic ranges for both AFP (0.01–200 ng mL⁻¹) and CEA (0.01–100 ng mL⁻¹), upper reportable limits remain inferior to conventional methods. Other proposed platforms used spICP-MS for multiplexed detection of breast cancer miRNAs (MiR-21, MiR-155 and MiR16) and gastric cancer biomarkers (CA724, CA199 and CEA).^198,199 Breast cancer detection was achieved via spICP-MS quantification of DNA modified Ag, Au and Pt NPs competitively displaced from sandwich structures by target miRNA. Alternatively, a multiplexed sandwich based immunoassay format, with antibody-labelled Ag, Au and Pt NP tags and spICP-MS detection, was used to quantify gastric cancer biomarkers. In this case the concentration of biomarker was inversely proportional to the number of unbound NPs after magnetic separation.

In addition to the above, cancer diagnosis is also a primary focus for quantitative analysis of biomarkers using LIBS and LA-ICP-MS. A sandwich immunoassay for the Human Epididymis protein-4 (HE4) ovarian cancer biomarker was developed using LIBS measurement of bound Si particles.²⁰⁰ The authors acknowledged limitations in the clinical utility of this assay with regards to lack of specificity due to non-specific cross-coupling and the limited dynamic range (0.0–0.0073 pmol L⁻¹) which necessitated serial sample dilution.²⁰⁰ Veverkova et al. reported adaptations to their previously described sandwich immunoassay for metallothionein with LA-ICP-MS readout.¹⁸³ Magnetic particles were modified by molecularly imprinted polymers and incubated with sample and spotted onto PVDF for LA-ICP-MS analysis.¹⁸³ Specificity was confirmed using anti-metallothionein antibody labelled Ag NPs as tags.¹⁸³ In a different approach, lanthanide labelled antibodies were employed for multiplexed quantitation of MMP-11 and CD45 in immunohistochemistry-assisted LA-ICP-MS.²⁰¹ Analysis of paraffin embedded breast cancer (n = 21) and healthy breast tissue samples (n = 4) showed overexpression of MMP-11 in metastatic tissue compared with other tissue types (P < 0.01) but no significant differences in the expression of CD45.²⁰¹ These results support previous findings relating to the potential of MMP-11 as a prognostic biomarker.²⁰¹

Further applications for elemental tagging include the use of nanoconjugates and quantum dots (QDs) for the detection of glutathione, myeloperoxidase and osteopontin.^202,203 Glutathione was quantified using either fluorescence or ICP-MS in a dual mode protocol with a g-C₃N₄:Tb/MnO₂ nanoconjugate. In the absence of glutathione, MnO₂ quenched fluorescence emission at 546 nm from Tb³⁺. Presence of glutathione reduced MnO_2, restoring Tb³⁺ fluorescence emission at 546 nm and generating Mn²⁺ quantified via ICP-MS. Whilst both readout methods demonstrated selectivity, superior LODs were obtained with ICP-MS quantitation (0.016 μmol L⁻¹) compared to fluorescence (0.17 μmol L⁻¹). Cao et al. employed MoS₂ and ZnS QDs for multiplexed determination of myeloperoxidase and osteopontin respectively in a sandwich immunoassay technique. Myeloperoxidase and osteopontin were captured on amino modified magnetic nanoparticles and subsequently labelled with secondary antibody labelled QDs. Following wash steps, QDs were dissociated from the complexes using formic acid and quantified via spICP-MS. Whilst results obtained were in agreement with a routine chemiluminescence assay, only 3 samples were used for this method comparison.

6.4 Multi-element applications

Non-targeted multielement analysis in a single sperm cell was achieved using single cell ICP-TOF-MS.¹⁰⁴ Complete mass spectra of pre-treated, dispersed sperm cells were obtained and datasets were interrogated by machine learning-based dimension reduction analysis. Data analysis was able to illustrate heterogeneity of each element within the sample and extract key information to create hierarchical clusters of elements based on their physiological role. Further data analysis, using a restricted cubic spline model, revealed a non-linear correlation between cellular Zn content and sperm quality, consistent with previous conclusions suggesting an optimal concentration range. Whilst this work had some limitations regarding sensitivity for particular elements, the methodology is applicable to other cell types to elucidate further understanding of elemental distribution patterns.

In contrast to the single cell approach, Liu et al. described a method for high throughput quantitative elementomics in biological samples such as human serum, plasma and urine.²⁰⁴ Quantification of 70 elements was achieved by ICP-MS in 5 min from a 50 μL sample volume. Reference ranges for all quantified elements were derived from analysis of plasma samples from 127 healthy Chinese adult participants.

6.4.1 Specimens analysed to investigate metallic implants and biomaterials. Again this year there were few reports relevant to this Update in the area of metallic implants. One paper of interest used different analytical approaches to study soluble and insoluble Cr species in periprosthetic tissues from 13 patients with failed Co/Cr/Mo implants.²⁰⁵ Soluble Cr was extracted, using an alkaline EDTA solution, from homogenised tissue that had been spiked with Cr^III to account for any experimental interconversion of Cr^III to Cr^VI. The supernatant was filtered (3 KDa) to remove NPs and protein complexes. Analysis of the supernatant by ion exchange HPLC-ICP-MS demonstrated that in all cases, Cr was present as Cr^III and no carcinogenic Cr^VI was detected. A MAD and ICP-MS method was then utilised to determine total Cr in: digested tissues; the alkaline extract, which also served as a quality control for the Cr speciation, and the solid residue remaining from the extraction. The Cr in the soluble fraction accounted for <1% of the total tissue Cr with the highest Cr content occurring in the insoluble residue (from 2.8 μg g⁻¹ to 5119 μg g⁻¹). Subsequent FTIR analysis showed Cr to be predominantly present in the form of CrPO₄ particles. The highest Cr tissue content was observed in patients with aseptic loosening of the joint as the indication for the revision. A more detailed study, which included extraction and separation of metal–protein complexes by blue native PAGE and quantitation of bands of interest by MAD and ICP-MS, was then performed in a patient with extensive metallosis. The overall distribution of Cr in this patient's tissues was: soluble Cr, 2.3 μg g⁻¹; Cr NPs in alkaline extract, 5.9 μg g⁻¹; insoluble Cr in sediment (CrPO₄), 5118 μg g⁻¹, albumin-bound Cr, 2.1 μg g⁻¹. A similar distribution pattern was also observed for Al, Co, Mo and Ti.

Two papers assessed longitudinal serum metal ion concentrations in patients with metal implants. Smolle et al.²⁰⁶ determined serum Ag concentrations by MAD and ICP-MS in patients with Ag-coated megaprostheses, where the Ag coating serves to create an antimicrobial environment around the implant. Patients had a baseline Ag measurement within several days of receiving the implant and then every six months thereafter for a median of 49.5 months. An initial increase in serum Ag concentrations compared with baseline was observed followed by stabilisation at a lower than baseline concentration (Ag concentration (median (interquartile range)) at the final follow-up vs. baseline: 7.4 (2.7 to 14.1) vs. 16.0 (9.1 to 29.1) ppb, (n = 26)). While there was no significant association of serum Ag concentration with time since surgery or renal function, increased serum Ag concentrations were significantly associated with the development of periprosthetic joint infection. From a clinical perspective, four patients developed localised argyria but there were no manifestations of systemic Ag toxicity identified within the cohort. Cundy et al.²⁰⁷ on the other hand, monitored serum Co, Cr, Ni and Ti concentrations over time in 56 children with metal spinal implants by way of HR-ICP-MS. The measurements were performed pre-operatively and then at various time points up to two years post-operatively. There was no significant trend in serum Ni concentrations over the study period. Concentrations of the other three metals rapidly increased within the first seven days and then peaked at 30 days. However, while Cr and Co serum concentrations then declined to either pre-operative or slightly increased concentrations respectively, Ti concentrations were either maintained or even increased at two years (median Ti concentration at two years vs. baseline: 2.338 vs. 0.452 ppb, n = 42). The authors highlighted concerns around the persistently raised serum Ti concentrations in the context of possible future pregnancies of the subjects, as Ti is known to cross the placenta.

6.4.2 Biological fluids and tissues. This year a spotlight has been shone on elemental profiling of urinary stones using XRF and LA-ICP-MS.^178,208,209 Bali et al.²⁰⁸ commented on the use of WDXRF for multi-element quantitation in 9 different stone samples following FTIR classification. Samples were prepared by pressing in boric acid and the concentrations of 23 different elements were measured by WDXRF. Good comparability between WDXRF and AAS analysis was demonstrated for As, Cr, Cu, Ni and Pb and spectroscopic results correlated with clinical details of the patients. Whilst no novel conclusions were reached regarding stone composition, the authors reckoned that WDXRF provided a suitable platform, with relatively simple sample preparation, for routine clinical analysis of stones. In a different approach, TXRF was used to quantify 23 elements in 30 low volume (few milligrams) urinary stone samples, collected from individuals living in arid areas in the middle east.²⁰⁹ Samples were prepared by MAD with UP HNO₃ and quantitative TXRF analysis was validated by comparison with ICP-OES measurements. Elemental correlations were interrogated using various statistical approaches including Pearson’s correlation, HCA and ANOVA. The number of quantified trace elements was found to decrease with increased calcium oxalate enrichment and U was detected in 3 samples, suspected to originate from exposure to untreated groundwater. Deng et al.¹⁷⁸ described the use of LA-ICP-MS to elucidate quantitative distribution patterns for Co, Cr, Cu, Fe, Mg, Mn, Sr and Zn in renal stones using the CaC₂O₄ precipitate CaOx-1 as a calibration standard. Elemental distribution maps showed an annular texture for Cu, Mg, Sr and Zn and homogeneous like distributions for Co, Cr and Fe, while Mn was conc. in the margin of the stone.

Analysis of pleural mesothelioma (PM) tissue samples by LA-ICP-MS revealed differences in Pt accumulation between non-tumorous and tumorous tissue compartments.¹⁸² Post-surgical tissue samples and concomitant serum samples were collected from 25 PM patients treated with Pt-based chemotherapy. Tissue samples were pre-treated with Au as a pseudo IS for Pt quantitation by ICP-MS. Whilst Pt concentrations were higher in non-tumorous tissue compared to tumorous tissue (1.23 μg g⁻¹vs. 0.83 μg g⁻¹ respectively, Mann–Whitney U test P = 0.0031), no clear correlation was observed with overall survival rates. In contrast, low serum Pt levels (<177.78 μg L⁻¹) (log rank test, P = 0.029) were associated with improved overall survival in Kaplan–Meier estimates.

Exhaled breath condensate (EBC) is emerging as a potentially useful non-invasive biological matrix for human biomonitoring. As part of the Human Biomonitoring for Europe (HBM4EU) group initiative, EBC was investigated as a potential matrix for monitoring exposure to Cr^VI in a multi-centre study (Finland, France, Italy, UK and the Netherlands).²¹⁰ Pre-working week and post-working week EBC samples were collected from 177 workers occupationally exposed to Cr^IV, along with control samples (n = 98) from a non-exposed cohort. Samples were stabilised with 0.5 mmol L⁻¹ EDTA (adjusted to pH 8 using 10% v/v NH₄OH) immediately after collection and Cr speciation analysis was performed using either LC, μLC or IC with ICP-MS depending on the instrumentation available at the participating laboratories. Of the three exposed groups (chrome platers, welders and surface treatment workers), EBC samples from chrome platers had the highest concentrations of Cr^IV, consistent with findings from other biological sample types. Although this indicates EBC may be a suitable matrix to assess occupational exposure, sample volumes were highly variable and concentrations of Cr^IV in EBC did not correlate well with concentrations in other sample types. Marie-Desvergne et al.²¹¹ investigated EBC as a non-invasive alternative to bronchoalveolar lavage (BAL) and bronchial washes (BW) to monitor lung particle burden. The concentrations of 12 elements in EBC, BAL and BW samples from 82 patients with interstitial lung diseases were determined using ICP-MS (EBC) or ICP-AES (BAL and BW). The authors noted an inverse correlation between the levels of Si in EBC compared to those in BAL, which they stipulated may be attributed to high lung load and low exhalation. However, it should be noted that the correlation coefficient was 0.515 which is extremely weak and further work would be required to expand on this.

Reliability of toenail samples as a non-invasive matrix for monitoring metal exposure has been questioned in previous reviews. Lin et al. documented a study looking at the intra-individual variation in elemental compositions of toenail samples collected at different time points from male participants in the Gulf Long-term Follow-up (GuLF) study.²¹² Study participants (n = 123) provided 2 toenail samples 2–4 years apart and ICP-MS analysis was employed to quantify 18 different metals. To improve consistency, standardised collection instructions were given to participants and toenail clippings were roughly cut to 25 mg, however significant variability in sample collection was observed (9–950 mg). Whilst sub-sample reliability was acceptable for all elements (Kendall’s W 0.72–0.90), intra-individual correlations between samples collected at different time points were poor (Spearman’s correlation 0.14–0.59) and comparisons with other biological samples were not performed making it difficult to ascertain clinical utility.

6.5 Progress for individual elements

6.5.1 Arsenic. Recent work has focussed on elucidating the intracellular metabolic pathways and protein binding of As species. Last year¹ we commented on the potential teratogenicity of As₂O₃ therapy in pregnant patients with acute promyelocytic leukaemia (APL). To further understand erythrocyte As accumulation in APL patients, binding mechanisms between different As species and Hb were investigated using SEC-ICP-MS.²¹³ Consistently with previous in vitro studies, MMA demonstrated higher binding affinity for Hb than DMA and iAs in erythrocytes from 9 APL patients treated with As₂O₃. Accurate mass TOF-MS analysis of MMA–Hb complexes confirmed Cys-104α and Cys-112β as the reactive binding sites for MMA. Men et al.¹¹⁶ established a novel method to track the metabolic fate of As^III in HepG2 cells, using an asymmetric serpentine microfluidic device coupled with CE-ICP-MS. Arsenobetaine and MMA were identified in HepG2 cell lysate, however only AB was detected in the excreted medium, indicating biotransformation of As^III first to MMA for detoxification and then to AB for non-toxic excretion.

A new methodology was reported for As exposure monitoring using HG-MIP-AES.¹²⁷ Generation of As^III-L-cysteine complexes from As^III, As^V, MMA and DMA in urine samples was achieved using a pre-reduction step with L-cysteine in an acid medium. Validation data showed acceptable inter-assay precision (6.9–15%), good recovery (95–97%) and correlation with HGAAS measurements (n = 30, R² = 0.9653), thus demonstrating suitability for biomonitoring purposes.

6.5.2 Calcium. The use of a naturally occurring stable Ca isotope ratio as a biomarker for bone Ca balance (BCaB) showed promise for monitoring paediatric renal patients on dialysis in a prospective proof-of-principle study.²¹⁴ Study participants (n = 137) from 5 European paediatric nephrology centres provided a 24 h urine sample, faecal sample and fasting blood sample and underwent other investigations including anthropometric studies, dual-energy X-ray absorptiometry (DXA) and peripheral quantitative CT (pQCT). Samples were analysed for Ca isotopes using MC-ICP-MS after Ca extraction and chemical purification. Positive linear correlations were observed between δ^44/42Ca_serum (‰) and ionized Ca (R² = 0.13–0.45), alkaline phosphatase (ALP) (R² = 0.0.14–0.38) and vitamin D in serum (R² = 0.09–0.17), whilst δ^44/42Ca_serum (‰) showed inverse relationships with serum concentrations of the bone resorption markers tartrate-resistant acid phosphatase 5b (TRAP5b) (R² = 0.31) and C-terminal telopeptide of type 1 collagen (CTX) (R² = 0.29). Bone densitometry measures, DXA and pQCT, also demonstrated weak correlations with δ^44/42Ca_serum (‰). Serum and urine δ^44/42Ca (‰) values were able to discriminate between different patient groups, however this was not possible with faecal samples. The group acknowledged that this was a proof-of-concept study and that further work was required to ascertain the clinical utility of this Ca isotope ratio as a biomarker for bone Ca content.

6.5.3 Copper. Various laboratory tests are available for the diagnosis and monitoring of Wilson's disease, many of which investigate Cu metabolism (serum total Cu, non-ceruloplasmin bound Cu, exchangeable Cu (CuEXC), albumin Cu, urinary Cu excretion, hepatic Cu content and test with radioactive or stable Cu isotopes).²¹⁵ Despite this, a recent systematic review highlights the need for further technological advancements, relating to Cu metabolism, for patients with diagnostic uncertainty, due to borderline results from currently available methods. Marino et al.²¹⁶ proposed the implementation of CuEXC together with the ratio of CuEXC to total Cu (REC) into first-line investigation panels, based on a small study of 54 Wilson's disease patients. This study found that a low REC (<15%) was useful for excluding the diagnosis of Wilson's disease in 6 patients who had been on long-term anti-Cu therapy, despite diagnostic doubts.

6.5.4 Lead. A candidate reference method for measurement of blood Pb was reported, using ICP-MS with Bi as an IS.²¹⁷ A bracketing calibration was applied, using standard solutions prepared from NIST SRM 3128, a primary calibration standard solution for Pb. The method was validated using Clinical and Laboratory Standards Institute (CLSI) standards, giving inter-assay RSDs of <1.88% and an LOQ of 1.1 μg L⁻¹. Analysis of 2021 EQA samples demonstrated good agreement between this new method and other routine methods, using GFAAS, tungsten coil AAS and ICP-MS, except for one low concentration sample (11.9 μg L⁻¹), highlighting a need to improve the accuracy of low concentration blood Pb measurements.

Use of tracer studies in clinical investigations has diminished over the years due to potential side effects caused by radioisotopes. Irrgeher et al.²¹⁸ presented a proof-of-concept study, using ²⁰⁴Pb as a stable isotope tracer in clinical studies. The participants (n = 42) received 250 mL drinking water spiked with 10 μg L⁻¹of ²⁰⁴Pb (below the Australian drinking water regulation) or placebo (no addition). Quantification of ²⁰⁴Pb in blood and urine was carried out by ICP-QMS and MC-ICP-MS with an isotope pattern deconvolution data processing method allowing tracer determination at concentrations <1 pg g^−1 204Pb. Although ultra-low LODs allow for clinical studies with tracer amounts below any level of concern, this type of methodology is unlikely to be considered for routine clinical studies where there is a drive towards more non-invasive analysis.

