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 preconcentration

• 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
  • 5.1 NPs as reagents
  • 5.2 Metabolism of NPs and use as biological tracers
  • 5.3 Measurement of NPs

• 6 Applications: clinical and biological materials
  • 6.1 Metallomics
  • 6.2 Imaging with MS and X-Rays
  • 6.3 Trace elements in pregnancy
  • 6.4 Elements as tags for indirect determinations
  • 6.5 Multielement applications
  • 6.6. Progress for individual elements

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

• 8 Applications: food and beverages
  • 8.1 Progress for individual elements
  • 8.2 Single and multielement applications in food and beverages

• Abbreviations

• Conflicts of interest

• References

---

Abstract

This update covers publications from the second half of 2019 to the middle of 2020. Techniques and applications relevant to clinical and biological materials, foods and beverages are discussed in the text, presenting key aspects of the work referenced, while the tables provide a summary of the publications considered. A new table has been introduced to include the large number of sample preparation papers that were seen. Reference intervals for quite specific groups and environmental settings were reported, reflecting focussed concerns rather than large general populations. Technical developments include applications of spICP-MS and ICP-QQQMS, particularly to NPs and elemental speciation. Measuring elements as tags for indirect clinical determinations is featured in this ASU as the number of such papers is gradually increasing. As in previous reviews, work featuring cancer and Alzheimer’s disease continues, but obesity, Cu metabolism and absorption of metals through skin are of current interest among the clinical applications of atomic spectroscopy. As meat-avoidance increases in many populations the elemental composition of vegetarian and vegan foods is attracting more attention. A study of As in rice-based infant formulas and foods reported levels above the EU limit in all samples, with much of it as iAs. On a similar theme, concentrations of elements in milk from yak and camel were measured alongside the less exotic species, cow, goat and buffalo. Elemental analysis of wines to discriminate authenticity is regularly included in these updates but this approach is now applied to saffron and coconut sugar, as foods which attract commercial interest.

---

1 Reviews

This latest update adds to that from last year¹ and complements other reviews of analytical techniques in the series of Atomic Spectrometry Updates from the previous year.^2–6

The scope of reviews addressing instrumentation is broader than has been noted in our recent updates. Curiously, none of interest features ICP-MS. Two refer to SIMS, focusing on quite different interests. The contribution from Gyngard et al.⁷ considered how isotope ratio measurements, coupled with achievable resolution of less than 50 nm, has enabled users of nanoscale SIMS to describe quantitative images of sub-organelle surfaces in different types of samples. Although not strictly elemental, TOF-SIMS also provides useful insight into cellular imaging. A tutorial review by Massonnet and Heeren⁸ focuses on the choice of beam and signal-enhancement methods that are best suited to the particular surface being investigated, in order to provide ideal images. Other technical approaches to provide images of tissue samples were recently reviewed and of particular interest are LIBS and X-ray techniques. The excellent review by Shah et al.⁹ presented a discussion of LIBS instrumentation, sensitivity, accuracy and sample preparation compared with other techniques. They pointed out advantages such as minimal sample preparation and speed of analysis but also poorer LODs compared to alternative methods. The review by Wang et al.¹⁰ is directed to biological soft tissues and how LIBS may be a valuable tool for cancer diagnosis and laser surgery. Matsuyama and colleagues¹¹ presented the history of X-ray imaging over several decades, leading on to submicron X-ray probes to give images of cellular organelles by SR-XRF spectrometry and XRD. Two further reviews of instrumental developments are those of Filatova et al.¹² and Martinez and Baudelet.¹³ For a number of years our ASUs have included one or two articles in which novel applications of HR-CS-ETAAS were reported. Filatova et al.¹² discuss the special advantages of the technique i.e. multi-element analysis, including analytes (atoms and molecules) with absorption outside the 190–900 nm range. The review of Martinez and Baudelet¹³ discusses calibration strategies for the analysis of biological samples by LIBS and LA-ICP-MS. They suggest that there is an increase in the availability of suitable RMs for this purpose, which will improve quantification. In a more focused review, Feng et al.¹⁴ describe the advantages of XRF spectrometry for analysis of foods: simple non-destructive preparation, high spatial resolution, multi-elemental measurements, and suggest that with recent developments, use of the technique is likely to increase.

Crucial to any analytical technique is the preparation of samples and, in the context of biological materials and food, analyte enrichment is a major component of many procedures. Our ASUs regularly discuss new approaches, and three recent reviews provide valuable overviews of different topics. Hashemi and Rezania¹⁵ summarise extraction of heavy metals with the aid of carbonaceous-based sorbents while Er et al.¹⁶ restrict their discussion to magnetic NPs as sorbents for clean-up and pre-concentration of elements in biological samples and foods. Locateni et al.¹⁷ provide a comprehensive review of the many varieties of on-line techniques, liquid–liquid, solid-phase, DLLME etc., comparing the features of each and the likely future developments.

With concern around food safety and limits for the concentrations of toxic elements, we have recognised increasing numbers of publications in recent years, relating to the procedures for analysis of foods. The elements that have attracted the most attention in recent years are As and Hg. In this review period, As, particularly the iAs species, features in many publications and reviews. Luvonga et al.¹⁸ suggest that diversity of metabolism in humans, as well as insufficient knowledge of the toxicology of all the As compounds and their concentrations in food types, inhibits setting regulatory limits for As in foods. Some would argue that there is sufficient knowledge concerning the toxicity of iAs and the other major As species and that regulatory limits are appropriate. In our last ASU¹ we described a method developed by four expert laboratories, to measure iAs in fish tissue with sufficient sensitivity to detect concentrations below the anticipated EU maximum allowable level. Nevertheless, it is necessary to be assured of safe concentrations in other foods and for levels to be monitored. The reviews by Luvonga et al.¹⁸ and Camurati and Salamone¹⁹ cover similar material as they discuss methods and legislation for As in marine species and seaweed, respectively. Zhang et al.²⁰ reviewed published work for As compounds in Chinese edible mushrooms. China is a ‘high-As’ region and is the major contributor to the world production of mushrooms. This comprehensive review summarises total As concentrations reported from different regions and species of mushroom, and compares values with those from other countries. Reports of As speciation show that the major species varies but, where determined, appears to be usually AB or iAs. The review also considers uptake from soils, changes occurring during processing and bioaccessibility. Apart from the work with As, the other element that featured during this review period is Cr and an extensive account of the analytical methodologies for Cr speciation in foods and other samples was prepared by Milačič and Ščančar.²¹ Alongside analysis of foods, safety can be linked to geographic origin, which can require demonstration of authenticity, a topic featured in a later section of this update. A systematic review of geographical authentication of origin was prepared by Katerinopoulou et al.²² These authors considered the techniques for analysis with special consideration given to multiple isotope ratio measurements and results of authentication studies published during 2015–2019.

Three reviews addressing different topics that appear in later sections of this update relate to advances in single-cell ultra-trace analysis,²³ asymmetrical flow field flow coupled to fractionation ICP-MS²⁴ and consideration of blood Ti concentrations as a marker for orthopaedic implant wear.²⁵

2 Metrology, interlaboratory studies, reference materials and reference ranges

The conclusions based on analytical work, be it for research purposes or for application in medicine, nutritional studies or food production and control, as covered by this review, rely on documented evidence, continuous monitoring and improvement in the quality of the measurement results. To this aim, both the development of reference methods, based on collaborative studies according to stated protocols, and of well-characterised RMs, are essential tools to assure measurement results are comparable across space and time, as they provide the segments of traceability chains linking individual results to stated references, such as the International System of Units.

In this year’s review, an interlaboratory study was reported by Cruijsen et al.²⁶ which aimed to extend the applicability of the previously assessed AOAC 2011.14 method beyond the tested matrices (chocolate milk powder, dietetic milk powder, infant cereal, peanut butter and wheat gluten). This collaborative exercise involved 14 laboratories, from 11 countries. The participants determined nine elements (Ca, Cu, Fe, K, Mg, Mn, Na, P and Zn), by ICP-AES, after microwave-aided digestion of the samples, in 24 items of milk, milk products, infant formula and adult nutritional products, as well as in the NIST SRM 1849a, a milk-based, hybrid infant/adult nutritional powder. Recovery of the certified values (mass fractions ± the 95% expanded uncertainty: Ca, 5253 ± 51 mg kg⁻¹; Cu, 19.78 ± 0.26 mg kg⁻¹; Fe, 175.6 ± 2.9 mg kg⁻¹; K, 9220 ± 110 mg kg⁻¹; Mg, 1648 ± 36 mg kg⁻¹; Mn, 45.59 ± 0.97 mg kg⁻¹; Na, 4265 ± 83 mg kg⁻¹; P, 3990 ± 140 mg kg⁻¹ and Zn, 151.0 ± 5.6 mg kg⁻¹) ranged from 98% to 101% across all laboratories. Precision, in terms of repeatability (within-day) and reproducibility (inter-laboratory), was assessed on the tested matrices, consisting of 18 nutritional products, including powders, ready-to-feed liquids, and liquid concentrates (of which 13 were fortified with Cu, Fe, Mn or Zn), and six dairy products, as liquids, powders, butter and processed cheese. For major elements, i.e. Ca, K, Mg, Na and P, both parameters satisfied the AOAC requirements (RSD_r = 5%, RSD_R = 10%) as stated in the AOAC document SMPR 2014.004. However, for the trace elements Cu, Fe, Mn and Zn, the achievable precision was strongly influenced by the concentration level and the type of matrix. Whereas, for the fortified nutritional product samples, the RSD_r was acceptable, and so was the RSD_R (except for Mn, for which the RSD_R was 12.2%), at the lower levels present in the non fortified nutritional products, both the RSD_r and RSD_R for Cu, Fe, Mn and Zn exceeded the requirements, varying, respectively, between 5.4% and 29.4% and between 13.2% and 82.8%. In the six dairy product samples, RSD_r values were acceptable for Zn, but ranged from 9.4% to 24.6% for Cu, Fe and the higher levels of Mn, whereas the RSD_R for all trace elements varied between 11.4% and 55.0%. Another group of researchers²⁷ explored the potential of ICP-MS to provide reference measurements for Na in human serum samples, after dilution 1 + 99 with 0.3% HNO₃. Sodium is routinely analysed in clinical laboratories, mainly using direct or indirect potentiometric techniques, however, matrix differences in the samples and/or inherent variability between methods and instrumentation from different manufacturers are a source of errors in everyday practice. Reference methods based on independent principles provide the means to highlight potential biases and standardize methods for routine use. The authors used a collision cell with a 3 mL min⁻¹ flow of He to overcome residual spectral interferences on either Na or Ge, that was chosen as the internal standard, as well as the method of standard additions for calibration. Under these conditions, the within-run and between-run precision (RSD), determined on three QC serum samples, at concentration levels of 115.1 mmol L⁻¹, 132.7 mmol L⁻¹ and 170.0 mmol L⁻¹, were less than 0.6%. The analysis of the NIST SRM 956d “Electrolytes in frozen human serum” at the following three levels of concentrations (±expanded uncertainty at 95% confidence level): 120.0 ± 0.7 mmol L⁻¹, 139.3 ± 0.9 mmol L⁻¹ and 158.7 ± 1.1 mmol L⁻¹, yielded values within the stated expanded uncertainties of the certified values. A comparison with the results obtained by the indirect ion selective electrode (ISE) method on the same samples in External Quality Assessment Schemes specifically designed for Reference Laboratories in Laboratory Medicine showed a correlation coefficient of 0.9914 (n = 40) and the mean difference between the results obtained with the two methods (ICP-MS vs. ISE) was −0.15%.

The availability of CRMs matching sample composition and concentration levels also depends on the development of analytical methods of higher accuracy, to allow the certification of CRMs with low uncertainty levels fit for the purpose of calibration and method validation. Among several areas of concern, the accurate measurement of potentially toxic elements in marine species is necessary for the purpose of both assuring food safety and monitoring pollution of the marine environment. Garcia and co-workers²⁸ developed and validated an analytical procedure for ID-SF-ICP-MS, to determine the total mass fractions of five trace elements (Cd, Cu, Hg, Pb, and Zn) and MeHg in a candidate CRM (IAEA-476) based on fish homogenate. To evaluate the measurement uncertainty of the results, they used the approach of mathematical modelling, taking into account all factors that may have an influence on the final result, in particular during the measurement step, such as isotopic equilibrium, instrumental background, spectral interferences, dead time, mass discrimination effects, and the repeatability of the measured isotope ratios. The analytical procedure was successfully applied in the certification campaign for the CRM IAEA 476, achieving excellent agreement with the certified values (±expanded uncertainty at the 95% confidence level) for, respectively: Cd, 21.2 mg kg⁻¹ ± 1.4 mg kg⁻¹; Cu, 28.0 10⁻³ mg kg⁻¹ ± 2.7 10⁻³ mg kg⁻¹; Hg, 0.578 mg kg⁻¹ ± 0.022 mg kg⁻¹; MeHg (as Hg), 0.523 mg kg⁻¹ ± 0.030 mg kg⁻¹; Pb, 0.644 mg kg⁻¹ ± 0.053 mg kg⁻¹; and Zn 53.6 mg kg⁻¹ ± 2.9 mg kg⁻¹. Both the marketing and the consumption of food supplements among the general population, whether justified or not, are constantly increasing. Although the amount of Cr present in a balanced diet already fulfils the nutritional needs of most, supplementation with this element has been suggested as beneficial in the prevention of a number of pathological conditions, including diabetes and CVD, as well as to help weight loss. As a consequence, Cr is often included in food supplements, and Cr picolinate, that shows better absorption and intracellular uptake than Cr^III, is the species commonly added. Chromium(VI), on the contrary, is a toxicant, with potential mutagenic and carcinogenic effects, although likely to be rapidly converted to Cr^III in the acidic environment of the stomach. To support the assessment of the marketed products, the NRCC has developed two multivitamin and mineral supplement candidate reference materials (VITA-1 and VITB-1). Mihai et al.²⁹ reported the steps undertaken to contribute to their certification for the content of Cr^VI and Cr picolinate, using HPLC-ICP-QQQMS. After crushing the supplement tablets to a fine powder, 50 mL of deionized water were added to 0.25 g of the powder and adjusted to pH 10 with NH₄OH in a plastic tube. After shaking (1 min) and centrifuging at 4000 rpm (10 min), the supernatant was filtered (0.2 μm, PTFE). An aliquot of 100 μL was analysed to measure Cr^VI, using a Hamilton PRP-X100 strong anion-exchange column and isocratic elution with 50 mmol L⁻¹ NH₄NO₃ at a flow rate of 1 mL min⁻¹, with an elution time of 8.2 min. For Cr picolinate determination, 0.25 g of the crushed powder were transferred to a 150 mL HDPE bottle with 100 mL of a 60 + 40 v/v solution of ACN–H₂O, sonicated for 40 min at 25 °C, then centrifuged at 4000 rpm (10 min) and the supernatant filtered (0.2 μm, PTFE). A 20 μL aliquot was loaded on an Agilent Eclipse XDB-C18 RP column and subsequently eluted with a 60 + 40 v/v ACN–H₂O solution as the mobile phase, within 7 min. The ICP-MS was equipped with a collision cell, using O₂ at a 30% and 40% flow, respectively, for Cr^VI and Cr picolinate determination. In both cases, quantification was achieved using the standard additions method. Recovery of spiked amounts was 99.4% ± 4.9% for Cr^VI and 97.7% ± 3.2% for Cr picolinate. The LODs for the multivitamin tablets were 0.13 μg g⁻¹ (Cr^VI) and 1.26 μg g⁻¹ (Cr picolinate), respectively.

Recent reviews have included papers reporting cross-sectional studies of reference ranges in various population groups. A notable publication by Baudry et al.³⁰ investigated the age-dependent changes of six trace elements in a healthy elderly German population (n = 219). Serum samples were collected from study participants at baseline (median age 58.32 years) and after 20 years (median age 77.60 years) and trace element concentrations determined via ICP-MS. A Wilcoxon sign-ranked test was used to evaluate the trace element profiles and the concentration ranges were reported using the median and interquartile range, due to non-normal distributions. Results showed decreases in Mn, Se and Zn and increases in Cu, Fe and I with age. Principal component analysis identified associations between changes in Mn and Zn and between increase in Cu and I, however these were described as mild correlations. Various sociodemographic, anthropomorphic and lifestyle factors were investigated but none were identified as strong determinants of trace element longitudinal variability except the baseline concentration and the use of dietary supplementation. The authors accept limitations in the sample size and the use of serum concentrations as markers of trace element status. As the study focussed on healthy subjects no inference can be drawn on the effect of disease on trace element status.

In this year’s review, six papers evaluated levels of selected elements in population groups in relation to environmental exposure. Judith et al.³¹ investigated U exposure in a nationwide Swiss population study (n = 1393) stratified for gender and age. The 24 h urinary U excretion was measured using HR-ICP-MS with ID calibration (²³⁸U : ²³³U). The median and 95^th percentile were 15 ng 24 h⁻¹ and 67 ng 24 h⁻¹ (interquartile range 7 to 31 ng 24 h⁻¹) and the authors concluded that a health risk from U exposure for the general adult population is unlikely. Gender was found to have a significant effect on U excretion, however there was no association with age, BMI, smoking status or other kidney related variables. The effect of albumin on U excretion was strongly dependent on the presence of diabetes mellitus. In work by Son et al.,³² age, gender, alcohol consumption, herbal medication usage, residential area, drinking water and occupation were identified as potential factors influencing blood Pb concentrations in an adult South Korean population (n = 6455). Blood Hg levels were positively correlated with serum Se levels and dietary fish intake in work by Ballesteros et al.³³ investigating a cohort of pregnant women in Spain (n = 141). Concentrations of total Hg were above the recommended limit set by the US EPA (6.4 μg L⁻¹) in 12% of participants and 31% reported levels above the limit of 3.5 μg L⁻¹, suggested by researchers from the Centre for Health Outcomes in Hawaii and the University of Hawaii.³⁴ Amid concerns regarding exposure of patients to Gd-based contrast agents (GBCA), Layne et al.³⁵ have proposed reference ranges for Gd in whole blood (<0.050 nmol L⁻¹), plasma (<0.057 nmol L⁻¹) and spot urine (<0.025 nmol mmol⁻¹ creatinine) from a group of 120 healthy volunteers with no previous exposure.

Two studies focussed on the exposure of children to non-essential metals. Soler-Blasco et al.³⁶ used AAS to evaluate Hg levels in hair samples from Spanish children, participating in the INMA (Environment and Childhood) birth cohort study (n = 405), at 4 and 9 years of age. Levels of total Hg were above or equivalent to the provisional tolerable weekly intake proposed by the World Health Organisation in 12% of the children studied and were found to be influenced by maternal BMI, smoking status, paternal employment and fish intake. The effect of different racial diets on levels of Cd, Hg and Pb were investigated by Gump et al.³⁷ in a small study of children (n = 295), from 9 to 11 years old. Whole blood levels of the metals were determined via ICP-MS and diet was assessed by 24 h dietary recall phone interviews. Although racial differences in blood levels of Pb, Hg and Cd could be accounted for by corresponding differences in diet, it was acknowledged that further study is required to confirm this due to regional differences in food consumption patterns.

3 Sample collection and preparation

3.1 Collection, storage and preliminary preparation

Timing of sample collection is important where there is known to be a diurnal, or similar, variation of analyte concentration or an acute response to exposure such as foods or drugs. These variations usually span a relatively short time interval, but a few studies have identified examples that reflect much longer patterns of change. In a possibly unique study, Sera et al.³⁸ reported the measurement of Al, As, Cr, Hg and Pb in beard samples collected every day from a single individual, for 12 years. The authors reported that their results, obtained by analysis using PIXE, showed a slow increase in the concentration of Hg with age, consistent with previous observations. They also found that, except for Cr, concentrations decreased from summer to winter. This annual seasonal variation was particularly marked for As and Hg, with winter levels at 50% and 86% of those in the summer, and was attributed to dietary changes during the year.

Avoidance of contamination is crucial to most investigations involving measurement of very low concentrations. Testing the equipment used to collect and store samples for trace element studies is recommended and widely practised. Many sources suggest that urine sample should be collected into acid-washed containers and, while experience has shown that this is probably unnecessary when the sample is to be analysed for Cu, it has never been formally documented. Robson et al.³⁹ therefore reported the investigations that they undertook which, using ICP-MS, confirmed no obvious contamination from non-acid washed containers. Hair samples are exposed to external sources of contamination before being collected. Washing, to provide a sample that is representative of the endogenous composition of the hair, is part of the preparation prior to analysis. Numerous protocols have been used, as has been noted in previous updates and Verrey and her co-workers⁴⁰ report on a three-step process using Triton, HNO₃ and then HCl, followed by acid digestion and analysis by ICP-MS. It is claimed that this approach is more effective than methods based on using surfactants and organic solvents. However, it should be remembered that over-aggressive washing may not only remove surface contamination, but also extract elements from within the hair and simple comparisons may be misleading.⁴¹ Containers for foods and drinks represent sources of contamination, especially where there is long-term storage. Safta et al.⁴² considered manufacturing defects in metal cans or the heat treatment operation applied during the canning process, and possible contamination into the contents. The cans contained either simulants, tomatoes or sardines and were stored for 18 months (simulants) or 36 months (foods). Concentrations of Al, Cd, Cu, Fe, Pb, Sn and Zn were measured using ICP-AES and the inner surface of the cans examined by SEM/EDXRD. High migration levels of Fe and Zn were found in cans filled with acidic food simulant or tomato paste or sardines in tomato sauce, coming from corrosion of the inner sides of the cans and the uncoated junction seals. Cd was only detected in cans filled with food. Low migration levels of Al, Cu, Pb and Sn were recorded.

Changes to concentrations during storage can involve losses as well as contamination. It has previously been shown that urinary SeSug1 is important for monitoring occupational exposure to Se, but that there is rapid degradation to methylselenic acid, even at −20 °C. In a series of experiments, Hildebrand and Goen⁴³ investigated the use of a bactericide (NaN₃), with and without adjustments to pH, to stabilise the SeSug1. Samples were stored at −20 °C and the selenosugar1 and methylselenic acid determined by LC-ICP-MS at intervals for up to 156 days. Stability was found to be possible with pH at 5.5 but to maintain this, 0.5% NaN₃ is also necessary to prevent changes due to bacterial activity.

3.2 Digestion, extraction and preconcentration

A microwave-assisted digestion method with dilute HNO₃ was developed⁴⁴ for the determination of Al, Ca, Cr, Cu, Fe, K, Mn, Mo and Ni in rice by ICP-OES. A full factorial design indicated that the optimum conditions (based on the residual carbon content as the figure of merit) were (for a 500 mg sample) digestion time 14 min, 5 mL of 1 mol L⁻¹ HNO₃ and 2.5 mL of 30% H₂O₂. The carbon content was measured by a spectrophotometric procedure, analogous to the determination of chemical oxygen demand by dichromate, but the results were expressed as a percentage efficiency of digestion (>89% was obtained) rather than as mg L⁻¹ of carbon, so comparison with other methods is difficult. It is not clear why the researchers did not determine carbon by AES. No information was provided about whether filtration or centrifugation was needed, but it seems likely that this would be the case, especially as the procedure proved unsuitable for the determination of As, for which low results were obtained. The method was validated by the analysis of NIST SRM 1568a (rice flour) and applied to the analysis of 33 locally purchased rice samples. Prior to analysis, samples (50 g) were ground for 1 min in a household food processor with a stainless-steel blade, sieved (“household sieve, thin type, with variable openings between 500 and 1000 μm”) and dried (65 °C, 72 h) to constant weight. No information was provided as to whether material was retained by the sieve. The researchers indicated that the low drying temperature was to avoid loss of As by volatilization, but did not cite a source to support this claim. They speculated that the low results obtained might be “loss by volatilization due to pressure release during digestion.” For some samples, they found the concentrations of fortified Ca and Fe to be much lower than those indicated on the package, which was interpreted as the result of poor production quality control. The LODs ranged from 0.009 mg kg⁻¹ (Mn) to 2 mg kg⁻¹ (Ca).