6.5.5 Mercury. A rapid, sensitive and robust ICP-MS method was developed for Hg biomonitoring in urine samples.²¹⁹ Stabilisation of Hg for 48 h in urine was achieved by 5-fold dilution with 5% HNO₃–0.625% HCl–0.25% thiourea (w/v). Non-spectral interferences were compensated for via¹⁹³Ir as the IS and analysis gave recoveries of 110–113.3% for ClinChek™ control samples and 95–104% for Hg in pooled urine samples. Boasting a reportable range from 65.6 ng L⁻¹ to 30 μg L⁻¹, the assay demonstrated excellent between batch precision (≤3.0%) and minimal carryover with diluent rinse times of just 30 s. The method demonstrated suitability for large scale exposure studies and the authors speculated it could be also used for accurate quantitation of other adsorptive elements.

6.5.6 Phosphorous. Whilst phosphoethanolamine (PEt) has been identified as a potential urine biomarker for hypophosphatasia, complexity of the phosphorous metabolome renders speciation analysis challenging. Lajin et al.²²⁰ described an HPLC-ICP-MS method to quantify phosphoethanolamine in human urine following derivatisation with fluorenylmethyloxycarbonyl chloride (fmoc-cl). Fast chromatographic separation (<4 min) of fmoc-PEt was achieved using a C18 column with an isocratic gradient of 9% 1,2 hexanediol (v/v) and 0.1% CH₃COOH (v/v). The method attained an LOD of 17 μg L⁻¹ and recoveries between 71% and 113% were reported for analysis of spiked urine samples.

6.5.7 Thorium. A simple, high sensitivity ICP-MS method was reported for Th quantitation in urine.²²¹ Urine samples were diluted with 100 ng L⁻¹ of ²³³U in 2% (v/v) HNO₃ and the memory effect of Th was reduced with a rinse solution of 0.025 mmol L⁻¹ oxalic acid–5% (v/v) HNO₃. With a reportable range between 0.77 ng L⁻¹ and 1000 ng L⁻¹, recoveries from spiked urine ranging from 93.5% and 98% and between run precision <6% RSD, the assay is suitable for biomonitoring in Th exposed populations (Table 2).

Table 2 Clinical and biological materials

Analyte	Matrix	Technique	Study aim, procedure and comments	Reference
As, Hg	Urine	HG-AFS	The study investigated the relationship between urine As concentrations and sperm DNA methylation status. A significant positive association was identified between urine As concentration and methylation defects of gene, MEG3. Urine Hg was not found to have any modulatory effect	305
Cr, Ni	Gingival crevicular fluid	ICP-MS	Metal concentrations were determined in gingival crevicular fluid from patients undergoing fixed orthodontic treatment who were using either fluoridated (n = 20) or non-fluoridated (n = 20) toothpaste. Metal ion release, which peaked at 30 days post-treatment and then declined by six months, was significantly increased in those using fluoridated toothpaste	306
Cr2O3, Mn3O4 and NiO NPs	Exhaled breath condensate (EBC), plasma, urine	spICP-MS	Particle sizes and concentrations of metal-oxide NPs were measured in stainless steel welders (n = 18) and controls (n = 15). In the workers, EBC concentrations of Cr2O3 NPs were significantly higher in post-shift compared with pre-shift samples (64 645 vs. 15 836 particles per mL). Larger Cr2O3 particles were detected in EBC with respect to in plasma (55–58 vs. 44 nm). No NPs were detected in urine	307
Cu	Urine	ICP-MS	A prospective cohort study in 693 renal transplant recipients explored the association between 24 h urine Cu excretion and long-term graft failure. Urine Cu excretion was positively associated with urinary excretion of both total protein and a biomarker of oxidative tubular damage (liver-type fatty-acid binding protein). High urine Cu excretions were also found to be associated with an increased risk of long-term graft failure	308
Cu (various (10))	Whole blood	ICP-MS	The association between blood Cu and chronic kidney disease (CKD) in an older Chinese population (n = 3285) was investigated. The odds ratio of CKD in the upper quartile of blood Cu concentration was 1.65 and there was a negative linear association between blood Cu and estimated glomerular filtration rate (eGFR). The findings also indicated that the positive association between blood Cu and CKD was strengthened by lower blood Mn concentrations	309
Hg (iHg and MeHg)	Hair	Double spike (species specific) ID-GC-ICP-MS	Species, MeHg and Hg^II, were simultaneously determined in 96 human hair samples, collected from Colombian gold mining communities. In most samples, MeHg was the predominant species, making up more than 80% of total Hg, although ten samples had a Hg^II content of greater than 30%. A higher Hg^II to total Hg ratio was measured in the hair of individuals involved in artisanal and small-scale gold mining activities compared with those who were not	310
I, Pb	Buffered protein solution, mouse neutrophil proteins	SEC-ICP-MS	Development of a sensitive method for separation and determination of metalloproteins in salt-rich matrices. Four I-labelled proteins were separated by SEC within 30 min and then detected with good reproducibility via the I-label by ICP-MS in high matrix introduction mode. The method was applied to Pb-binding proteins from neutrophils extracted from the bone marrow of mice exposed to Pb(C2H3O2)2	311
Pb isotope ratios	Blood	MC-ICP-MS	Lead isotope ratios were measured both in children (n = 6), who had been exposed to Pb, as well as in the potential environmental Pb sources. This allowed the likely source of Pb exposure to be determined for each individual	312
Se	Plasma, vitreous fluid	ICP-MS	Concentrations of Se were measured in patients with type 2 diabetes mellitus, both with (n = 40) and without (n = 20) diabetic retinopathy, as well as in non-diabetic controls (n = 20). In those with proliferative diabetic retinopathy, plasma Se concentrations were significantly higher than in all other groups and vitreous concentrations were higher than in controls. Intravitreal injection of anti-vascular endothelial growth factor led to a reduction in vitreous Se concentrations in the proliferative retinopathy group	313
Th	Urine	ICP-MS	A rapid, high-throughput method was developed for urine Th, which utilised 0.025 mol L^−1 oxalic acid and 5% HNO3 (v/v) to eliminate memory effects, and ^233U as the IS. The LOD achieved was 0.77 ng L^−1	221
UO2–peptide complexes	UO2–peptide contact solution	HILIC-ICP-MS/ESI-MS	Development of a single-step method to quantify and identify complexes formed between UO2 and multi-phosphorylated biomimetic peptides, which involved simultaneous coupling of ICP-MS and ESI-MS to HILIC separation. Concentrations of U were quantified using an external calibration, with Bi as the IS. The method enabled the effect of peptide structure on its affinity towards UO2 to be investigated	314
U isotope ratios	Urine	ICP-MS	A rapid method for determination of ^235U/^238U following dilution of 1 mL of urine in 2% HNO3. Between run precision ranged from 0.82 to 5.12%. Agreement with calculated values in spiked urine CRMs was within 6%. The LOD for ^235U was 0.42 ng L^−1, which was equivalent to ∼200 ng L^−1 of total U, based on a ^235U/^238U ratio of 0.002	315
Various (5)	Urine	Extraction chromatography-ICP-MS	A high-throughput method for determination of ultra-trace concentrations of ^241Am, ^239Pu, ^237Np, ^232Th and ^238U. The sample run time was 23 min and only 0.5 mL of urine was required. The LODs ranged from 0.015 ng L^−1 (Am) to 4.5 ng L^−1 (Th)	316
Various (5)	Urine	ICP-MS	The association between urine concentrations of As, Cd, Cu, Se and Zn and the risk of renal impairment and chronic kidney disease (as assessed by eGFR) was investigated in a Chinese population (n = 2210)	317
Various (39)	Whole blood, urine	ICP-MS	A study assessing the suitability of 13 elemental ISs for determination of 26 different elements. Reaction gas, O2, was used to eliminate polyatomic interferences. A large number of experimental conditions were investigated using the factorial design of experiments method. It was observed that the most suitable IS identified by this empirical approach was not always that with the closest mass proximity to the element being measured	318
Various (12)	Whole blood	ICP-MS	A prospective cohort study in patients with sepsis in which concentrations of various metals/metalloids were measured and compared with healthy controls. Concentrations of blood Ca, Cr and Cu were higher in septic patients compared with controls, while concentrations of As, Fe, Hg, Mg, Mn, Pb, and Zn were lower. Critically ill septic patients had lower blood concentrations of Fe and Mn versus non-critically ill patients and binary regression logistic analysis showed these to be independent risk factors for critical illness in sepsis	319
Various (23)	Serum	ICP-MS	A rapid method for measurement of 23 elements in serum following sample dilution in 0.5% HNO3, 0.02% Triton X and 2% MeOH. The ISs used were Bi, In, Sc, Tb and Y. Collision gas, He, was used to eliminate polyatomic interferences. Between-day precision was <12.19% and recoveries of a spiked standard were 89 to 110%. The LODs achieved ranged from 0.0004 to 0.2232 μg L^−1	320
Various (19)	Peritoneal fluid	ICP-QQQ-MS	The developed method used a sample flow rate of 100 μL min^−1, a dilution factor of 1 : 4 to reduce matrix effects and He collision gas to avoid interferences on Cu, Fe and Zn. The LODs ranged from 0.05 μg L^−1 for La to 6 μg L^−1 for Zn	321
Various (5)	Human teeth	LA-ICP-MS	In 353 children from a longitudinal birth cohort in Mexico City, As, Cd, Cr, Li and Pb were determined in deciduous teeth at sampling points corresponding to weekly increments, and then the data related to serum cystatin C concentrations (a marker of renal function), collected when the children were 8 to 12 years old. Higher Pb exposure in the late third trimester was associated with increased eGFR in childhood and increased exposure to a mixture of metals in the late second/early third trimester was observed to be associated with decreased eGFR	322
Various (7)	Hair, nails	LA-ICP-MS	A method for determination of Cu, Mn, Pb, Sr, U and Y in hair and nails. Washing steps and pre-ablation of the samples to remove external contamination was optimised. A dried droplet calibration approach using element enriched filter paper was validated using a pressed pellet of hair CRM mixed with Na2B4O7; recoveries for the various elements compared with concentrations obtained from solution-based ICP-MS ranged from 97.2 to 105.4%	323
Various (6)	Sperm cells (mouse)	scICP-MS	Elements, Cr, Fe, Mg, Mn, P and Zn, were determined in single sperm cells at different stages of the spermatogenesis cycle by scICP-MS, employing a dwell time of 0.1 ms. The content and distribution of these elements was found to change through the process of spermatogenesis	324

---

7. Applications: drugs and pharmaceuticals, traditional medicines and supplements

Activated charcoal can be used as a dietary supplement at around 100 mg per day but can also be a useful treatment for nonspecific antidote for poisoning, when is given in large doses (15 g for children and 50 g for adults). As such the levels of elemental impurities according to the ICH Q3D guidelines are particularly of interest. Zergui and team²²² analysed 60 samples of commercially available activated charcoal by ICP-MS for Al, As, Cd, Co, Cr, Cu, Mn, Ni, Pb and Zn following microwave digestion. The estimated daily intake was calculated for each element in adults of 70 kg and children of 15 kg. The most abundant elements in the samples were found to be Al (0.73 μg g⁻¹), Cd (0.60 μg g⁻¹) and Cr (0.34 μg g⁻¹), at these concentrations when taken as a supplement the levels are of no concern, but if it administered as an antidote then levels of Al may be of concern. The authors propose that special care should be taken when preparing activated charcoal from plant materials to choose those from areas with a low prevalence of metal contamination in the soil.

When considering elemental impurities, products of geological origin materials are of the highest concern. Quartz album, commonly used in traditional Chinese medicine, is a mined mineral product and as such the levels of toxic elements are important. In a study by Xu²²³ 30 elements were investigated using ICP-MS. Thirteen samples from different regions of China were digested in HF using MAD. Calcium, Fe, K, Mg, Mn and Ti were the most abundant elements and samples from Fengyang County, Anhui Province, were found to have the lowest levels of toxic elements such as As, Cd, Hg and Pb. Chemometric analysis was carried out on obtained data, and by use of PCA analysis orthogonal PLS, it was found that geographical origin could be established using fingerprint elements of Al, Ba, Be, Cu, Ga, K, Li, Mo, Na, Se, Si, Sn, and Sr, which can be used to ensure patient safety by use of the quartz albums from areas with the lowest risk of toxic element contamination.

Cyanobacteria such as spirulina are commonly used as a dietary supplement and have a relatively high Fe content, but the bioavailability of this iron is less clear. In a study conducted by Isani et al.²²⁴ Fe measurements and SEC were used to establish the presence of free and bound Fe, particularly in phyocyanin. Fractions from the SEC study were centrifuged and the pellets, as well as the untreated samples, underwent MAD. The supernatants from the SEC study were analysed directly. Protein analysis was carried out using SDS-PAGE. Total Fe content varied considerably from 353 to 1459 μg g⁻¹ across the sample set, and all samples showed a higher presence of Fe (>70%) in the insoluble phase of the separated samples, which is likely to be in the inorganic form as ferrihydrite. In the soluble phase, it was shown that the presence of Fe was related to high molecular mass proteins, such as phyocyanin, but also low molecular mass proteins, proposed to be mycosporine-like amino acids.

It has been several years since the implementation of ICH-Q3D for the control of elemental impurities in medicines was introduced. Since the introduction of the regulations, an abundance of papers have been published on the various approaches for the analysis of the elemental impurities. Aleluia et al.²¹ have reviewed the many papers on this topic. One of the key observations, perhaps unsurprisingly was that ICP-MS and ICP-OES are the analytical methods of choice for performing ICH-Q3D type studies and they also suggested that in the future, in-line analysis and greener methodologies should be considered.

Due to the high recombination rate and quenching effects with oxygen, Hg can be problematic when being determined by LIBS. Wang²²⁵ discussed using argon to reduce the oxygen effect as well as NP enhancement to improve performance of Hg analysis in medicinal herbs. In this paper Au NPs were used to enhance the detection of Hg in tangerine peel. A method was optimised using a range of NP sizes (10, 20, 40 and 60 nm). In the selected method, it was found that 40 nm particles gave the best performance, which is proposed to be due to optimal inter-particle distance and large surface area. Under optimised conditions the method was shown to give a 6.19 times improvement in spectral intensity and a reduction in RSD from 20.63% to 4.99%.

8. Applications: foods and beverages

In this section, papers highlighting progresses in the area of food and beverages analysis are discussed. In addition, the technical details and most relevant findings of a number of other papers in this area are summarised in Table 3, complementing those reported in Table 1. Over the last couple of years there have been an increasing number of papers published on the subject of chemometrics to determine origin, authenticity and adulteration of food products. We have, therefore, introduced into this Update a new table (Table 4) summarising the matrix, analytes, statistical methods and purpose of the studies to simplify reference to these papers.