Several research groups have reported on a comparison of sample preparation procedures. For the determination of the mineral content of infant formula by ICP-OES,⁴⁵ ultrasound-assisted extraction was selected as the most suitable, as it was simple, fast and low-cost in comparison with the other procedures (microwave-assisted digestion and dry ashing). Sample (2 g) was extracted with 2 mL HNO₃ and 1 mL H₂O₂ in a capped 50 mL tube in an ultrasonic bath for 30 min at 60 °C. Following cooling, the solution was made up to 25 mL with H₂O and filtered (0.45 μm). The method was validated by the analysis of NIST SRM 1849a (infant/adult nutritional formula). LODs for Ca, Cu, Fe, K, Mg, Mn, Na, P and Zn ranged from 0.0001 to 0.03 mg per 100 mL. The researchers also reported on some in vitro studies of the bioaccessibilities of these elements. In a study of As speciation in mushrooms, several sample pre-treatment procedures were evaluated.⁴⁶ It was concluded that microwave-assisted extraction (with H₂O) was superior to titanium probe ultrasound-assisted extraction (also with H₂O). Bioaccessibilities were also evaluated by a standard enzyme-assisted procedure, described as a “modified unified BARGE method.” Quantification of As species in the extracts was performed by HPLC (AEC)-ICP-DRC-MS and extracts obtained after enzyme-assisted extraction were analysed by SEC with UV-Vis and ICP-DRC-MS detection. The researchers found As^III, As^V, AB, DMA, and MMA in a very broad range of concentrations that depended on the place of their growth, and also identified protein-bound As (via monitoring of SO⁺ at m/z 50) in all mushrooms, as well as at least one unknown As compound in Boletus edulis. The procedure for total As, determined by ICP-MS (with O₂ reaction gas) following microwave-assisted digestion (quartz vessels and 200 °C) with HNO₃–H₂O₂, was validated by the analysis of the Nuclear Chemistry and Technology Institute (Warsaw, Poland) Control Material CS-M-3 (microelements in mushroom powder). Several sample preparation methods were evaluated²⁴ for the determination of SeMet in selenized yeast by asymmetric flow FFF ICP-MS. The seven procedures were extraction with methanesulfonic acid (with and without reflux step), room temperature extraction with HCOOH, alkaline extraction with SDS, and H₂O extraction (both manual and ultrasound-assisted). The researchers found that alkaline extraction with SDS gave the highest accuracy for the determination of SeMet in a CRM, NRC SELM-1 (selenized yeast), since it minimized hydrolysis of the protein peptide bonds optimally required for the FFF separation. The effect of filtration was examined but found not to have a significant effect on the SDS results. The performances of three regenerated cellulose FFF membranes (5, 10 and 500 kDa) were compared and no significant differences between the results obtained with 5 and 10 kDa membranes were observed, whereas results for the 500 kDa membrane were significantly low. The researchers showed that instrument response was independent of the calibration compound (either iSe or SeMet) and obtained a LOD of 0.5 μg L⁻¹ (for SeMet as Se) with monitoring at m/z 78. It was concluded that this kind of FFF for speciation purposes is a viable alternative to LC, allowing sufficient interaction between the analyte and the field to achieve separation but without the protein degradation commonly observed in traditional chromatography.

Arsenic species were rapidly and efficiently extracted⁴⁷ from powdered herbs by a dual-frequency ultrasonic enzymatic procedure. Samples (0.3 g) and cellulase in 10 mL Tris–HCl buffer (pH 8.2) were irradiated simultaneously at 40 kHz in an ultrasonic bath and at 20 kHz by insertion of an ultrasonic probe at 30 °C. Following centrifuging and filtering, 5 mL of extract was diluted to 10 mL with mobile phase. A 95% extraction efficiency was achieved in just 6 minutes. Inorganic As species were separated by AEC (Hamilton PRP X-100 column, isocratic elution with phosphate buffer at pH 8.2) and detected by AFS following post-column HG. The connection featured two gas–liquid separators. There were no significant differences between measured and certified values for BMEMC CRMs GBW(E)090066 (Salvia) and GBW(E)090067 (Paeoniae Radix Rubra). The LODs were 0.7 and 2.5 μg kg⁻¹ for As^III and As^V, respectively.

In the ultrasound-assisted extraction of As species from freeze-dried samples problems were encountered.⁴⁸ For H₂O in an ultrasonic bath (30 min, 35 W, 44 kHz, 20 °C to ca. 40 °C during sonication), efficiencies, ranging from 29% to 106%, were low for most samples. The addition of up to 90% MeOH did not improve the efficiency and caused problems with enhancements of the responses for several species. The determination of 11 species (As^III, As^V, DMA^V, MMA^V, AB, AC, TMA, TMAO, dimethyldithioarsinic acid, monomethylmonothioarsinic acid, and trimethylarsine sulfide) in fish, crustacean, mussel, cephalopod and mushroom by HPLC-ICP-MS was reported. The separation was achieved by RP ion-pair chromatography on a C8 stationary phase with a mobile phase containing 1.2 mmol L⁻¹ TMAH, 4 mmol L⁻¹ malonic acid, 6.2 mmol L⁻¹ sodium butane-1-sulfonate, and 0.05% (v/v) MeOH. The LODs ranged from 1.4 ng g⁻¹ to 4.0 ng g⁻¹. Method validation involved analysis of CRM ERM BC211 (rice). Thio-arsenic species dimethyldithioarsinic acid was found in mussel (Mytilus chilensis) and cuttlefish (Sepia pharaonis) samples at very low concentrations.

A large number of methods incorporating pre-concentration by either LLE or SPE were published in the recent review period, many of which were motivated by the need to decrease the LOD of a procedure in which analytes were quantified by FAAS. These and related studies are summarized in Table 1. Much of the work described in these articles relates to finding the optimum conditions, usually meaning those that produce the highest sensitivity, which were assumed to also produce the lowest LOD, and so the relevant parameters were varied according to some optimization strategy (often just a single cycle of the alternating-variable search). However, it should not be overlooked that these pre-concentration procedures also separate the analytes from potentially interfering matrix components. In a review (168 references) of carbon-based sorbents and their nanocomposites for the pre-concentration of potentially toxic metal ions,¹⁵ numerous applications to the analysis of clinical and biological materials, foods and beverages were cited, though there were also many applications to the analysis of environmental waters. The reviewers concluded that carbonaceous sorbents and their composites have several outstanding features including chemical and thermal stability, large sorption capacity, the possibility of functionalization, and large surface area. Among the various materials, graphene and carbon nanotubes are increasingly popular due to the ease of preparation and functionalization.

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

Analytes	Matrix	Technique	Procedure (LLE) or substrate (SPE)	Reagents (LLE) or substrate coating (SPE)	LOD in μg L^−1 (unless stated otherwise)	Validation	Ref.
As, Cd, Pb	Chocolate, mussels, red wine, rice	ICP-OES	DLLME. A supramolecular solvent was generated by combining THF with either 1-decanol, or 1-undecanol. After separation, the solution was diluted with acetic acid to decrease the viscosity	APDC	2.4 (As) 0.6 (Cd) 2 (Pb)	Spike recovery	49
As	Rice	HG-AAS	LLE of ion-pair complexes into a DES (proline–malic acid mixture (1 + 1 mol ratio)) followed by dispersion with THF. Following phase separation (3500 rpm for 3 min), the aqueous phase was decanted and the remaining phase was diluted with acidic ethanol	7-(Diethylamino)-4-hydroxy-3-oxophenoxazin-10-ium-1-carboxamidechloride (celestin blue, CB)	2 and 3 ng L^−1, respectively	NIST SRM 1573a (tomato leaves) and SRM 1568a (rice flour) and spike recovery	171
Cd	Chocolate, mussels, red wine, rice	ICP-OES	See As ref. 49				49
Cd	Water, tap and sparkling	MIP-OES	DLLME into (a) CHCl3 (extractant) and EtOH (disperser) or (b) 1-decanol (extractant) and THF (self-assembly agent and disperser solvent)	DDTP	1	Spike recovery	74
Cd	Water (drinking fountain)	FAAS	SPE granular activated carbon and eluted with 5 mL of 1 M HNO3	Quercetin (a polyphenol derived from plants). All three analytes were found in all 22 samples	0.8 (Cd), 8 (Ni), 7 (Pb)	NIST SRM-1643e. (Trace elements in water) and spike recoveries	186
Cd	Rice, wheat, tea	ETAAS	SPE metal–organic framework composite	Chitosan/thiol functionalised	0.03 (Pb) and 0.008 (Cd)	CRM GBWE080684 (rice), GBW10011 (wheat), and GBW10016 (tea)	187
Co	Mussel, human hair	ICP-MS	Dispersive micro-SPE with fibrous TiO2@g-C3N4 nanocomposites (graphitic carbon nitrides)	The surface functional groups, such as –NH2, –NH– and =N–, can selectively adsorb the target ions	0.1 pg mL^−1 (Co) and 1 pg mL^−1 (Ni)	CRM GBW 07308 (stream sediment) GBW 07601 (human hair) and GBW 08571 (mussel)	188
Co	Food, tap water, river water, mineral water, tuna fish, tomato, potato, macaroni, cabbage, red lentil, black and green teas, mint, baby rice powder, biscuit, coffee, bovine liver, gluten-free biscuit, meat, beef and soil	ICP-OES	SPE Amberlite XAD-4	Anoxybacillus kestanboliensis	0.04 (Co) and 0.06 (Hg)	CRM NCSZC 73014 (tea leaves), NCS DC78301 (river sediment) and DORM2 (dogfish muscle)	189
Cr^III Cr^VI	Food, rice, sausage, tap water	FAAS	LLME choline chloride–phenylethanol DES with air-assisted emulsification	1-(2-Pyridylazo)-2-naphthol	0.4	Spike recovery	179
Cr	Honey	ETAAS	DLLME of Cr^III as [CrCl2(H2O)4]Cl into a magnetic ionic liquid (MIL) trihexyl(tetradecyl)phosphonium ([P6,6,6,14]FeCl4) with acetonitrile as dispersant	No chelating agent needed. Ash dissolved in 1.0 mol L^−1 HCl	5 ng L^−1	Spike recovery	190
Cu	Grapes, apples, tea, water, and rice	FAAS	Thin film microextraction into poly m-phenylenediamine/carbon nanotube electrospun nanofiber	No chelating agent needed. The sample ash was dissolved in HNO3 + H2O2 and the pH adjusted to 6.5	0.3	206 BG 326 (ore polymetallic gold Zidarovo-PMZrZ) and spike recoveries	191
Cu	Tea, dietary supplements	FAAS	SPME on magnetic graphene oxide	Pyrocatechol violet	4.0	CRM NCS ZC73032 (celery), CRM025-050 (metals in soil), TMDA-64.2 (fortified water) and spike recoveries	192
Cu	Edible oils (sunflower, olive, flaxseed, corn) and baby oil	FAAS	LLME with triethylamine was the extraction solvent, with 7.5 mol L^−1 HNO3 as the hydrophilicity-switching trigger	No chelating agent needed	7	Spike recoveries	193
Cu	Urine, sweat, dialysis solution and water samples	FAAS	LLE into a supramolecular solvent (THF and decanoic acid)	1-(2-Pyridylazo)-2-naphthol was used as a chelating agent in an ultrasound-assisted procedure. The concentrations in all samples were below the LOD	7	TMDA 53.3 (fortified water) and TMDA 64.2 (fortified water) and spike recoveries	67
Cu^I	Biological serum and cell homogenate, cell membrane components	ETAAS	LLE into n-pentanol	Serum and cell homogenate, cell membrane components were deproteinized with trichloroacetic acid, then adjusted to pH 9 and 0.05% 2,2′-biquinoline in N-pentanol added. The organic layer was evaporated, and the residue dissolved in HNO3 + H2O2	0.04	Spike recovery	194
Fe	Hair	FAAS with slotted quartz tube atomiser	Switchable solvent LLE into N,N-dimethlybenzylamide (DMBA) and ultrapure H2O with the addition of dry ice for protonation, and NaOH for deprotonation	A Schiff base ligand (E)-2,4-dibromo-6-(((3-hydroxyphenyl)imino)methyl)phenol was synthesised	3	Spike recoveries (only one sample with Fe concentration below the LOD)	64
Hg	Food, tap water, river water, mineral water, tuna fish, tomato, potato, macaroni, cabbage, red lentil, black and green teas, mint, baby rice powder, biscuit, coffee, bovine liver, gluten-free biscuit, meat, beef and soil	ICP-OES	See Co ref. 189				189
Mn	Coffee and wastewater	FAAS	LLME with DES with a choline chloride–phenol mixture as the extraction solvent, and THF as the emulsifier	3-[[(2-Hydroxyphenyl)imino]methyl]-2-naphthalenol	0.5	Spike recoveries	68
Ni	Water (drinking fountain)	FAAS	See Cd ref. 186				186
Ni	Mussel, human hair	ICP-MS	See Co ref. 188				188
Ni	Crab, oyster and rice	ETAAS	Magnetic SPE by EDTA bound nano silk fibroin attached to Fe3O4	No ligand needed. The analytical procedure was not described, though the SPE substrate and sample preparations were	0.002	Spike recoveries	195
Ni	Chamomile tea and coffee	FAAS with slotted quartz tube atomiser	DLLME into CHCl3 with isopropyl alcohol as the disperser. After evaporation to dryness, the residue was dissolved in HNO3	Ni was complexed with the Schiff base ligand derived from the reaction of 5-bromosalicylaldehyde and o-phenylenediamine	5	Spike recoveries	66
Pb	Chocolate, mussels, red wine, rice	ICP-OES	See As ref. 49				49
Pb	Water (drinking fountain)	FAAS	See Cd ref. 186				186
Pb	Food, rice, wheat, tea	ETAAS	See Cd. ref. 187				187
Pb	Milk	FAAS with slotted quartz tube atomiser	DLLME into a DES (prepared by mixing choline chloride with phenol)	Dithiozone. The DES was injected into a solution containing the Pb–dithizone complex. Then THF was added as an emulsifier and the phases separated by centrifugation. The upper THF layer was analysed	9	Spike recoveries	196
Pb	Milk	FAAS	Magnetic dispersive SPE with Fe3O4@SiO2@3-chloropropyltriethoxysilane@o-phenylendiamine NPs. Extracted Pb was dissolved in 1.1 mol L^−1 HNO3	Chelating agent not needed	0.1	Spike recoveries	197
Pd	Water (tap and well)	FAAS	DLLME into CHCl3. After centrifuging, the organic phase was diluted with HNO3 + ethanol	N-(3-Chloro-4-fluorophenyl)-N-phenylthiourea solution added to sample acidified to 0.3 mol L^−1 HNO3. Pd was not detected in any samples	2	Spike recoveries	198
Sb species Sb^III Sb^V residual, digestible, and total	Milk	ICP-MS	Dispersive micro-SPE with fibrous TiO2@g-C3N4 nanocomposites (graphitic carbon nitrides)	Samples were prepared with artificial gastric juice to determine bioaccessible species. The selective adsorption of Sb^III Sb^V on the substrate was investigated and Sb^III was eluted completely with 1.5 mol L^−1 HNO3	0.4 pg mL^−1	CRM GBW 10017 (milk powder)	199
Se	Rice	HG-AAS	See As ref. 171				171
Se	Tea, bergamot and mint	FAAS with slotted quartz tube atomiser	DSPE on ZrNPs. Se was eluted with concentrated HCl	No chelating agent needed. Hot water extracts were analysed, but Se concentrations were below the LOD	5	Spike recoveries	112
Se	Food mushroom, honey, cinnamon, leek, corn flour, kiwi, grape, molasses, rice, ground beef, and egg	ETAAS	DLLME into CHCl3 with THF as disperser	Extraction into dibenzyldithiocarbamate at pH 2 was selective for Se^IV over Se^VI	1	CRM LGC 6010 (hard drinking water), SRM 1577b (bovine liver) and SRM 1570a (spinach leaves)	180
Se	Tea	FAAS with slotted quartz tube atomiser	DSPE on zirconium nanoparticles. Se was eluted with concentrated HCl	No chelating agent needed. Hot water extracts were analysed, but Se concentrations were below the LOD	5	Spike recoveries	65
Zn	Mineral water and isotonic sports drinks	HG-AFS	SPE on a polyurethane foam column loaded with PAR from pH 8 samples (carbonate/bicarbonate buffer). Elution was with 3% HCl	PAR chelating agent preloaded on column	0.03	NRCC CASS-4 (sea water) and spike recoveries	95
Various (5)	Tea	ETAAS	DES of 1-decyl-3-methylimidazolium chloride and n-butanoic acid was the extraction solvent at 50 °C. After separation and cooling, the floating solid was dissolved in acetonitrile	DDTP complexes of As, Cd, Cu, Hg and Pb	0.1 (As), 0.005 (Cd), 0.03 (Cu), 0.1 (Hg) 0.03, (Pb)	CRM GBW 10052 (green tea)	177
Various (10)	Tea, black and green, infusions and leaf	ICP-OES	SPE by RP (Discovery DSC-18) and strong cation-exchange extraction (Discovery DSC-SCX) columns connected in series. Analytes were eluted with 2.0 mol L^−1 HCl	Operationally defined chemical fractionation of Al, Ba, Ca, Cu, Fe, Mg, Mn, Ni, Sr and Zn	Not given	Spike recoveries and comparison of results with those of a reference method	247

---

In a comparison of LLE and SPE for the pre-concentration of As, Cd and Pb⁴⁹ from the digests of chocolate, mussels, and rice with HNO₃, the APDC complexes of the analytes were either dispersively extracted into THF (with the help of a surfactant, 1-decanol or 1-undecanol) or retained in the inner walls of a PTFE knotted tubular reactor (id 0.5 mm, 334 cm length) followed by elution with an organic solvent (MeOH, EtOH, 1-propanol, CH₃COOH or ACN). The method was also applied directly to undiluted red wine. The researchers concluded that the DLLME procedure was better because of the 3-fold lower LODs resulting from the higher enhancement factor (ratio of the sensitivity with pre-concentration to that without pre-concentration), consumption index (the ratio of the sample volume to the enhancement factor) and sample throughput. The LODs were 2 μg L⁻¹ (As), 0.6 μg L⁻¹ (Cd) and 2 μg L⁻¹ (Pb), which were not low enough to detect the elements in any of the samples except the mussels, which were also analysed by ICP-MS as part of the validation. Samples were also spiked (600 μg kg⁻¹); recoveries ranged from 96% to 101%. The researchers pointed out that, even with the 50-times dilution of the sample preparation (0.5 g and 25 mL), the LODs were low enough to detect concentrations at the limits set by EU Regulation 1881/2006. Calibration standards were matrix matched, which for wine was 12% EtOH and 1000 mg L⁻¹ potassium, and for the other materials was 10% HNO₃.

Liu et al.⁵⁰ developed a method that featured vapour–solid pre-concentration in a dielectric barrier discharge trap, consisting of three concentric tubes, for the determination of As in whole blood by AFS. Blood (0.25 mL) was deproteinised with 10 mL of 3% HNO₃ in a 50 mL centrifuge tube, and after centrifuging, 2 mL was transferred to a batch HG reactor containing borohydride in alkaline solution. The gases were swept into the DBD with Ar + O₂ and the arsenic oxides were trapped on the quartz tube surface from a discharge at 11 kV. After switching to Ar + H₂, As species were released at 13 kV and transported to the hydrogen diffusion flame atomiser of the AFS instrument, producing a rapid transient signal of base-width approximately 1 s. The calibration range was 0.05–300 μg L⁻¹ and the LOD was 7 pg, which, in a 0.25 mL sample, corresponds to 0.03 μg L⁻¹. The method was validated by comparison of the results with those obtained the by ICP-MS following microwave-assisted digestion in acid, and by spike recoveries from five samples, all of which contained As at concentrations ranging from 2.7 to 4.4 μg kg⁻¹. Visual inspection of the results shows no significant differences between the two methods, and for a single spike of 10 μg kg⁻¹ to 5 samples, recoveries ranged from 96% to 107%.

4 Progress with analytical techniques

4.1 Mass spectrometry

This review period has seen an increased number of papers using single particle ICP-MS, both to determine elemental content of individual cells and to quantify numbers and sizes of NPs. Two papers considered the sample introduction systems for single cell analysis with the aim of minimising the need for specialist equipment. Zhou et al.⁵¹ described a bespoke oil-free passive microfluidic sample introduction system, where the ability to control cell throughput rate and the time interval between adjacent cells allowed optimal use of a slower ICP-QMS instrument capable of 10 ms dwell times. The reported LODs for dissolved Au and AuNP colloidal solution (1.42 and 1.83 μg L⁻¹ respectively) were low compared with pneumatic nebulisation and there was increased detection efficiency of single cells (∼70%). In a different approach, Tanaka et al.⁵² used a standard nebuliser, with no microfluidic device, alongside a total consumption spray chamber coupled to ICP-QQQMS for analysis of Mg, P, S and Zn in single yeast cells. While good agreement was reported for elemental concentrations compared with bulk digestion and ICP-MS detection, transport efficiencies for single cells were poor (from 8% to 13%), which may limit the utility of this set up.

With the aim of achieving adequately low LODs for application of spICP-MS to physiologically relevant SiO₂ NP toxicokinetic studies, interference removal and background reduction for Si measurement were the focus of another piece of work.⁵³ Removal of polyatomic interferences, [¹⁴N¹⁴N]⁺ and [¹²C¹⁶O]⁺, at m/z 28 along with interferences on the other Si isotopes, was investigated using reaction gases, CH₄, NH₃, H₂ (10% in Ar) and O₂. Methane at a flow rate of 0.8 mL min⁻¹ was found to be optimal. Background Si concentrations were reduced by 20% through replacement of the sample introduction system with non-quartz components and use of rigorous decontamination procedures yielded a further 10-fold decrease. The study reported detection limits for total Si in tissues that were one to two orders of magnitude lower than similar studies (LODs, from 0.2 μg g⁻¹ to 0.5 μg g⁻¹, and LOQs, from 0.7 μg g⁻¹ to 1.8 μg g⁻¹, for all except one of the tissues studied). Single particle ICP-MS analysis of enzymatically digested liver tissue with a dwell time of 3 ms achieved a concentration LOD of 0.02 μg g⁻¹ and a size LOD of 350 nm for SiO₂ NPs. The size LOD was not adequately low to detect all SiO₂ NPs present in the tissues as seen by SEM.

The wider adoption of ICP-QQQMS for superior interference removal is reflected in the technique featuring in an increased number of publications this year. A study characterising biogenic SeNPs in single yeast cells involved detection of [⁸⁰Se¹⁶O]⁺, and [³¹P¹⁶O]⁺ as an elemental cell marker, in O₂ reaction mode to confirm the presence of intracellular Se.⁵⁴ Transport efficiency for spICP-MS (69 ± 3%) was improved by employment of a high performance concentric nebuliser with an inner capillary tube of 110 μm and a low volume spray chamber, using a sheath gas flow. Elemental content of Se was measured with an LOD of 0.16 fg cell⁻¹ and was observed to differ between individual cells. The percentage of cells with incorporated Se was calculated by comparing the number of cell events for both Se and P. In the second part of the work, a range of SeNP sizes (40 to 250 nm, median 103 nm) was calculated to be present within the lysed cells. Smaller NPs, below the LOD of the spICP-MS technique (20 nm with a dwell time of 0.1 ms), were observed by HR-TEM and confirmed by HPLC-ICP-MS. In another study, ICP-QQQMS with species unspecific ID for accurate measurement of Se improved performance of previously described speciation methods for selenoproteins, using a column switching system based on SEC and AF chromatography, and selenometabolites, by way of AEC.⁵⁵ Isotope dilution with ⁷⁴Se was introduced post column and Se was determined using a mixture of H₂ (2 mL min⁻¹) and O₂ (40%) gases. Impressively low LODs from 0.1 ng g⁻¹ to 0.4 ng g⁻¹ Se for selenoproteins and selenometabolites were achieved. The method was validated using a CRM, supplemented by spiking with selenometabolites, before being applied to analysis of serum from patients with lung cancer vs. healthy controls.

Fu et al.⁵⁶ explored the use of N₂O gas as an alternative to more commonly used reaction gases in development of a method suited to high sample throughput for determination of elemental biomarkers in serum of patients with hepatocellular carcinoma. While O₂ reaction gas yielded good signals for the oxides of Fe, Mn, Ni and Se, only low signals were observed for [CuO]⁺ and [ZnO]⁺, necessitating use of a second gas mode, e.g., NH₃/He. Reaction gas, N₂O, at a flow rate of 0.36 mL min⁻¹, allowed all six elements to be detected as oxides with comparable or increased sensitivity with respect to O₂ or NH₃/He gases, thus avoiding the need for two gas modes. A bovine CRM was analysed and the results compared with those obtained by SF-ICP-MS to validate the approach. However, the very low LODs reported (from 0.008 μg L⁻¹ to 0.25 μg L⁻¹) should be viewed with caution as the description of the calculation method provided suggests two approaches to calculating LODs may have been confused, leading to an underestimation of the LODs achieved.