Table 3 Food and beverages

Analytes	Matrix	Technique	Study aim, procedure and comments	Ref.
Al, Cd, Pb	Soups	ICP-OES	A study to investigate the levels of 3 toxic elements in a variety of instant soups. 130 soups were analysed by ICP-OES following dry ashing and dissolution in dilute HNO3. Concentrations were used to assess risk in a variety of age groups, by type of soup (vegetable, meat or poultry). It was noted that for children ages 3–10, the estimated daily intake of Pb and Al could exceed 15% of tolerable weekly intake from vegetable soups	325
Sb	Water, tea and honey	ICP-MS	GO and MIL for dispersive SPE was developed for iSb analysis. Sb^v was measured in the aqueous phase following SPE and total Sb was measured by direct analysis of the samples. Sb^III was calculated by difference. The method was shown to have an extraction efficiency of 99.7% and LOD of 5 ng L^−1 for Sb^III and 3 ng L^−1 for Sb^V	75
As, Co, Cr, Cu, Fe, Se, Zn, Ca and Pb	Edible beetles	ICP-MS	Nutritional elements as well as As and Pb were studied in 2 species of edible beetles. The nutritional elements were studied as an alternative food source as these insects are rich in essential minerals as well as useful fats and proteins. The levels of Pb and As were found to be below toxic levels for humans	326
iAs, As	Mushrooms	ICP-MS	Speciation of As in A. blazei and T. matsutake mushrooms. As^III, As^V, AB, AC, DMA, MMA, TMA^+ and TMAO were separated using anion and cation exchange chromatography. Good recoveries were obtained for spiked samples (89–94%). All samples tested were found to have a total As above the limit of 0.5 mg kg^−1 but only 3 samples exceeded this limit for iAs	171
Tl(I), Tl(III)	Beverages	GFAAS	A two-step extraction of Tl species using of 1-(2-pyridylazo)-2-naphthol and dicyclohexano-18-crown-6 from filtered soft drinks and water pH adjusted to 6.0. Good recoveries obtained (90–110%). LODs 1.9 ng L^−1 and 2.5 ng L^−1 for Tl^III and Tl^I respectively, showing improved LODs for Tl detection by AAS methods	90
Cd	Water and fish	SQT-FAAS	Use of a SQT showed a 1467 times increase in sensitivity for Cd compared to FAAS. Optimised methods were used to assess Cd and gave good recoveries for reference materials. The reported LOQ was 0.075 ng mL^−1 lower than ICP-MS and ICP-OES for some applications	121
Cd and Pb	Herbs and spices	GFAAS	Examination of common culinary herbs and spices for Cd and Pb. Due to low amounts used most do not pose a risk, but levels of Pb in watercress, jiaogulan, celery, basil and dill may pose a risk for sensitive groups	327
As, Cd, Hg and Pb	Canned tuna	ICP-MS	222 samples from 36 countries (split into developed and non-developed) were analysed following MAD digestion. iAs assumed to be 3% of total As levels. All samples were within relevant safety limits, except 3 samples with elevated Pb levels. The authors propose that daily consumption of 120 g of tuna is above the target hazard quotient for Hg exposure, and 1 meal of 120 g of tuna per week can increase risk of cancer from exposure to iAs	328
Al, As, B, Cd, Co, Cr, Cu, Mn, Mo, Pb, Sb, Sr, Ti, Tl, V, Zn	Raspberry fruit and leaves	ICP-MS, ICP-OES	Metal content in fruit and leaves of organic, conventional and wild cultivated raspberries were compared. The study aimed to indicate if cultivation type reduced exposure to heavy metals, but was inconclusive with elevated Cd, Mn, V and Zn in organic fruits and Cd, Cu, Mn, Zn in wild fruits compared to conventionally grown fruits	329
Al, As, Cd, Co, Cr, Cu, Mn, Ni, Pb and Zn	Coffee	ICP-MS	Instant, ground and whole bean coffees were analysed for toxic elements with a view to perform a health impact assessment. Packaged coffees were found to have higher levels of these elements, suggesting packaging processes have an impact	330
Various	Various	Various	A useful review of current preparation and analysis techniques used in food analysis. A overview of standard methodologies and relative strengths and weaknesses of each	7
As, Ba, Co, Cr, Cd, Cu, Hg, Li, Mn, Mo, Ni, Pb, Se, Sr, Ti and V	Egg	ICP-MS	Method optimisation for analysis of egg for minor nutritional components by ICP-MS	56
Cd, Cr, Mn and Ni	Oils	EDXRF	Reverse phase liquid/liquid extraction with EDXRF was shown to be suitable for the analysis of 4 elements. Allowing for a quicker analysis with lower chemical usage than ICP-OES with a suitable LOD of <3 μg kg^−1 for all elements	331
Ca, Cu, Fe, Mg, Mn and Zn	Meat and plant based burgers	ICP-OES	Simulated in vivo digestion and MAD followed by ICP-OES was used to determine mineral contents and bioavailability in meat/meat substitute burgers. Fe and Zn were found to be lower in plant based burgers, soy and mycoprotein burgers did have comparable levels of Fe, but with lower bioavailability	332
Al, Cd, Cr, Cu, Fe, Li, Ni, Pb and Zn	Tuna	ICP-OES	Samples of tuna from the Canary Islands were ashed and analysed by ICP-OES, method verified with CRMs. In Thunnus thynnus Cd was shown to be of most concern, and if the recommended 750 g of fish was eaten by an adult, then exposure to Cd would exceed EU guidelines	333
Spectral	Rice	XRF, ICP-MS	Comparative study of seeds to identify location of elements within rice structure, as well as linking this to genetic and molecular mechanisms that control distribution of elements in rice grains	334
Ca, Cr, Cu, Fe, K, Mg, Mn, Mo, Na, P, S and Zn	Spelt	ICP-MS	Comparison of 12 wholemeal and normal spelt flours. MAD followed by ICP-MS showed that differences in element abundance and content were seen between the sample types. Spelt flour: S > K > P > Mg > Ca > Na > Fe > Zn > Mn > Mo > Cu > Cr, and in spelt wholemeal: K > P > S > Mg > Ca > Fe > Mn > Na > Zn > Cu > Mo > Cr	335
Zn	Rice	XRF, ICP-OES	Comparison of Zn determination by XRF and ICP-OES in rice on 200 samples. Good correlation was seen between the techniques, with an R^2 = 0.83	162
Ag, Al, Ba, Cd, Cr, Cu, Fe, K, Li, Mg, Mn, Ni, Pb	Oils	ICP-OES	A comparison of dilute and shot and fast dispersive liquid–liquid aerosol phase extraction (DLLAPE) for analysis of oils by ICP-OES. The DLLAPE approach can potentially offer a rapid preparation technique compared to other extraction methodologies for direct analysis of metals in oils and fats by ICP-OES	63
Al, As, B, Ca, Cd, Co, Cr, Cu, Fe, Hg, K, Li, Mg, Mn, Na, Ni, Pb, P, Sb, Se, Sn, Sr and Zn	Plants	ICP-MS	Elemental composition of 4 culinary herbs valued for their medicinal properties were measured. Allium orientale Boiss., Eremurus spectabilis M. Bieb., Anchusa officinalis L. and Arum elongatum Steven. Samples underwent MAD and analysis by ICP-MS. The plants were shown to be rich in minerals, and to have low levels of toxic elements, except for high levels of Al in A. orientale and A. elongatum	336
tAs, iAs, As species	Animal and fish tissue	HPLC-ICP-MS	In an interlaboratory study, a consensus method was used to measure total As and As species in a range of animal and fish CRMs. Labs followed a consensus method, as well as performing in-house preparation if wanted. Collated data showed that the consensus method gave good reproducibility, and show improved reproducibility when compared to the results obtained from in-house methods	40
As, Ca, Co, Cr, Cu, Fe, Pb, Se, Zn	Coleoptera	ICP-MS	Insects are a valuable source of protein, fat and minerals, but also can accumulate toxic metals. In this study the authors recognise the nutritional value of water insects and state that the standard cultivation of these insect types poses a low risk of excess exposure to toxic elements	326
As, Cd, Hg, Pb	Baby foods	ICP-MS	Samples of ready to eat baby foods were analysed for heavy metals following MAD. Samples with rice were found to have higher As content, leafy vegetable food were found to have more Cd and root vegetable based foods were shown to have the highest levels of Pb	337
REE, Sc and Y	Mushrooms	ICP-MS	2235 samples from 22 locations in Poland, across 10 years were analysed for several elements. The measured levels were found to be sufficiently low as to pose no risk to consumers	338
Cd	Cocoa and chocolate	ICP-MS	349 cocoa containing products were analysed for Cd to understand the risk to consumers. They were estimated to contribute to 7–9% of dietary Cd exposure, but it was also noted that the bioavailability of Cd was 2–5-fold lower than that of wheat flours	339
Ag NP	Beef	spICP-MS	Two digestion techniques were assessed for the measurement of Ag NPs in ground beef. Enzymic digests and alkaline digestion in TMAH were compared. TMAH digests gave 95% recovery, enzyme digests had a lower recovery of 60%, although both methods had no effect on NP size, so deemed suitable for this analysis	340

---

Table 4 Applications related to origin and authenticity of food and beverages

Analytes	Matrix	Technique	Statistical analysis	Purpose	Study aim, procedure and comments	Ref.
Al, Ba, Ca, Cu, Fe, K, Li, Mg, Mn, Na, P, Sb, Se and Zn	Olive Oil	ICP-OES	PCA	Origin	A variety of different olive oils, from different regions as well as non-olive oils underwent ICP-OES analysis and PCA. It was shown that Italian olive oils were typically higher in Fe, and oils from different regions from both Spain and Portugal were shown to distinct profile to allow for geographical discrimination	341
Ag, Al, As, Ba, Bi, Br, Ca, Cd, Ce, Cl, Co, Cr, Cu, Fe, Hg, K, La, Mg, Mn, Mo, Nb, Nd, Ni, P, Pb, Pr, Rb, S, Sb, Se, Si, Sn, Sr, Th, Ti, U, V, W, Y, Zn and Zr	Prawns	XRF	PCA	Origin	Using a handheld scanner, prawns from various sites around the Australian coast were analysed for 41 elements. It was shown that raw samples were better for provenance, and could discriminate site with 87.5% accuracy, 98% accuracy for jurisdiction and 100% for production method	342
			Compositional data analysis (CoDA)
			Redundancy analysis (RDA)
Al, B, Cr, K, Mn, Na, P and Pb	Wine	MIP-OES	SVM, DT and Logistic Regression (LR)	Origin	Discrimination between Brazilian and Argentinian wine was accurately achieved using the SVM, DT and LR. B, Cr and K were found to be the key indicators, even with these elements removed, discrimination was still found to be accurate to 80–90%	343
Al, B, Ba, Ca, Co, Cr, Cu, Fe, K, Mg, Mn, Na, Ni, Pb, Rb, Si, Sr, Ti and Zn	Wine	ICP-MS	PCA, orthogonal partial least squares discriminant analysis (OPLS-DA and O2PLS-DA), ANOVA and multifactor analysis (MFA)	Origin	69 wine samples from 4 Chinese regions, 3 vintages, and 3 grape maturity levels. Region had the greatest influence on elemental distribution, followed by vintage and grape maturity. K was found to negatively impact determination of origin, so was removed for these assessments	344
As, Cd, Co, Cr, Cu, Fe, Mg, Mn, Na, Ni, Pb and Zn	Sesame seeds	ICP-OES	ANOVA, PCA, LDA	Origin	93 samples from 3 main sesame producing regions in Ethiopia were analysed. LDA was found to give 100% accuracy in determination of origin	345
Various (60)	Peppers	ICP-MS (51 elements), ICP-OES (9 elements)	OPLS-DA, Variable Importance in Projection (VIP), Canonical Discriminant Analysis (CDA)	Origin/authenticity	REEs are strong indicator elements to identify non-Korean peppers being sold in the market as being of domestic origin. CDA could correctly distinguish between 100% of analysed samples. Twenty-six elements (Al, As, Cs, Cu, Dy, Er, Eu, Fe, Gd, Ge, Ho, Ir, La, Nb, Nd, Pd, Pr, Rh, Sm, Sr, Th, Ti, Tm, U, Yb and Zn) were shown to be key in determining geographical origin from OPLS-DA modelling and obtained 98.7% accuracy	346
Various (75) ^87Sr : ^86Sr	Durum wheat	SF-ICP-MS	PCA, SVM	Origin/adulteration	Origin of durum wheat, as either from Italy, or the rest of the world, was assessed using Sr isotope ratio of Sr and elemental composition. The isotope ratio did not provide clear discrimination, but was used for a tiered approach where the ^87Sr : ^86Sr ratio was input into a second step of SVM classification modelling based on the Al, Mn, Mo, P, S, Ti, Y and Zn	347
Various (24) REEs (14)	Wine	HR-ICP-MS	PCA, PLS-DA	PDO protection	PDO of sparkling Italian wines was investigated using data for 24 elements (Cd, Co, Cr, Cu, Fe, Ga, Hf, Hg, Mn, Mo, Nb, Ni, Pb, Re, Sb, Sn, Ta, Th, Tl, U, W, V, Zn, Zr), 14 REEs (Ce, Dy, Er, Eu, Gd, Ho, La, Lu, Nd, Pr, Sm, Tb, Tm, Yb) and isotope ratios for Pb and Sr (^208Pb : ^206Pb, ^207Pb : ^206Pb, ^206Pb : ^204Pb, ^208Pb : ^207Pb, ^87Sr : ^86Sr). Also compared to red and white wines from the same region. 95% classification, Sr ration provided a powerful tool for micro-scale geographical origins	348
Isotopes ratios for Pb and Sr		MC-ICP-MS
42 elements	Honey	ICP-MS, ICP-OES	PCA	Geographical and botanical origin	This study found that the fingerprint of soil, described by the lanthanide distribution, is transmitted unaltered from soil to flowers, with a slight fractionation on the heavier lanthanides (from Dy to Lu) occurring in the passage from flowers to honey	330
As, Cd, Co, Cr, Cu, Fe, Hg, I, Mn, Mo, Ni, Pb, Se and Zn	Cheese	ICP-MS	PCA, HCA	Manufacture	A variety of Gallican cheeses were studied, produced industrially, from organic sources and artisanal. The study showed that distinction between artisanal and organic cheese was not clear, but these were clearly different to the industrial cheese. Key discriminating elements were Cu, I, Mn and Se, the authors proposed this arose from routine supplementation of dairy cattle in industrial farms. No differences were seen between smoked and unsmoked cheese	349
B, Ba, Ca, Cu, Fe, K, Mg, Mn, Na, Ni, P, Sr and Zn	Peppers	ICP-OES	PCA, SIMCA	PDO Protection	Fifty Senise and Altino bell peppers and 20 commercial samples were analysed. The modelling using SIMCA successfully differentiated between the Senise and Altino bell peppers despite close geographical origins, with 100% sensitivity and selectivity. The model was also able to reject samples not from the designate PDO regions	350
Spectral	Tomato and peppers	EDXRF	PLS-DA	PDO protection	EDXRF was found to be a useful tool in discrimination of geographical region and cultivation technique (organic or conventional). The two sample types needed to be considered separately as the models varied. Ca, Fe, K and Zn were key components in evaluation of geography and cultivation	351
As, Br, Ca, Cl, Cr, Cu, Fe, Hg, K, Mn, Na, Ni, P, Pb, Pr, Rb, S, Sb, Se, Sm, Sr, Ti, V, Y and Zn	Sea bream	XRF	RF	Production	This study utilised XRF as a simple analytical technique to assess elemental composition in sea bream muscle from earth pond, caged and wild fish from 2 locations. 90% discrimination was obtained with the key elements being As, Br, Mn, Pb, Rb, S and Se	352
46 elements	Traditional medicine	ICP-MS	PCA, HCA, PLS-DA	Origin	Three traditional medicines (Atractylodis macrocephalae Rhizoma, Lilii Bulbus, and adlay) from different regions in China were analysed. Classification by PLS-DA was found to be near 100% in identification of species and origin with K, Mg, Mn, Na, Rb and Zn being the key components in identification	55
ICP-MS: Al, B, Ba, Ca, Ce, Co, Cs, Cu, Fe, K, Mg, Mn, Mo, Na, P, Rb, Sr, Zn; INAA: Br, Cl, Cr, Fe, Sc, Se, Zn, Hg, Na	Honey	ICP-MS, INAA	PCA, CA	Origin	The results have demonstrated that elemental analysis may be useful in discriminating honey sourced from Montana against honey sourced from North and South Dakota	353
Ca, Cu, Fe, K, Mn, Na, P and Zn	Rice flour	Time resolved-LIBS	SVM, LDA, naïve Bayes algorithm (NB)	Adulteration	Study demonstrated significant improvements in predictions compared to standard LIBS but was not sufficient to fully identify adulteration at lower levels	354
Ag, Al, As, B, Ba, Be, Ca, Cd, Ce, Co, Cr, Cs, Cu, Dy, Er, Eu, Fe, Ga, Gd, Ho, K, La, Lu, Mg, Mn, Na, Nd, Ni, P, Pb, Pr, Rb, S, Se, Sm, Sr, Th, Tl, Tm, U, V, Yb and Zn	Olive oil	ICP-QQQ-MS	PCA	Origin	Limited outcomes from the statistical analysis, utilising cleanroom and ICP-QQQ-MS to better identify extra-virgin olive oil from different regions	355
Spectral	Rice	LIBS, hyperspectral imaging (HSI)	SVM	Origin	Use of LIBS or HSI with SVM has shown to give an accuracy of 93% and 88% respectively. This study combined the methods giving an improved accuracy of 99.9%	356
Spectral		LIBS	ENET (elastic net-regularized multinomial classifier)	Authenticity/origin/maturity	Handheld LIBS to investigate cheese, vanilla, balsamic vinegar, coffee and spices. Method was shown to have a prediction accuracy of 81–98% in detection of food authenticity, origin or maturity (dependent of matrix). Liquid samples were supported on nitrocellulose paper for analysis	357
26 elements, As, Cd, Cs, Cu, Fe, iAs, Mo, Ni, Rb, Se, Zn and Mn identified as most informative	Rice	ICP-MS and LC-ICP-MS	SVM	Origin	640 samples analysed from 5 regions in Brazil. A reduced number of elements were shown to predict origin and type of rice with accuracy of between 91–99%	358
Spectral	Meats	LIBS-Raman	RF, BPNN	Species	LIBS-Raman used to identify pork, beef and mutton. 1200 samples measured, accuracy of 99.4%	138
Al, As, Ba, Bi, Ca, Cd, Co, Cr, Cu, Fe, Ga, K, In, Li, Mg, Mn, Na, Ni, P, Pb, Se, Sr, Zn	Pollen	ICP-OES	CDA	Origin/season	71 samples analysed, K and P were found to be key signifiers for origin. 90% accuracy in assigning apiary, and 100% accuracy for seasonal origin	359
Al, B, Ba, Ca, Cd, Co, Cr, Cs, Cu, Fe, Ga, Hg, K, Mg, Mn, Na, Ni, Pb, Rb, Sr, Tl and Zn	Honey	ICP-MS	PCA, LDA, DT	Origin	352 honeys from 34 countries were analysed. Discrimination of honey from New Zealand was found to be 92.4% accurate using DT modelling with the primary discriminating elements being Ba, Cs and Rb	360
60 elements, 24 for modelling	Kimchi	ICP-MS and ICP-OES	PLS-DA, OPLS-DAPCA, CDA, LDA	Origin	Discrimination between Chinese and Korean kimchi was established with OPLS-DA providing the highest accuracy and prediction, with 100% prediction being achieved	361
Al, As, B, Ba, Ca, Ce, Co, Cs, Cu, Fe, Ge, K, La, Li, Mg, Mn, Mo, Na, Nd, Ni, P, Rb, Se, Sr, Ti, Tl, Zn and Zr	Jujube	ICP-MS	PCA, CA, DA	Origin	13 varieties of red jujube from 4 Chinese provinces. PCA was unable to distinguish but CA and DA were effective in discriminating region of production	362
Al, B, Ba, Ca, Co, Cu, Fe, K, Mg, Mn, Mo, Na, Ni, P, Si, Sn, Sr, Ti and Zn, vitamin and antioxidants	Prickly pear	ICP-MS (UPLC-MS/MS)	LDA, s, multivariate ANOVA (MANOVA)	Origin/variety	74 samples collected from 4 Mediterranean countries, as well as 3 varieties from within Cyprus. Vitamin, antioxidant and mineral analysis were used to identify fruits by nation, and by variety within the Cypriot group. The applied methods gave an accuracy of 86.5% in identification of origin and 90.7% accuracy for variety	363
Ca, K, Mg, Se, and Zn	Nuts	AAS, AES, GFAAS	PCA, CA, DA	Variety	120 samples of 10 different nut types were analysed. PCA was able to distinguish type with a 99% accuracy with 4 component analysis. The study also indicated the value of nut in a balanced diet, and how different nuts can provide a different balance of essential minerals	364
41 elements	Coffee	ICP-MS, ICP-OES, ICP-QQQ-MS	LDA	Origin/variety	76 samples of coffee from different continents, countries and species were analysed in this study. Accuracy was found to be 93.3% for continent, 97.8% for country and 100% for species. Key elements for species discrimination were Ba, Ca, Cd, Rb and Sr, for origin Rb, S, Sr and Tm were key indicators	365
^87Sr/^86Sr	Tea	MC-ICP-MS, ICP-MS	PLS-DA	Origin/variety	26 tea samples were analysed for Sr ratio, elemental composition and VOC by SIFT-MS. Sr ratio was able to identify 6 of the 11 regions alone, when combined with elemental profile, 10 of 11 regions were specified, with 2 regions in Sri Lanka showing similar inorganic profiles. Teas of similar processing regime were also clustered together	366
Ag, Al, As, B, Ba, Be, Ca, Cd, Co, Cr, Cu, Fe, Hg, K, Mg, Mn, Mo, Na, Ni, Pb, Sb, Se, Si, Ti, Tl, V, Zn + polyphenols	Herbal infusions	ICP-MS	PCA, DA	Authenticity	Elemental and phenolic components were used to determine authenticity of herbal infusions. High levels of Mn and Rb were identified as key markers for conventional tea, polyphenols marked out other infusions such as Rooibos	367
B, Ba, Ca, Ce, Co, Fe, K, La, Li, Mg, Mn, Mo, Pb, Rb, Sr and Zn and δ^13C	Egg	ICP-MS, IRMS	PLA-DA	Farming method	Egg samples from barn and backyard hens were separated into yolk and white and analysed to establish whether rearing method could be determined statistically. Discrimination was found to be 100% in egg yolk and 95% in egg white. The study also concluded that there was a low risk of exposure to toxic elements according to EDI levels	368

---

8.1 Progress for individual elements

8.1.1 Arsenic. Arsenic speciation is, as ever, an area of much interest and research, particularly speciation of As in foods. Utilising HPLC-ICP-MS D'Amore and team²²⁶optimised a heated water bath extraction for use on a variety of foods. The acid and temperature conditions of the preparation were sufficient to oxidise iAs to As^V whilst maintaining AB, DMA and MMA, the extracts were then analysed using HPLC-ICP-MS using 50 mmol L⁻¹ NH₄HCO₃ as the mobile phase. The chromatography allowed for separation within 7 min and was validated giving linearity of R² ≥ 0.99, LODs of 0.025 μg kg⁻¹ to 0.106 μg kg⁻¹ and LOQs of 0.075 μg kg⁻¹ to 0.321 μg kg⁻¹. The precision RSD was ≤7.3% and recovery ranged from 81% to 118%. Good z scores were obtained in interlaboratory studies. This data suggests that the method was fit-for-purpose and provides an effective and quick assessment for As species across a variety of food matrices.

8.1.2 Cadmium. Cocoa is an economically important crop, but as a natural product can introduce a risk of exposing consumers to toxic metals, such as Cd. To understand the mechanisms of Cd uptake in cocoa plants Blommaert et al.²²⁷ utilised MC-ICP-MS to look at ¹¹⁴Cd : ¹¹⁰Cd ratio. Cadmium species in the cocoa plant materials were measured by XAS and LA-ICP-MS was used to map Cd locations in branches. The distribution of Cd species was found to be mainly in the stem and branch (77%), roots (14%) and leaves (9%), with only <1% in the fruit, most of which located in the husk, particularly as heavier isotopes. The speciation work indicated that most of Cd in the nibs was stored as Cd–phytate, which the authors suggested may have a low bioavailability and warranted follow-up studies.

8.1.3 Mercury. In a comprehensive study, samples of wild fish and seafood from the western Mediterranean Sea were assessed for their total Hg levels. Samples were analysed using a direct vapour AAS analyser. A total of 1345 specimens were analysed in this study, and it was found that only 13 species (sardines, anchovies, pickerel, salema, blue whiting, black-spot seabream, gilthead seabream, painted comber, brown meagre, surmullet, pearly razorfish, common dolphinfish and squid) had total Hg levels below the EU maximum allowed limit of 0.5 μg g⁻¹. The authors²²⁸ suggested that if all Hg was present as MeHg then a typical diet would expose consumers to 49% to 70% of the provisional tolerable weekly intake of MeHg. However, it is also mentioned that consumption of fish has many health benefits and consumption of Mediterranean fish is of lower risk that fish from other regions.

8.1.4 Selenium. Infant formula and milks from Brazil and Belgium were analysed for total Se and Se species, by de Paiva et al. ²²⁹ For total Se, samples underwent MAD based on the AOAC method 2015.06 and the digests were analysed using ICP-QQQ-MS. The use of O₂ allowed a mass shift from m/z 78 to m/z 94 (SeO⁺), thus improving sensitivity, under these conditions, an LOQ of 10 μg kg⁻¹ was achieved. Speciation was performed using LC-ICP-MS, measuring Se^IV, Se^VI and SeMet. The samples analysed were shown to give huge variability against the label claim, with 55% to 317%, of the stated amount being measured. The authors also noted that plant based infant formulae typically had lower Se levels and were typically fortified with Se^IV which has a lower bioavailability than other forms of Se.

Direct and indirect analysis of Se species in wheat was investigated by Subirana et al. ²³⁰ The study looked at accumulation of Se species in the roots, shoots and seeds of wheat which was biofortified with Se^IV, Se^VI or a mixture of both. Enzymatic digestion followed by HPLC-ICP-MS identified 5 key species, namely Se^IV, Se^VI, SeCyst, SeMeCys, and SeMet. Direct analysis was carried out using XANES and the key components were identified as Se⁰, Se^IV, Se^VI, C–Se–Se–C and C–Se–C. The authors commented that the combined use of these two techniques provided useful information on the transport and conversion of Se in the wheat plant and will aid in the targeting of how to best fortify wheat with Se to obtain the highest concentrations of bioaccessible forms such as SeCyst and SeMeCys without inducing metal toxicity in the plant.

Hydride generation is a commonly used analytical tool for determination of Se. Due to the chemistry of Se, chemical speciation can be performed by measuring both total Se and Se^IV on a pre-treated sample, thus allowing Se^VI to be calculated by difference. Multi-syringe FI-AFS was used to speciate Se in teas, with no sample pre-treatment.²³¹ Both the sample and KI, as the reducing agent, were pumped into the system, under UV light to reduce all Se to Se^IV. The Se^IV fraction present in the sample was then measured directly with the KI valve closed and the UV treatment turned off. The LOQ was found to be 0.18 μg L⁻¹ for both species. The analysis of spiked samples gave recoveries ranging from 91% to 111%. The concentrations of Se^IV and total Se in the teas varied from 0.2 μg L⁻¹ to 0.6 μg L⁻¹ and from 0.4 μg L⁻¹ to 2.0 μg L⁻¹, respectively.

8.2 Single and multi-element applications in food and beverages

8.2.1 Dietary intake studies. The nutritional and toxic metal content of baby food from jars was studied by González-Suárez et al. ²³² Meat, vegetable, fruit and mixed baby foods were obtained from the local market. The samples underwent drying, ashing and analysis by ICP-OES for Al, B, Ba, Ca, Cd, Co, Cu, Fe, K, Li, Mg, Mn, Mo, Na, Ni, Pb, Sr, V and Zn. The method was verified using various CRMs which had recoveries between 94% to 103%. The study found high levels of Li in all sample types at around 300 ± 100 mg kg⁻¹. Exposure to Al, Cd, Ni and Pb were found to be higher than recommended exposure limits, in some cases by a significant amount, for example and infant consuming 250 g of vegetable based food per day could consume 870% of the weekly permissible intake of Ni. The authors suggested that care should be taken when using these convenience foods to limit risks from metal exposure.

In the 6th Chinese total dietary study samples from 12 food categories from 24 regions were analysed for macro and micro nutrients by ICP-MS. The method LOD was determined to be Na 125 μg kg⁻¹, K 80 μg kg⁻¹, Ca 150 μg kg⁻¹, Mg 15 μg kg⁻¹, P 200 μg kg⁻¹, Mn 1 μg kg⁻¹, Fe 8 μg kg⁻¹, Zn 4 μg kg⁻¹, Cu 2 μg kg⁻¹, Se 0.25 μg kg⁻¹, Mo 0.5 μg kg⁻¹, Cr 1 μg kg⁻¹. The study found that plant foods were the main source of macro and micro nutrients across the entire population, accounting for between 68% and 96% of total intake. Most minerals were consumed in appropriate quantities, but too much Na was common and so was insufficient intake of Ca, Mg, Se and Zn. The findings show that dietary guidance is required to encourage healthier diet patterns.

8.2.2 Human milk and infant formula. A growing number of the population taking to vegetarian and vegan diets, and within this group are women who are lactating, and it is important to understand the impact of these diets on the nutritional composition of breast milk. Milk from 63 participants following either vegan, vegetarian or omnivore diets were analysed for nutritional and toxic elements by ICP-OES (Ca, K, Mg, Na and P) and ICP-MS (As, Cd, Cr, Cu, Fe, I, Mn, Mo, Pb, Se and Zn). In general few differences were seen between the 3 types of diet, with the notable exception of Se, which was shown to be consistently higher in vegan and vegetarian diets. The authors suggest that with a larger dataset and use of chemometric modelling it may be possible to establish links between diet and small variations in metal content of breast milks.

8.2.3 Dairy products. In a study by Rihawy et al.,²³³ ion beam methods were utilised to study several elements in milks from 4 species of Syrian mammals, namely, cow, goat, camel and sheep. Samples were analysed using both PIXE and PIGE for Br, Ca, Cl, K, Mg, Na, P, Rb, S, Sr, and Zn, which was combined with Rutherford Backscattering (RBS) to determine matrix elements, with the samples undergoing freeze-drying and milling prior to analysis. Concentration values of major elements (Ca, K, Mg, Na, P and S) were in the range 1000–10 000 μg g⁻¹ and trace (Br, Cl, Rb, Sr and Zn) elements in the range 5–100 μg g⁻¹. The authors noted that PIXE was more efficient in detecting the heavier elements, whilst PIGE was better with the lighter elements. Though this was a limited study, with only two samples for each species, the results could offer a promising approach to simple nutritional screening of milk samples.

Donkey milk was the subject of a study by Fantuz et al. ²³⁴ The team investigated the distribution of 14 elements (B, Cd, Co, Cr, Cu, Li, Mn, Mo, Pb, Rb, Se, Sr and Ti) in the fat, casein, whey and aqueous phases of the milk. Samples were separated by ultra-centrifugation and ultra-filtration to obtain the 4 phases that then underwent MAD followed by ICP-MS. The profiles of individual elements in the different phases were varied, but commonly, very low analyte concentrations were seen in the fat phase, with only Mo and Cr having more than 10% of the total content in the fat phase with 13.6% and 10.4% respectively. Boron and Li were only found in the aqueous phase, and the other elements had varying distributions in the casein, whey and aqueous phases, with the casein phase being favoured by the majority of elements, particularly Zn (95% of the total Zn content).

In a comparative study of cow and buffalo milk,²³⁵ samples were gathered from animals kept on the same farm with the same diet to reduce the variability of the study. They underwent MAD followed by ICP-MS for B, Ca Co, Cu, Fe, K, Li, Mg, Mn, Mo, Na, P, Rb, Se, Sr, Ti and Zn determination. As similar diets and geography was applied for both species, difference in mineral content are more likely to be due to physiology of the animals. It was shown that concentrations of B, Ca, Cu, Fe, Li, Mg, P, Sr, Ti and Zn were higher in buffalo than cow milk whereas K, Li, Mo, Na, and Rb were lower, though interestingly the Ca/P ratio for both species was the same.

8.2.4 Cereals. As one of the worlds most consumed food products, rice is always a popular topic for study, particularly when investigating origin, authenticity and organic status. Adulteration and mislabelling is commonplace due to demand for particular varieties and origins. Wadood²³⁶ and team have collated the latest studies into rice authenticity utilising a variety of analytical and statistical methodologies: ICP-MS, ICP-OES and IRMS are most commonly employed for analysis and PCA, HCA (Hierarchical Cluster Analysis), ANN, KNN (K-nearest neighbour), LDA, PLS-DA, SIMCA, and SVM. The collated information provides a useful summary to the current progress in this field.

The ability to carry out analysis in the field is particularly useful for food production and safety, but often there is a compromise with detectability with portable equipment. Utilising catalytic pyrolysis followed by miniature AAS, Liu et al.¹²² were able to demonstrate LODs of 1.1 μg kg⁻¹ and 0.3 μg kg⁻¹ for Hg and Cd using a 20 mg grain sample within 5 min. With safety limits of 100 μg kg⁻¹ for Cd and 20 μg kg⁻¹ for Hg, this technology is sufficient to satisfy these demands. Solid samples are heated directly using a metal–ceramic heater based electrothermal vaporizer, an on-line catalytic pyrolysis furnace charged with alumina was used to remove organic interferences, a composite Pt/Ni trap utilised to enrich the gas phase sample before flowing into a miniature AA spectrometer utilising folded path detection. The method was validated using rice, wheat and corn reference materials, both neat and spiked. Good recoveries were seen across all matrices with recoveries for spiked samples between 96% and 111%.

Magnetic solid-phase extraction is a useful tool for pre-concentration of samples, particularly those with expected low levels on analytes. Tian and team²³⁷ optimised an automatic extraction of Cd, Cu, Mn, Pb and Zn in cereals and feeds using carboxyl-functionalized magnetic beads. Key conditions were pH, eluent composition and absorption time. Using the optimised system, pre-concentration factors of between 40 and 200 were obtained, the LODs were 0.16 ng L⁻¹ for Cd, 11.54 ng L⁻¹ for Cu, 8.84 ng L⁻¹ for Mn, 6.01 ng L⁻¹ for Pb and 2.71 ng L⁻¹ for Zn. Loading capacity for the magnetic beads was found to be between 152 mg g⁻¹ and 426 mg g⁻¹. A range of samples of grain and feed were compared using traditional MAD preparation, which demonstrated the pre-concentration method gave suitable recoveries between 80% and 110%.

8.2.5 Vegetables, fruit, mushrooms and nuts. Mercury content of beetroot and beetroot supplements were measured by direct thermal decomposition-gold amalgamation cold vapour atomic absorption spectrometry. Conventional and organically grown beetroots were prepared to give unpeeled, peeled and skin portions of the vegetables. Samples were freeze-dried and analysed directly, the method was validated using CRMs, which returned between recoveries between 93% and 102% against certified vales, and afforded an LOQ of 0.96 ng g⁻¹. None of the samples were found to have levels of Hg which may be of concern, but it was shown that higher levels of Hg are seen in the skin of the beetroot. In the supplements the levels of Hg were also low, but had significantly higher variability. The authors²³⁸ suggested prudent use of these supplements to avoid excessive exposure to Hg.

8.2.6 Fish and seafood. Arsenic speciation in canned fish was the focus of a study carried by Hoyne and team. ¹⁷⁰ The canning process may be responsible for interconversion of AB into inorganic species, which may have a serious impact on human health. Canned tuna fish in both brine and oil from the local market were used for As speciation using IC-ICP-MS. Samples for IC-ICP-MS were extracted in an aqueous solution of 10 mmol L⁻¹ of (NH₄)₂CO₃ in 0.05% EDTA. Total As was measured by direct ICP-MS following MAD. The method was verified using CRM BCR-627 tuna fish muscle. Limits of quantification between 0.07 mg kg⁻¹ and 0.59 mg kg⁻¹ were achieved. The measured values of AB, DMA and total As were comparable to the CRM reference values. In the majority of the commercial samples AB was the largest contributor to the total As content (48–85%), whereas iAs was not detected in these studies. The author concluded that speciation of As is required to give a true understanding of the As content in fish, and that AB is stable through the processing of canned tuna and does not necessarily cause an excess of iAs.

An interesting review of atomic spectroscopy analysis of seaweeds was published by Lindenmayer et al. ²² The review set out the concerns for heavy metals in seaweeds, both as a health and environmental concern and discussed national and international limits particularly for As, Cd, Hg and Pb. The authors then covered sample preparation and the variety of available analytical techniques used to perform these analysis, as well as summarising some of the prior work carried out by other researchers in this field.

8.2.7 Drinking water and non-alcoholic beverages. In a comprehensive study carried out by Godebo et al.²³⁹ a variety of fruit juices, non-dairy milks, soft carbonated drinks and teas were analysed for Al, As, B, Ba, Ca, Cd, Co, Cr, Cu, Fe, K, Li, Mg, Mn, Mo, Na, Ni, Pb, Sb, Se, Sr, Th, Tl, U and Zn. The measured levels were compared to national and international guidance limits in drinking water to assess potential risk. Homogenised samples underwent hotplate digestion in reverse aqua-regia, followed by a secondary peroxide digest. The obtained solutions were analysed by ICP-MS for the 25 elements listed. The levels of most elements were below the recommended limits, except for Al, B, Mn and Ni which exceeded the limits in 40%, 17%, 37% and 38% of samples, respectively. Fruit juices were found to have the highest levels of nutritional and toxic elements, and the authors suggested that consumption of fruit juice should not be excessive, particularly in young children and infants.

8.2.8 Authenticity. In a summary review paper, Mazarakioti et al.²³ neatly collated some of the current work conducted where ICP-MS is the primary tool for investigating food authenticity and origin. The paper discussed the basic principle for ICP-MS and various matrices where the method has been applied for authenticity studies. It also called out the value of isotope ratio analysis to complement ICP-MS analysis. With 175 different papers mentioned, the review is a useful guide to inorganic monitoring for authenticity applications.

In another review, Varrà et al. ²⁴ looked at isotope fingerprinting as a modern tool in traceability and food safety in the animal-derived food chain. The review covered 135 studies looking at isotopic ratio as a measurement of authenticity in fish and seafood, meat, eggs, milk, and dairy products. Ratios of δ¹³C and δ¹⁵N were commonly looked at as well as Sr ratio and Pb isotope modelling. The authors concluded that isotopic modelling is a useful tool in the arsenal for food fraud prevention.