The interference removal capabilities of ICP-QQQMS facilitate accurate determination of heteroatoms by ICP-MS, extending the scope of the technique to quantify proteins and other biomolecules. A paper this year reported a strong AEC speciation method for the determination of the Cu transport protein, caeruloplasmin, in serum using simultaneous detection of Cu and S to monitor elution of the metal and protein moiety.⁵⁷ Optimal detection of S was achieved as [³²S¹⁶O]⁺ and [³⁴S¹⁶O]⁺ in triple quadrupole mode with O₂ reaction gas at a flow rate of 0.31 mL min⁻¹, while better figures of merit were observed with detection of ⁶³Cu⁺ and ⁶⁵Cu⁺ in single quadrupole mode. For both elements, ⁸⁹Y was employed as the internal standard. The method was validated using standard solutions, serum from patients with and without Wilson’s disease, and quality control materials. The LOD and LOQ for caeruloplasmin were 0.1 μg L⁻¹ and 0.4 μg L⁻¹ respectively, which are considerably lower than those for immunoturbidimetric assays. Hu et al.⁵⁸ reported measurement of P concentrations by SF-ICP-MS to directly detect microRNA, a class of short RNA whose abnormal expression has been detected in various tumours. Use of medium resolution mode allowed ³¹P⁺ to be distinguished from polyatomic interference [¹⁵N¹⁶O]⁺. Coupled with an improved amplification strategy (two-stage hybridization chain reaction), adequate sensitivity was achieved for the detection of low concentrations of microRNA, with an LOD of 13 fM. The method was applied to human cervical and liver cancer cell lines with comparative analysis by RT-PCR.

In the routine clinical laboratory, doubly charged ⁵⁶Gd²⁺ interference on ⁷⁸Se measurement is a relevant issue in patients who have undergone administration of Gd-based contrast agents for imaging investigations. Elimination of this interference may be achieved with ICP-QQQMS by O₂ mass shift but not with a standard ICP-QMS instrument. To address this, Wilschefski et al.⁵⁹ evaluated a matrix independent interference correction equation. By utilising “narrow peak mode”, a higher resolution setting, the signal at m/z 78.5 corresponding to ¹⁵⁷Gd was detected and multiplied by the relative isotopic abundance, ¹⁵⁶Gd : ¹⁵⁷Gd, before being deducted from the ⁷⁸Se signal. Internal standard, Te, was employed to mitigate the tendency to under recover light element Se in the presence of heavy Gd. The calculation was validated using serum and commercial QC or external quality assessment samples spiked with Gd at 2 to 20 mg L⁻¹. Average recoveries were 97.4 to 106.5% and results from quality control or external quality assessment materials were within specified ranges. Precision was excellent with RSDs < 4.3%.

Three publications over this review period describe the use of laser ablation as a sampling technique for high throughput quantitative ICP-MS analysis. In the study by Zheng et al.⁶⁰ LA-ICP-MS was combined with use of a novel PDMS microwell array to produce a method for measuring AgNP concentrations in single cells. Detection of ³¹P allowed cell-containing microwells to be distinguished from empty microwells. The optimised LA parameters included a fluence of 6.0 J cm⁻¹ and spot size 60 μm, operating in single shot analysis mode. Single cell standards were manufactured using an inkjet printer loaded with Ag standard solution. The approach achieved an LOD and LOQ of 0.2 and 0.7 fg Ag cell⁻¹ respectively. Following validation of the results against bulk digestion of a known number of cells, the method was applied to 500 single cells, of which 487 had measurable concentrations of Ag (from 0.80 fg cell⁻¹ to 383 fg cell⁻¹, corresponding to between 18 and 8705 AgNPs per cell). Another application of LA-ICP-MS was a screening method for direct analysis of Br and I in dried serum spots using a non-matrix matched calibration. In the optimised method, line mode was utilised with spot size 60 μm, ablation energy 2.69 J cm⁻¹ and scan speed 60 μm s⁻¹. Analytical figures of merit were clinically relevant LODs of 0.23 mg L⁻¹ and 0.03 mg L⁻¹ and RSDs of <7.2 and <8.7% at 500 μg L⁻¹ for Br and I respectively. Recoveries in serum with and without spiked Br vs. a standard digestion ICP-MS method, employing a matrix matched calibration, ranged from 81.5% to 118%, supporting the validity of this approach. Finally, LA-ICP-MS was employed to determine 15 elements characteristic of gunshot residue from nasal mucus collected on a novel sampling device.⁶¹ The method offered shorter swab sampling times with respect to traditional methods. Optimised parameters for LA were spot size 160 μm, ablation energy 40% and scan rate 40 μm s⁻¹. Detection by ICP-MS was performed under wet plasma conditions by mixing the aerosol generated by LA with 1 ng g^{−1 205}Tl standard solution prior to nebulisation. Following normalisation of the signals against ¹³C, criteria for identification of gunshot residue particles were at least 61 counts for ¹²¹Sb, 299 for ¹³⁷Ba and 2741 for ²⁰⁸Pb.

Determination of Cu isotopic ratios using MC-ICP-MS was the focus of two clinical papers this year. The first compared the isotopic composition of Cu in serum from patients with age-related macular degeneration vs. healthy controls and found light isotope enrichment in the disease group.⁶² To correct for bias arising from instrumental mass discrimination, a sample-standard bracketing sequence was used, incorporating NIST SRM 976 as the external standard, and all solutions were adjusted to 20 μg L⁻¹ Cu with Ga as the internal standard at 50 μg L⁻¹. In the second study, following separation by SEC, Cu isotopic analysis of four intracellular protein fractions of a neuronal human cell line exposed in vitro to U was undertaken.⁶³ A modified sample-standard bracketing approach used ERM©-AE633 i-CRM as the external standard, with Cu concentrations matched to within 50% difference, and internal standardisation with Zn. Precision of the isotope ratios was better than 0.15‰. Differences in δ⁶⁵Cu values were reported between the four protein fractions.

4.2 Atomic absorption and atomic emission spectrometry

A comprehensive review of recent advances in atomic absorption and emission spectrometry, complements the work discussed in this update.²

Although AAS is not extensively employed for applications within the scope of this year’s update it has been noted that there is a continuing interest where it is possible to take advantage of the speed of analysis and simplicity of the technique. This is best seen in relation to publications that refer to devices that provide for an increase in sensitivity. One approach is to use the slotted quartz tube, for which a resurgence in interest is noted. Atsever and colleagues⁶⁴ used this device to determined Fe in hair samples. The preparation involved switchable solvent liquid extraction and, with the slotted tube, the LOQ for the FAAS measurement was 8.6 μg L⁻¹, a sensitivity increase of 92-fold. Two further reports involved the analysis of green tea⁶⁵ and chamomile tea and coffee.⁶⁶ Both groups employed DSPE procedures for the initial sample preparation. For the analysis of green tea, Karlidag et al.⁶⁵ used ZrNPs, prior to measurement of Se by slotted quartz tube FAAS while Saylan et al.⁶⁶ prepared a Schiff base ligand from the reaction of 5-bromosalicylaldehyde and o-phenylenediamine to form a coordinate nickel complex. The reported enhancement for the Se analysis was approximately 420-fold and 66-fold for the Ni procedure. Another analyte enhancement process from the past, the knotted reactor was re-assessed by Martinez-Rubio et al.⁴⁹ who compared performance characteristics with those given by a DLLME method. Following preliminary treatment, an analyte–reagent complex is pumped through the PTFE-tubing knotted reactor. Changes in the flow direction caused by the knots push the analyte complex particles towards the tubing walls. The complex is subsequently eluted for quantitative analysis. These workers applied the two procedures to measure concentrations of As, Cd and Pb in food samples by ICP-AES. It was found that the knotted reactor gave a 10-fold improvement in the LODs whereas a 40-fold enhancement was possible with the DLLME method.

Applications featuring liquid–liquid or solid–liquid analyte enrichment followed by AAS continue to be reported and examples are these are summarised in Table 1. Most refer to ETAAS but two publications using FAAS are of note. Concentrations of Cu in urine, sweat, dialysis solution and water, and Mn in coffee, were measured by Uzcan and Soylak⁶⁷ and Tışlı et al.,⁶⁸ respectively. The reported LODs and (enrichment factors) were 7.30 μg L⁻¹ (40-fold) for Cu and 0.52 μg L⁻¹ (93-fold) for Mn. The results for Cu are unimpressive and concentrations in real samples were all below the LOD.

As previously noted (Section 1), AAS is no longer widely used for analysis of clinical and food/beverage samples but each year a few niche applications are reported in which CS-AAS offers useful advantages compared to alternative techniques. Within this category, the work of Pires et al.⁶⁹ is of interest. Using a HR-CS-GF-system, Cl, Fe and Si were simultaneously determined by AAS (Fe and Si) and molecular absorption (Cl as the InCl molecule at 267.2181 nm). A Pd–Mg chemical modifier was used and the temperatures for vaporisation and atomisation were optimised. The method was applied to the analysis of beer samples and provided LOQs of 0.17 mg L⁻¹, 6.7 μg L⁻¹ and 0.26 mg L⁻¹ for Cl, Fe and Si, respectively. Concentrations found in ten beer samples were 106–277 mg L⁻¹ for Cl, <20 μg L⁻¹ for Fe and 15–37 mg L⁻¹ for Si. In two other reports HR-CS-ETAAS was employed for the measurement of Cd and Fe in cereal flakes⁷⁰ and Pb in dietary supplements.⁷¹ The optimised conditions and LODs were reported. These appear to be similar to those given in other published studies.

Deng and co-workers⁷² developed a portable instrument which was used for a simple procedure to monitor Cd contamination in rice. The analyser consists of a tungsten-coil trap with vaporisation into an atmospheric pressure glow discharge plasma, to allow measurement by AES. Cadmium, in milligram amounts of dried sample, was vaporised and trapped on the cold tungsten coil. The Cd was released by heating the coil, into a stream of Ar–H₂, and introduced into the GD plasma. The entire process was complete within 3 min and the LOD, for a 10 mg sample, was 2.6 pg. Accuracy was demonstrated by analysis of a rice CRM and by comparing results with those given by microwave digestion and ICP-MS.

Some interesting applications using AE spectrometric techniques were reported during this review year. Maung and Beauchemin⁷³ measured F in biological CRMs with an ETV-ICP-AES procedure. Solid samples, 2 mg, in graphite boats were placed into the ETV furnace and the F pyrolysed at 250 °C with vaporisation at 2200 °C. To compensate for sample load effects in the plasma, Ar emission at 404.442 nm was used as an internal standard while CRM DUWF-1 (Durum wheat flour) was used for calibration. Determination of Cd in water featured in the work of Serrano et al.⁷⁴ The objective of this project was to investigate spectral and non-spectral interferences in MIP-AES, derived from the organic matrix when samples are prepared by DLLME. Using CHCl₃ and supramolecular solvent based on 1-decanol and THF, interferences from carbon-based molecular band emissions were evident at wavelengths greater than 328 nm. Investigations of the plasma conditions and the spectrometer settings enabled the interferences to be mitigated. Adopting an analyte wavelength free from interference, and with an optimised nebuliser gas flow rate, Cd was successfully determined in tap, sparkling and synthetic sea water. Enrichment factors of 46 and 42 were possible from CHCl₃ and supramolecular based solvent procedures, respectively. The advantages of DLLME and improved aerosol generation and transport gave an LOD for Cd of 1 μg L⁻¹. Two further reports present work in which samples were analysed using MIP-AES. Carvalho et al.⁷⁵ determined concentrations of Cu, K, Mg, Mn, P and Zn in instant soups. The samples were simply microwave digested and aqueous calibrator solutions used for quantification. The results of analyses of three NIST SRMs, and from recovery experiments, demonstrated the accuracy of the procedures. The LODs, 0.09 mg kg⁻¹, 4.9 mg kg⁻¹, 1.0 mg kg⁻¹, 0.04 mg kg⁻¹, 5.4 mg kg⁻¹ and 0.88 mg kg⁻¹, for Cu, K, Mg, Mn, P and Zn, respectively, were more than adequate for the concentrations found in the 19 commercial samples analysed. The highest concentrations were found in chicken with noodles and rib soups. The concentrations of 11 elements in Mexican red wines were also measured by MIP-AES.⁷⁶ Samples were simply diluted from five-fold to 100-fold depending on the initial concentration. Three calibration strategies were compared, conventional standards addition, standard dilution analysis (SDA) and multi-energy calibration (MEC). Standard dilution analysis was based on two preparations for each sample: one diluted 1 + 1 with 2% HNO₃, the second, diluted as for the first and then spiked to double the concentration. For MEC, the paired samples were prepared in the same way, but the spike contained all analytes. Concentrations for all elements were determined by standard additions, trace elements also by SDA, and the major elements also by MEC. The SDA and MEC approaches offer simpler preparation than is required for conventional standard additions while still providing for matrix matching. The concentrations found in the Mexican wines were reported according to their winery regions and compared to maximum allowable levels in various states and other countries. The mineral content of infant formula was the focus of a study by Fioravanti et al.⁴⁵ The work involved both method development and validation, using ICP-AES and an in vitro study of the ‘dialysability’ of elements from the formula. Similar results were obtained from ultrasound assisted extraction and microwave digestion when an infant food CRM was analysed. Therefore, the ultrasound procedure was used for the experimental work. A simulated digestion model gave an indication of bioaccessibility of elements from the formula. Concentrations of Ca, Cu, Fe, K, Mg, Mn, Na, P and Zn in five infant formulae were all above the limits stated in the FAO/WHO Codex Stan 72-1981, revision 2007. Dialysable fractions from the formulae were generally similar in the four skimmed milk formulae but very much lower in the soybean-based formula.

4.3 Laser induced breakdown spectroscopy

Within this update period, a number of review articles dedicated to LIBS were published, providing useful resources for researchers in this growing field. Shah et al.⁹ produced a comprehensive tutorial style review of LIBS covering the current state of the art, technical considerations, advantages over other techniques and calibration approaches. A number of applications were also presented which included food and biological samples. With over 200 references, it provided a thorough assessment of LIBS. Martinez and Baudelet¹³ considered the challenges and future directions of calibration strategies for LIBS as a solid sampling technique, alongside LA-ICP-MS. This article focussed specifically on the elemental quantification of biological materials. Wang and co-workers¹⁰ considered LIBS as a tool for disease detection and diagnosis in soft tissue, with particular emphasis on sample preparation and data processing. A number of applications were discussed and potential future directions highlighted such as real time analysis for laser surgery.

Over the past year, the use of LIBS for clinical applications has shown significant development. Modlitbová et al.⁷⁷ utilised LIBS as a detector for immunoassays with nanoparticle tags as labels. The feasibility and experimental set up was established using Ag and AuNPs with 96 well microtiter plates. Once optimised, the approach was applied to a sandwich immunoassay for HSA using streptavidin-coated AgNP labels. The LOD was comparable to those of traditional fluorescent methods at 10 ng mL⁻¹, yet LIBS has a number of advantages such as multiplexing, immediate signal response and automation, offering a strong alternative detector for immunoassays. Chu and co-workers⁷⁸ combined LIBS with NPs in a different manner in order to elucidate the biological half-life of inorganic–organic hybrid nanomaterials. Bovine serum albumin was used as a drug delivery carrier which enabled the formation of nanoclusters using two inorganic NPs, MnO₂ and AlO(OH), encapsulated in the protein. The nanoclusters were separately injected into mice and blood samples taken at various time points over 48 h. The whole blood was digested in 1 + 1 HNO₃–H₂O₂ using an ultrasonic bath for 5 min. An aliquot (20 μL) was dried on a silicon pellet before LIBS analysis. Aqueous calibrants were prepared by the same method. The spectra required additional processing utilising both polynomial and Lorentzian fitting to correct for the background spectral interferences on the main analytical lines: Mn I 403.076 and Al I 394.406 nm. The data enabled calculation of the half-life of the nanomaterials. The blood samples were also analysed by ICP-MS for additional confirmation which showed agreement within 5%. The work demonstrated the applicability of LIBS for rapid analysis of blood samples in pharmacokinetic studies. Gondal et al.⁷⁹ reported an investigation of the elemental content in cancerous colon tissue compared to healthy tissue using LIBS. It was found that the lines corresponding to Ce, Cr, Hg and Pb were observed in the cancerous tissues but not detected in healthy specimens. The elemental concentrations were then confirmed by ICP-OES: 3.2 μg L⁻¹ Ce, 14.6 μg L⁻¹ Cr and 2.7 μg L⁻¹ Pb; Hg was not detected, which may have been a result of the open vessel digestion procedure employed. The workers then applied calibration-free quantitative LIBS by estimating the electron density, plasma temperature and relative number density of neutral and singly ionic species from the spectral lines, by assuming that the plasma was optically thin and local thermodynamic equilibrium was attained. Following the mathematics, concentrations could then be obtained: 3.1 μg L⁻¹ Ce, 13.4 μg L⁻¹ Cr, 7.61 μg L⁻¹ Hg and 3.1 μg L⁻¹ Pb, achieving good comparability with the ICP-OES data. The ease and speed of LIBS analysis combined with the ability to detect toxic elements at low concentration levels was clearly demonstrated.

Continuing the trend from previous updates,^1,80the application of LIBS as a tool for the rapid analysis of food products has continued with particular attention on fraud detection and geographical authentication. Honey was the focus of two studies^81,82 within this review period, with both publications highlighting that honey required no sample preparation as a significant advantage over other analytical techniques. Nespeca et al.⁸¹ compared honey from different botanical sources with sweeteners and fortified samples using LIBS. A calibration set was prepared by mixing the honey with sweetener at incremental levels. Statistical pre-treatment of the data was undertaken to focus on the variables which explained more than 80% of the variance. After optimisation of the PLS-DA models, adulterants detection achieved 100% accuracy, noting that Ca and Fe were key discriminating elements. The authors suggested the method could be extended to other ‘high density’ food products such as condensed milk, ketchup or mayonnaise. Peng et al.⁸² analysed acacia honey adulterated with high fructose corn syrup and rape honey at varying levels. The LIBS data were evaluated using univariate statistical analysis and PLSR, showing that emissions from Ca, K, Mg and Na were the most significant variables. The root-mean-square error ranged between 4.8% and 8.9% indicating good accuracy. Olive oils were the subject of a geographical origin study by Gyftokostas and co-workers,⁸³ who applied machine learning algorithms to the LIBS analysis of 36 olive oils from Crete, Greece. The optimal model was LDA combined with pre-processing using PCA to focus on the variables with the highest discrimination capability, minimising the computational time required for the model. Despite the small geographical footprint, a discrimination accuracy of 94.0% ± 1.1% was achieved. The authors postulate the technique could be extremely valuable if expanded to more regions of Greece and other countries. Zhangcheng et al.⁸⁴ assessed the suitability of LIBS for fraud identification in traditional Chinese medicines. Due to the price of saffron, adulteration with radish strands and corn silk had been observed, therefore samples were analysed directly by LIBS followed by PCA statistical treatment. A number of elements were detected in saffron, namely Al, Ba, Ca, Fe, K, Mg, Mn, Na, P, Si, Sr and Ti, which were confirmed by complementary ICP-MS analysis, in addition to C observed at 247.856 nm. For differentiation purposes, Al, C, Ca, Fe, Mg, P and Si were the most relevant and clear clusters were obtained for the three sample types, highlighting the power of the technique despite no formal quantitation via LIBS.

Infant milk formula was the focus of a study by Markiewicz-Keszycka et al.⁸⁵ who utilised LIBS with PLSR to develop a calibration model for Mg quantification. Commercial samples were mixed with MgCl₂ at different doping levels and were first analysed by AAS following microwave-assisted acid digestion to obtain reference values for the calibration and validation sample sets. For LIBS analysis, the powders were pressed into pellets and measured directly. The model was tested with a validation sample set, achieving regression coefficients from 0.85 to 0.94, although it was noted that at the highest Mg levels, saturation of the detector may have occurred leading to non-linearity. However the method provided a rapid approach for infant milk formula analysis with very little sample preparation.

The application of multiple analytical techniques to investigate diseases in food crops were demonstrated in two works by Sharma and co-workers^86,87 who used LIBS, WDXRF and FTIR for rice and papaya. In the first paper,⁸⁶ rice infected with false smut disease, caused by a fungus which leads to reduced grain yield and quality, was investigated. Both LIBS (qualitative) and WDXRF (quantitative) provided complementary elemental data showing that levels of Ca, Cu, Fe, Mg and Si were decreased in diseased rice grains whilst K increased. The spectra from FTIR showed changes in the organic components of the rice, particularly in the starch fingerprint region, contributing to the effect of crop quality. The second study⁸⁷ followed the same principles, but the researchers focussed on the parasitic root knot nematode in the papaya plant which causes yield reductions and eventually kills the plant. The diseased plants showed increases in Ca, Mg and Si indicating the activation of the plant’s defence mechanisms whereas K and P decreased. As with the rice, FTIR demonstrated molecular changes with the starch significantly affected by the presence of the nematode. Both investigations showed how LIBS could be used qualitatively to rapidly determine elemental levels to provide data for assessing crop quality.

4.4 Vapour generation procedures and atomic fluorescence spectrometry

There were four reports of As speciation by HPLC with post-column HG in this review period. In a study of the elemental composition and As speciation in Brazilian rice varieties selected for biofortification, Freire et al.⁸⁸ determined Al, As, Cd, Co, Cr, Fe, Mg, Mn, Ni, Rb, Sb, Sr and Zn by ICP-MS following microwave-assisted digestion. For As speciation, sample (0.2 g) was extracted with10 mL of 2% HNO₃ with heating from 25 to 95 °C (0.75 h) and holding at 95 °C for 1.5 h. After cooling, the extracts were filtered (0.20 μm cellulose) and 100 μL was injected for separation by anion-exchange HPLC (Hamilton PRP X-100, with 10 mmol L⁻¹ phosphate buffer at pH 8.5 + 2% MeOH). The eluent was merged with streams of borohydride and of HNO₃ containing 0.5 μg L⁻¹ Ga as internal standard (m/z 69) and the mixture delivered to the nebuliser of the spectrometer. No details of the nebuliser or spray chamber were given. Visual inspection of the chromatograms with and without HG, shows an increase is sensitivity of about 3–4 times (based on peak height), some peak broadening (not enough to cause peak overlap), and the peristaltic pump pulsations. The LODs were 0.004 μg L⁻¹ for DMA^V, 0.003 μg L⁻¹ for MMA^V and As^III, and 0.01 μg L⁻¹ for As^V, which were some 30–50 times lower than those for the same instrumentation without the HG, and lower than those reported in 15 other publications. The procedure was validated (both for As and nine other elements) by the analysis of CRM NIST SRM1568a (rice flour), which is not certified for the arsenic species content, though there are numerous reports of the speciation analysis of this material in the literature. They also looked for all the lanthanoids together with Sc, Th and Y, many of which were not detected, though Sc was found in all samples at double- and triple-digit ng g⁻¹ concentrations. de las Torres et al.⁸⁹ determined As species in strawberries, grown in arsenic-enriched hydroponic media, by a procedure involving a similar HPLC separation (the mobile phase was a 20 mmol L⁻¹ phosphate buffer at pH 5.8 with no MeOH) following extraction of the lyophilized fruit with hot H₂O and a C18 cartridge clean-up. The separated species that are borohydride-active were detected by AFS following post-column derivatisation. The authors did not provide LODs, but were able to detect As^III, As^V, MMA^V DMA^V, at concentrations down to 0.01 mg kg⁻¹, which for the 50-times dilution in the sample preparation (200 mg in 10 mL) corresponds to a concentration of 0.2 μg L⁻¹. The method was validated by the accurate determination of total As in two CRMs NIST SRM 1568b (rice flour) and LGC 7162 (strawberry leaves) and by the determination of As species in 1568b. Cheng et al.⁴⁷ also separated As species by AEC on a Hamilton PRP-X100 column (isocratic elution with phosphate buffer at pH 8.2) following enzymatic extraction by a novel dual ultrasound-assisted extraction procedure (described in more detail in Section 3.2). Although the coupling between the HG manifold and the spectrometer contained two gas–liquid separators, the additional peak broadening did not significantly affect the chromatographic separation; baseline resolution was achieved for the first three peaks, corresponding to As^III, DMA^V and MMA^V and for concentrations up to 7 μg L⁻¹. The peak for As^V was, as is typical in this chromatography, well separated from the other early eluting compounds. The LODs were 0.7 and 2.5 μg kg⁻¹ for As^III and As^V, respectively. Although LODs were not give for DMA and MMA, they would have been similar to the value for As^III. There were no significant differences between measured (sum of inorganic species) and certified values for BMEMC CRMs GBW(E)090066 (Salvia) and GBW(E)090067 (Paeoniae Radix Rubra). The procedure was applied to four herb samples, in all of which both inorganic arsenic species were detected, and in two of which DMA was detected, and from which the recoveries of spikes (concentration not specified, but possibly at 0.33 μg g⁻¹) ranged from 92% to 103%. For the determination of As species in samples containing high concentrations of chloride (urine or seawater) by HPLC-HG-AFS, Yu et al.⁹⁰ devised a procedure in which two Hamilton PRP-X100 columns were connected in series and eluted with 35.0 mmol L⁻¹ (NH₄)₂HPO₄ (pH 6.00). Up to 10 g L⁻¹ chloride could be tolerated. The eluent was merged with a stream of thiourea to reduce As^V to As^III, thereby increasing the sensitivity. The analysis time was shortened (from about 12 min to about 9 min) by flow programming in which the flow-rate was doubled to 2.0 mL min⁻¹ for several minutes at the start and at the end of the chromatographic run. The LODs were 0.4 μg L⁻¹, 0.9 μg L⁻¹, 0.7 μg L⁻¹ and 1 μg L⁻¹ for As^III, As^V, MMA^V and DMA^V, respectively. The method was validated by the analysis of CRMs from the National Standard Substances Center (China) GBW(E)080231(seawater) for As^III, and GBW09115 (urine) for As^III, MMA^V and DMA^V. Low values were obtained for the concentration of As^III, which in the case of the urine material was significantly different from the certificate value.