9. Abbreviations

AA	atomic absorption
AAS	atomic absorption spectrometry
AB	arsenobetaine
AC	arsenocholine
ACN	acetonitrile
AES	atomic emission spectrometry
AF	atomic fluorescence
AF4	asymmetric flow-field flow fractionation
AFS	atomic fluorescence spectrometry
ANN	artificial neural network
ANOVA	analysis of variance
APDC	ammonium pyrrolidine dithiocarbamate
APGD	atmospheric pressure glow discharge
ASU	Atomic Spectrometry Update
ASV	anodic stripping voltammetry
BARGE	bioaccessibility research group of Europe
CA	cluster analysis
CDA	canonical discriminant analysis
CE	capillary electrophoresis
CNT	carbon nanotube
CPE	cloud point extraction
CRC	collision/reaction cell
CRM	certified reference material
CS	continuum source
CV	cold vapour
CVG	chemical vapour generation
DA	discriminant analysis
DBD	dielectric barrier discharge
DBS	dry blood spot
DCM	dichloromethane
DDTC	diethyldithiocarbamate
DDTP	o,o-diethyldithiophosphate
DES	deep eutectic solvent
DLLME	dispersive liquid–liquid microextraction
DLS	dynamic light scattering
DMA	dimethylarsenic
DMSO	dimethysulfoxide
DNA	deoxyribonucleic acid
DRC	dynamic reaction cell
DSPME	dispersive solid phase micro extraction
DT	decision tree
EDTA	ethylenediaminetetraacetic acid
EDXRF	energy dispersive X-ray fluorescence
eGFR	estimated glomerular filtration rate
EQA	external quality assessment
ERM	European reference material
ESI	electrospray ionisation
ETAAS	electrothermal atomic absorption spectrometry
EtOH	ethanol
EtHg	ethylmercury
ETV	electrothermal vaporisation
EU	European Union
FAAS	flame atomic absorption spectrometry
FFF	field flow fractionation
FI	flow-injection
fs	femtosecond
FTIR	Fourier-transform infrared spectroscopy
GC	gas chromatography
GD	glow discharge
GF	graphite furnace
GFAAS	graphite furnace atomic absorption spectrometry
GO	graphene oxide
Hb	haemoglobin
HCA	hierarchical cluster analysis
HG	hydride generation
HGAAS	hydride generation atomic absorption spectrometry
HILIC	hydrophilic interaction liquid chromatography
HPLC	high performance liquid chromatography
HR	high resolution
IAEA	International Atomic Energy Agency
iAs	inorganic arsenic
IC	ion chromatography
ICH	international council on harmonization (of technical requirements for pharmaceuticals for human use)
ICP	inductively coupled plasma
ICP-QMS	ICP-quadrupole MS
ID	isotope dilution
id	internal diameter
iHg	inorganic mercury
IL	ionic liquid
INAA	instrumental neutron activation analysis
IR	infrared
IRMS	isotope ratio mass spectromety
IS	internal standard
iSb	inorganic antimony
ISE	ion selective electrode
ISO	international organization for standardisation
KED	kinetic energy discrimination
LA	laser ablation
LC	liquid chromatography
LDA	linear discriminant analysis
LIBS	laser-induced breakdown spectroscopy
LIF	laser-induced fluorescence
LLE	liquid liquid extraction
LLME	liquid liquid microextraction
LOD	limit of detection
LOQ	limit of quantification
LPME	liquid phase microextraction
LR	logistic regression
m/z	mass-to-charge ratio
MAD	microwave-assisted digestion
MAE	microwave-assisted extraction
MC	multicollector
MeHg	methylmercury
MeOH	methanol
MeSeCys	methylselenocysteine
MIBK	methyl isobutyl ketone
MIL	magnetic ionic liquid
MIP	microwave-induced plasma
MNP	magnetic nanoparticle
MMA	monomethylarsenic
MOF	metal–organic framework
MS	mass spectrometry
MS/MS	tandem MS
NIES	national institute for environmental studies
NIST	national institute of standards and technology
NMIJ	national metrology institute of Japan
NMR	nuclear magnetic resonance
NP	nanoparticle
NRCC	national research council of Canada
od	outer diameter
OES	optical emission spectrometry
OPLS-DA	orthogonal partial least squares discriminant analyses
PAGE	polyacrylamide gel electrophoresis
PCA	principal component analysis
PDO	protected designation of origin
PIGE	particle induced gamma-ray emission
PIXE	particle-induced X-ray emission
PLS	partial least squares
PLS-DA	partial least squares discriminant analysis
PT	proficiency testing
PTFE	poly(tetrafluoroethylene)
PVA	polyvinylalcohol
Q	quadrupole
QA	quality assurance
QC	quality control
QD	quantum dot
QMS	quadrupole mass spectrometry
QQQ	triple quadrupole
QT	quartz tube
REE	rare earth element
RF	random forest
RM	reference material
RNA	ribonucleic acid
RP	reversed phase
RSD	relative standard deviation
SAGD	solution anode glow discharge
sc	single cell
SD	standard deviation
SDS	sodium dodecylsulfate
SEC	size exclusion chromatography
SEM	scanning electron microscopy
SeCys	selenocysteine
SeCys2	selenocystine
SeMeCys	selenomethylcysteine
SeMet	selenomethionine
SF	sector field
SI	Système International d'unités – international system of units
SIMCA	soft independent modelling of class analogy
SIMS	secondary ion mass spectrometry
sp	single particle
SPE	solid phase extraction
SPME	solid phase microextraction
SQT	slotted quartz tube
SR	synchrotron radiation
SRM	standard reference material
SVM	support vector machine
SVR	support vector regression
SXRF	synchrotron XRF microscopy
TEM	transmission electron microscopy
THF	tetrahydrofuran
TIMS	thermal ionisation mass spectrometry
TMAH	tetramethylammonium hydroxide
TMA^+	tetramethylarsonium cation
TMAO	trimethylarsine oxide
TMSO	trimethylselenonium ion
TOF	time-of-flight
TXRF	total reflexion XRF
UAE	ultrasound-assisted extraction
UP	ultrapure
UPLC	ultra high performance liquid chromatography
UV	ultraviolet
UV-VIS	ultraviolet-visible spectrophotometry
VG	vapour generation
vs.	versus
WD	wavelength dispersive
XANES	X-ray absorption near-edge structure
XRD	X-ray diffraction
XRF	X-ray fluorescence

Conflicts of interest

There are no conflicts to declare.