To estimate exposure to iAs from rice-based infant food in Australia, Gu et al.⁹¹ determined iAs by HG-ICP-MS following LLE. After digestion with 50% HClO₄ (80 °C for 1 h), As^V was reduced to As^III with HBr and hydrazine sulfate. Concentrated HCl was added, and the AsCl₃ species extracted into CHCl₃, then back-extracted into dilute HCl. Total As was determined by ICP-MS following digestion with concentrated HNO₃ at 100 °C for 2 h, dilution, filtration (0.45 μm) and further dilution. The sample LODs were 0.05 mg kg⁻¹ and 0.004 mg kg⁻¹, for iAs and total As, respectively; the value for iAs is rather high compared with values routinely achieved by other methods, possibly due to the dilution involved in the sample preparation, which is 200-times for the total As method, though not enough information was provided to calculate this for the iAs method. Unfortunately, the analysis of a CRM, NCS ZC73031 (carrot), for total As had a relative measurement error of −26%. The material contains 0.11 ± 0.02 mg kg⁻¹ total As. Accurate analysis of another RM, AGAL40 (about which no further information was provided) for iAs was achieved, though the concentration of 6.43 mg kg⁻¹ is rather high compared with the values encountered in the real samples. The researchers concluded that their estimates of dietary exposure showed that infants consuming large amounts of rice pasta or crackers have an increased risk of the health impacts associated with excess intake of iAs.

The simultaneous determination of As, Hg, Sb and Se by HG-AFS required both compromise pre-reduction and instrument operating conditions.⁹² A number of Chinese herbal foods, including mint and liquorice, were dried, ground and digested (500 mg) in HNO₃ at room temperature for 24 h. Hydrogen peroxide was added, and the mixture heated at 120 °C for 6 h, when the volume was decreased to 2 mL. After cooling, 5 mL of 50% HCl was added and the solution heated for 30 min to reduce Se^VI to Se^IV. After cooling, the solution was diluted to 25 mL with the addition of 1 mL 3% thiourea, and 30 min allowed for the reduction of As^V to As^III and of Sb^V to Sb^III. The method was validated by analysis of three CRMs, GBW07601 (human hair), GBW07604 (poplar leaves) and GBW10016 (tea). No statistical analysis was presented, but visual inspection of the results shows no significant differences. The LODs were 0.05 μg L⁻¹, 0.006 μg L⁻¹, 0.03 μg L⁻¹ and 0.05 μg L⁻¹ for As, Hg, Sb and Se, respectively. The procedure was also applied to the analysis of radix et rhizoma rhei, herba ephedrae and radix codonopsis.

The determination of As in blood by HG-AFS with gas–solid pre-concentration in a DBD trap⁵⁰ is discussed under pre-concentration in Section 3.2.

There are two reports of the determination of MeHg by GC-AFS. The MeHg content of rice plants was determined in a study of the effect of amending mercury-contaminated soil with selenium as selenite.⁹³ Powdered sample (50 mg) was extracted with 2 mL of 25% (w/w) KOH–MeOH by shaking for 4 h at 60 °C followed by dilution to 10 mL with H₂O and centrifugation at 4000 rpm for 30 min. The supernatant was ethylated with NaBEt₄ (citrate buffer pH 4.1–4.5) and the methylethylmercury formed quantified by “cold vapour atomic fluorescence spectrometry (CVAFS, Brooks Rand, USA) following USEPA method 1630.” No information about any instrumentation was provided, but as the LOD was stated to be 0.002 μg L⁻¹ for “GC-AFS” it is assumed that the method was not just a purge-and-trap procedure, but also included GC separation. As the sample preparation involves a 200-fold dilution, it may be calculated that the sample LOD is 0.4 μg kg⁻¹. The concentration of MeHg in the control grain was around 55 μg kg⁻¹, which decreased to about 30 μg kg⁻¹ when selenium was applied. The method was validated by the analysis of CRMs ERM-CC580 (estuarine sediment) and DORM-3 (fish muscle) for which relative measurement errors of −8% and −16% were obtained, both of which may be significantly different from zero. Recoveries of spikes (at 50 μg kg⁻¹) to the grain samples were also low (89 ± 4%, n = 6). Results were also presented for other parts of the plant (roots, straw, hull) and the spatial distribution and relative concentrations of some nutrient elements (Cu, Fe, Se and Zn) in grains were determined by SR-XRF and it was concluded that supplementation with Se increased the concentration of Cu, Fe and Zn. Total Hg was determined by ICP-MS. Santa-Rios et al.⁹⁴ determined MeHg in whole blood and dried blood spots by a procedure almost identical to that just described. Samples (40 μL) were mixed with 600 μL of 4 M HNO₃ and 0.02% L-cysteine and heated at 60 °C for 24 h. A subsample (300 μL) of the cooled digest was diluted 1 + 99 with H₂O, adjusted to pH 4.0–4.5 using citrate buffer and ethylated using 1% NaBEt₄ in according to US EPA method 1630. The Hg species were purged, trapped on Tenax, thermally desorbed, separated on a GC column and detected by AFS following post-column pyrolysis to elemental Hg. All of these steps were carried out by a Tekran series 2700 instrument. The LODs for whole blood were 2 μg L⁻¹ (iHg) and 0.2 μg L⁻¹ (MeHg), and for dried blood spots were 1 μg L⁻¹ (iHg) and 0.3 μg L⁻¹ (MeHg). The method was validated by analysis of CRMs reference material PC-B-M1510 (MeHg concentration 2.56 ± 1.25 μg L⁻¹) and PC-BM1112 (MeHg concentration 8.68 ± 0.26 μg L⁻¹) and by spike recoveries. The researchers concluded that measurement of MeHg in dried capillary blood reflects the concentrations in venous whole blood, so that the analysis of dried blood is suitable for assessing MeHg exposure (but more work is required to assess iHg exposure). They also studied the stability of the dried blood spot samples and concluded that MeHg and total Hg concentrations are “relatively stable” after one year of storage at room temperature.

A method for the determination of Zn by HG-AFS, following pre-concentration by SPE, has been developed⁹⁵ and applied to the analysis of mineral waters and isotonic sports drinks. The sample pH was adjusted to 8 with a carbonate/bicarbonate buffer, and the Zn retained on a polyurethane column preloaded with PAR. Following elution with 3% HCl, the Zn was measured by AFS following HG in an Aurora Lumina 3300 atomic fluorescence spectrometer fitted with a high-intensity antimony HCL with a quartz tube atomiser. Few details of the instrument operation are provided, so it is not clear whether an antimony lamp was really used or what the HG conditions were. The optimization studies reported were all concerned with the SPE methodology. The enrichment factor was 89, producing a LOD of 0.03 μg L⁻¹, and the method was validated by the analysis of NRCC CASS-4 (seawater, containing 0.38 ± 0.06 μg L⁻¹ Zn) for which the percentage relative error was −16%, (no statistical analysis was provided, but visual inspection indicates that this is not significant) and by spike recoveries (at concentrations of 30, 40 and 60 μg L⁻¹ from three samples that contained 34, 40 and 65 μg L⁻¹, respectively) that ranged from 101–113%. The procedure was applied to the analysis of 10 real samples, in all of which Zn was detected at concentrations ranging from 30 to 80 μg L⁻¹.

For the electrochemical HG-AFS determination of Cd in waters (tap, bottled and environmental) Liu et al.⁹⁶ further developed the method previously reported for the determination of As. The apparatus consists of two pyrolytic graphite tubes (electrothermal atomisers for AAS) connected in series. The constant voltage (18 V) power supply was an ac to dc wall socket adapter for a smart phone or computer. The system did not have an ion-exchange membrane and only needed a pump to transport the feed solution and waste. Since no signal was detected with the flame off, it was deduced that the volatile Cd species was the hydride and not free atoms. To minimize decomposition of the hydride in the system, all connecting tubing was kept as short as possible. Under the optimized conditions, the electrochemical generation efficiency (based on the determination of the Cd concentration emerging from the reactor) was 38%, and the LOD was 0.05 ng mL⁻¹, which increased to 0.4 ng mL⁻¹ in the presence of thiourea, added if needed to remove the interference of copper, iron, nickel and zinc. The method was validated by the accurate analysis of three environmental water CRMs, GBW08608 (batch 17 081), GSB07-1185-2000 (batch 201431) and GSB07-3186-201 (batch 200933), from the Institute for Environmental Reference Materials of the National Ministry of Environmental Protection (China) that contained 12 ng mL⁻¹, 15 ng mL⁻¹ and 140 ng mL⁻¹, respectively. The procedure was applied to the analysis of one tap and one bottled water, in neither of which was Cd detected, but from which spike recoveries at 2 ng mL⁻¹ were significantly different from 100%.

There are two reports of CVG with atomic spectrometries (AES and AAS) other than AFS. For the determination of Hg in traditional Chinese medicines, a CV-AES method was developed⁹⁷ in which the Hg vapour generated by reaction with borohydride was entrained in an argon carrier stream that flowed through a stainless-steel hollow needle electrode (1 mm id, 1.5 mm od, 15 mm length) into a plasma sustained in the 5 mm gap between this electrode and a solid tungsten needle electrode (1 mm diameter, 20 mm length) by an ac power supply (31 kHz, 8 kV peak voltage, and 30 mA maximum current). The electrodes were sheathed in a quartz tube to confine the discharge and prevent ingress of air. The emission intensity at the 253.65 nm line was monitored resulting in an LOD of 0.2 ng mL⁻¹. Sample (0.2 g) was mixed with 5.0 mol L⁻¹ HCl (5 mL), irradiated with ultrasound (10 min at room temperature), centrifuged (4000 rpm), and the supernatant diluted to 10 mL with H₂O. The method was validated by the accurate analysis of a CRM BMEMC GBW 08042 (material unknown) and by spike recoveries ranging from 93% to 116% at 1 μg g⁻¹ from seven real samples (in all but two of which Hg was detected). The problems of accurate determination of Se at low concentrations in beans (<5 ng g⁻¹) by ETAAS have been investigated.⁹⁸ Spectral interferences from the molecular bands of PO and NO as well as from iron lines (195.950 nm, 196.061 nm, and 196.147 nm) close to the Se lines at 196.026 nm and 203.985 nm cause inaccuracies. The researchers were unable to overcome the interferences either by judicious use of chemical modification or by using CS-AAS, achieving LODs of 24 ng g⁻¹, 33 ng g⁻¹ and 29 ng g⁻¹ for Ir, Ru, and Pd/Mg modifiers, respectively. However, for separation of the analyte from the matrix by HG with in-atomiser trapping on Ir and the use of HCL-source ETAAS, the LOD was 30 pg g⁻¹, an improvement of approximately 1000-fold. The researchers noted that part of the problem for the CS instrument was the very low emission intensity from the short-arc xenon lamp at the Se wavelength. The method was validated by accurate analysis of a CRM GBW10010 (rice) from the Institute of Geophysical and Geochemical Exploration, Langfang, China, which contains 61 ± 15 ng g⁻¹ Se. No results were presented for the “14 different bean samples (Phaseolus vulgaris L.)” that were used in the study.

4.5 X-ray spectrometry

A comprehensive review of recent advances in X-ray spectrometry³ 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 below. In addition to this, a review of in vivo measurement systems using XRF spectrometry and NAA was published by Chettle and McNeill⁹⁹ which covered the development of eight such instruments for biomedical purposes.

Gruber et al.¹⁰⁰ utilised TXRF spectrometry for quantitative analysis of various biological samples, namely liver, parenchymal and non-parenchymal liver cells, and bone marrow-derived macrophages. Either Ga or Ti was added as an internal standard for single point calibration; the selection of which depended on the sample type as Ga suffered from Zn overlap especially in liver tissues and some samples types were found to contain Ti (bone marrow-derived macrophages). The method was validated for Fe as an exemplary element through the use of NIST SRM 1577c bovine liver achieving a recovery of 99 ± 10% and complementary analysis by GF-AAS (0.9993 ± 0.0014 gradient), with linearity demonstrated across 3 orders of magnitude. Additional recovery data was presented for NIST SRM 1577c for Cu, Mn, Rb, Se and Zn, with Cu, Mn and Zn achieving recoveries of 92–106%. However, Rb and Se showed significant differences compared to the reference value, although it was noted these were close to the LOD. The work demonstrated good accuracy with improved total analysis times compared with traditional microwave digestion and atomic spectroscopy with comparable detection limits. Almeida and co-workers¹⁰¹ considered the use of a portable TXRF system to investigate Alzheimer’s disease in mouse models. The levels of essential elements such as Cu, Fe and Zn in the brain are known to change in Alzheimer’s patients and may interact with the formation of beta-amyloid plaques. Brain tissue from mice with induced Alzheimer’s caused by injection of beta-amyloid soluble oligomers at two concentration levels were analysed. The elemental levels of Cu, Fe, K, P, Rb, S and Zn were determined and found to quickly change which may explain cognitive issues independently of amyloid plaques. An unusual approach was taken by Gao et al.¹⁰² (in Chinese) who analysed food stuffs by TXRF spectrometry after simple preparation as a slurry. Three solutions were compared, namely ultrapure H₂O, 3 mol L⁻¹ HNO₃ and 1% Triton X-100, with different sample masses and analysed for As, Ca, Cu, Fe, K, Mn, P and Zn. The evaluation was performed using CRMs and found that 20 mg of sample dispersed in ultrapure water with Ga as the internal standard was optimal. Several other food-based CRMs were then measured, achieving recoveries in region of 80–120% for most elements with <15% RSD which was satisfactory considering the overall analysis time.

A number of papers described the use of EDXRF and WDXRF spectrometry for clinical and food applications. Zdiniakova and de la Calle¹⁰³ employed EDXRF spectrometry to assess the feasibility of elemental analysis as an authentication tool for coconut sugar, due to its high price compared with conventional sugars. Although IRMS is capable to differentiate coconut (C3 plant) from cane (C4 plant) with ease, it is much harder to separate coconut from beet (C3). The levels of Ca, Cl, Cr, Cu, Fe, I, K, Mg, Mn, Na, P, Se and Zn were compared and statistically analysed using SIMCA. It was not possible to discriminate beet from cane as only one beet sample was measured but the model was able to classify 8 of the 11 coconut sugars from 10 cane sugars. The authors acknowledged that geographical origin may also play a role which was not possible to control in the study due to availability. Two publications by Sharma and co-workers^86,87 included WDXRF spectrometry along with LIBS and FTIR for the investigation of diseases in food crops. These works are discussed further within the Section 4.3 of this review. Sichangi et al.¹⁰⁴ investigated the use of EDXRF and EDXRF scattering spectrometry as a cancer diagnostic tool. Robust multivariate chemometrics were applied to the spectra for calibration (based on spiked paraffin wax standards) and for the discrimination of diseased tissues. The accuracy of the developed method was demonstrated with NIST SRM 1566b Oyster Tissue, achieving recoveries within ±6% of expected values for Co, Cu, Fe, Mg, Mn, Na and Zn. The approach was then applied to control and cancerous tissues (mammary and prostate) from dogs. In both tissue types, the levels of Cu, Fe, Mg and Zn were significantly higher in malignant tumours compared with benign, with the ratios ranging from two to five times higher. Furthermore, positive correlations were found between Cu and Mg (r = 0.999) and between Mn and Fe (r = 0.999) in prostate samples whereas for mammary tumours, negative relationships were observed for Cu and Mg (r = −0.994) and Mn and Fe (r = −0.974). The statistical models were also able to successfully discriminate between normal and cancerous tissues. Furthermore, speciation analysis via inspection of K_α1 and K_α2 sub peaks indicated a higher valency for Cu, Fe and Mn (as Cu²⁺, Fe³⁺ and Mn⁷⁺) in malignant tumours compared to benign. The work demonstrated how a significant amount of information can be extracted from EDXRF spectrometry data which could be applied as a diagnostic tool.

This review would not be complete without mention of XRF spectrometry for the analysis of bone. Nguyen et al.¹⁰⁵ focussed on the quantification of La and Gd by KXRF spectrometry, with Ba and I inferred from sensitivity factors derived from the La calibration line and Monte Carlo simulations. Tibiae from human cadavers were obtained from individuals with no known exposure to either La (e.g. in form of La carbonate) or Gd-based contrast agents to understand background levels in bone. The ex vivo approach was calibrated using phantoms and achieved a MDL of 0.4 μg g⁻¹ La and 0.5 μg g⁻¹ Gd, but the results for all samples were below these limits. In order to consider future work with live subjects, by correcting for the expected tissue density, the in vivo MDLs were estimated to be 3.9 μg g⁻¹ La and 5.3 μg g⁻¹ Gd which could be suitable for long term exposure studies for patients exposed to La or Gd containing drugs. Furthermore, levels of Ba and I were estimated, finding a mean of 3.4 ± 0.8 μg g⁻¹ for Ba and, interestingly, a large range of I concentrations across the patients which was due to the use of I-based contrast agents in four individuals: unexposed mean −0.5 ± 0.3 μg g⁻¹ and exposed 27.8 ± 28.4 μg g⁻¹ (note negative values are often obtained due to effects from processing the spectra and background overcorrection). An unusual application of XRF spectrometry was presented by Izumoto and co-workers¹⁰⁶ who established Pu could be rapidly determined in blood spots even in the presence of U (up to 500 times higher). In the event of an incident at nuclear facilities, any potential exposure to workers would typically be assessed by detection of α particles in the blood stream, however this work demonstrated a simple and rapid alternative.

The use of SR-μXRF spectrometry was highlighted in two diverse biological applications. Jiang et al.¹⁰⁷ determined the distribution of Ni in skin to investigate Ni-induced dermatitis. Guinea pigs were used as models and were split into various groups based on NiSO₄ concentration and time post-exposure. Analysis of skin tissue cross-sections by SR-μXRF spectrometry showed that Ni increased significantly up to 200 μm deep then decreased at further depth in the early phase group (24 h after 2^nd exposure) compared with the late phase (72 h). By 500 μm depth, the levels had stabilised and were similar between the groups but strongly elevated when compared to controls. Furthermore, the application of μXANES revealed that Ni was not present as Ni²⁺ but bound within proteins. The work provided insight into Ni-induced dermatitis. In an archaeological application, Lorentz et al.¹⁰⁸ analysed ancient hair from the 3^rd millennium BC by SR-μXRF spectrometry. The samples were from a large urban site known for metal working in Iran. The data revealed differences in Cu levels and distribution between the samples which indicated biogenic origin rather than from environmental exposure. The approach could be used to understand more about the elemental levels in other populations from historical hair samples.

A number of papers within this review period have employed PIXE to analyse food, beverages and clinical materials. Debastiani et al.¹⁰⁹ considered the topic of elemental profiles in coffee as a function of water temperature. Drip coffee was prepared using ground beans at various temperatures between 20 °C and 80 °C, with the solutions, ground coffee and spent material analysed by PIXE. Perhaps unsurprisingly, the elemental levels of Ca, Cl, Cu, K, P, Rb and Si increased as the water temperature increased although at different rates and solid-to-solution ratios. In particular, Cl and K had the highest extraction factor at all temperatures. Another work utilising PIXE was published by Kuhn and co-workers¹¹⁰ to compare Himalayan salt with Brazilian salt. The researchers manually separated the “red” grains from the “white” grains in the Himalayan salts and analysed these separately, whereas the Brazilian salts were all white and of marine origin. In general, differences were found for the elements detected (Al, Br, Ca, Fe, K, Mg, Mn, S, Si, Sr and Ti), with the “red” particles containing significantly higher concentrations of Fe, K, Mg, Si and Ti, whereas it was difficult to differentiate between the white grains due to the similarities or large ranges observed. Virk et al.¹¹¹ investigated the effect of acute As exposure on the elemental levels in rat liver tissues by PIXE analysis whilst also examining the soluble protein fractions with centrifugal molecular weight cut off filters. The liver tissue itself was found to contain 6 ppm of As in comparison to the control group which was <1 ppm. The protein fractionation showed that the As was predominantly found in the ≥50 kDa portion. In the As treated group, the concentrations of Ca, Mn and S were found to increase in the 10–50 kDa fraction compared to the control. The work provided an insight into the metabolic process of As but further work is required to characterise these proteins.

5. Nanomaterials

In the context of this update, nanoparticles are considered here where they are used as reagents in experiments, their metabolism or use as tracers and their determination.

5.1 NPs as reagents

Procedures using NPs for sample preparation are proving to be increasingly popular. Er et al.¹⁶ reviewed applications of magnetic NPS for SPE methods, emphasising that their small size and magnetic properties facilitate their use. Practical examples of NPs used for DSPE were reported by the group of Karlidag. In their work ZrNPs were used in assays to measure concentrations of Sb in bergamot and mint teas, and Se in green tea.^65,112

Immunoassays are used extensively in clinical laboratories to determine concentrations of proteins, drugs and many other metabolites. The antibody used in an assay carries a label that can be measured by a suitable technique, such as fluorescence, to give a quantitative result. Recent publications show that NPs may be used as antibody labels and examples of assays where metal-NPs were measured by ICP-MS, have been included in our recent updates. In the report by Modlitbová et al.⁷⁷ the authors describe how LIBS affords an alternative technique for the quantification of these assays. Their work involved determination of HSA in a sandwich immunoassay with streptavidin-coated AgNPs. Reactions were performed using 96-well microtitre plates with LIBS readings made directly from the plate. It is still hard to see these particular applications becoming widely used, but they may have a role to play in niche situations.

5.2 Metabolism of NPs and use as biological tracers

As human exposure to NPs is increasing with their use in industrial processes and domestic, medicinal and food products, knowledge of metabolism and potential toxicity is ever more relevant. To optimise their potential application to cancer therapy Kruszewska et al. set out to determine the intracellular localization and transformation of AuNPs.¹¹³ Using the MCF-7 breast cancer cell lines and NPs of 5, 10, 20 and 50 nm diameter, Au was measured, by ICP-MS, in serum protein, cell membranes, cytosol and nuclei, to investigate transfer of NPs into cell compartments. Less than 5% of the Au entered the cells from the 5 nm particles compared to 50–60% from the larger NPs. Of the internalised Au, approximately half was with cell membrane and the remaining portion in cytosol, organelles and nuclei. Slightly different distributions were seen with the normal cell line, MCF-10a. Further analyses using CE-ICP-MS, HPLC-ICP-MS and spICP-MS were applied to the cytosol. The AuNP forms were at least 30 nm, and could not be resolved either by HPLC or CE, suggesting a large number and range of species inside the cell. There was no evidence to suggest formation of ionic gold. Fahmy et al.¹¹⁴ elected to investigate the biodistribution and toxicity of CuNPs in a rat model. The CuNPs, prepared and characterised by these workers, were injected into rats and the brains removed 2 days later. Distribution of Cu in parts of the brain was determined by ICP-MS and markers of oxidative stress were increased in the same regions. The injected dose was quite high at 15 mg kg⁻¹ and the i.v. route is non physiological but the results provide a pointer as to where further studies could be directed. In a more straightforward experiment.