References

1. M. Patriarca, N. Barlow, A. Cross, S. Hill, A. Robson and J. Tyson, J. Anal. At. Spectrom., 2023, 38(3), 496–577.
2. C. Vanhoof, J. R. Bacon, U. E. A. Fittschen and L. Vincze, J. Anal. At. Spectrom., 2022, 37(9), 1761–1775.
3. S. Carter, R. Clough, A. Fisher, B. Gibson and B. Russell, J. Anal. At. Spectrom., 2022, 37(11), 2207–2281.
4. J. R. Bacon, O. T. Butler, W. R. L. Cairns, O. Cavoura, J. M. Cook, C. M. Davidson and R. Mertz-Kraus, J. Anal. At. Spectrom., 2023, 38(1), 10–56.
5. E. H. Evans, J. Pisonero, C. M. M. Smith and R. N. Taylor, J. Anal. At. Spectrom., 2023, 38(5), 974–999.
6. R. Clough, C. F. Harrington, S. J. Hill, Y. Madrid and J. F. Tyson, J. Anal. At. Spectrom., 2023, 38(7), 1339–1371.
7. L. El Hosry, N. Sok, R. Richa, L. Al Mashtoub, P. Cayot and E. Bou-Maroun, Foods, 2023, 12(4), 895.
8. K. Shi, X. Tu, Y. Hu and X. Zhang, Appl. Spectrosc. Rev., 2023, 1–36,  DOI:10.1080/05704928.2023.2205920.
9. M. Ibourki, O. Hallouch, K. Devkota, D. Guillaume, A. Hirich and S. Gharby, J. Food Compos. Anal., 2023, 120, 105330.
10. V. Andruch, A. Kalyniukova, J. Płotka-Wasylka, N. Jatkowska, D. Snigur, S. Zaruba, J. Płatkiewicz, A. Zgoła-Grześkowiak and J. Werner, Microchem. J., 2023, 189, 108509.
11. M. F. Lanjwani, M. Tuzen, M. Y. Khuhawar, M. R. A. Mogaddam and M. A. Farajzadeh, Crit. Rev. Anal. Chem., 2022 DOI:10.1080/10408347.2022.2111655.
12. I. Hagarová and L. Nemček, Nutrients, 2022, 14(17), 3530.
13. N. Ullah and M. Tuzen, Crit. Rev. Anal. Chem., 2022, 1–19,  DOI:10.1080/10408347.2022.2128635.
14. I. Morales-Benítez, P. Montoro-Leal, J. C. García-Mesa, J. Verdeja-Galán and E. I. V. Alonso, TrAC, Trends Anal. Chem., 2022, 157, 116777.
15. D. Clases and R. Gonzalez de Vega, Anal. Bioanal. Chem., 2022, 414(25), 7337–7361.
16. D. Clases and R. Gonzalez de Vega, Anal. Bioanal. Chem., 2022, 414(25), 7363–7386.
17. A. B. S. da Silva and M. A. Z. Arruda, J. Trace Elem. Med. Biol., 2023, 75, 127086.
18. M. Vladitsi, C. Nikolaou, N. P. Kalogiouri and V. F. Samanidou, Methods Protoc., 2022, 5(4), 61.
19. C. Davison, D. Beste, M. Bailey and M. Felipe-Sotelo, Anal. Bioanal. Chem., 2023, 415(28), 6931–6950.
20. Z. Peng, P. Yin, D. Li, Y. Chen, K. Shu, Q. Fan, Y. Duan and Q. Lin, J. Innovative Opt. Health Sci., 2024, 17(01), 2330008.
21. A. C. M. Aleluia, M. d. S. Nascimento, A. M. P. dos Santos, W. N. L. dos Santos, A. de Freitas Santos Júnior and S. L. C. Ferreira, Spectrochim. Acta, Part B, 2023, 205, 106689.
22. R. Lindenmayer, L. Lu, F. Eivazi and Z. Afrasiabi, Appl. Sci., 2023, 13(8), 4764.
23. E. C. Mazarakioti, A. Zotos, A. A. Thomatou, A. Kontogeorgos, A. Patakas and A. Ladavos, Foods, 2022, 11(22), 3705.
24. M. O. Varrà, E. Zanardi, M. Serra, M. Conter, A. Ianieri and S. Ghidini, Molecules, 2023, 28(11), 4300.
25. M. Patriarca, N. Barlow, A. Cross, S. Hill, A. Robson, A. Taylor and J. Tyson, J. Anal. At. Spectrom., 2022, 37(3), 410–473.
26. N. Lemke, A. H. El-Khatib, T. Tchipilov, N. Jakubowski, M. G. Weller and J. Vogl, Anal. Bioanal. Chem., 2022, 414(15), 4441–4455.
27. S. Escudero-Cernuda, C. Garcia de Pablos, F. C. Celis, M. L. Fernandez-Sanchez and J. R. Encinar, Anal. Chim. Acta, 2023, 1251, 341002.
28. T. Narukawa, Anal. Sci., 2023, 39(7), 1191–1197.
29. P. K. Yadav, S. S. Tripathy, H. Chandra, L. Taneja, C. Kochar, A. Krishna, J. Pokhariyal, D. D. Toppo, S. Raina, N. Singh, B. Singh and R. K. Kotnala, Mapan, 2023, 38(4), 815–825.
30. L. Hendriks, R. Brunjes, S. Taskula, J. Kocic, B. Hattendorf, G. Bland, G. Lowry, E. Bolea-Fernandez, F. Vanhaecke, J. J. Wang, M. Baalousha, M. von der Au, B. Meermann, T. R. Holbrook, S. Wagner, S. Harycki, A. Gundlach-Graham and F. von der Kammer, Nanoscale, 2023, 15(26), 11268–11279.
31. C. Zhang, Y. Yan, J. Zeng, J. Liu, N. Dong and C. Zhang, Clin. Chim. Acta, 2023, 544, 117331.
32. K. W. Liao, P. C. Chen, W. C. Chou, I. Shiue, H. I. Huang, W. T. Chang and P. C. Huang, Int. J. Hyg. Environ. Health, 2023, 252, 114195.
33. R. Ruiz, C. Estevan, J. Estevez, C. Alcaide, M. A. Sogorb and E. Vilanova, Int. J. Mol. Sci., 2023, 24(9), 8127.
34. F. Wikstrom, C. Olsson, B. Palm, N. Roxhed, L. Backlund, M. Schalling and O. Beck, J. Pharm. Biomed. Anal., 2023, 227, 115269.
35. M. Perrais, A. Thomas, M. Augsburger and S. Lenglet, J. Anal. Toxicol., 2023, 47(2), 175–181.
36. A. Breton, C. M. Cirtiu, N. Fleury, A. Lajeunesse and J. Rudge, Bioanalysis, 2023, 15(2), 71–81.
37. R. M. Coyte, T. H. Darrah, E. Barrett, T. G. O'Connor, J. W. Olesik, C. M. Salafia, R. C. Shah, T. Love and R. K. Miller, Placenta, 2023, 131, 98–103.
38. L. Copeland-Hardin, T. Paunesku, J. S. Murley, J. Crentsil, O. Antipova, L. Li, E. Maxey, Q. Jin, D. Hooper, B. Lai, S. Chen and G. E. Woloschak, Sci. Rep., 2023, 13(1), 7806.
39. N. Beauval, A. Leroyer, M. Hisbergues, D. Allorge, P. Maboudou, M. Howsam and C. Nisse, J. Trace Elem. Med. Biol., 2022, 74, 127080.
40. Z. Gajdosechova, P. Grinberg, K. Kubachka, M. Wolle, A. Raab, J. Feldmann, R. Sim, Á. H. Pétursdóttir, T. Matoušek, S. Musil, B. Wozniak, S. Springer, N. W. Sadiq, H. Gurleyuk, C. H. Palmer, I. Pihillagawa Gedara and Z. Mester, Environ. Chem., 2023, 20, 5.
41. E. Matsumoto, S. Ito and T. Nishimura, Mar. Biotechnol., 2023, 25(5), 677–682.
42. S. Soares, L. M. B. Moraes, F. R. P. Rocha and A. Virgilio, J. Food Compos. Anal., 2023, 115, 104942.
43. I. Hagarova, Chem. Listy, 2022, 116(7), 423–431.
44. H. Ahmed, K. M. S. Almohammed, M. N. Janjua and N. Alhafez, Int. J. Environ. Anal. Chem., 2022, 1–14,  DOI:10.1080/03067319.2022.2070011.
45. A. A. Brouziotis, A. Giarra, A. Marano, A. Di Nunzio, F. Lombardo, M. Guida, G. Libralato and M. Trifuoggi, Front. Environ. Sci., 2022, 10 DOI:10.3389/fenvs.2022.942441.
46. Y. Liu, Y. Yang, Y. Y. Xia, J. V. de Seymour, D. Z. Zhao, Y. M. Li, H. Zhang and T. L. Han, J. Anal. Methods Chem., 2022, 2022, 5742468.
47. S. Aydinoglu, Braz. J. Pharm. Sci., 2022, 58, e21094.
48. I. Bankaji, R. Kouki, N. Dridi, R. Ferreira, S. Hidouri, B. Duarte, N. Sleimi and I. Caçador, Separations, 2023, 10(1), 40.
49. Y. Chen, N. Guo, L. Zhang, K. Hu, J. Yang and K. Shi, J. Radioanal. Nucl. Chem., 2023, 332(4), 1071–1081.
50. G. K. Deshwal, L. G. Gomez-Mascaraque, M. Fenelon and T. Huppertz, Molecules, 2023, 28(10), 3988.
51. M. R. Mehrasebi, M. Peyda, R. Feizolahi, A. Taromi, P. Homayuni, S. Fathi and Z. Farahmandkia, Environ. Health Eng. Manage. J., 2023, 10(2), 191–196.
52. M. Welna, A. Szymczycha-Madeja and P. Pohl, Molecules, 2023, 28(8), 3374.
53. J. Lee, Y.-S. Park and D. Y. Lee, Lwt, 2023, 173, 114351.
54. M. S. Lemos and K. G. F. Dantas, Biol. Trace Elem. Res., 2023, 201(7), 3513–3519.
55. X.-Z. Wang, Y.-Y. Chang, Y. Chen, H.-L. Wu, T. Wang, Y.-J. Ding and R.-Q. Yu, Microchem. J., 2022, 183, 107937.
56. W. Z. Fang, H. F. Tang and J. Zhou, J. Anal. Chem., 2023, 78(6), 728–736.
57. X. Zhang, X. Y. Jiang, S. Chen, Y. L. Yu and J. H. Wang, Anal. Chem., 2023, 95(4), 2375–2381.
58. V. P. Kolotov, D. N. Dogadkin, V. E. Zaichik, V. I. Shirokova and N. N. Dogadkin, J. Anal. Chem., 2023, 78(3), 324–329.
59. G. C. Leal, F. Rovasi, M. Maziero, P. C. do Nascimento, L. M. de Carvalho and C. Viana, J. Food Compos. Anal., 2022, 112, 104651.
60. M. Maziero and C. Viana, Spectrosc. Lett., 2022, 55(8), 534–545.
61. Z. Bunina, V. Varchenko, K. Bryleva, V. Chornyi and K. Belikov, ChemistrySelect, 2022, 7(29), e202201457.
62. T. G. S. Guimarães, F. S. Costa, I. M. N. R. Menezes, A. P. R. Santana, D. F. Andrade, A. Oliveira, C. D. B. Amaral and M. H. Gonzalez, J. Mol. Liq., 2023, 381, 121801.
63. A. Torres, I. Romero, R. Sánchez, A. Beltrán, G. Guillena and J.-L. Todolí, Talanta Open, 2023, 7, 100234.
64. X. Yang, C. Yan, R. Zhang, Y. Sun, Z. Li, Y. Liu, S. Yang, L. Shen and X. Wen, Anal. Chim. Acta, 2022, 1221, 340109.
65. X. Y. Ding, X. Teng, Z. X. She, Y. Li, Y. Y. Liu, Y. Zhuang and C. C. Wang, RSC Adv., 2022, 12(50), 32526–32533.
66. Y. Xiao, Y. He, C. Ji, M. Z. Hua, W. Liu, S. Yang, D. Chen, W. Zheng and X. Lu, Food Chem., 2023, 404(Pt A), 134680.
67. D. Ozdes, C. Duran, H. Bektas and E. Mentese, J. Serb. Chem. Soc., 2023, 88(3), 301–312.
68. C. Duran, D. Ozdes, E. Mentese and H. Bektas, Anal. Lett., 2024, 57(2), 225–236.
69. A. S. Farias, H. M. Santos, A. L. S. d. Silva Junior, V. H. C. d. Silva, R. B. e. S. Mendonça, J. P. Coutinho, I. P. Lôbo and R. M. d. Jesus, Food Sci. Technol., 2022, 42, e01321.
70. P. Mayotha, P. Chantiratikul and W. Siriangkhawut, J. Food Meas. Charact., 2023, 17(1), 607–614.
71. F. Iaquinta, J. Santander, M. Piston and I. Machado, Braz. J. Anal. Chem., 2022, 9(37), 129–139.
72. Y. Hao, G. Yin, X. Wang, Z. Jiang, G. Zhang, Z. Feng and Q. Sun, Int. J. Anal. Chem., 2022, 2022, 9299416.
73. M. Calleja-Gómez, J. M. Castagnini, E. Carbó, E. Ferrer, H. Berrada and F. J. Barba, Separations, 2022, 9(10), 302.
74. L. Chirita, E. Covaci, M. Ponta and T. Frentiu, Anal. Methods, 2023, 15(13), 1734–1746.
75. M. N. Oviedo, M. B. Botella, E. F. Fiorentini, P. Pacheco and R. G. Wuilloud, Spectrochim. Acta, Part B, 2023, 199, 106591.
76. J. R. Chen, W. H. Tsai and C. K. Su, Anal. Chim. Acta, 2023, 1271, 341489.
77. Y. Lu, X. Wei, M. Chen and J. Wang, Anal. Chim. Acta, 2023, 1251, 340993.
78. M. Tuchtenhagen, M. Stiboller, B. Witt and T. Schwerdtle, J. Anal. At. Spectrom., 2023, 38(3), 587–594.
79. W. Shi, Y. Hou, Z. Zhang, X. Yin, X. Zhao and L. Yuan, Horticulturae, 2023, 9(2), 161.
80. M. S. Leoncio and E. E. Garcia, Biol. Trace Elem. Res., 2023, 201(12), 5825–5838.
81. B. Kang, H. Liu, G. Chen, H. Lin, S. Chen and T. Chen, Food Chem., 2023, 408, 135192.
82. F. L. F. da Silva, F. E. H. Lima, D. M. Andrade Neto, F. L. de Menezes, L. Fechine, A. Akhdhar, L. P. D. Ribeiro, A. R. A. Nogueira, P. B. A. Fechine, G. S. Lopes and W. O. Matos, Food Chem., 2023, 412, 135461.
83. Ş. Tokalıoğlu, S. Shahir, B. F. Şenkal and E. T. Akgül, J. Food Compos. Anal., 2023, 118, 105210.
84. S. E. O'Sullivan, S. S. Jensen, G. N. Nikolajsen, H. Z. Bruun, R. Bhuller and J. Hoeng, J. Water Chem. Technol., 2023, 45(3), 256–269.
85. S. Chen, Y. Wang, J. Yan, C. Wang and D. Lu, J. Food Meas. Charact., 2023, 17(4), 3745–3752.
86. J. T. Yan, S. Zhang, C. L. Wang, D. B. Lu and S. Z. Chen, Anal. Lett., 2023, 57(1), 17–29.
87. H. Shirkhanloo, M. Habibnia, A. Rashidi, A. Faghihi Zarandi and M. Dehghani Mobarake, Food Chem., 2023, 403, 134394.
88. J. Rakhtshah, H. Shirkhanloo and M. Dehghani Mobarake, Food Chem., 2022, 389, 133124.
89. S. Z. Chen, C. H. Zhang, J. T. Yan, C. L. Wang and D. B. Lu, J. Iran. Chem. Soc., 2022, 19(11), 4389–4396.
90. S. Chen, J. Yan, C. Wang, C. Zhang and D. Lu, J. Food Compos. Anal., 2023, 117, 104967.
91. S. Chen, J. Yan, C. Zhang, C. Wang and D. Lu, Anal. Lett., 2023, 56(9), 1488–1501.
92. Z. K. You, X. L. Li, J. Huang, R. Q. Chen, J. Y. Peng, W. W. Kong and F. Liu, Molecules, 2023, 28(6), 2777.
93. Y. J. Zhu, S. X. Ma, G. Y. Yang, H. W. Tian and D. M. Dong, Front. Phys., 2023, 11, 1179574.
94. D. V. Danilov, P. Y. Sharanov and N. V. Alov, Pharm. Chem. J., 2023, 57, 306–310.
95. M. Janeda and M. Slachcinski, Anal. Biochem., 2023, 670, 115129.
96. J. Y. Cai, S. Liu, S. Chen, Y. L. Yu and J. H. Wang, Anal. Chem., 2023, 95(26), 9813–9821.
97. A. Morales-Rodriguez, M. Perez-Lopez, E. Puigpelat, A. Sahuquillo, D. Barron and J. F. Lopez-Sanchez, J. Chromatogr. A, 2022, 1684, 463549.
98. A. Jerse, H. Amlund, S. L. Holdt and J. J. Sloth, Environ. Chem., 2023, 20(2), 95–104.
99. R. Sim, J. Feldmann, D. B. Stengel and Á. H. Pétursdóttir, Environ. Chem., 2023, 20, 55.
100. R. Li, Y. Pan, C. Sun, C. Lin, S. Chen, Y. Wu and F. Fu, Anal. Methods, 2023, 15(14), 1802–1811.
101. F. Costa, C. L. Mieiro, M. E. Pereira and J. P. Coelho, Mar. Pollut. Bull., 2022, 179, 113736.
102. E. C. da Silva Junior, M. Babaahmadifooladi, K. Folens, A. R. dos Reis, L. R. G. Guilherme, T. Van de Wiele, L. Jacxsens and G. Du Laing, J. Food Compos. Anal., 2023, 118, 105162.
103. X. Tian, H. Jiang, M. Wang, W. Cui, Y. Guo, L. Zheng, L. Hu, G. Qu, Y. Yin, Y. Cai and G. Jiang, Anal. Chim. Acta, 2023, 1240, 340756.
104. X. W. Tian, X. Li, N. Liu, W. B. Cui, L. N. Zheng, Y. Y. Guo, Y. W. Liu, L. G. Hu, M. Wang, Y. Cai, Y. G. Yin, Y. Cai, G. B. Jiang and L. Jin, Chem. Commun., 2023, 59(38), 5709–5712.
105. R. Alvarez-Fernandez Garcia, L. Gutierrez Romero, J. Bettmer and M. Montes-Bayon, Nanomaterials, 2022, 13(1), 12.
106. P. Menero-Valdes, A. Lores-Padin, B. Fernandez, C. D. Quarles Jr, M. Garcia, H. Gonzalez-Iglesias and R. Pereiro, Talanta, 2023, 253, 123974.
107. H. Chang, J.-M. Zhu, X. Wang and T. Gao, J. Anal. At. Spectrom., 2023, 38(4), 950–962.
108. N. J. Norman, J. Ghali, T. L. Radzyukevich, J. A. Heiny and J. Landero-Figueroa, Microchem. J., 2023, 185, 108222.
109. J. Aoki, S. Wakaki, H. Ishiniwa, T. Kawakami, T. Miyazaki, K. Suzuki and Y. Takagai, Anal. Chem., 2023, 95(11), 4932–4939.
110. J. H. Dong, J. Z. Liu, P. J. Xing, S. Y. Li, J. S. Piao, H. F. Tao, L. J. Li, H. T. Zheng, Z. Zhang, S. R. Han, X. Liu and Z. L. Zhu, Anal. Chem., 2023, 95(15), 6271–6278.
111. T. Liu, E. Bolea-Fernandez and F. Vanhaecke, Spectrochim. Acta, Part B, 2022, 198, 106546.
112. Q. He, C. C. Li, X. Y. Kong, M. G. Zhao and J. Zhang, J. Anal. At. Spectrom., 2022, 37(9), 1876–1882.
113. G. Lan, X. Li, J. Yao, X. Yu, Q. Liu, C. Qiu and X. Mao, Front. Nutr., 2023, 10, 1201801.
114. F. Li, X. Lei, H. Li, H. Cui, W. Guo, L. Jin and S. Hu, J. Anal. At. Spectrom., 2023, 38(1), 90–96.
115. S. Kim, H.-C. Shin, H. Kim, S. H. Hur, N. Kim, S. U. Park, H. J. Kim and Y.-K. Kim, Anal. Lett., 2024, 57(1), 130–137.
116. X. Men, C. X. Wu, X. Zhang, X. Wei, W. Q. Ye, M. L. Chen, T. Yang, Z. R. Xu and J. H. Wang, Anal. Chim. Acta, 2022, 1226, 340268.
117. J. Zajda, P. Borowiecki and M. Matczuk, Talanta, 2023, 264, 124749.
118. M. Pan, Y. Lu, L. Feng, X. Zhou, J. Xiong and H. Li, Anal. Chem., 2022, 94(34), 11753–11759.
119. F. A. Redeker, J. E. Lesniewski, G. Hahm, W. P. McMahon and K. Jorabchi, Anal. Chem., 2022, 94(34), 11865–11872.
120. M. Atasoy, ACS Omega, 2023, 8(25), 22866–22875.
121. H. Karababa, M. Atasoy, D. Yildiz, I. Kula and M. Tuzen, ACS Omega, 2023, 8(7), 7063–7069.
122. T. Liu, J. Liu, X. Mao, X. Jiang, Y. Zhao and Y. Qian, Foods, 2023, 12(9), 1778.
123. Y. Liu, C. Yang, P. J. Xing, X. Liu, J. Z. Liu and Z. L. Zhu, J. Anal. At. Spectrom., 2023, 38, 742–750.
124. M. Gorska, J. Weiss and P. Pohl, Anal. Methods, 2023, 15(14), 1775–1789.
125. A. S. Martins, H. A. Dantas Filho and K. G. Fernandes Dantas, J. Anal. At. Spectrom., 2023, 38(5), 1007–1015.
126. B. M. Fontoura, A. I. Barros and J. A. Nobrega, Anal. Methods, 2023, 15(2), 154–163.
127. P. Pizzorno, L. Falchi, N. Mañay, M. Pistón and V. Bühl, Spectrochim. Acta, Part B, 2023, 201, 106630.
128. Eurachem Guide: the Fitness for Purpose of Analytical Methods – A Laboratory Guide to Method Validation and Related Topics, ed. B. Magnusson and U. Örnemark, 2nd edn, 2014, available from https://www.eurachem.org.
129. P. Vazquez-Quintal, J. Barrón-Zambrano, S. Medina-Peralta, Y. Moguel-Ordoñez, J. Nelson and D. Muñoz-Rodríguez, Anal. Lett., 2023, 56(14), 2249–2261.
130. J. A. Carter, P. J. Gray and T. I. Todorov, J. Anal. At. Spectrom., 2023, 38(4), 818–826.
131. N. Lerner, H. Avraham, N. Leffler, I. A. Weinstock and O. Zeiri, Anal. Methods, 2023, 15(12), 1594–1600.
132. J. Hassan and M. H. Karbasi, Curr. Anal. Chem., 2022, 18(9), 999–1002.
133. L. He, Y. Lin, P. Chen, Y. Su, Y. Li and C. Zheng, J. Hazard. Mater., 2022, 439, 129607.
134. H. H. Ma, S. Q. Shi, D. Zhang, N. Deng, Z. L. Hu, J. G. Liu and L. B. Guo, Foods, 2022, 11(21), 3398.
135. K. Wei, G. Teng, Q. Wang, X. Xu, Z. Zhao, H. Liu, M. Bao, Y. Zheng, T. Luo and B. Lu, Foods, 2023, 12(8), 1710.
136. W. T. Zhou, E. L. Wan, D. P. Tian and Y. Z. Liu, Microw. Opt. Technol. Lett., 2023, 65(5), 1331–1338.
137. L. Jiao, C. Sun, N. Yan, C. Yan, L. Qu, Q. Wang, S. Zhang and L. Ma, Anal. Lett., 2023, 56(16), 2625–2636.
138. H. Sun, C. Song, X. Lin and X. Gao, Spectrochim. Acta, Part B, 2022, 194, 106456.
139. L. Ren, Y. Tian, X. Yang, Q. Wang, L. Wang, X. Geng, K. Wang, Z. Du, Y. Li and H. Lin, Food Chem., 2023, 400, 134043.
140. W. H. Yan, X. Y. Yang, X. Geng, L. S. Wang, L. Lu, Y. Tian, Y. Li and H. Lin, Spectrosc. Spectral Anal., 2022, 42(12), 3714–3718.
141. M. Perez-Rodriguez, A. Mendoza, L. T. Gonzalez, A. Lima Vieira, R. G. Pellerano, J. A. Gomes Neto and E. C. Ferreira, Foods, 2023, 12(2), 365.
142. Q. Yan, Y. Tian, Y. Li, H. Lin, Z. Jia, Y. Lu, J. Yu, C. Sun, X. Bai and V. Detalle, Plasma Sci. Technol., 2023, 25(4), 045509.
143. R. R. Gamela, B. M. Fontoura, V. I. C. Costa, D. V. Babos and E. R. Pereira-Filho, Anal. Lett., 2023, 56(6), 944–957.
144. W. L. Wang, Y. C. Liu, Y. W. Chu, S. Y. Xiao, J. F. Nie, J. L. Zhang, J. W. Qi and L. B. Guo, Biomed. Opt. Express, 2022, 13(12), 6778–6790.
145. Y. Wang, Y. Bu, B. Yang and Y. Cai, J. Anal. At. Spectrom., 2023, 38(7), 1469–1477.
146. Z. Feng, S. Li, T. Gu, X. Zhou, Z. Zhang, Z. Yang, J. Hou, J. Zhu and D. Zhang, Molecules, 2022, 27(19), 6438.
147. P. Winnand, M. Ooms, M. Heitzer, M. Lammert, F. Holzle and A. Modabber, Oral Oncol., 2023, 138, 106308.
148. Priyanka, U. VK, P. Mitra, S. N, J. Baptist, K. K and R. Ongole, Heliyon, 2022, 8(10), e10946.
149. X. Wen, Z. Hu, J. Nie, Z. Gao, D. Zhang, L. Guo, S. Ma and D. Dong, J. Anal. At. Spectrom., 2023, 38(5), 1108–1115.
150. B. Štádlerová, L. Sagapova and S. Musil, J. Anal. At. Spectrom., 2023, 38(6), 1213–1223.
151. Z. Slejkovec, A. Stajnko, D. Mazej, M. J. Hudobivnik, O. Mestek, B. Lajin, W. Goessler, J. T. van Elteren and I. Falnoga, Anal. Bioanal. Chem., 2023, 415(2), 317–326.
152. Z. F. Liu, Z. L. Zhu, L. L. Jin and S. H. Hu, At. Spectrosc., 2022, 43(6), 465–473.
153. K. Greda, M. Welna, A. Szymczycha-Madeja and P. Pohl, Talanta, 2023, 253, 124072.
154. P. M. Carvalho, E. Marguí, A. Kubala-Kukuś, D. Banaś, J. Machado, D. Casal, D. Pais, J. P. Santos and S. Pessanha, Spectrochim. Acta, Part B, 2022, 198, 106548.
155. I. Allegretta, G. Squeo, C. E. Gattullo, C. Porfido, A. Cicchetti, F. Caponio, S. Cesco, C. Nicoletto and R. Terzano, Food Chem., 2023, 401, 134124.
156. S. Salles, R. Salles, M. S. G. Pavao, S. C. Cardoso and M. P. Stelling, BMC Cancer, 2023, 23(1), 322.
157. Y. Luo, T. Paunesku, O. Antipova, Y. Liu, N. J. Zaluzec, Z. Di, G. Woloschak and S. Chen, Metallomics, 2022, 14(9), mfac025.
158. S. B. Zhou, Q. M. Cheng, D. C. Weindorf, B. Y. Yang, Z. X. Yuan and J. Yang, J. Anal. At. Spectrom., 2022, 37, 2461–2469.
159. D. E. B. Fleming, M. G. Kaiser, B. D. Rankin and K. M. M. Schenkels, J. Trace Elem. Med. Biol., 2023, 77, 127136.
160. A. J. Specht, D. W. Steadman, M. Davis, S. M. Bartell and M. G. Weisskopf, Sci. Total Environ., 2023, 880, 163197.
161. L. Zhu, Y. Teng, W. D. Han, L. H. Yin and X. Teng, Analyst, 2023, 148, 1116–1122.
162. S. Sultana, H. A. Khatun, M. Faruquee, M. M. U. Islam, H. J. Tonny and M. R. Islam, Foods, 2023, 12(5), 1044.
163. A. Fuentes-Cervantes, J. Ruiz Allica, F. Calderon Celis, J. M. Costa-Fernandez and J. Ruiz Encinar, Nanomaterials, 2023, 13(6), 1132.
164. V. Kantorová, G. Krausová, I. Hyršlová, M. Loula, O. Mestek and A. Kaňa, Spectrochim. Acta, Part B, 2023, 199, 106592.
165. C. Suarez-Oubina, P. Herbello-Hermelo, N. Mallo, M. Vazquez, S. Cabaleiro, I. Pinheiro, L. Rodriguez-Lorenzo, B. Espina, P. Bermejo-Barrera and A. Moreda-Pineiro, Anal. Bioanal. Chem., 2023, 415(17), 3399–3413.
166. S. Feng, S. Shen, Y. Yao, M. Liang, Y. Chen and H. Liu, J. Sep. Sci., 2023, 46(4), e2200575.
167. O. Breuer, O. Brodin, A. Razaghi, D. Brodin, B. Gammelgaard and M. Bjornstedt, Biomedicines, 2023, 11(2), 295.
168. K. Flouda, J. M. Dersch, C. Gabel-Jensen, S. Stürup, S. Misra, M. Björnstedt M and B. Gammelgaard, Anal. Bioanal. Chem., 2016, 408(9), 2293–2301.
169. T. Narukawa, T. Suzuki, S. Okabayashi and K. Chiba, Anal. Methods, 2023, 15(2), 240–246.
170. T. F. M. Hoyne, L. V. Vieira, O. A. Heringer, G. P. Brandão, J. R. d. Souza and M. T. W. D. Carneiro, J. Food Compos. Anal., 2023, 116, 105051.
171. S. Chen and L. Liu, J. Food Compos. Anal., 2023, 115, 105001.
172. Y. Luo, G. Chen, X. Deng, H. Cai, X. Fu, F. Xu, X. Xiao, Y. Huo and J. Luo, Separations, 2022, 9(9), 242.
173. S. Strekopytov, K. Billimoria and H. Goenaga-Infante, J. Anal. At. Spectrom., 2023, 38(3), 704–715.
174. D. Metarapi, A. Schweikert, A. Jerse, M. Schaier, J. T. van Elteren, G. Koellensperger, S. Theiner and M. Sala, Anal. Chem., 2023, 95(19), 7804–7812.
175. M. Schaier, S. Theiner, D. Baier, G. Braun, W. Berger and G. Koellensperger, JACS Au, 2023, 3(2), 419–428.
176. S. B. Seiffert, M. Elinkmann, E. Niehaves, A. Vennemann, D. Mozhayeva, S. Kroger, M. Wiemann and U. Karst, Anal. Chem., 2023, 95(15), 6383–6390.
177. M. T. Westerhausen, M. Bernard, G. Choi, C. Jeffries-Stokes, R. Chandrajith, R. Banati and D. P. Bishop, Anal. Methods, 2023, 15(6), 797–806.
178. H. Deng, H. Xu, J. Z. Zhou, D. S. Tang, W. Q. Yang, M. Hu, Y. Zhang and Y. Q. Ke, Anal. Bioanal. Chem., 2023, 415(9), 1751–1764.
179. C. K. Pereira, V. M. Neves, G. M. Hidrich, H. Faccin, D. Pozebon and V. L. Dressler, Braz. J. Anal. Chem., 2023, 10(38), 79–91.
180. S. Braeuer, T. Van Helden, T. Van Acker, O. Leroux, D. Van Der Straeten, A. Verbeken, J. Borovicka and F. Vanhaecke, Anal. Bioanal. Chem., 2022, 414(25), 7517–7530.
181. M. Stiborek, L. Jindrichova, S. Meliorisova, A. Bednarik, V. Prysiazhnyi, J. Kroupa, P. Houska, B. Adamova, J. Navratilova, V. Kanicky and J. Preisler, Anal. Chem., 2022, 94(51), 18114–18120.
182. A. Tisza, T. Klikovits, M. Benej, S. Torok, B. Szeitz, Z. Valko, M. A. Hoda, B. Hegedus, M. Bonta, W. Nischkauer, K. Hoetzenecker, A. Limbeck, K. Schelch, V. Laszlo, Z. Megyesfalvi and B. Dome, Br. J. Clin. Pharmacol., 2023, 89(11), 3364–3374.
183. K. Veverkova, K. Pavelicova, M. Vlcnovska, M. Vejvodova, V. Horak, V. Kanicky, V. Adam, T. Vaculovic and M. Vaculovicova, J. Anal. At. Spectrom., 2023, 38(8), 1618–1625.
184. P. Bucker, R. M. Buzzi, K. Akeret, L. Mosberger, H. Richter, M. Sperling, M. Hugelshofer, D. J. Schaer and U. Karst, Metallomics, 2022, 14(8), mfac057.
185. S. Ma, Z. Zhang, Z. Xu, X. Qiu and W. Hang, J. Anal. At. Spectrom., 2023, 38(3), 595–602.
186. J. Lovric, N. Najafinobar, M. E. Kurczy, O. De Castro, A. Biesemeier, L. von Sydow, M. Klarqvist, T. Wirtz and P. Malmberg, Anal. Chem., 2022, 94(37), 12798–12806.
187. F. Penen, R. Raavé, A. Kip, S. Heskamp and P. Malmberg, Microchem. J., 2022, 181, 107860.
188. B. Busser, A. L. Bulin, V. Gardette, H. Elleaume, F. Pelascini, A. Bouron, V. Motto-Ros and L. Sancey, J. Neurosci. Methods, 2022, 379, 109676.
189. F. Forste, L. Bauer, C. Streeck, M. Radtke, U. Reinholz, D. Kadow, C. Keil and I. Mantouvalou, Anal. Chem., 2023, 95(13), 5627–5634.
190. G. Deng, M. Vu, M. Korbas, V. F. Bondici, C. Karunakaran, D. Christensen, H. A. Bart Lardner and P. Yu, Food Chem., 2023, 421, 135661.
191. T. Kim, W. S. Lee, M. Jeon, H. Kim, M. Eom, S. Jung, H. J. Im and S. J. Ye, Med. Phys., 2023, 50(1), 529–539.
192. M. Szczerbowska-Boruchowska, P. Stec, M. Czyzycki, Z. Szczerbowski, R. Simon, T. Baumbach and A. Ziomber-Lisiak, Spectrochim. Acta, Part A, 2023, 293, 122468.
193. A. P. Morrell, R. A. Martin, H. M. Roberts, H. Castillo-Michel, J. F. W. Mosselmans, K. Geraki, A. T. Warfield, P. Lingor, W. Qayyum, D. Graf, M. Febbraio and O. Addison, Metallomics, 2023, 15(6), mfad030.
194. X. Zhan, J. Zhou, Y. Jiang, P. An, B. Luo, F. Lan, B. Ying and Y. Wu, Biosens. Bioelectron., 2023, 229, 115229.
195. Q. Kang, B. Chen, M. He and B. Hu, Anal. Chem., 2022, 94(37), 12934–12941.
196. Y. Wen, X. W. Zhang, Y. Y. Li, S. Chen, Y. L. Yu and J. H. Wang, Anal. Chem., 2022, 94(46), 16196–16203.
197. Y. L. Zhu, J. K. Wang, Z. P. Chen, Y. J. Zhao and R. Q. Yu, Microchim. Acta, 2023, 190(1), 20.
198. M. Jiang, J. Zhou, X. B. Xie, Z. L. Huang, R. Liu and Y. Lv, Anal. Chem., 2022, 94(44), 15433–15439.
199. Z. L. Huang, X. Zhao, J. Y. Hu, C. C. Zhang, X. B. Xie, R. Liu and Y. Lv, Anal. Chem., 2022, 94(37), 12899–12906.
200. R. Karunanithy, S. Ratnasingam, T. Holland and P. Sivakumar, Bioconjugate Chem., 2023, 34(3), 501–509.
201. D. Johnson, D. Clases, M. L. Fernandez-Sanchez, N. Eiro, L. O. Gonzalez, F. J. Vizoso, P. A. Doble and R. Gonzalez de Vega, Metallomics, 2022, 14(8), mfac052.
202. Q. X. Sun, S. Q. Zhang, X. Wei, T. Yang, J. H. Wang and M. L. Chen, Anal. Chim. Acta, 2022, 1221, 340100.
203. Y. P. Cao, H. P. Chen, Y. L. Wei, Y. Y. Wei, J. Liang, Z. Z. Pan and B. Y. Deng, At. Spectrosc., 2023, 44(2), 65–75.
204. J. Liu, L. Peng, Q. Wang, X. D. Wang and H. Tang, Talanta, 2022, 250, 123720.
205. R. Pechancova, J. Gallo, D. Baron, D. Milde, P. Antal, Z. Slobodova, K. Lemr and T. Pluhacek, J. Biomed. Mater. Res., Part B, 2023, 111(2), 271–283.
206. M. A. Smolle, M. Bergovec, S. Scheipl, W. Gossler, F. Amerstorfer, M. Glehr, A. Leithner and J. Friesenbichler, Sci. Rep., 2022, 12(1), 13041.
207. P. J. Cundy, G. Antoniou, B. J. C. Freeman and W. J. Cundy, Spine, 2022, 47(17), 1241–1247.
208. V. Bali, Y. Khajuria, A. K. Pathak, U. Kumar, P. K. Rai, C. Ghany and V. K. Singh, X-Ray Spectrom., 2023, 1 DOI:10.1002/xrs3338.
209. A. A. Shaltout, O. H. Abd-Elkader, P. Lassen and U. A. E. Fittschen, X-Ray Spectrom., 2023, 52(6), 336–349.
210. E. Leese, K. Jones, B. Bocca, R. Bousoumah, A. Castano, K. S. Galea, I. Iavicoli, M. E. Lopez, V. Leso, S. Ndaw, S. P. Porras, F. Ruggieri, P. T. J. Scheepers, T. Santonen and HBM4EU chromates study team, Toxicol. Lett., 2023, 375, 59–68.
211. C. Marie-Desvergne, M. Dubosson, L. Leclerc, C. Campo, D. Bitounis, V. Forest, J. Pourchez, M. Cottier, J. M. Vergnon, A. Tarantini and V. Chamel-Mossuz, J. Breath Res., 2023, 17(1), 016008.
212. J. J. Y. Lin, L. J. Koffman, M. W. Tehrani, R. Chen, S. G. Han, D. P. Sandler, K. G. Lawrence, W. B. Jackson 2nd, A. S. Dickerson, G. Ramachandran, L. S. Engel and A. M. Rule, J. Exposure Sci. Environ. Epidemiol., 2023, 33(6), 945–953.
213. C. L. Gao, L. W. Lin, J. Li, M. L. Wu, J. Lv, S. Tian and X. Hai, Toxicol. Lett., 2023, 380, 31–39.
214. R. Shroff, A. D. Lalayiannis, M. Fewtrell, C. P. Schmitt, A. Bayazit, V. Askiti, A. Jankauskiene, J. Bacchetta, S. Silva, N. Goodman, L. McAlister, L. Biassoni, N. Crabtree, A. Rahn, D. C. Fischer, A. Heuser, A. Kolevica and A. Eisenhauer, Kidney Int., 2022, 102(3), 613–623.
215. G. Gromadzka, M. Grycan and A. M. Przybylkowski, Diagnostics, 2023, 13(11), 1830.
216. Z. Marino, C. Molera-Busoms, C. Badenas, J. Quintero-Bernabeu, M. Torra, X. Forns and R. Artuch, J. Inherited Metab. Dis., 2023, 46(5), 982–991.
217. X. Y. Fan, Q. Li, H. L. Fang, Y. Ju, Z. G. Jin, H. Y. Li and X. Q. Zhang, Clin. Chim. Acta, 2023, 538, 164–168.
218. J. Irrgeher, T. Berger, A. Tchaikovsky, C. Tschegg, G. Gouya, P. Lechner, A. Retzmann, C. Opper, C. Firbas, M. Freissmuth, K. Peschel-Credner, K. Anderle, C. Meisslitzer, M. Wolzt and T. Prohaska, Anal. Bioanal. Chem., 2023, 415(2), 255–268.
219. M. Winter, F. Lessmann and V. Harth, Anal. Methods, 2023, 15(16), 2030–2038.
220. B. Lajin and W. Goessler, Anal. Chem., 2023, 95(22), 8706–8710.
221. Y. Z. Liu, G. Xiao and R. L. Jones, J. Radioanal. Nucl. Chem., 2022, 331(9), 3957–3964.
222. A. Zergui, M. L. Joseph and C. B. Inkale, BioMetals, 2023, 36(5), 969–974.
223. C. Xu, X. Zhu, Y. Ma, B. Yan, S. Dai, X. Wang, J. Zhao, Y. Wang, S. Wu, Q. Zhao, X. Bao, W. Yang, F. Fang and S. Liu, Pharmacogn. Mag., 2023, 19(3), 592–602.
224. G. Isani, E. Ferlizza, M. Bertocchi, T. Dalmonte, S. Menotta, G. Fedrizzi and G. Andreani, Int. J. Mol. Sci., 2022, 23(22), 13949.
225. Y. Wang, Y. Bu, Y. Cai and X. Wang, J. Anal. At. Spectrom., 2023, 38(1), 121–130.
226. T. D'Amore, O. Miedico, C. Pompa, C. Preite, M. Iammarino and V. Nardelli, Life, 2023, 13(2), 511.
227. H. Blommaert, A. M. Aucour, M. Wiggenhauser, C. Moens, P. Telouk, S. Campillo, J. Beauchene, G. Landrot, D. Testemale, S. Pin, C. Lewis, P. Umaharan, E. Smolders and G. Sarret, Front. Plant Sci., 2022, 13, 1055912.
228. M. Capodiferro, E. Marco and J. O. Grimalt, Environ. Pollut., 2022, 314, 120274.
229. E. L. de Paiva, A. Ruttens, N. Waegeneers, G. D. Laing, M. A. Morgano, K. Cheyns and A. P. Arisseto-Bragotto, Food Res. Int., 2023, 164, 112289.
230. M. A. Subirana, R. Boada, T. Xiao, M. Llugany and M. Valiente, Physiol. Plant., 2023, 175(1), e13843.
231. L. A. Portugal, E. Palacio, R. Frizzarin, J. H. S. Neto, V. Cerda and S. L. C. Ferreira, Talanta, 2023, 252, 123897.
232. S. González-Suárez, S. Paz-Montelongo, D. Niebla-Canelo, S. Alejandro-Vega, D. González-Weller, C. Rubio-Armendáriz, A. Hardisson and Á. J. Gutiérrez-Fernández, Appl. Sci., 2022, 12(16), 8044.
233. M. S. Rihawy, D. Halloum, A. Wassouf, M. Alwazzeh and K. Abbas, Appl. Radiat. Isot., 2023, 193, 110622.
234. F. Fantuz, S. Ferraro, L. Todini, R. Spurio, A. Fatica, F. Marcantoni and E. Salimei, J. Dairy Sci., 2022, 105(8), 6422–6430.
235. F. Fantuz, L. Todini, S. Ferraro, A. Fatica, F. Marcantoni, M. Zannotti and E. Salimei, Beverages, 2022, 8(3), 51.
236. S. A. Wadood, J. Nie, C. Li, K. M. Rogers, A. Khan, W. A. Khan, A. Qamar, Y. Zhang and Y. Yuwei, J. Food Compos. Anal., 2022, 112, 104677.
237. W. Tian, Y. Liu, S. Wang, J. Ye, H. Liu, Y. Wang and M. Zhou, Foods, 2022, 11(24), 3944.
238. J. Brzezińska-Rojek, M. Rutkowska, J. Ośko, P. Konieczka, M. Prokopowicz and M. Grembecka, J. Food Compos. Anal., 2022, 114, 104828.
239. T. R. Godebo, H. Stoner, M. Pechilis, H. Taylor-Arnold, J. Ashmead, L. Claman, L. Guest, W. Consolati, O. DiMatteo, M. Johnson, K. Cowden, D. Shaferman, E. Gordon, H. Dillman, N. Phan, A. Tegegn, S. V. Garrido and E. Heard, J. Food Compos. Anal., 2023, 119, 105230.
240. I. López-García, M. J. Muñoz-Sandoval and M. Hernández-Córdoba, Spectrochim. Acta, Part B, 2023, 202, 106643.