Chang and colleagues¹¹⁵ describe how bimetallic AuNPs were used to improve analysis and quantification of circulating tumour cells. Current techniques use markers that exhibit considerable non-specific binding to other cells. The AuNP were coated with leukocyte membranes and conjugated with epithelial cell adhesion molecule (EpCAM) antibodies. These new reagents demonstrate high specificity for EpCAM-positive circulating cancer cells, with low background binding to leukocytes. The bound cancer cells, in blood samples, were separated and Au measured by ICP-MS.

Two studies were aimed at investigating the interactions between NPs and surrounding tissues. The model employed by Matter and colleagues¹¹⁶ involved metal oxide NP hybrids of crystalline CeO₂ and biodegradable ceramic bioglass. Topically applied test materials remained at the application site and were taken into macrophages and accumulated within lipid tissues. The samples were analysed by XRF spectrometry and by ICP-MS. Donahue et al.¹¹⁷ validated a spICP-MS procedure to quantify NP size distributions and aggregation in situ. The authors opine that the technique could be applied to the preparation of colloidally stable NP formulations for bioanalytical assays and nanomedicine.

Compounds of Gd are extensively used for MRI imaging. Santelli et al.¹¹⁸ note that Gd₂O₂SNPs may also be useful for imaging by X-ray computed tomography and photoluminescence imaging, in addition to MRI. In this review, these techniques were used to investigate toxicity, biodistribution and excretion of these NPs following intravenous injection into rats. Administration of NPs up to 400 mg kg⁻¹ appear to be very well tolerated with a long-lasting imaging signal associated with a slow hepatobiliary clearance with more than 60% excreted through the faeces after five months and very little urinary excretion. The NPs accumulate mainly in liver and spleen and to a lesser extent lungs and bones. These results have relevance to recommendations from the European Medicines Agency relating to the use of Gd agents in MRI. Separate in vitro work showed that the NPs are insoluble in pure water and human plasma but corrosion/degradation phenomenon appears in acidic conditions such as in cell lysosomes.

5.3 Measurement of NPs

As already seen, various techniques are used to measure concentrations of NPs in biological samples and in foods. When TiO₂ is added to foods it is identified on packaging as E171 with no indication as to the amount. This prompted Sungur et al.¹¹⁹ to measure concentrations in chewing gums, chocolates and white coloured foods, using ICP-AES. Concentrations of TiO₂ were from 3 to 2400 mg kg⁻¹. The NP diameters were also determined, by TEM, and were found to vary from 30 to 310 nm. Because of their antimicrobial properties AgNPs are also widely included in consumer products with inevitable incorporation into tissues and cells. Zheng et al.⁶⁰ elaborated a system to isolate 16HBE (bronchial epithelium cultured) cells onto a polydimethylsiloxane (PDMS) microwell array. As many as 60% of the microwells contained a single cell, from which measurement of Ag was achieved by LA-ICP-MS. Concentrations were from 0.80 to 383 fg per cell with a log-normal distribution. The results demonstrate the potential for single cell analysis to the study of the biological effects of metal NPs.

6 Applications: clinical and biological materials

6.1 Metallomics

A comprehensive review of recent advances in inorganic speciation complements the work with clinical and biological materials, foods and beverages covered in this update.⁴ Discussion of work relating to NPs is included in Section 5. In our previous ASU reviews, much of the work considered in this section addressed advances in speciation techniques, especially for As, Hg and Se. During the last year, it appears that more of the work is aimed at metabolism, employing some of the previously described analytical techniques.

A multi-institutional project to investigate Se metabolism in turkey liver described, among other Se species, what the authors quaintly call “selenosugar-decorated proteins”.¹²⁰ Turkey poults were defined as Se-deficient, adequate-Se or high-Se, after 28 days feeding diets with 0, 0.4 μg g⁻¹ and 5 μg g⁻¹ Se (as Na₂SO₃). Total Se concentrations in the liver were then approximately 0.2 μg g⁻¹, 2 μg g⁻¹ and 13 μg g⁻¹. Se-containing species were identified using HPLC-ICP-MS and HPLC-ESI-MS/MS. No SeMet was found in any samples. The level of SeCys in the high-Se group was twice that in the adequate-Se group despite the much larger difference in total Se. These results indicate that turkeys cannot synthesize SeMet from iSe and that most of the increase in total Se is not due to synthesis of selenoproteins. The bulk of the Se was found to be present as low M_r species (glutathione-, cysteine- and methyl-conjugates of the selenosugar seleno-N-acetyl galactosamine), together with general proteins decorated with selenosugars, via mixed-disulfide bonds. These selenosugar-decorated proteins constituted around 50% of the water-soluble Se.

A second Se project, by Takahashi and Ogra¹²¹focused on biliary excretion and enterohepatic circulation. In this work, rats were given 10 μg Se by i.v. injection, as iSe, SeMet or SeCN. The bile duct was previously cannulated and the bile collected for 60 minutes. In a separate experiment, rats received 10 μg ⁸²iSe by i.v. injection. Bile and urine were collected for 24 h. Using HPLC-ICP-MS and HPLC-ESI-MS, a biliary metabolite was identified as selenodiglutathione. The amount formed from SeMet was much less than from iSe and SeCN implying that there is little metabolism of the organic species via this route. From the experiment using ⁸²iSe around 20% and 60% of the administered dose was recovered in the bile and urine, respectively. Sham-operated rats similarly given ⁸²iSe, excreted 84% of the dose in the urine. With comparable total recoveries from the experimental and sham-operated rats it is concluded that biliary excreted Se is reabsorbed in the gut, re-metabolised and finally excreted in the urine.

A small number of plants have been found to be Se hyperaccumulators. Neptunia amplexicaulis is a leguminous plant found in Central Queensland, Australia and concentrations of Se in the leaf, in excess of 4000 μg g⁻¹, have been reported. Harvey et al.¹²² used XRF microscopy, SEM, LC-MS and SR-XAS to determine the Se distribution throughout the plant and the elemental speciation. The Se, as SeMet and MeSeCys was mainly in young leaves, flowers, pods and taproot with highest concentrations, up to 13 600 μg g⁻¹ total Se in young leaves. In a related species, Neptunia gracilis, the distribution of Se was similar but concentrations were very much lower.

Returning to advances in speciation techniques Solovyev and colleagues⁵⁷ established a procedure to accurately quantify the serum Cu-containing protein, caeruloplasmin. The widely used immunoassays detect the biologically active holoprotein and also the apoprotein, and can overestimate the caeruloplasmin. Solovyev et al.⁵⁷ developed a procedure with Q-STAT strong AEC-ICP-MS/MS which was used to examine samples from patients with Wilson’s disease and healthy controls and also to evaluate methods to measure the so-called exchangeable or non-caeruloplasmin-bound Cu fraction in serum.

6.2 Imaging with MS and X-Rays

The impact of a new LA cell optimised for bio-imaging (NWR Image 266 nm LA system with a TwoVol2 chamber combined with a duel concentric injector in the torch) was described by Greenhalgh et al.¹²³ which revealed potential matrix effects when analysing lung cancer cell models treated with cisplatin. A correlation was observed between the Pt signal from LA-ICP-MS and carbon. Through the utilisation of multiple complementary techniques (ion beam analysis, PIXE and elastic backscattering spectrometry), it was concluded that the enhancement was likely due to the presence of carbon deposits and not just matrix effects, which may also impact the biodistribution of cisplatin. The combination of multiple instruments for imaging was a recurring theme during this review period, with other studies taking this approach discussed subsequently.

It is infrequent to feature food products within the imaging section however a paper by Wu et al.¹²⁴analysed rice from the world’s largest Sb mining region by LA-ICP-MS, SR-μXRF and μXANES. The soil from the rice paddies in Xikuangshan, China, was found to contain significant levels of Sb (5.91–322.35 mg kg⁻¹) and to a lesser extent As (0.01–57.21 mg kg⁻¹). Through the combination of analytical techniques, it was shown that both elements were present but at different levels within the plant, following the order of root > shoot > husk > grain. Despite the higher concentration of Sb in the soil, As was found to accumulate in the grain to a much larger degree, revealing the higher level of bioavailability for As. Furthermore, speciation data demonstrated that DMA was present as the major species, representing >60% of the total content. The use of LA-ICP-MS and SR-μXRF showed that As particularly concentrated in the rice husk, bran and embryo and, whilst Sb followed a similar pattern, it was not found in the endosperm of the rice grain. The work provided an insight into the relationship of As and Sb in food products grown in close proximity to mining activities. Li et al.⁹³ also focussed on rice but in Hg-contaminated areas. The ability of Se to reduce the amount of MeHg in rice grains was shown in previous lab scale experiments, so in this work, the effect of Se supplementation in rice paddies from these regions was investigated. Total levels of Se and Hg, along with other nutritional elements such as Cu, Fe and Zn, were determined by CV-AFS and ICP-MS. The distribution within rice grains was then assessed by SR-XRF spectrometry, finding that Se significantly reduced the total Hg and MeHg concentrations in rice. Specifically, Hg decreased in the embryo and endosperm whilst levels of Fe, Cu, Zn and Se increased in the grains and embryos. The work provided evidence of the actions of Se in Hg-contaminated areas which could be used for food production.

A study by Jagielska et al.¹²⁵ focused on sample preparation procedures for imaging analysis of tissues. Chicken livers were used as a model sample to investigate the effects of different drying techniques prior to LA-ICP-MS analysis. Unprocessed tissue slices were compared against different drying methods, namely (i) oven at 50 °C until constant mass, (ii) freeze-dried at −40 °C and 0.12 mbar for 3 or (iii) freeze-dried at −40 °C and 0.12 mbar for 17 hours. Samples were also digested for total analysis by solution ICP-MS. A number of biologically relevant elements (Ca, Cl, Cu, Fe, K, Mg, Mn, Na, P, S, Sr and Zn) were determined, finding that Ca, Cu, Mg, Mn, P, Sr, S and Zn decreased in concentration whereas Cl, K and Na were enhanced in the freeze-dried tissue compared with unprocessed. Similar trends were found for the oven dried samples but to a lesser extent. It was suggested that Cl, K and Na were more mobile, migrating to the surface during moisture removal, leading to higher signals when measured by LA-ICP-MS. Furthermore, elemental losses were found if homogenisation was applied prior to drying and ICP-MS analysis. The authors concluded that the drying of tissues can significantly affect the elemental levels which potentially creates difficulties to compare studies where different processing techniques were applied. However, the question of how to resolve this was left unanswered.

Within this review period, a significant number of papers featuring alternative MS techniques for imaging applications were published, providing valuable data in a number of biological applications. Pour et al.¹²⁶ used TOF-SIMS to determine Zn distribution and concentration changes in the hippocampus following traumatic brain injury in rat models. An additional benefit of TOF-SIMS was the ability to monitor organic ions such as the phosphatidylcholine head group (C₅H₁₅PNO₄⁺) and cholesterol (C₂₇H₄₅⁺). Regions of interest were further scanned with a resolution of 2 μm at the white/grey matter interface. The workers ascertained that ZnOH₃⁺ and ZnO₂H⁺ were the predominant species observed, both of which increased in the injured group, although no changes to the organic species were found. Comparisons were made with conventional fluorescence staining using fluozin-3 which showed increases in free Zn within the injured zone but in different regions to those identified from the TOF-SIMS analysis. It was concluded that fluorescence staining only reacts with free Zn whereas TOF-SIMS was more likely to determine the protein bound Zn due to the difficulty to ionise Zn²⁺ in its free state by TOF-SIMS. The combination of both analytical approaches has shown differences in the mechanism of Zn metabolism and mobility in response to brain injury which appeared to support the theory that free vesicular Zn accumulation was not the only mechanism for Zn toxicity but that Zn bound proteins may also play a role. This complementary approach could be useful to provide data for other forms of central nervous damage such as from seizures, ischemia, and strokes. Massonnet and Heeran⁸ provided a concise review of TOF-SIMS. Although mainly for molecular and cellular imaging, mention was made to the applicability for inorganic analysis, for example, imaging of the heart for Na and K distribution was highlighted. Andersen and co-workers¹²⁷ published the application of MALDI MS imaging for Zn determination in prostate tissues as a biomarker for cancer detection. Prostate epithelial cells are known to produce high levels of Zn and citrate which decrease with cancer development, therefore offering a potential diagnostic tool and insight into the disease progression/metabolism. As with TOF-SIMS, the use of MALDI MS enabled the detection of molecules and in this work, ZnCl₃⁻ and citrate, alongside aspartate and N-acetylaspartate as citrate precursors, were measured in healthy and prostate cancer tissues. Complementary analysis by LA-ICP-MS confirmed the Zn results. The study demonstrated correlation between Zn, citrate and aspartate, additionally verifying the reduced levels in prostate cancer tissues compared to healthy samples. The approach provided a novel application of organic MS imaging for Zn however the authors noted that other tissue types would be more challenging to analyse due to the lower concentrations of Zn levels. Ender et al.¹²⁸ applied nanoSIMS to the investigation of As in brown seaweed (Laminaria digitata), a known As hyperaccumulator to further understand the distribution of As at the cellular level. This was combined with TEM for identification of the cellular structures and HPLC-ICP-MS for species identification. The nanoSIMS technique achieved a resolution of 300 nm and was able to detect additional elements: carbon (¹²C₂⁻), nitrogen (¹²C¹⁴N⁻), phosphorus (³¹P⁻), sulfur (³²S⁻) and arsenic (⁷⁵As⁻). The total As mass fraction was 117 mg kg⁻¹ with 53% identified as iAs, 32% as arsenosugars and 1.5% as arsenolipids via solution-based ICP-MS. NanoSIMS imaging found the majority of As was distributed in the cell wall and cell membrane with virtually none inside the cell. This was considered unexpected given the high level of hydrophilic species identified and raised questions regarding the speciation analysis. It was postulated that the iAs and arsenosugars may be bound to polymeric carbohydrate alginates or fucoidans in the seaweed which requires further investigation. The work of He et al.¹²⁹ provided an insight into the formation process of collagen IV (a basement membrane protein) utilising nanoSIMS. As the process is mediated by bromide and peroxidasin to form the sulfilimine cross-links between methionine and hydroxylysine, Br could be used to track this reaction. Kidney tissues from mice and human subjects (controls and nephropathy patients) were imaged via a number of secondary ions: ¹²C⁻ or ¹³C⁻, ¹²C¹⁴N⁻ or ¹³C¹⁴N⁻, ³¹P⁻, ³²S⁻, ⁷⁹Br⁻, ⁸¹Br⁻, ¹²⁷I⁻. The results demonstrated the location of the processes in the basement membranes and also the role of Br. By comparing relative Br levels, thickening of the glomerular basement membrane was confirmed in diabetic nephropathy patients, with enhanced levels of Br present. Although I was measured, it did not appear to show any changes or differences between healthy and diseased tissues. In combination with LC-MS/MS, it was also found that bromination of a tyrosine residue in collagen IV was possible which may be a potential explanation for increased levels of bromotyrosine commonly observed in the urine of diabetic patients.

The use of synchrotron-based XRF for bioimaging of brain tissues at the nanoscale in Parkinson’s disease models was reported by Lemelle and co-workers.¹³⁰ The nano-imaging beam line at the European Synchrotron Radiation Facility was utilised to determine the elemental levels in the substantia nigra brain region. The workers applied a semi-quantitative calibration approach to estimate Ca, Fe, P, S and Zn in the acquired images from control rat brains and those overexpressing alpha-synuclein to model Parkinson’s disease. By scanning in “fine mode” at 50 × 50 nm², with some regions analysed further in 25 nm steps, it was possible to gain substantial information on the elemental levels within the cell structures. It was found that Fe and S were increased in the cytoplasmic granules compared to the cytoplasm but Fe was markedly increased in the disease models, further confirming the dysregulation of Fe homeostasis at the organelle level. Gherase and Fleming¹³¹ have reviewed the role of synchrotron-based analysis via XRF and XAS specifically for the bioimaging of human tissues. The article included an overview of the techniques then examined published works covering (i) cell biology and cancer; (ii) brain and the nervous system, (iii) bone, teeth and internal organs and (iv) hair, skin and nails. With over 100 references, it is a useful resource for human studies utilising these methods.

6.3 Trace elements in pregnancy

Within this review period, a number of articles investigating the clinical utility of monitoring trace elements and heavy metals during pregnancy were published. Multi-element studies in placental tissue¹³² and amniotic fluid¹³³ demonstrated associations between trace element concentrations and risks of foetal neural tube defects (NTD). Yin et al. measured concentrations (at ng g⁻¹ or μg g⁻¹ dw placental tissue) of six trace elements (Co, Fe, Mn, Mo, Se and Zn) in the placental tissue from 408 foetuses with NTD and 593 non-malformed foetuses. Inductively coupled plasma-OES was used to measure Fe and Zn concentrations while the levels of Co, Mn, Mo and Se were determined by ICP-MS. Multivariable logistic regression identified associations between high levels of Mn, Se and Zn with increased risk of NTD, whilst higher Co concentrations correlated with decreased risk of NTD. Ovayolu et al. evaluated concentrations of Al, As, Cd, Co, Cr, Cu, Hg, Mn, Mo, Ni, Pb, Sb, Sn and Zn in amniotic fluid samples from a much smaller cohort of pregnant women (n = 75). In contrast to results reported by Yin et al. in placental tissue, Ovayolu et al. observed possible associations of low levels of Mo and Zn and high levels of Al, Hg, Sb and Sn with increased risk of NTD. No significant differences were reported for As, Cd, Co, Cr, Cu, Mn, Ni and Pb between affected (n = 36) and unaffected (n = 39) groups.

Three studies reported associations between elemental concentrations during pregnancy and birth outcomes such as decreased gestational age and weight/length. Increased prenatal U exposure was linked to decreased gestational age and increased risk of preterm birth in a large study including 8500 women in Wuhan, China.¹³⁴ Creatinine corrected U concentrations were determined using ICP-MS analysis of spot urine samples and linear regression models were used to identify associations with continuous birth outcomes. Although increased U concentrations were associated with adverse birth outcomes, the group were unable to report specific exposure windows at which mothers would be at risk. Howe et al.¹³⁵ reported an association between prenatal exposure to As and lower birth weights in a lower income Hispanic population in Los Angeles. This finding correlates with a study in Chinese women included in our last update,¹ however it was only observed for As measured in hair samples (n = 167). The study did not reveal any significant correlation between blood (n = 176) or urine (n = 296) As levels in pregnancy and birth weight. Lower infant birth length was also correlated to prenatal urinary I levels <150 μg L⁻¹ in a prospective cohort of 214 pregnant women in Rio di Janeiro, Brazil.¹³⁶

In addition to looking at neonatal outcomes, de Morais et al.¹³⁶ demonstrated that a urinary I level of less than 250 μg L⁻¹ is an independent risk factor for gestational diabetes mellitus (GDM) and hypertensive disorders of pregnancy (HDP). Median urinary I concentrations were determined from analysis of six spot urines for each participant. The association between Sb exposure during pregnancy and GDM was reported in our last ASU.¹ Complementary to previous work, Wang et al.¹³⁷ have shown correlations between higher prenatal levels of Ni and Sb and increased risk of GDM in a study of 2090 women in Tongji, China. Levels of As, Cd, Co and V were also evaluated by Wang et al.¹³⁷ and Bayesian kernel machine regression identified a positive joint effect of the six metals on the risk of GDM. Positive associations between Sb exposure and increased GDM risk were also reported in a study of 1789 pregnant women enrolled in the Birth Cohort Study on Prenatal Environments and Offspring Health in Guangzhou, China.¹³⁸

6.4 Elements as tags for indirect determinations

Spectrometric elemental analysis is increasingly being investigated as an alternative to conventional readout methods such as fluorescence, chemiluminescence and electrochemiluminescence which have long since been the mainstay of routine clinical immunoassays. Jiang et al.¹³⁹ describe an element tagged immunoassay for quantitative detection of the cancer biomarker carcinoembryonic antigen, which has been evaluated according to the Clinical Laboratory and Standards Institute. Target antigen first binds to biotinylated antibodies (Ab) captured on streptavidin coated magnetic microspheres. After binding of secondary Eu labelled Ab conjugates and subsequent wash steps, Eu is hydrolysed from the complex with 1% HNO₃ and quantified using ICP-MS. Analysis of 469 clinical samples showed good correlation with a commercial electrochemiluminescence immunoassay (ECLIA), however it should be noted that the precision and reporting range of the proposed assay are inferior to currently available commercial options. Despite the rapid analysis time, this method more closely reflects manual ELISA methodologies previously superseded by automated immunoassays: therefore, practicalities would currently preclude its implementation into routine clinical use.

Modlitbová et al.⁷⁷ successfully used LIBS as a novel readout method for nanoparticle-based immunoassays. In the method, termed “Tag-LIBS”, fluorescein isothiocyanate–biotin–Ab conjugates were incubated with HSA previously captured on an Ab-coated microtiter plate. Streptavidin coated AgNP were then added, after which LIBS scanning was performed in the dry state. Human serum albumin was quantified down to a LOD of 10 ng mL⁻¹. Total analysis time for the assay is in excess of 4 hs due to long incubation times: however, this paper provides proof of concept for further development. The authors highlight the potential for automation of the readout platform and expect Tag-LIBS to be used for detection of many different types of biomarkers in NP-based immunoassays.

In addition to emerging elemental tagging and Tag-LIBS, Hu et al.⁵⁸ reported a tag-free methodology for biosensing of miRNA using intrinsic isotope detection. Biotin-modified hairpin DNA (ON1), bound to the surface of streptavidin coated microtiter plates, recognises target miRNA. Two sequential hybridisation chain reactions (HCR) are then initiated incorporating hairpin nucleotides ON2, ON3, ON4 and ON5 to generate a double HCR amplified bilevel DNA structure. Intrinsic ³¹P isotopes are quantified using ICP-MS following acid digestion. Analysis of two cancer cell lines demonstrated good agreement with previous data for miRNA content and assay specificity was demonstrated by anti-interference studies incorporating non-target miRNA.

6.5 Multielement applications

6.5.1 Specimens analysed to investigate metallic implants and biomaterials. Compared to the number of publications in recent years there are few of note in the current review period. Blood and serum Co and Cr measurements are established for biomonitoring in patients with Co and Cr-containing implants where some deteriorate in situ, with release into surrounding tissue and a potential for toxicity. Swiatkowska et al.¹⁴⁰ propose similar markers for monitoring patients with implants composed of, or include a component containing Ti. From a series of 95 patients, who had received the same design of hip implant inserted by the same surgeon, blood and plasma Ti concentrations were measured by HR-ICP-MS. From their results, the reference limits for blood and plasma Ti were proposed as below 2.20 and 2.56 μg L⁻¹, respectively, for both males and females.

Last year’s ASU¹ included a report concerning titanium-dental implants and peri-implantitis. Concentrations of Ti in the inflammatory tissue were much greater than in control samples although it was not possible to say if Ti was a cause or a consequence of the inflammation. Using a mini pig maxilla model, He et al.¹⁴¹ investigated release of Ti and Zr from dental implants. Samples of bone tissue adjacent to the implants were analysed by ICP-AES and ICP-MS 12 weeks after implantation. The mean Ti concentration, 1.67 mg kg⁻¹ is similar to that reported last year (1.2 ± 0.9) while the mean Zr concentration was 0.59 mg kg⁻¹. Additional serial measurements by LA-ICP-MS were taken from immediately adjacent to an implant to 4 mm away. These showed highest levels close to the top of the screw thread. Separate cell culture experiments found evidence for DNA damage from TiNPs and ZrO₂NPs. However, the EC₅₀ concentrations were very high at 13.96 and 810 mg mL⁻¹ respectively.

6.5.2 Biological fluids and tissues. This section of the review begins with the long-standing health issue of obesity, considering adipose tissue as a potential target for obesogenic pollutants such as toxic metals. Freire et al.¹⁴² presented the first exploratory study into the elemental composition of adipose tissue, identifying Ni, Pb, Sn and Ti in all 228 samples from an adult study population in Southern Spain. Arsenic was only identified in 51% samples from the same cohort. Evaluation of socio-demographic data showed associations with age, geographical area of residence, social class, dietary habits and occupation history and the levels of different elements observed, whilst smoking did not appear to have a strong predictive effect. The authors noted limitations of the study, including a hospital-based design and insufficiencies in the questionnaire used to collect exposure data. Nevertheless, this study is likely to be the first of many investigating exposure patterns and metal concentrations in obesity and obesity-related diseases.

Isotopic signatures of endogenous elements have gained recent attention as potential biomarkers for cancer and neurodegenerative disorders. After previously looking at the cellular level, Paredes et al.⁶³ reported isotopic analysis of Cu, U and Zn in subcellular protein fractions of neuronal cells. Protein fractions were separated using SEC from neuron-like cells previously exposed to 10 μM U. Isotope ratios were determined from the protein fractions by MC-ICP-MS. Copper isotopic signatures differed between the four main protein fractions, demonstrating a potential new strategy to investigate Cu metabolic processes. However, the authors accept that further analytical developments are required, as isotope ratios for U and Zn could not be accurately determined.