241. G. Song, X. Guo, Q. Li, J. Liao, D. Wang, T. Yuan, L. Li, R. Fang, M. Zhang, Q. Shen, F. Zheng and J. Gong, Food Control, 2023, 152, 109846.
242. N. Ebrar Karlidag, M. Toprak, R. Demirel, B. Tugba Zaman and S. Bakirdere, Food Chem., 2022, 396, 133669.
243. F. Shah and M. Ghafoor, Separations, 2023, 10(2), 124.
244. A. Islam and S. Rais, Food Chem., 2023, 405(Pt A), 134751.
245. L. R. Castilho, V. dos Santos Fernandes Corrêa, M. A. Rosa, A. C. do Lago, G. de Fátima Lima Martins and E. Costa Figueiredo, Food Anal. Methods, 2023, 16(6), 1017–1029.
246. H. Alsaeedi, H. Ahmad, M. F. Altowairqi, A. A. Alhamed and A. Alsalme, Polymers, 2023, 15(3), 532.
247. S. Amini, S. H. Kandeh, H. Ebrahimzadeh and P. Khodayari, Food Chem., 2023, 420, 136122.
248. E. Pourbasheer, S. Morsali, S. Ansari, B. Mirtamizdoust, H. Vojoudi and M. R. Ganjali, Curr. Anal. Chem., 2022, 18(8), 907–913.
249. T. D. Dinh, Q. L. Nguyen, M. D. Vu, T. M. H. Tran, T. H. N. Tran, M. H. Nguyen and T. D. Pham, Colloid Polym. Sci., 2023, 301(10), 1197–1207.
250. K. Parvizzad, S. M. Sorouraddin and M. A. Farajzadeh, J. Food Compos. Anal., 2022, 112, 104654.
251. A. A. Gouda, R. El Sheikh, H. M. El Sayed, A. M. Khedr, S. A. Al Ezz, W. Gamil and M. Hamdy, J. Appl. Spectrosc., 2022, 89(3), 567–578.
252. T. Xia, X. Yang, R. Zhang, A. Huang, K. Hu, F. Hao, Y. Liu, Q. Deng, S. Yang and X. Wen, Talanta, 2023, 256, 124316.
253. P. Ali Mohammadzadeh Baghaei, M. R. Afshar Mogaddam, M. A. Farajzadeh, A. Mohebbi and S. M. Sorouraddin, J. Food Compos. Anal., 2023, 117, 105125.
254. A. Islam and H. Javed, J. Appl. Polym. Sci., 2022, 139(43), e53065.
255. Z. X. Li, T. Che, M. H. Yang and X. J. Hu, Food Addit. Contam. Part A Chem. Anal. Control Expo. Risk Assess., 2022, 39(11), 1828–1842.
256. H. Haghighat, Microchem. J., 2022, 181, 107770.
257. A. Zanganeh, H. R. Ghasempour, M. K. Koohi and N. Karimi, Res. Chem. Intermed., 2023, 49(5), 2123–2140.
258. E. S. Aydin, B. T. Zaman, H. Serbest, F. Kapukiran, F. Turak and S. Bakirdere, J. Food Compos. Anal., 2022, 114, 104833.
259. H. E. H. Ahmed, O. Ozalp and M. Soylak, J. Food Compos. Anal., 2023, 118, 105163.
260. O. Ozalp and M. Soylak, Talanta, 2023, 253, 124082.
261. A. Elik, Sustainable Chem. Pharm., 2022, 29, 100801.
262. B. Kartoğlu, A. Bahçivan, S. Erarpat, A. Bayraktar and S. Bakirdere, J. Food Compos. Anal., 2023, 121, 105373.
263. A. Alsalme, J. Phys. Chem. Solids, 2023, 180, 111433.
264. A. Sharifi, R. Hallaj and S. Bahar, J. Anal. Methods Chem., 2023, 2023, 5424221.
265. F. A. Borges, L. M. Costa, C. R. T. Tarley, G. de Fatima Lima Martins and E. C. Figueiredo, Food Chem., 2023, 413, 135676.
266. R. Kalantari, A. Moghimi and F. Azizinezhad, J. Appl. Spectrosc., 2023, 90(3), 686–695.
267. A. C. Carneiro, J. Meneguin, P. dos Santos, M. Corazza, M. C. Prete, A. W. Rinaldi and C. R. Tarley, J. Braz. Chem. Soc., 2022, 33(8), 958–968.
268. A. T. Bişgin, J. Mol. Liq., 2023, 384, 122207.
269. S. Bodur, S. Erarpat, Ö. F. Tutar and S. Bakırdere, J. Food Compos. Anal., 2023, 117, 105144.
270. N. Manousi, A. Kabir, K. G. Furton and A. N. Anthemidis, Anal. Chem., 2022, 94(38), 12943–12947.
271. N. Manousi, A. Kabir, K. G. Furton and A. Anthemidis, Molecules, 2023, 28(5), 2103.
272. K. Hanifar, Y. Q. Almajidi, H. Sanaan Jabbar, A. Alexis Ramirez-Coronel, F. M. A. Altalbawy, A. F. Almulla, A. Turki Jalil, S. A. Awad and L. Andres Barboza-Arenas, J. Food Prot., 2023, 86(7), 100102.
273. S. M. Uvaisova, M. A. Babuev and U. M. Damyrov, J. Anal. Chem., 2023, 78(5), 566–571.
274. M. Behbahani, G. Rabiee, S. Bagheri and M. M. Amini, Microchem. J., 2022, 183, 107951.
275. N. Altunay, M. Tuzen and M. R. Afshar Mogaddam, Food Addit. Contam. Part A Chem. Anal. Control Expo. Risk Assess., 2022, 39(12), 1963–1973.
276. M. Mahmoudian, A. H. M. Sarrafi, E. Konoz and A. Niazi, ChemistrySelect, 2022, 7(48), e202203738.
277. N. E. Karlidag, R. Demirel, H. Serbest, F. Turak and S. Bakirdere, Anal. Methods, 2023, 15(1), 56–62.
278. T. Zarnoushe Farahani, G. Bagherian, M. Arab Chamjangali and M. Ashrafi, SN Appl. Sci., 2023, 5(7), 183.
279. O. Sagsoz, B. Arvas, B. T. Zaman, C. Yolacan and S. Bakirdere, Environ. Monit. Assess., 2023, 195(2), 313.
280. G. B. Grecco, K. F. Albini, L. S. Longo, M. A. Andreo, B. L. Batista, F. R. Lourenço and L. A. Calixto, J. Iran. Chem. Soc., 2023, 20(2), 371–380.
281. P. Farisi, M. R. Afshar Mogaddam, M. A. Farajzadeh and M. Nemati, J. Food Compos. Anal., 2023, 117, 105107.
282. P. A. M. Baghaei, S. M. Sorouraddin, M. A. Farajzadeh and M. R. A. Mogaddam, J. Iran. Chem. Soc., 2022, 20(2), 339–349.
283. F. Sotouneh, M. R. Jamali, A. Asghari and M. Rajabi, Microchem. J., 2023, 190, 108622.
284. A. A. A. Yaqin, S. Suherman, M. Mudasir and D. Siswanta, Indones. J. Chem., 2022, 22(6), 1469–1479.
285. A. Shishov, I. Gurev and A. Bulatov, J. Food Compos. Anal., 2023, 119, 105247.
286. F. Cajamarca Suquila, L. Bertoldo, E. Lins and C. Tarley, Separations, 2023, 10(2), 122.
287. J. R. K. Silveira, L. C. Brudi, S. R. Waechter, P. A. Mello, A. B. Costa and F. A. Duarte, Microchem. J., 2023, 184, 108181.
288. Y. Bakircioglu Kurtulus, D. Bakircioglu and N. Topraksever, J. Food Compos. Anal., 2022, 112, 104682.
289. S. M. Sorouraddin, K. Parvizzad and M. A. Farajzadeh, Anal. Sci., 2023, 39(1), 23–31.
290. A. S. Al-Wasidi, A. M. Naglah, F. A. Saad and E. A. Abdelrahman, Arabian J. Chem., 2022, 15(10), 104178.
291. I. A. Wysocka, A. M. Rogowska and P. Kostrz-Sikora, Environ. Pollut., 2023, 323, 121289.
292. J. L. Liu, Q. Han, E. Y. Hu, C. Yang and M. M. Yin, J. Anal. Chem., 2023, 78(3), 303–309.
293. S. S. Alharthi and H. M. Al-Saidi, Arabian J. Chem., 2022, 15(9), 104094.
294. M. Soylak, I. Ungur and O. Ozalp, Instrum. Sci. Technol., 2023, 51(4), 447–464.
295. G. J. Shabaa, F. A. Semysim, R. K. Ridha, E. A. Azooz and E. A. J. Al-Mulla, J. Iran. Chem. Soc., 2023, 20(9), 2277–2284.
296. P. Chaikhan, Y. Udnan, R. J. Ampiah-Bonney and W. C. Chaiyasith, J. Food Compos. Anal., 2023, 118, 105214.
297. R. Zhang, C. X. Yan, X. F. Yang, K. Hu, F. F. Hao, S. C. Yang, Q. W. Deng, Z. J. Duan, Y. Liu and X. D. Wen, Anal. Chim. Acta, 2023, 1251, 340992.
298. J.-L. Liu, X.-M. Zhou, M. Sun, A.-Q. Jia, H.-T. Shi and Q.-F. Zhang, Arabian J. Chem., 2023, 16(7), 104866.
299. P. Y. Quintas, E. Agostini, A. L. Wevar Oller and L. B. Escudero, J. AOAC Int., 2023, 106(6), 1542–1549.
300. T. A. Jouybari, H. A. Jouybari, F. Hosseini, M. Nesari and N. Fattahi, Environ. Sci. Pollut. Res. Int., 2023, 30(9), 24553–24561.
301. Y. Y. Ni, W. T. Bu, K. Xiong, S. Hu, C. T. Yang and L. G. Cao, Talanta, 2023, 262, 124710.
302. N. Altunay, B. Hazer, M. F. Lanjwani, M. Tuzen, H. Ul Haq and G. Boczkaj, Food Chem., 2023, 425, 136523.
303. V. Georgiev, I. Dakova and I. Karadjova, Molecules, 2022, 27(17), 5516.
304. D. S. C. Neves, A. S. Souza and L. R. de Lemos, J. Food Compos. Anal., 2023, 118, 105206.
305. Z. X. Lu, C. Zhao, J. Yang, Y. F. Ma and M. Qiang, Environ. Geochem. Health, 2023, 45(6), 3055–3068.
306. A. Pritam, A. Priyadarshini, K. Hussain, A. Kumar, N. Kumar and A. Malakar, J. Pharm. BioAllied Sci., 2021, 13(6), S1588–S1590.
307. B. Bocca, V. Leso, B. Battistini, S. Caimi, M. Senofonte, M. Fedele, D. M. Cavallo, A. Cattaneo, P. Lovreglio and I. Iavicoli, Environ. Res., 2023, 216(Pt 3), 114736.
308. M. Yepes-Calderon, D. Kremer, A. Post, C. G. Sotomayor, U. Seidel, P. Huebbe, T. J. Knobbe, K. Luersen, M. F. Eisenga, E. Corpeleijn, M. H. de Borst, G. J. Navis, G. Rimbach and S. J. L. Bakker, Am. J. Nephrol., 2023, 54(9–10), 425–433.
309. F. J. Guo, Y. Y. Lin, L. Meng, L. H. Peng, H. F. Zhang, X. Zhang, M. J. Jin, J. B. Wang, Y. M. Zhang, M. L. Tang and K. Chen, Clin. Nutr., 2022, 41(12), 2720–2728.
310. L. Suarez-Criado, P. Rodriguez-Gonzalez, J. Marrugo-Negrete, J. I. G. Alonso and S. Diez, Environ. Res., 2023, 231, 115970.
311. Y. Tang, L. Liu, Q. Nong, H. Guo, Q. Zhou, D. Wang, Y. Yin, J. Shi, B. He, L. Hu and G. Jiang, J. Chromatogr. A, 2022, 1677, 463303.
312. S. R. Shaw, S. Scott, M. Pell, J. Raiche-Gill, C. Tomasallo and J. Meiman, J. Environ. Health, 2023, 85(9), 8–19.
313. C. Wang, R. Ran, X. Jin and X. Zhu, Medicine, 2022, 101(39), e30877.
314. L. Abou-Zeid, A. Pell, M. G. Cortes, H. Isnard, P. Delangle and C. Bresson, Anal. Chim. Acta, 2023, 1242, 340773.
315. G. Xiao and J. Button, J. Radioanal. Nucl. Chem., 2023, 332(1), 185–191.
316. Y. Liu, G. Xiao and R. L. Jones, Anal. Chem., 2022, 94(51), 18042–18049.
317. Y. Yu, W. Meng, H. Kuang, X. Chen, X. Zhu, L. Wang, H. Tan, Y. Xu, P. Ding, M. Xiang, G. Hu, Y. Zhou and G. H. Dong, Sci. Total Environ., 2023, 882, 163100.
318. T. Lanot, M. Lavit, P. Gandia and S. El Balkhi, J. Trace Elem. Med. Biol., 2023, 78, 127178.
319. B. Zhang, Q. Yang, X. Wang, B. Jia, L. Cheng and H. L. Zeng, J. Trace Elem. Med. Biol., 2023, 75, 127098.
320. G. Zhang, F. Zhang, W. Liu, C. Liu, J. You, M. Tian, T. Cao, J. Jiang, Z. Yang, H. Wu and W. Wu, BMC Chem., 2023, 17(1), 34.
321. A. Lopez-Botella, R. Sanchez, J. L. Todoli-Torro, M. J. Gomez-Torres, I. Velasco and M. Acien, Toxics, 2023, 11(5), 399.
322. A. P. Sanders, C. Gennings, M. Tamayo-Ortiz, S. Mistry, I. Pantic, M. Martinez, G. Estrada-Gutierrez, A. Espejel-Nunez, L. T. Olascoaga, R. O. Wright, M. M. Tellez-Rojo, M. Arora and C. Austin, Environ. Int., 2022, 166, 107361.
323. Y.-N. Chan, J. T.-S. Lum and K. S.-Y. Leung, Microchem. J., 2023, 188, 108503.
324. Y. Wu, Y. Chao, Y. Miao, Y. Li, T. Xu, S. Li and J. Peng, Anal. Chim. Acta, 2023, 1255, 341054.
325. D. Niebla-Canelo, A. J. Gutierrez-Fernandez, C. Rubio-Armendariz, A. Hardisson, D. Gonzalez-Weller and S. Paz-Montelongo, Foods, 2022, 11(23), 3810.
326. Z. Aydogan, Environ. Monit. Assess., 2023, 195(1), 249.
327. A. Winiarska-Mieczan, K. Jachimowicz, M. Kwiecien, R. Krusinski, S. Kislova, L. Sowinska, Z. Zasadna and D. Yanovych, Biol. Trace Elem. Res., 2023, 201(7), 3567–3581.
328. S. Ulusoy, Mar. Pollut. Bull., 2023, 187, 114518.
329. M. Kotuła, J. Kapusta-Duch and S. Smoleń, Appl. Sci., 2022, 12(15), 7610.
330. Z. Anissa, B. Sofiane, A. Adda and J. Marlie-Landy, Biol. Trace Elem. Res., 2023, 201(11), 5455–5467.
331. V. J. Ferreira, V. A. Lemos and L. S. G. Teixeira, J. Food Compos. Anal., 2023, 117, 105098.
332. G. O. Latunde-Dada, N. Kajarabille, S. Rose, S. M. Arafsha, T. Kose, M. F. Aslam, W. L. Hall and P. A. Sharp, Nutrients, 2023, 15(12), 2732.
333. E. Lozano-Bilbao, I. Delgado-Suarez, S. Paz-Montelongo, A. Hardisson, J. J. Pascual-Fernandez, C. Rubio, D. G. Weller and A. J. Gutierrez, Foods, 2023, 12(7), 1438.
334. Z. W. Ren, M. Yang, B. A. McKenna, X. M. Lian, F. J. Zhao, P. M. Kopittke, E. Lombi and P. Wang, Plant Physiol., 2023, 191(3), 1520–1534.
335. L. Sinkovic, V. Toth, M. Rakszegi and B. Pipan, J. Elem., 2023, 28(1), 27–39.
336. I. Ekin, Istanbul J. Pharm., 2022, 52(2), 179–186.
337. P. J. Gray, Food Addit. Contam.: B Surveill., 2023, 16(2), 79–85.
338. M. Medyk, J. Falandysz and I. C. Nnorom, Environ. Sci. Pollut. Res. Int., 2023, 30(14), 41473–41484.
339. R. Vanderschueren, J. Doevenspeck, L. Goethals, M. Andjelkovic, N. Waegeneers and E. Smolders, Food Chem. Toxicol., 2023, 172, 113599.
340. A. Chalifoux, M. Hadioui, N. Amiri and K. J. Wilkinson, Molecules, 2023, 28(11), 4442.
341. A. Giacomino, P. Inaudi, G. Silletta, A. Diana, S. Bertinetti, E. Gaggero, M. Malandrino, F. Stilo and O. Abollino, Molecules, 2023, 28(1), 153.
342. J. C. Martino, J. Crawford, P. Gadd, J. Sammut, N. Saintilan and D. Mazumder, Food Control, 2023, 151, 109813.
343. C. N. Carneiro, F. J. V. Gomez, A. Spisso, M. F. Silva, J. L. O. Santos and F. S. Dias, Biol. Trace Elem. Res., 2023, 201(9), 4590–4599.
344. X. Han, H. C. Lu, Y. Wang, X. T. Gao, H. Q. Li, M. B. Tian, N. Shi, M. Y. Li, X. L. Yang, F. He, C. Q. Duan and J. Wang, Food Res. Int., 2023, 163, 112165.
345. W. A. Hika, M. Atlabachew and M. Amare, Food Chem.: X, 2023, 17, 100545.
346. S. H. Hur, H. Kim, Y.-K. Kim, J. M. An, J. H. Lee, S. Kim, Y.-H. Choi and H. J. Kim, J. Food Compos. Anal., 2023, 116, 105076.
347. C. Monti, D. Cavanna, I. Rodushkin, A. Monti, A. Leporati and M. Suman, Cereal Chem., 2023, 100(2), 522–531.
348. M. Rapa, M. Ferrante, I. Rodushkin, C. Paulukat and M. E. Conti, Br. Food J., 2023, 125(7), 2644–2662.
349. E. F. de Oliveira Filho, M. Miranda, T. Ferreiro, C. Herrero-Latorre, P. Castro Soares and M. Lopez-Alonso, Molecules, 2022, 27(15), 4938.
350. F. Di Donato, A. Biancolillo, M. Foschi, V. Di Cecco, L. Di Martino and A. A. D'Archivio, J. Food Compos. Anal., 2023, 115, 104948.
351. D. Galvan, J. C. de Andrade, L. Effting, C. A. Lelis, F. L. Melquiades, E. Bona and C. A. Conte-Junior, Food Control, 2023, 143, 109326.
352. R. Mamede, I. A. Duarte, I. Cacador, P. Reis-Santos, R. P. Vasconcelos, C. Gameiro, P. Canada, P. Re, S. E. Tanner, V. F. Fonseca and B. Duarte, Foods, 2022, 11(19), 3081.
353. T. M. Weilert, C. L. Ray, J. A. Gawenis and J. D. Brockman, J. Radioanal. Nucl. Chem., 2022, 331(12), 4971–4981.
354. Z. Huang, R. Bi, S. Musil, A. H. Petursdottir, B. Luo, P. Zhao, X. Tan and Y. Jia, Sci. Total Environ., 2022, 847, 157429.
355. C. Telloli, S. Tagliavini, F. Passarini, S. Salvi and A. Rizzo, Food Chem., 2023, 402, 134247.
356. Y. Y. Liu, S. Y. Zhao, X. Gao, S. Y. Fu, C. Song, Y. P. Dou, S. Z. Song, C. Y. Qi and J. Q. Lin, RSC Adv., 2022, 12(53), 34520–34530.
357. X. Wu, S. Shin, C. Gondhalekar, V. Patsekin, E. Bae, J. P. Robinson and B. Rajwa, Foods, 2023, 12(2), 402.
358. L. Suchecki Barnet, G. Harumi Yamashita, M. J. Anzanello and D. Pozebon, Food Chem., 2023, 402, 134208.
359. S. Valverde, J. A. Tapia, A. Pérez-Sanz, A. V. González-Porto, M. Higes, J. J. Lucena, R. Martín-Hernández and J. Bernal, J. Food Compos. Anal., 2023, 119, 105235.
360. M. N. C. Grainger, H. Klaus, N. Hewitt, H. Gan and A. D. French, Biol. Trace Elem. Res., 2024, 202(2), 754–764.
361. S. Hye Hur, H. Kim, Y. K. Kim, J. M. An, J. Hye Lee and H. Jin Kim, Food Chem., 2023, 423, 136235.
362. Y. M. Lang, Y. Cheng, H. R. Yang, Q. Z. Chen, Q. Liu, X. T. Wang, P. S. Bian and X. X. Yang, Qual. Assur. Saf. Crops Foods, 2022, 14(4), 178–187.
363. A. P. Louppis, M. S. Constantinou, M. G. Kontominas, F. Blando and G. Stamatakos, J. Food Compos. Anal., 2023, 119, 105219.
364. R. Markiewicz-Zukowska, A. Puscion-Jakubik, M. Grabia, J. Perkowski, P. Nowakowski, J. Bielecka, J. Soroczynska, G. Kangowski, J. M. Boltryk and K. Socha, Foods, 2022, 11(20), 3152.
365. F. Vezzulli, M. C. Fontanella, M. Lambri and G. M. Beone, J. Sci. Food Agric., 2023, 103(9), 4303–4316.
366. M. Reyrolle, G. Bareille, E. N. Epova, J. Barre, S. Berail, T. Pigot, V. Desauziers, L. Gautier and M. Le Bechec, Food Chem., 2023, 423, 136271.
367. L. P. Puig, M. C. Boque, A. V. Ferrer, L. Fernandez-Ruano, J. L. L. Blasco and M. A. Cladera, Food Chem., 2023, 428, 136767.
368. G. Cristea, A. Dehelean, R. Puscas, A. R. Hategan and D. A. Magdas, Molecules, 2023, 28(2), 503.

---