Moving away from the subcellular level, elemental analysis of brain tissue remains an active area of research for Alzheimer's disease (AD). This year Scholefield et al.¹⁴³ have added to the evidence base by analysing the effects of age, sex, Braak stage, brain weight and post-mortem delay (PMD) on the elemental composition of brain tissue. Concentrations of Ca, Cu, Fe, K, Mg, Mn, Na, Se and Zn were measured using ICP-MS in cingulate gyrus (18 AD and 18 controls) from brain banks in Manchester, UK and Auckland, New Zealand. The same nine elements were also measured in the cerebral cortex, cerebellum and hippocampus of forty healthy adult Wistar Han rats with varying PMDs (0, 24, 48 and 72 h). Although no specific conclusions were reached regarding elemental composition and AD, the report concludes no impact on metal concentrations from pre-analytical variables such as age, sex, Braak stage, brain weight and PMD. Concentrations of Ca, Fe, K, Mg, Mn, Se and Zn were found to vary across different brain regions.

Midander et al.¹⁴⁴ provided a proof of concept method to investigate how metals permeate the skin. Exposure to Co, Cr and Ni was monitored by ICP-MS analysis of receptor medium from a liquid permeation test cell. Artificial sweat spiked with 2 mmol L⁻¹ Co, Cr and Ni, full thickness skin from stillborn piglets and phosphate buffered saline were used as the donor medium, biological membrane and receptor medium respectively. Recirculation of receptor medium between the diffusion cell and the mass spectrometer enables continuous monitoring of skin permeation in multi-element analysis. Steady increases in all three metals were observed, with higher increases for Ni than for Co and Cr. Absorption of Ni was lower in co-exposure experiments with Co and Cr than in single exposure.

Complementary to their own previous work, Stojsavljevic et al.¹⁴⁵ reported on the elemental content of thyroid tissue in patients with Hashimoto’s thyroiditis. Inductively coupled plasma mass spectrometry was used to determine concentrations of As, Cd, Cu, Mn, Ni, Pb, Se, U and Zn in blood, urine and thyroid tissue samples from 51 female patients and 52 healthy female donors. Arsenic and Pb levels were elevated in thyroid tissue and blood but decreased in urine from patients while the converse was true for Se. This is consistent with the hypothesis that thyroid tissue retention of As and Pb in Hashimoto’s thyroiditis in turn causes extrusion of Se from thyroid tissue.

Table 2 summarises this year contributions in the area of clinical and biological material analysis.

Table 2 Clinical and biological materials

Element	Matrix	Technique	Sample treatment/comments	Ref.
Ag	Serum	LIBS	A novel LIBS method for detection of streptavidin-coated AgNPs in a sandwich immunoassay for serum albumin. The LOD was 10 ng mL^−1	77
Ag	Human bronchial epithelial cells (16HBE)	LA-ICP-MS	Following trapping of single cells by a PDMS microwell array, LA-ICP-MS was used to quantify AgNPs in each cell with a high throughput. The average single cell Ag content was in good agreement with results from bulk cell digestion. The LOD and LOQ were 0.2 and 0.7 fg cell^−1 respectively	60
Al	Blood	LIBS	A study to determine half-lives of inorganic–organic hybrid nanomaterials, MnO2–BSA and AlO(OH)–BSA, in mice using LIBS detection. The half-lives calculated were within 5% of those obtained by ICP-MS, confirming the validity of the approach	78
As	Blood	HG-AFS	In a rapid method for ultra-trace detection of As, blood was first extracted in 3% HNO3 before the supernatant was mixed with 5 g L^−1 KBH4 in 1.5 g L^−1 KOH in a modified gas liquid separator. Samples were then introduced by HG into an in situ DBD trap AFS set up. Analytical figures of merit were an LOD of 7 pg, an RSD of 3.9% (n = 11) and spiked recoveries of 96 to 107%. The method was validated against microwave-aided digestion and ICP-MS	50
As	Blood, urine and hair	ICP-MS, LC-ICP-MS	In a low income Hispanic cohort, a doubling of hair As content at 14 ± 4 weeks gestation was associated with lower birthweights. The trend was reflected in blood but not in urine. The association between urine MMA and gestational age at birth was found to be influenced by pre-pregnancy BMI	135
As	Urine (CRM GBW09115) and seawater (CRM GBW(E)080231)	LC-HG-AFS	A method to determine four As species in samples containing up to 10 g L^−1 of Cl utilised two coupled AEC columns and elution with 35 mmol L^−1 (NH4)2HPO4 at pH 6.0 to eliminate Cl interference	90
As	Breast milk	ICP-MS, LC-ICP-MS, LC-ESI-MS	Breast milk samples (n = 15) from a single mother, collected over a 3 day period after consumption of a salmon meal, showed peak concentrations of 1.72 μg kg^−1 total As and 0.45 μg kg^−1 arsenolipids at 8 h post meal. The arsenolipids were demonstrated to consist of mainly As hydrocarbons and AB accounted for the majority of the remaining As in the milk	200
Au	HeLa cells and RAW 264.7 macrophages	spICP-MS	A novel oil-free microfluidic sample introduction system for use with spICP-MS for the determination of AuNP in single cells. The approach achieved a single cell detection efficiency of 70% and an LOD of 1.42 μg L^−1	51
Au	MCF-7 cells	ICP-MS, spICP-MS, CE-ICP-MS, LC-ICP-MS	Different ICP-MS couplings and techniques were employed to characterise the distribution and chemical nature of AgNPs within human tumour cells	113
Au	MCF-7 cells	ICP-MS	A method using novel biomimetic immunomagnetic Au hybrid NPs with ICP-MS detection for quantification of circulating tumour cells	115
Br	Dried serum spots	LA-ICP-MS	Dried serum spots made from 5 μL of diluted serum were analysed for Br and I by LA-ICP-MS using a non-matrix matched calibration. The LODs achieved were 0.23 and 0.03 mg L^−1 for Br and I respectively and RSDs were <10%. Results were within 81.5 to 118% of those obtained by digestion followed by ICP-MS using matrix matched standards	201
Br	Nails	ICP-MS	Three methods for preparation of nail samples prior to ICP-MS determination of Br and I were compared. The final microwave induced combustion method efficiently digested sample masses of up to 100 mg and eliminated interferences relating to C content and memory effects. Recoveries ranged from 93% to 102%, and the RSD was <8%	202
Br	Human and mouse kidneys	NanoSIMS	Enrichment of ^79Br and ^81Br was observed in basement membranes of human and mouse kidneys by nanoSIMS. In peroxidasin knockout mice, there was an 85% reduction in Br enrichment, indicating peroxidasin activity is largely restricted to basement membranes in mammalian tissues	129
Ca	Tooth cementum	TOF-SIMS	Imaging of Ca content of tooth cementum from a single subject using TOF-SIMS, revealed low Ca intensities at growth layers corresponding to six pregnancies	203
Cd	Blood	ICP-MS	Measurement of blood Cd, Hg and Pb concentrations in a child cohort (n = 295, aged 9 to 11 years) revealed significantly higher blood Pb and Hg concentrations in African American children compared with European Americans. This finding was attributed to differences in diet	37
Cd	Serum	ICP-MS	A study comparing two approaches to estimating dietary Cd intake in 51 Italian subjects; a food frequency questionnaire with quantification of Cd in food and beverages vs. a model for estimation of dietary Cd intake from serum Cd concentrations. The two methods showed some agreement in certain sub-groups	204
Cl	Sweat	ICP-MS	An ICP-MS method for sweat Cl determination in the routine clinical laboratory. The RSD was <5% and accuracy was 99.7 to 103.8%. The LOD and LOQ were 3.2 and 7.0 mmol L^−1 respectively. Results were in agreement with those obtained by coulometric titration	147
Co	Blood	ICP-MS	No correlation was found between blood Co and Cr concentrations and physical activity parameters in a cohort of patients with MoM hip prostheses (n = 62)	205
Co	Hair CRM, mussel CRM, stream sediment CRM and water	ICP-MS	A new adsorbent, consisting of fibrous TiO2@g-C3N4 nanocomposites, was used for dispersive micro-SPE of Co and Ni prior to their determination by ICP-MS. Following optimisation of the sample pre-treatment, an enrichment factor of 100 and LODs of 0.12 and 1.34 pg mL^−1 for Co and Ni, respectively were achieved	188
Co	Phosphate buffered saline and skin	ICP-MS	An experimental set up to investigate skin penetration of Co, Cr and Ni ions demonstrated that absorption of Ni was faster in single rather than combined metal exposure but that this was not the case for Co or Cr. Analysis of the skin barrier post exposure showed that similar amounts of all three metals were retained in skin	144
Cr	Blood	ICP-MS	See Co, ref. 205	205
Cr	Phosphate buffered saline and skin	ICP-MS	See Co, ref. 144	144
Cr	Colon tissues	LIBS	Cancerous and non-cancerous colon tissues were analysed by LIBS for Cr, Hg and Pb. The heavy metals were detected in the cancerous tissue at concentrations of 13.4 (Cr), 7.1 (Hg) and 3.1 (Pb) μg L^−1. The accuracy of the LIBS method was validated against ICP-OES	79
Cu	Serum	LC-ICP-QQQMS	A strong AEC-ICP-QQQMS method for quantification of caeruloplasmin achieved an LOD of 0.1 μg L^−1 and an LOQ of 0.4 μg L^−1. The method was applied to investigate the accuracy of an exchangeable Cu method based on ultrafiltration	57
Cu	Serum, cell homogenates and water	ICP-MS	A method for determination of Cu^I in the presence of ten-fold higher Cu^II concentrations involved deproteinisation of the sample with trichloroacetic acid before mixing of the supernatant with glycine–NaOH buffer (pH 12.5) and addition of 0.05% 2,2′-biquinoline in N-pentanol. Following separation and evaporation of the organic layer, the residue was digested in HNO3 and H2O2 prior to ICP-MS analysis. The LOD was 0.04 μg L^−1, RSDs were <5% and recoveries were 95 to 102%. When applied to serum of patients with cervical cancer, concentrations of Cu^I were higher than in controls	194
Cu	Urine	ICP-MS	Minimal Cu concentration bias was demonstrated between 24 h urine samples collected into plain and acid washed containers, confirming that acid washed containers are not required for urine Cu measurement	39
Cu	Urine, artificial sweat, dialysis fluid and water(CRMs TMDA 53.3 and TMDA 64.2)	FAAS	A supramolecular solvent-based LPME method for Cu determination in a variety of matrices. The chelating agent was 1-(2-pyridylazo)-2-naphthol and the supramolecular solvent was prepared using THF and decanoic acid. Following optimisation of the procedure, a pre-concentration factor of 40 was achieved and the LOD and LOQ were 7.3 and 24.2 μg L^−1 respectively	67
Cu	Hair	SR-XRF	Distribution of Cu was established by SR-XRF in 10 μm sections of ancient human hair (n = 8) and compared to modern controls (n = 2). Different Cu content was observed between individuals with one young female having high Cu levels located within the hair cortex	108
Cu	Rat brain	ICP-MS	Characterisation of the distribution of CuNPs in brain tissue from rats that had been injected with 15 mg kg^−1 13 nm Cu NPs vs. controls included ICP-MS measurement of total Cu content in different brain regions. Significant increases in brain Cu concentration were found in the study group	114
Cu	Neuron-like human cells	SEC-MC-ICP-MS	Isotope variations of Cu were determined in four protein fractions isolated from lysed cells by way of SEC-MC-ICP-MS	63
Eu	Serum	ICP-MS	Validation of an immunoassay for serum CEA using Eu labelled antibodies and ICP-MS detection. The LOD was 0.83 ng mL^−1 and RSDs ranged from 4.7 to 17.6%	139
Fe	Blood	MC-ICP-MS, AAS	The physiological requirement for Fe in females of reproductive age (n = 61) was calculated by measuring ^58Fe abundance by MC-ICP-MS and total Fe by AAS after the subjects had consumed 50 mg of stable isotope ^58Fe over a period of two weeks. Results obtained for average requirements (11 to 13 mg d^−1) were slightly lower than current recommended values in China	155
Fe	RBCs	MC-ICP-MS	A study measuring the rate of Fe incorporation into RBCs in Chinese children (n = 57). After each subject was given 30 mg ^57Fe over five days, stable isotope ratios were periodically determined in RBCs over 90 days. Peak erythrocyte Fe incorporation was observed at day 60 with a higher incorporation rate observed for girls compared with boys	206
Fe	Hair	FAAS	A switchable solvent based LPE method was developed for the determination of Fe in hair. Following addition of pH 4.0 buffer and 0.015% complexing agent, the switchable solvent, synthesised from N,N-dimethlybenzylamide and dry ice, was added. Prior to analysis by FAAS, addition of NaOH facilitated deprotonation of the switchable solvent. An enhancement factor of 92 was achieved with an LOD and LOQ of 2.6 and 8.6 μg L^−1 respectively. Recoveries of spiked samples were 91 to 113%	64
Fe	Human brain	ICP-MS, SEC-ICP-MS	Distribution of Fe in ten anatomical structures from post mortem human brains (n = 7) with no apparent neuropathology was examined by ICP-MS. Use of SEC-ICP-MS revealed that approximately 60% of the Fe was associated with transferrin	153
Fe	Liver (SRM 1577c bovine liver) and bone marrow-derived macrophages	TXRF GFAAS	A rapid TXRF method for Fe determination with a one point calibration and internal standardisation using Ga and Ti for cell and liver tissue analysis respectively. Recovery experiments for Fe in homogenates gave recoveries of 99.93 ± 0.14% and LODs were reported to be in the low pg range. Results were compared against those obtained by GFAAS	100
Fe	Human neuroblastoma (SH-SY5Y) cells	CE-ICP-MS	Quantitative measurement of Fe^II and Fe^III by CE-ICP-MS in cells using conductivity-pH-stacking to improve peak shapes and reaction gas, NH3, for ICP-MS interference removal. The method achieved an LOD of 3 μg L^−1, RSDs were ≤3.5% and recoveries were 105% for Fe^II and 97% for Fe^III	154
Gd	Blood, plasma and urine	ICP-MS	A reference range study based on 120 healthy subjects not previously exposed to Gd-based contrast agents in the matrices listed. Reference intervals proposed (97.5^th percentile) were: whole blood, <0.050 nmol L^−1; plasma, <0.057 nmol L^−1 and spot urine, <0.025 nmol mmol^−1 creatinine	35
Gd	Human brain	ICP-MS	Analysis of brain autopsy samples (n = 10) from children who had received at least one dose of Gd-based contrast agent revealed highest Gd concentrations in the globus pallidus. Lower levels of Gd were observed in those who had received macrocyclic rather than linear ionic agents. The findings were consistent with those reported in adults	149
Gd	Mouse brain	ICP-MS, SR-XRF	Rats with experimentally induced autoimmune encephalomyelitis (n = 6) had higher brain Gd concentrations with respect to healthy controls (n = 4) following administration of 2.5 mmol kg^−1 Gd-based contrast agent ((mean (SD): 5.3 (1.8) vs. 2.4 (0.6) μg g^−1)). Small hotspots, located in areas of high inflammatory activity, were observed by nanoSR-XRF	148
Hg	Blood	ICP-MS	See Cd, ref. 37	37
Hg	Blood and serum	CV-AAS, ETAAS	Measurement of blood total Hg and serum Se concentrations by CV-AAS and ETAAS respectively in 141 Spanish women of childbearing age found mean concentrations of 2.89 μg L^−1 (Hg) and 73.06 μg L^−1 (Se). Concentrations of Hg were positively associated with fish intake and Se, specifically with tuna intake	33
Hg	Blood and dried blood spots	GC-CV-AFS	Validation of a method for MeHg and iHg was undertaken using 49 paired whole blood and dried blood spot samples. LODs in whole blood and dried blood spots respectively were 0.2 and 0.3 μg L^−1 for MeHg and 1.1 and 1.9 μg L^−1 for iHg. Good agreement between the two matrices was obtained for MeHg but not for iHg	94
Hg	Hair	AAS	Total Hg was measured in hair in a Spanish birth cohort (n = 405) at 4 and 9 years old. Mean Hg at 9 years old was 0.89 μg g^−1 and Hg content fell on average by 22% between 4 and 9 years old	36
Hg	Rat brain extracellular fluid	ICP-MS	An online microdialysis-microfluidic-based photocatalyst-assisted vaporisation device coupled to ICP-MS was developed for the in vivo quantification of Hg in rat brain extracellular fluid after the administration of MeHg and thimerosal. Under optimal conditions, the LODs were 2.7 and 1.7 ng L^−1 for MeHg and thimerosal respectively	207
Hg	Colon tissues	LIBS	See Cr, ref. 79	79
I	Dried serum spots	LA-ICP-MS	See Br, ref. 201	201
I	Urine	ICP-MS	Median urine I, assessed from six spot samples collected during pregnancy, in a cohort of 214 Brazilian women, was 219.7 μg L^−1, with 17.2% and 38.7% having concentrations <150 μg L^−1 and ≥250 μg L^−1 respectively. Urine I ≥ 250 μg L^−1 was found to be an independent risk factor for gestational diabetes and hypertensive disorders of pregnancy while urine I < 150 μg L^−1 was associated with lower infant birth length	136
I	Amniotic fluid, CSF and breast milk	ICP-MS	A rapid ICP-MS method was validated for determination of I in the matrices listed. The LOD and LOQ were 0.233 and 0.778 μg L^−1 respectively, RSDs ranged from 3.2 to 9.2% and recoveries were 97.7 to 109.8%	152
I	Nails	ICP-MS	See Br, ref. 202	202
In	Epithelial cells (16HBE) and macrophages (RAW264.7)	ICP-MS	Concentrations of In were measured in lung cells following treatment with In2O3 NPs. In the epithelial cells, In concentrations were ten-fold higher than in macrophages and serious cell damage was observed	151
La	Serum	ICP-MS	Technique ICP-MS was used as a detector in a sandwich immunoassay, involving a La-labelled monoclonal anti-human IgG1 antibody, to measure serum infliximab. Analytical parameters were an LOD of 0.4 μg mL^−1, measuring range 1.0 to 50 μg mL^−1 and RSDs <10%. Results were compared to those obtained by ELISA and cell-based assays	208
Na	Serum (SRM 956d)	ICP-MS	A serum Na candidate reference method involved 100-fold dilution of serum in 0.3% ultrapure HNO3 prior to ICP-MS detection of Na with Ge as the IS. The total RSD was <0.67% and trueness was assessed using a CRM. Results obtained had a mean deviation of −0.15% vs. the routine indirect ion selective electrode method	27
Ni	Artificial saliva	ICP-MS	There was no difference in Ni ion release into artificial saliva from NiTi orthodontic archwires when coated with AgNPs as a new bactericide treatment vs. non-coated wires	209
Ni	Hair CRM, mussel CRM, stream sediment CRM and water	ICP-MS	See Co, ref. 188	188
Ni	Phosphate buffered saline and skin	ICP-MS	See Cr, ref. 144	144
Ni	Skin tissue	SR-XRF, XANES	Higher concentrations of Ni were measured in the top skin layer in the early sensitisation phase compared with the late stage in Guinea pigs challenged with different concentrations of NiSO4	107
P	HeLa and HepG2 cancer cell lines	HR-ICP-MS	Determination of P by HR-ICP-MS was used to quantify micro-RNA, amplified using a novel concatenated hybridisation chain reaction. The LOD was 13 fM	58
Pb	Blood	ICP-MS	See Cd, ref. 37	37
Pb	Blood	AAS	Mean blood Pb in 6455 participants of a Korean environmental health survey was 19.5 μg L^−1, which represented a decrease from 1999 (45.8 μg L^−1). Smoking and drinking alcohol were associated with higher blood Pb concentrations	32
Pb	Blood and toenails	ICP-MS	Measurement of Pb concentrations in blood and toenails in a cohort of 12 year old children (n = 224) revealed that blood Pb was significantly associated with decreased intelligence but toenail Pb was not	157
Pb	Blood and brain tissue	ICP-MS	A case report of fatal Pb poisoning with Pb encephalopathy caused by retained bullet fragments. Blood Pb was 1650 μg L^−1 at hospital admission and at autopsy, brain tissue Pb was 3.04 μg g^−1 (cortex) and 2.70 μg g^−1 (nucleus lentiformis)	158
Pb	Bone	SF-ICP-MS, MC-ICP-MS	Analysis of acid digested femora from Roman (n = 30) and pre-Roman (n = 70) remains showed 70-fold higher Pb concentrations in the Romans and Pb isotope ratios were indicative of widespread Pb pollution affecting this population	156
Pb	Colon tissues	LIBS	See Cr, ref. 79	79
Pt	Mouse cochlear	ICP-MS	A higher cochlear uptake of Pt was reported in mice given cisplatin compared with those given carboplatin or oxaliplatin, consistent with the increased ototoxicity of this drug	210
Pt	Ovarian cancer SKOV3 cells	ICP-MS	In work to investigate the chemo-sensitivity of this cell line to short term treatment with cisplatin and cisplatin-loaded biodegradable NPs, ICP-MS was utilised to measure intracellular concentrations of Pt. Higher efficacy was observed with higher Pt concentrations for the cisplatin-loaded NPs	159
Pt	Non-small cell lung cancer explant models	LA-ICP-MS	Imaging of cisplatin treated tumours was performed by LA-ICP-MS using a low aerosol dispersion ablation chamber. A correlation was observed between the Pt signal and the presence of C deposits within lung tissue	123
Pu	Blood	XRF	In a method designed for rapid assessment of Pu contamination in wounds, XRF analysis was performed on blood collected onto filter paper. The signal intensity of the Pu L alpha peak was not affected by U at a mass ratio 500 times greater than Pu	106
Sb	Plasma	ICP-MS	In a large Chinese population study (n = 4733), an increased plasma Sb concentration at baseline was associated with higher bilirubin concentrations three to five years later. The plasma Sb concentration was also found to be correlated with elevations in both total and direct bilirubin	146
Sb	Urine	ICP-MS	A study in a large Chinese birth cohort (n = 1789) found increased risk of gestational diabetes and increased blood glucose concentrations in the highest quartile of creatinine corrected urine Sb concentrations. Maternal age was reported to modify the relationship between urine Sb and gestational diabetes	138
Se	Blood and serum	CV-AAS, ETAAS	See Hg, ref. 33	33
Se	Plasma	ICP-MS	Evaluation of a correction equation to eliminate ^156Gd double charged interference on ^78Se. Recoveries of Se in Gd-spiked plasma and QC/EQA samples were 97.4 to 106.5% with RSDs <4%	59
Se	Serum	XRF	Serum (n = 166) from 33 patients with COVID-19 infection had lower Se concentrations compared with a reference population (mean (SD): 50.8 (15.7) vs. 84.4 (23.4) μg L^−1), with 43.4% falling below the 2.5^th percentile of the reference group. The Se concentrations were significantly higher in serum from surviving COVID patients with respect to non-survivors (mean (SD): 53.3 (16.2) vs. 40.8 (8.1) μg L^−1)	160
Se	Serum	SEC-LC-ID-ICP-QQQMS, AEC-ID-ICP-QQQMS	Method for speciation of selenoproteins and selenometabolites based on SEC and affinity chromatography with ID-ICP-QQQMS detection and an AEC-ID-ICP-QQQMS method for speciation of other selenometabolites. The methods were applied to serum from patients with lung cancer and healthy controls	55
Se	Urine and plasma	ICP-MS	Urine Se, corrected for specific gravity, osmolality and creatinine, was assessed against plasma Se as a biomarker of Se status in 741 females of reproductive age and 665 school-age children	162
Se	Urine	LC-ICP-MS	The effect of long term storage at −20 °C of selenosugar 1, the main Se elimination product in human urine, was assessed by simultaneous determination of SeSug1 and methylselenic acid by LC-ICP-MS. Treatment of urine with 0.1% NaN3 at a pH of 5.5 maintained stability of SeSug1 for more than three months	43
Se	Urine	LC-ICP-MS	Three chromatography methods (RP, AEC and CEC) coupled to ICP-MS were employed for the determination of 11 Se species in urine, achieving LODs of between 0.03 and 0.10 μg L^−1. An improved LOD of 0.02 μg L^−1 for trimethyl Se was obtained through use of on-line post column derivatisation. Reported RSDs were 2.7 to 10.6% and recoveries were 87 to 108%	164
Se	Urine and bile	LC-ICP-MS, LC-ESI-Q/TOF MS	A selenometabolite found in bile of rats that had been administered SeMet, SeCN or Se^IV, was identified as selenodiglutathione using the two techniques listed. Further investigation of enterohepatic circulation of Se was performed using Se measurements in urine and bile	121
Se	CSF	LC-ICP-MS	Speciation analysis of Se in CSF from patients with Parkinson’s disease (n = 75) and age-matched healthy controls (n = 68) detected eight different species with no differences observed between the groups. A highly significant correlation between total Se, selenoprotein P and human serum albumin bound Se was demonstrated	161
Se	Turkey liver	LC-ICP-MS, LC-ESI-QQQMS	A six-fold increase in liver Se was observed in turkeys fed a high vs. adequate Se diet and this was shown to consist mainly of low Mr selenometabolites and selenosugars with no SeMet detected	120
Se	Lamb tissues (muscle, heart liver)	LC-ICP-MS, LC-ESI-MS/MS	Five seleno-compounds, including SeMet, SeCys2 and Se-methyl–SeCys, along with Se^IV and Se^VI were separated and quantified by LC-ICP-MS in tissues of lambs fed either an iSe or organic Se enriched diet. Identification of the seleno-compounds was verified by LC-ESI-MS/MS	163
Se	Yeast cells	spICP-QQQMS, LC-ICP-MS	sp-ICP-QQQMS was applied first to determination of Se content of single yeast cells and then, following lysis of the cells, to detection of SeNPs with complementary analysis by LC-ICP-MS	54
Si	Tissues	ICP-MS, spICP-MS	A method for determining Si in tissues involved solubilisation of SiO2 NPs by microwave dissolution in HNO3, H2O2 and HF and ICP-MS analysis with CH4 reaction gas for elimination of spectral interferences. The LODs for Si in most tissues were 0.2 to 0.5 μg g^−1. spICP-MS was used to assess tissue deposition of the SiO2 NPs	53
Ti	Blood and plasma	HR-ICP-MS	A study to determine reference ranges for blood and plasma Ti in patients with a well-functioning MoM hip implant (n = 95). The upper reference range limits derived were 2.20 and 2.56 μg L^−1 for blood and plasma respectively	140
Ti	Simulated body fluid	ICP-MS	Metal ion release of Ti and Zr was assessed for antimicrobial dual coated TiZr implants as part of a wider characterisation study	211
Ti	Mini pig bone tissues (maxillae)	ICP-MS, ICP-OES, LA-ICP-MS	Release of Ti and Zr metal ions from Ti and ZrO2 implants into surrounding bone tissue was investigated in mini pigs and increased concentrations of Ti and Zr were detected near the implants (1.67 and 0.59 mg kg^−1 bone weight respectively). The spatial distribution of the metal ions was assessed by LA-ICP-MS	141
Ti	Mice gut tissues	ICP-MS	Following oral administration of 40 mg kg^−1 of TiO2 in mice, the absorption of TiO2 particles was assessed in gut tissues by ICP-MS and microscopy methods. Results showed that early intestinal uptake of TiO2 particles mainly occurred through the villi of the small intestine	165
Tl	Blood, urine and hair	ICP-MS	A case study on the fatal poisoning of three individuals in the same family reported blood and urine Tl concentrations prior to death much higher than those considered lethal: 3.4 to 10 μg mL^−1 and 16.3 to 42 μg mL^−1 respectively. Hair Tl content at autopsy was 5.72 to 10.38 ng mg^−1	212
U	Urine	SF-ICP-MS	Uranium was quantified using ID-SF-ICP-MS in 24 h urine collections taken as part of a Swiss population study (n = 1393). Median and 95^th percentile U excretions were 15 and 67 ng 24 h^−1 respectively. Place of residence significantly influenced urinary U excretion although did not fully explain the elevated excretion rates observed	31
U	Urine	ICP-MS	Analysis of urine from 8500 pregnant Chinese women gave a mean U concentration of 0.03 μg L^−1. After adjusting for potential confounders, increased urine U was found to be associated with a significant decrease in gestational age of delivery and increase in preterm delivery	134
U	Urine	ICP-MS	Estimation of ^234U activity in 24 h urine collections (n = 105) from occupationally exposed workers was based on measurement of ^235U and ^238U by ICP-MS and average ^234U : ^238U isotopic ratios obtained by alpha spectrometry. However agreement with direct measurement by alpha spectrometry was unsatisfactory	166
Various (4)	Blood	ICP-MS	Chinese children from a waste recycling area (n = 73) had higher blood concentrations of As, Cd, Hg and Pb compared with children from other areas (n = 74)	213
Various (68)	Human cord blood	ICP-MS	A metabolomics study involved correlation of single element concentrations and element ratios in placental cord blood samples (n = 168) with pregnancy outcomes. Ratio P : Cu provided the best diagnostic power (91.7%) in discriminating pre and post term deliveries	214
Various (45)	Plasma	ICP-MS	A Spanish population study (n = 419) measuring 45 elements, including rare earth elements, found median plasma concentrations of 7.7 (Bi), 0.19 (Ce) and 0.16 (Y) ng mL^−1. A positive correlation between a number of elements and advancing age was observed	215
Various (6)	Serum	ICP-MS	Age dependent changes in serum Cu, Fe, I, Mn, Se and Zn over 20 years were investigated in 219 healthy participants. Overall, a decrease in Mn, Se and Zn, and an increase in Cu, Fe and I, were observed with increasing age	30
Various (6)	Serum and bovine serum (SRM 1598a)	ICP-QQQMS	Elements, Cu, Fe, Mn, Ni, Se and Zn, were determined in serum of patients with hepatocellular carcinoma vs. healthy controls by way of ICP-QQQMS with N2O reaction gas. Results were verified against SF-ICP-MS and a CRM was used to validate the accuracy of the method	56
Various (7)	Serum	ICP-MS, MC-ICP-MS	Investigation of potential homeostatic alterations of essential mineral elements, Cu, Fe, K, Mg, Na, P and Zn, and Cu isotope ratios, in serum from 20 patients with age-related macular degeneration and 20 controls, revealed higher serum concentrations of P and Zn and lower δ^66Cu values in the disease group	62
Various (20)	Serum and immune T cells	ICP-MS	Results obtained from an ICP-MS method for 20 elements in serum (10 μL) and 12 elements in sorted immune T cells (<250 000) from naïve and tumour bearing mice indicated a systematic tumour effect on the elemental profiles in both matrices	216
Various (6)	Urine	ICP-MS	The association of creatinine corrected urine As, Cd, Co, Ni, Sb and V in early pregnancy and the risk of developing gestational diabetes was investigated in 2090 Chinese women. While positive associations were found with urine concentrations of As, Co, Ni, Sb and V and risk of gestational diabetes, Ni was shown to have the greatest effect both individually or as part of a metal mixture	137
Various (18)	Urine	ICP-OES	A study comparing internal standardisation (using Ge or Pd) and multi-energy calibration strategies for direct analysis of urine reported recoveries of 80 to 120% and RSDs <7.5% with the internal standardisation approach but poor recoveries and higher LODs with the multi-energy calibration approach	217
Various (19)	Urine and blood	SF-ICP-MS	In an Italian longitudinal study over three years in workers at a waste-energy plant, involving measurement of Pb in blood and 18 elements in urine, a general decrease in concentrations of most elements was seen over time, implying no significant occupational exposure	218
Various (19)	Urine and hair	ICP-MS	Analysis of pooled spot urine samples and hair from 29 pregnant Canadian women, residing in an area of intense natural gas exploitation, found higher Mn concentrations in both urine (0.49 μg L^−1) and hair (0.16 μg g^−1), and higher median hair concentrations for Ba (4.48 μg g^−1), Al (4.37 μg g^−1) and Sr (4.47 μg g^−1) with respect to reference range data	219
Various (37)	Hair	ICP-MS	A washing method for removing external contamination from hair, involving three steps (Triton, HNO3 and HCl), was evaluated in non-spiked and spiked hair samples (n = 10) and was deemed to be more effective than methods based on solvents and surfactants	40
Various (16)	Rat serum, faeces, urine and liver	ICP-MS	Rats fed a low Zn diet exhibited changes in urinary and faecal excretion of 12 of the elements studied compared with rats fed a Zn replete diet. Novel findings were increased faecal Ag excretion, decreased serum Ag concentration and decreased urinary As and Cr	220
Various (9)	Blood, urine and thyroid tissue	ICP-MS	Concentrations of essential elements (Cu, Mn, Se and Zn) and toxic elements (As, Cd, Ni, Pb and U) were measured in patients with Hashimoto’s thyroiditis and healthy controls. In the study group, elevated As and Pb, and lower Se concentrations, were observed in blood and thyroid tissue and urine Se concentrations were increased	145
Various (35)	Serum, ruminal fluid and tissues (liver, muscle, testis)	ICP-MS, LC-MS/MS	As part of a comprehensive study, in which more than 145 metabolites were identified and quantified in various bovine biofluids and tissues, concentrations of 35 elements were determined	221
Various (13)	Breast milk	ICP-MS	Breast milk concentrations of Cu, Fe, K and Zn decreased through the course of lactation in Malaysian Malay mothers (n = 20) who provided samples at three consecutive two week intervals. There was a significant correlation between maternal intake of Fe, K, Na and Se and breast milk concentrations of these elements	222
Various (15)	Nasal mucous	LA-ICP-MS	An LA-ICP-MS method was validated for determining elements associated with gunshot residue in nasal mucous, collected using a novel sampling swab device, and found to be suitable for forensic analysis	61
Various (14)	Amniotic fluid	ICP-MS	Significantly lower mean concentrations of Mo and Zn and higher mean concentrations of Al, Hg, Sb and Sn were measured in amniotic fluid from pregnancies affected by neural tube defects (n = 36) with respect to normal pregnancies matched for gestational age (n = 39)	133
Various (6)	Placental tissue	ICP-OES, ICP-MS	Concentrations of six elements, Co, Fe, Mn, Mo, Se and Zn, were measured in a placental tissue from 408 pregnancies. Higher concentrations of Mn, Se and Zn were found to be associated with an increased risk of neural tube defects, while higher Co concentrations were associated with decreased risk	132
Various (23)	Bovine liver (NIST SRM1577b) and green algae (NIES no. 3 Chlorella)	ICP-OES	A multi-channel type concentric grid nebuliser for efficient online standard addition was validated in the CRMs listed. Signal intensities were two- to three-fold higher compared with a conventional nebuliser	223
Various (12)	Chicken liver	LA-ICP-MS	Three procedures for sample stabilisation (oven drying, short freeze-drying and long freeze-drying) were compared for their suitability for subsequent LA-ICP-MS analysis using chicken liver as a model. Freeze-drying led to reduced content of Ca, Cu, Mg, Mn, P, S, Sr and Zn, whereas Cl, Na and K content was enhanced	125
Various (4)	Rat liver	PIXE	Different Mr post mitochondrial liver protein fractions, separated using ultra centrifugation were analysed with PIXE to map the distribution of elements, As, Ca, Mn and S, in As supplemented rats and controls. Sequestration of As by higher Mr proteins (≥50 KDa) was observed	111
Various (5)	Rat brain	SR-XRF, TEM	A quasi-correlative method for analytical nanoimaging of the Substantia nigra, based on TEM and SR-XRF. Increased Fe was observed in cytoplasmic granules when alpha-synuclein protein, associated with Parkinson’s disease, was overexpressed	130
Various (7)	Mouse brain	TXRF	Elements, Cu, Fe, K, P, Rb, S and Zn, were determined in mouse brains, following intracerebroventricular injection of amyloid beta soluble oligomers at different doses to induce Alzheimer’s disease vs. controls. Differences in elemental concentrations in some brain regions between the treated groups were observed	101
Various (9)	Human brain	ICP-MS	Concentrations of nine essential metals (Ca, Cu, Fe, K, Mg, Mn, Na, Se and Zn) were compared in the cingulate gyrus region of post mortem brains from patients with Alzheimer’s disease and controls. Results were found to be independent of age, sex, delay to post mortem analysis, brain weight or Braak stage	143
Various (7)	Dog tissues	EDXRF	Direct determination, using oyster tissue validated methods, of Co, Cu, Fe, Mg, Mn, Na and Zn, and associated speciation (Cu, Fe and Mn), in malignant and benign soft body tissues. Elemental speciation pattern recognition was reported to discriminate malignant vs. benign tissues to 97% accuracy	104
Various (5)	Adipose tissue	HR-ICP-MS	Metals, As, Ni, Pb, Sn, and Ti, were determined in adipose tissue samples (n = 228) with Ni, Pb, Sn and Ti being detected in all subjects and As in 51%	142
Various (4)	Bone	XRF	Elements, Gd and La, were determined in post mortem human tibiae with LODs of 0.5 and 0.4 μg g^−1 respectively. Concentrations of Ba and I were estimated from the experimental La calibration line and Monte-Carlo derived sensitivity factors	105
Various (8)	Human bones	ICP-MS, CV-AAS	Elemental concentrations of Ba, Ca, Cu, Fe, Mn, Pb and Sr, measured by ICP-MS, and Hg by CV-AAS, in cortical and trabecular femoral tissues (n = 87) from a Medieval and post Medieval burial ground, showed differences between nobles and townspeople	224
Various (25)	Tooth dentine	ICP-MS, AES	Increased concentrations of toxic elements, Li, Pb and Sn, and essential elements, B, Ba, K, Mg S and Sr, with increasing age were observed in 150 human coronal tooth dentine samples. Concentrations of K and Pb in molars and premolars showed the highest variation with age	225
Various (5)	Yeast, RBCs and green algae	sp-ICP-MS	The elemental composition, including Fe, Mg, P, S and Zn, of three types of single cells determined using SP-ICP-MS was compared to bulk digestion and analysis by ICP-MS. Good agreement was seen for yeast and green algae but not for RBCs	52
Zn	Rat lungs, brain, liver, kidney, RBCs and plasma	ICP-MS EDXRF TEM	Following exposure by inhalation at two air concentrations of ZnO NPs, Zn content in tissues was determined by ICP-MS, along with ZnO characterisation using TEM and EDXRF spectrometry. Exposure to the lower ZnO NPs concentration resulted in significant increases of liver Zn, while exposure at the higher level significantly increased Zn concentrations in lungs and brain	226
Zn	Rat brain	TOF-SIMS	A TOF-SIMS method to investigate the distribution of Zn in the hippocampus region of rat brains with and without traumatic brain injury. The two Zn species, [ZnOH3]^+ and [ZnO2H]^+, were visualised and the images were compared to those obtained by Zn probe-based fluorescence spectroscopy	126
Zr	Mini pig bone tissues (maxillae)	ICP-MS, ICP-OES, LA-ICP-MS	See Ti, ref. 141	141

---

6.6. Progress for individual elements

6.6.1 Antimony. We have commented on a large prospective study investigating associations between plasma Sb levels and liver dysfunction in 4733 middle-aged and elderly Chinese adults taken from the Dongfeng-Tongji cohort.¹⁴⁶ Plasma Sb was determined at baseline (2008–2010) by ICP-MS and serum liver enzymes were measured at follow up in 2013. Spline regression demonstrated a linear association between Sb and elevated bilirubin levels, however associations with liver enzymes (ALP, ALT and AST) were not statistically significant. The size of this study appears to be its greatest power, however there are some notable limitations. Samples for liver enzymes and Sb were collected 4 years apart and concentrations were obtained from single measurements despite known variability in Sb measurements. This study is likely to be the first of many investigating exposure to Sb in terms of liver disease risk.

6.6.2 Chlorine. Sweat chloride remains the gold standard diagnostic marker for cystic fibrosis and the use of ICP-MS for sweat chloride analysis is no longer novel. However, Marvelli et al.¹⁴⁷ are the first to publish the use of ICP-MS for analysis of chloride in sweat samples collected via the Gibson and Cooke technique. The method boasts an LOQ of 7.0 mM with linearity up to 224.5 mM and acceptable precision (<6% RSD). Accuracy between 99.7% and 103.8% was achieved by repeated measurements of two commercial quality control samples and one low level control prepared from a certified chlorine solution. Agreement with coulometric titration (r = 0.993, bias −0.9 units) was determined using 50 serum samples from newborns with chloride concentrations 10 to 131 mmol L⁻¹. Whilst this is excellent agreement, the authors do not comment as to why serum rather than sweat samples were used. Quantification of sweat samples from 50 healthy volunteers gave a mean Cl⁻ concentration of 15.7 mmol L⁻¹ and concentrations of 65.6 mmol L⁻¹ and 81.2 mmol L⁻¹ were obtained for two patients with cystic fibrosis. The authors also suggest that reliable quantitation could be obtained on smaller samples (<75 mg), but further studies would be required to confirm sample viability as this limit reflects collection rather than analysis capability.

6.6.3 Copper. Fahmy et al.¹¹⁴ described the effect of CuNPs on brain activity in adult male Wistar rats. Treated and control groups of rats (n = 10 in each group) were injected with 15 mg kg⁻¹ CuNPs and saline solution respectively over 2 days. The rats were then sacrificed and brains dissected and the Cu content of different brain regions was determined using ICP-OES. The treated group showed increased levels of Cu in the cortex, cerebellum, striatum, thalmus and hippocampus but not in the midbrain or medulla. Injection of CuNPs was also associated with changes in oxidative stress status and cholinergic neurotransmission. The authors point to the importance of toxicity assessments when evaluating the use of CuNP.

Moving on from Cu toxicity in the brain, Paredes et al.⁶³ present an analytical development allowing for the determination of Cu isotopic ratios at the protein level in neuronal cells. Isotopic ratios were determined by MC-ICP-MS in neuronal cell protein fractions separated by SEC. Although isotopic determination was unsuccessful for U and Zn in the same study, the authors hope this preliminary work will enable further developments in the field.

6.6.4 Gadolinium. Interest in Gd has continued in this review period as concerns surrounding the use of Gd-based contrast agents (GBCAs) in medical imaging persist. Two studies have investigated brain tissue retention of Gd following administration of GBCAs. Wang et al.¹⁴⁸ used healthy control mice and experimental autoimmune encephalomyelitis (EAE) induced female SJL/J mice to determine the effect of inflammation on Gd retention. Experimental autoimmune encephalomyelitis (n = 4) and control (n = 4) mice were administered with 2.5 mmol kg⁻¹ Gd–DTPA over 10 days and were then sacrificed either 1 day or 10 days after the last Gd application. For each mouse brain, one hemisphere was processed for Gd quantitation using ICP-MS and the other for histology and SR-XRF. Analysis of brain tissue by ICP-MS demonstrated higher levels of Gd in EAE (mean 5.3 μg g⁻¹) than in control mice (mean 2.4 μg g⁻¹) and quantifiable concentrations of Gd were present in brain tissue from both groups up to 10 days after administration. NanoSR-XRF identified 160 nm Gd hotspots in locations with high inflammatory activity identified by conventional H&E staining. Despite the small number of animals included in the study, this work highlights neuroinflammation as a potential risk factor for patients undergoing GBCA-enhanced MRI.

Stanescu et al.¹⁴⁹ evaluated brain tissue Gd retention in a small cohort of paediatric patients following contrast MRI. Postmortem brain samples were obtained for ten patients, aged from 1 year 11 months to 13 years at the time of death, with previous GBCA exposure. Gadolinium was quantified in brain tissue by microwave assisted acid digestion and ICP-MS analysis and concentrations were normalised by the cumulative dose of Gd received. Detectable Gd was present in all decedents and levels were highest in those subjected to multiple exposures. In line with previous studies, ICP-MS normalised Gd ratios were higher in children receiving linear agents than in those who only received macrocyclic agents (Wilcoxon rank sum P = 0.004).

The use of GBCAs in medical imaging also presents analytical challenges for the quantitation of other elements as Gd is a known interferent in clinical elemental analysis. Experimental quantitation of Gd interferences was investigated this year by a series of Gd spiking experiments.¹⁵⁰ In this work, Day et al. demonstrated the space charge effect of Gd on internal standard elements Ga, Ge, Rh and Y by analysing urine samples spiked with 1000 ppm Gd. Analysis of spiked samples demonstrated positive interference in Se measurements from Gd at concentrations >10 000 ppm due to the double charge effect of ¹⁵⁶Gd on ⁷⁸Se. The group also illustrated how the polyatomic interference of GdCl, causes falsely elevated ¹⁹⁵Pt concentrations. Wilschefski et al.⁵⁹ have also investigated Gd interferences and have documented a correction calculation to account for the double charge effect in Se analysis. The calculation utilises a higher resolution analysis mode to measure m/z 78.5 (¹⁵⁷Gd²⁺) and incorporates the isotopic abundance of ¹⁵⁶Gd compared to ¹⁵⁷Gd. Calculated recoveries from 97.4% to 106.5% compared to target values were obtained for samples spiked with 2–20 mg L⁻¹ Gd.

6.6.5 Indium. Li et al.¹⁵¹ investigated the cytotoxic mechanisms of In₂O₂ nanoparticles (IO-NP) and In₂O₂ fine particles (IO-FP) on pulmonary epithelial and macrophage cells. Epithelial cells (16HBE) and macrophages (RAW) were treated with 2 μg mL⁻¹, 20 μg mL⁻¹ and 200 μg mL⁻¹ IO-NP or IO-FP for 24 h in serum free medium. Inductively coupled plasma mass spectrometry was used to quantify cellular In content following acid-digestion. Results demonstrated ten-fold higher concentrations of In in 16HBE cells compared to RAW cells. Release of In from IO-NP in 16HBE cells inhibits of phagocytosis and migration of macrophages whilst stimulating cytokine release from RAW cells.

6.6.6 Iodine. In the previous review¹ we reported on a simple ICP-MS method for determining I in serum and urine. This year Zou et al.¹⁵² described a rapid ICP-MS method for I quantification in amniotic fluid, breast milk and CSF with an LOQ of 0.778 μg L⁻¹. In amniotic fluid, breast milk and CSF samples, median concentrations were 176.3 μg L⁻¹ (n = 53), 136.0 μg L⁻¹ (n = 547) and 81.8 μg L⁻¹ (n = 150), respectively. Intermediate precision (as RSD) over 5 days was 3.3–8.0% with recoveries from 97.7% to 109.8% across all samples. All three sample types demonstrated large individual variation and weak correlations with I levels in matched serum/plasma samples. Further studies are required to determine the clinical utility of I concentrations in amniotic fluid, breast milk and CSF.

6.6.7 Iron. In this review period, two papers reported on the distribution and quantification of different Fe species in biological samples. McAllum et al.¹⁵³ evaluated regional Fe distribution and ferroprotein profiling in human brain tissues. Total Fe concentrations were measured in different brain regions using ICP-MS. Soluble Fe-binding proteins were quantified using SEC-ICP-MS and insoluble Fe-binding proteins using ICP-MS following nitric acid digestion. Total Fe concentrations showed heterogeneity across different regions, however cytosolic and membrane compartmentalisation remained consistent. Quantitative profiling of Fe-binding proteins in healthy human brain showed a general association between Fe distribution and ferritin expression. Simultaneous quantitation of Fe^II and Fe^III in cell lysates was achieved by Michalke et al.¹⁵⁴ using CE-ICP-MS. The method achieved a LOD of 3 μg L⁻¹ with precision measurements up to 3.5% and accuracy from 97% to 105% for both redox species.

Lu et al.¹⁵⁵ presented a metabolic study looking at physiological Fe requirements in women of reproductive age. Women of reproductive age, with no previous pregnancies, from Hebei, China were given 50 mg ⁵⁸Fe over 2 weeks and were followed up for 800 days. Whole blood total Fe and ⁵⁸Fe were quantified using AAS and MC-ICP-MS respectively following microwave-aided acid digestion. Estimated Average Requirement (EAR) and recommended intake (RNI) were calculated using a formula method (n = 21) and a linear regression method (n = 33). The two methods were comparable and yielded calculated EAR (11–13 mg per day) and RNI (15–18 mg per day) lower than current Chinese recommended values of 15 mg per day and 20 mg per day, respectively.

6.6.8 Lead. Publications of interest evaluated Pb exposure in two very distinct populations. Scott et al.¹⁵⁶ used magnetic sector ICP-MS to investigate the elemental composition of 30 femora from three Roman cemeteries in Londinium. Lead concentrations were 70-fold higher in Londinium femora than in 70 femora dating to the pre-Roman Iron age. Data are consistent with previously identified widespread Pb pollution and the hypothesis that Pb intoxication may have contributed to declining populations. In contrast to the archaeological study, Dantzer et al.¹⁵⁷ explored the use of toenails as biomarkers for Pb exposure in children. Blood and toenail samples were collected from a population of 224 12 year old children in the Cincinnati Childhood Allergy and Air Pollution Study and Pb measured using ICP-MS with HNO₃ digestion. Full Scale Intelligence Quotient (FSIQ) on study participants was assessed using the Wechsler Intelligence Scale for Children-4^th edition. A weak statistically significant association was observed between logarithms of blood and toenail Pb concentrations (R² = 0.49, p < 0.001). While blood Pb was significantly associated with decreased FSIQ (p < 0.001), no significant correlation was observed between toenail Pb and FSIQ (p = 0.192). The clinical utility of toenails as biomarkers for Pb exposure and cognitive health remains unclear.

Lelievre et al.¹⁵⁸ documented an unusual case study of Pb intoxication from homemade bullets following an attempted homicide. Despite surgical removal of the bullets, the 50 year old woman presented with worsening systemic and cerebral Pb poisoning due to persisting Pb fragments in her abdomen and back. A blood Pb concentration of 1650 mg L⁻¹ was determined using ICP-MS and post-mortem Pb concentrations in brain tissue, along with blood brain barrier disruption, confirmed Pb encephalopathy.

6.6.9 Mercury. Santa-Rios et al.⁹⁴ described a method for measuring MeHg in dried blood spots using GC-CVAFS. The method was validated against US EPA criteria, using paired dried blood spots and whole blood samples from 49 volunteers. Precision (RSD) was 13% and 0.4% for dried blood spots and whole blood samples, respectively, at concentrations >1 μg L⁻¹ but was more variable at lower concentrations. The LODs for dried blood spots and whole blood samples were 0.3 μg L⁻¹ and 0.2 μg L⁻¹, respectively. Good correlation between the two sample types was demonstrated for MeHg (r: 0.85–0.95) and Bland–Altman analysis showed that in the majority of cases (82–98%) bias between the two MeHg measurements met the 35% maximum allowable difference. Although they also investigated iHg, the authors note that further work is required to assess iHg exposure using dried blood spots.

6.6.10 Platinum. Studies by Bortot et al.¹⁵⁹demonstrated increased efficacy of cisplatin-loaded biodegradable nanoparticles (Cis-NP) compared to conventional cisplatin in reducing the tumour burden from epithelial ovarian cancer. In vitro investigations used SKOV3 cell lines due to their representation of cisplatin resistance. Varying concentrations of Cis-NP and free cisplatin were administered to SKOV3-luc cells, which were subsequently assessed for Pt content, cytotoxicity, morphology and RNAseq transcriptome analysis. In vivo animal studies used an SKOV3-luc cells’ xenograft model to inject mice with saline (n = 6), free cisplatin (n = 6) and Cis-NP (n = 7). Tumour growth was monitored by bioluminescence for 2 weeks. In vitro intracellular Pt content, quantified by ICP-MS, was substantially higher 72 h following treatment with Cis-NP compared with free cisplatin. Cytotoxicity in Cis-NP treated cells was higher than in cells treated with free cisplatin and these results correlated with in animal studies showing decreased tumour burden with Cis-NP compared to free drug.

6.6.11 Selenium. Selenium has attracted a lot of interest in this review period and various developments have been reported pertaining to biomarker potential, metallomics and analytical methods. Three groups investigated the potential use of Se species as biomarkers for specific disease states. Perhaps one of the most topical papers in this year’s review, Moghaddam et al.¹⁶⁰ report on an association between Se deficiency and COVID-19 mortality risk. Total Se and selenoprotein P were quantified by XRF in 166 serum samples from 33 COVID-19 patients. Surviving COVID-19 patients had higher Se levels than non-survivors (Se: 53.3 ± 16.2 vs. 40.8 ± 8.1 μg L⁻¹, selenoprotein P: 3.3 ± 1.3 vs. 2.1 ± 0.9 μg L⁻¹). The authors conclude that Se status can provide diagnostic information for COVID-19 and support Se supplementation, however they acknowledge the mechanism for Se deficiency remains unexplained. In addition to developing methods for quantification of selenoproteins and selenometabolites in serum, Callejon-Leblic et al.⁵⁵ identified the ratios of the two species as potential biomarkers for lung cancer. Ratios of various selenoproteins and selenometabolites were significantly different between the serum of lung cancer patients (n = 48) and healthy control patients (n = 39). Maass et al.¹⁶¹ also investigated Se speciation in the CSF of patients with Parkinson’s disease, however analysis was unable to yield any potential biomarkers.

Phiri et al.¹⁶² evaluated the use of urine Se as a biomarker to assess Se status at a population level. ICP-MS was used to determine Se concentrations in spot urine samples from 1406 women of reproductive age and 665 school aged children across Malawi in whom plasma Se status had already been measured. Urine Se concentrations, adjusted for hydration status using creatinine, specific gravity and osmolarity, gave reasonable predictions for variability in plasma Se between population clusters (R²: 0.52–0.61). The R² values for variation between individuals and within households were much smaller, therefore urine Se concentrations may be an option for assessing Se status in populations but not in individuals.

Two papers report developments in the understanding of Se metallomics. Takahashi et al.¹²¹ identified the biliary Se metabolite, selenodiglutathione and performed experiments in rats to demonstrate maintenance of Se status in the body via enterohepatic circulation. Katarzyna et al.¹²⁰ measured selenometabolites in turkey livers to investigate the effect of high dietary supplementation with inorganic Se. Results showed that high Se supplementation increased selenosugar formation leading to increased selenosugar-decorated proteins and selenosugar linked to low M_r thiols.

Finally, two groups reported analytical advances for the determination of Se species in different biological tissues. Gawor et al.¹⁶³ used HPLC-ICP-MS to develop a protocol for quantitative analysis of SeMet, SeCys₂ and Se-methyl–Se-cysteine in biological tissues. This protocol was applied to liver, muscle and heart tissue from lambs fed with inorganic (Na selenate) and organic (Se-enriched yeast) Se compounds. Hildebrand et al.¹⁶⁴ reported a method to determine 11 Se species in urine based on the separate coupling of three different chromatographic techniques with ICP-MS. Reverse phase chromatography was used to separate SeSug1, SeSug2, SeMet, MeSeCys, seleno-D,L-ethionine, methylselenic acid and methylselenoglutathione. Selenate (Se^VI) and Se^IV were identified using AEC, whereas CEC was applied to allow the determination of the Se species methyl-2-amino-2-deoxy-1-seleno-b-D-galactopyranoside (SeSug3) and TMSe. Detection limits between 0.03 μg Se L⁻¹ and 0.10 μg Se L⁻¹ were reported for all Se species with intermediate precision (as RSD) between 2.7% and 10.6% and recoveries ranging from 87% to 108%. The same group published further work demonstrating increased stability of SeSug1 following the addition of NaN₃ as a bactericide and sufficient buffering to pH 5.5.⁴³

6.6.12 Titanium. Comera et al.¹⁶⁵ performed in vivo and ex vivo studies to investigate the routes of TiO₂ (E171) absorption in the gut. Mice (n = 8 for each condition) were given either H₂O or a 40 mg kg⁻¹ TiO₂ suspension and were sacrificed after either 4 or 8 h. Laser-reflective confocal microscopy and ICP-MS, after sample digestion with HNO₃, were used to determine the intestinal uptake of TiO₂. Absorption of TiO₂ peaked at 4 h in the jejunal and ileal villi and after 8 h in the jejunal Peyer’s patches. Absorption was lower in the colon and TiO₂ was detectable in the blood at 4 and 8 h post oral ingestion. Ex vivo experiments exposed jejunal loops to pharmacological inhibitors of paracellular tight junction permeability and of transcellular endocytic passage. Whilst various inhibitors were investigated, a decrease in TiO₂ absorption was only observed in the presence of the paracellular permeability blocker triaminopyrimidine (66% reduction after 30 min, p < 0.001 vs. control). The authors conclude that the main routes of TiO₂ uptake in the gut are via paracellular tight junction permeability and jejunal villus absorption via a goblet-cell associated passage.

6.6.13 Uranium. Albendea et al.¹⁶⁶ investigated the use of ICP-MS as an alternative to alpha spectroscopy for individual monitoring of U in 24 h urine samples from occupationally exposed workers in Spain. Direct quantitation of ²³⁴U via ICP-MS is not feasible due to its low mass proportion but measurement is required due to its high specific activity and potential radiotoxicity. Despite good comparability between ICP-MS and alpha spectroscopy achieved for ²³⁵U and ²³⁸U, high variability in the ²³⁴U : ²³⁸U ratios precludes the use ICP-MS for ²³⁴U estimation.

7 Applications: drugs and pharmaceuticals, traditional medicines and supplements

Results of analyses of several sample types were reported during this review period. Some were given simply to demonstrate that a newly developed procedure could be applied to real samples but in a few, the primary objective was to assess the content of samples available for sale. Reboredo et al.¹⁶⁷ determined, by EDXRF, the elemental composition of 6 plant-based food supplements available in Portugal. The iAs content for hemp protein is such that the supplier’s recommended daily consumption would be 1.75 μg kg⁻¹ body weight per day which is well in excess of the European Food Safety Authority limit of 0.64 μg kg⁻¹ body weight per day. Arsenic in Maca, Ashwagandha and Camu-Camu were all around the 95^th level of this limit. Consumption of Hemp protein also leads to a daily intake of Mn above the daily reference intake. The workers also report that the recommended intake of 28 g of organic and conventionally grown Goji berries would provide 315.3 and 324.8 μg per day, respectively. The Cd and Cr content of 21 Brazilian herbal tincture samples were measured using ETAAS, by Lopes et al.¹⁶⁸ The content of Cd and Cr in these tinctures, intended for conditions such as dyspepsia, occasional intestinal constipation, gastritis and breathing problems were below the LODs in about half the samples tested. The highest concentrations found were 42.24 μg kg⁻¹ Cd and 12.43 μg kg⁻¹ Cr. The authors report the maximum allowable limits proposed by several authorities are all much higher than the measured concentrations reported here. Figueiredo et al.¹⁶⁹ point out that in some communities up to 50% of the population use some form of food supplements, usually without any control or medical supervision and during extended periods of time. Following the lead from the United States Pharmacopeia and the European Medicines Agency for impurities in pharmaceuticals, these workers used WDXRF to measure concentrations of 11 elements in 25 food supplements intended for weight loss. Most of the measured results were within the limits for pharmaceuticals but one sample had high concentrations of Mn and Pb and in another two, the concentrations of Cr were also above the limits for pharmaceuticals. Examples of differences between the measured concentrations and the values stated on the product labels were also found. The authors recommend greater control over food supplement labelling and application of the same quality standards for food supplements as for pharmaceuticals.

8 Applications: food and beverages

In addition to the discussion of individual papers below, Table 3 provides a summary for this year’s contributions in this area.

Table 3 Food and beverages

Element	Matrix	Technique(s)	Sample preparation/comments	Ref.
Ag	Fish liver tissue (rainbow trout)	spICP-MS	4 reagents were evaluated for the extraction of AgNPs from rainbow trout liver tissue, prior to analysis by spICP-MS, consisting of proteinase K or TMAH, both with or without CaCl2. Only the mixture of TMAH–CaCl2 successfully solubilised the liver. The method was applied in an in vivo dietary study, comparing fish fed either AgNO3 or AgNPs and showed no significant difference for particle number concentration, mean particle size or particle mass concentration	227
Al	Wine (white and red wine)	LC-ICP-MS	Al^3+, AlF, oxalate and citrate Al complexes in red and white wines were determined by LC-ICP-MS using a collision cell and a 6 mL min^−1 He flow as the cell gas. The separation was achieved in 10 min by CEC, using gradient elution, with a NH4NO3 solution as the mobile phase, at pH 3.00 ± 0.01, and 2 mL min^−1 flow rate. The results for total Al and Al species were compared with those of other methods	182
As	Rice	HG-AAS	Following microwave-aided digestion and reduction, As^III and Se^IV were extracted (DES and UAE) and determined by HG-AAS. The LODs were 2 and 3 ng L^−1, respectively and the method was validated by the analysis of SRMs and spike recovery	171
As	Chocolate, mussels, red wine, rice	ICP-AES	Samples were dissolved in HNO3 by microwave-aided digestion. Compromise conditions for the DMLLE of the three (As, Cd, Pb) APDC complexes into THF produced LODs of 2, 0.6 and 2 μg L^−1, respectively corresponding to enhancements of 35–45-fold	49
As species	Strawberries	CVG-AFS	Arsenic species were extracted from lyophilized fruits by microwave-assisted extraction at 90 W for 4 min. After centrifuging (5300 rpm for 5 min), clean-up (C-18 cartridge), and filtering (0.45 μm), species were separated by AEC (Hamilton PRPX-100 with 20 mmol L^−1 KH2PO4^2− + K2HPO4^− at pH 5.8) and detected after post-column HG with borohydride. Total As was determined after acid digestion by HG-AFS. The method was validated using NIST SRM 1568b (rice flour) and spike recoveries	89
As species	Rice	HPLC-HG-ICP-MS, ICP-MS	Total toxic elements (Al, As, Cd, Sb), essential elements (Co, Cr, Fe, Mn, Mg, Ni Se, Zn) and other elements (Rb, Sr, Th and REEs) were prepared by closed-vessel microwave-assisted digestion and determined by ICP-MS. For speciation analysis, 0.2 g of sample was extracted with 10 mL of 2% HNO3 (water bath from 25 to 95 °C for 0.75 h and held at 95 °C for 1.5 h). After cooling, the extracts were filtered (0.20 μm). Species were separated on a Hamilton PRP-X100 column with mobile phase 10 mmol L^−1 HPO4^2− + H2PO4^− at pH 8.5 with 2% MeOH, and detected after post-column HG. The method was validated by the analysis of NIST SRM 1568a (rice flour). The LODs were DMA 0.004 μg L^−1, MMA and As^III 0.003 μg L^−1, and As^V 0.01 μg L^−1	88
As species	Rice-based infant food	ICP-MS, HG-ICP-MS	For total As, samples were dried (60 °C for 48 h) and ground, then digested (1.0 g) with 6.0 mL 70% HNO3 at 100 °C for 2 h, diluted to 25 mL and filtered (0.45 μm). A 1.25 mL aliquot was diluted to 10 mL. For iAs, the sample (1.0 g) was extracted with 7 mL 50% HClO4 at 80 °C for 1 h. To 5 mL of extract, 4 mL 10 M HCl, 0.5 mL 48% HBr and 0.5 mL 3% hydrazine sulphate were added and As determined via HG after extraction into CHCl3. The LODs in the samples were 4 and 50 μg kg^−1, for total As and iAs, respectively. The fate of DMA was not investigated	91
As species	Vegetables	HPLC-ICP-MS	Closed vessel microwave-assisted extraction from 500 mg of freeze-dried sample, with 10 mL of 2% HNO3 at 90 °C for 17 min, was followed by AEC (Hamilton PRPX-100 with phosphate mobile phase) and detection at m/z 75. The extraction efficiencies for total arsenic species from 24 market vegetable samples ranged from 91.4–106%, with As^III and As^V predominant at concentrations from 5–225 μg kg^−1	228
As species	Seafood	HPLC-ICP-MS	As species were extracted into 5 mL H2O from freeze-dried samples (0.25 g) by ultrasound assisted extraction (30 min, 35 W, 44 kHz), followed by centrifugation (20 min at 20 000 g), filtration (0.45 μm) and separation by RP C-8 with a mobile phase of 1.2 mmol L^−1 TMAH, 4 mmol L^−1 malonic acid, 6.2 mmol L^−1 sodium butane-1 sulfonate, and 0.05% MeOH. The LODs for 11 As species (As^III and As^V, DMA, MMA, AB, AC, TMA, TMAO, dimethyl-dithioarsinic acid, monomethyl-monothioarsinic acid, and trimethyarsine sulfide) ranged from 1.4 to 4.0 μg kg^−1	48
As species	Mushrooms	HPLC-ICP-MS	Species were extracted (ultrasound assisted, microwave assisted and enzymatic assisted extraction), separated by AEC and quantified at m/z 91 (AsO^+), with ^73Ge^+ as internal standard. Sulfur (as ^50SO^+) was also monitored as an indicator of peptides and proteins. SEC (Superdex 75 10/300 GL) with UV-Vis detection followed by ICP-MS allowed detection of protein bound As	46
As species	Seafood	Numerous	A comprehensive and authoritative review with 308 references. There are almost a hundred organoarsenicals identified in marine dietary sources. The development of standards, especially for organoarsenicals with confirmed toxicities, should be prioritised	18
As species	Rice	HPLC-ICP-MS	Moisture was removed by heating (65 °C, 48 h). For total As, 150–200 g was ground and 0.2 g was dissolved (MAD) in 6 mL HNO3, diluted to 20 mL, and then 1 + 9 with water. The method was validated by analysis of NIST SRM 1568b (rice flour). Arsenic species were extracted (1.5 g in 15 mL 2% HNO3) by heating at 95 °C for 1.5 h, followed by dilution to 50 mL, filtration (5 μm) and separation by AEC [PRP-X100 column, mobile phase 20 mmol L^−1 NH4H2PO4 + (NH4)2HPO4 at pH 5.6]. The speciation method was not validated	229
As species	Rice	HPLC-ICP-MS, SR-μXRF, μXANES	Unpolished rice was ground under liquid nitrogen and As species were extracted with 1% HNO3 and determined by HPLC-ICP-MS, AEC Hamilton PRP-X100 with 4 mmol L^−1 NH4H2PO4, 4 mmol L^−1 (NH4)2SO4 and 4 mmol L^−1 NH4HCO3, at pH 9 in 2% EtOH. Speciation and distribution in rice grain and soil was also studied with SR-μXRF and μXANES	124
As species arsenolipids	Salmon, human milk	HPLC-ICP-MS, ICP-MS/MS	Total As, lipid-soluble As + specific arsenolipids, and water-soluble As species were determined. Total As was analysed after MAD with HNO3, using Ge as internal standard. Collision cell mode (He) and the addition of 1% CO2 in argon enhanced the As signal. The LOD was 0.01 μg L^−1 and the method was validated with two RMs. Lipid-soluble As involved a 9–10 step extraction (with dichloromethane) and a clean-up (with Florisil) procedure followed by RP-HPLC-ICP-MS and RP-HPLC-ESI-MS. Water-soluble As species (AB, AsC, DMA, TETRA, TMAO) remained after the arsenolipid extraction were separated by CEC (mobile phase pH 2.8) and detected by ICP-MS/MS, operated in oxygen reaction mode	200
Ca	Infant formula powder	LIBS, AAS	PLSR models, based on LIBS, Raman and FTIR spectral data, respectively, were developed to quantify Ca in infant formula powder samples and the results compared with those obtained by AAS	230
Cd	Rice, CRM	ETV-GD-AES, ICP-MS	A miniaturized GD-AES coupled to ETV with a tungsten coil trap was applied for the direct determination of Cd in 1–15 mg rice samples in 3 min. The ETV procedure included sample drying and pyrolysis, followed by vaporisation of Cd on a cold tungsten coil. Cd was then released by heating and transported by an Ar–H2 stream into the plasma. The LOD was 0.26 μg kg^−1 for 10 mg of rice sample. The method was validated with a rice CRM and by comparison with ICP-MS applied to microwave-aided acid digested samples	72
Cd	Beverages, food, serum	ICP-MS	To improve the assessment of individual dietary intake of Cd, a new biomarker was proposed, based on measurements of serum Cd and a food frequency questionnaire, accompanied by measurements of Cd in food and beverages, and tested on a group of 51 subjects	204
Cd	Food CRMs (rice, wheat and tea)	ETAAS	An SPE procedure, based on a chitosan/thiol modified metal–organic frameworks composite as adsorbent, was developed to determine Cd and Pb in food samples by ETAAS, with LODs of 0.008 μg L^−1 (Cd) and 0.033 μg L^−1 (Pb), respectively	187
Cd	Drinking water, CRMs	HG-AFS	An electrochemical VG system was applied to determine the Cd content of drinking and environmental water samples, using HG-AFS. An LOD of 0.05 ng mL^−1 was achieved. The RSD was 3.2%	96
Cd	Herbal tinctures	ETAAS	For the direct analysis, the alcoholic samples were diluted with 0.1% Triton X-100 and 0.2% HNO3. Calibration standards also contained 7% alcohol. The chemical modifiers were Pd and Mg(NO3)2. Microwave-assisted digestion was part of a reference method. LODs (μg kg^−1) were: direct method 0.5 (Cd), 1 (Cr); reference method 0.06 (Cd), 0.3 (Cr). No significant differences were observed for the analysis of 21 samples	168
Cd	Chocolate mussels, red wine, rice		See As, ref. 49	49
Cd	Drinking water	FAAS	Sample pre-concentration on quercetin, (a polyphenol derived from plants), modified using granular activated carbon, to measure the concentrations of Cd, Ni and Pb in drinking fountain water using FAAS	186
Cd	Water (tap water, sparkling water, synthetic seawater)	MIP-AES	A novel attempt to couple DLLME to MIP-AES revealed spectral and non-spectral carbon-based interferences due to the components of DLLME extractants, such as CHCl3, 1-decanol and THF, affecting wavelengths above 328 nm. The optimised conditions for the determination of Cd in water samples (tap, sparkling and synthetic seawater) lead to enrichment factors of 46 and 42, depending on the extractant used and an LOD (1 μg L^−1) equivalent to that reported for DLLME-ICP-AES	74
Cd	Drinking water	LIBS	The determination of Cd in drinking water was achieved using LIBS, after pre-concentration on a chelating resin, with an LOD of 3.6 μg L^−1	231
Cd	Cereal flakes	HR-CS-AAS	Cd (0.01–2 μg L^−1) and Fe (10–500 μg L^−1) were simultaneously determined in acid digested samples of different cereal flakes by means of HR-CS-AAS, using Pd–Mg(NO3)2 as a chemical modifier	70
Ce	Radish	spICP-MS	Radish plants were cultivated hydroponically in presence of CeO2 and CuO NPs to study the bioaccessibility, they were analysed before and after simulated gastrointestinal digestion for total Ce and Cu. Samples (0.05 g dry mass) were acid digested in a microwave oven, with 5 mL HNO3–3.5 mL H2O2 mixture then diluted to 25 mL with water. LOQs were 0.39 μg L^−1 (Ce) and 4.3 μg L^−1 (Cu), respectively	173
Cl	Beer	HR-CS-ETAAS, ICP-AES	Cl, Fe and Si were determined in beer samples using HR-CS-ETAAS with Pd(NO3)2–Mg(NO3)2 as a chemical modifier. Cl was detected as InCl which formed in the presence of In and vaporised at 1800 °C, followed by the atomisation of Fe and Si, at 2600 °C. LODs and LOQs (respectively) were: 0.05 and 0.17 mg L^−1 Cl, 0.08 and 0.26 mg L^−1 Si, and 2.0 and 6.7 μg L^−1 Fe. Recoveries ranged from 85% to 120%. This simultaneous procedure was compared with ICP-AES (Fe and Si) and IC (Cl)	69
Co	CRMs (stream sediment, human hair and mussel) environmental water samples	ICP-MS	The pre-concentration of Co and Ni from environmental water samples and CRMs, prior to analysis by ICP-MS, was achieved by dispersive micro-SPE using fibrous TiO2@g-C3N4 nanocomposites as a new adsorbent. Quantitative retention was obtained in the pH range 5.0–8.0, followed by elution with 1.0 mL of 0.5 mol L^−1 HNO3. LODs were 0.12 pg mL^−1 (Co) and 1.34 pg mL^−1 (Ni) and RSDs at 1.0 ng mL^−1 were 3.9% (Co) and 4.8% (Ni)	188
Co	Food, water, soil	ICP-AES	Co and Hg were determined in a variety of food and water samples, by ICP-AES, after pre-concentration, by a factor of 80, using Amberlite XAD-4 loaded with Anoxybacillus kestanboliensis. The achieved LODs were 0.04 ng mL^−1 (Co) and 0.06 ng mL^−1 (Hg). The matrices tested included soil, tap, river and mineral water and several types of food	189
Cr	Honey	ETAAS	2.0 g were mineralised with 2 mL of 50% H2SO4 by step-wise heating in a muffle furnace finishing at 750 °C for 6 h. The ash was dissolved in 50 mL of 1 mol L^−1 HCl. Following DLLE, the magnetic ionic liquid phase was separated by a magnetic rod and diluted in CHCl3	190
Cr	Herbal tinctures	ETAAS	See Cd, ref. 168	168
Cr species	Rice, sausage, tap water	AAS	Solid samples were extracted with conc. HCl at 60 °C. DES extraction of Cr^III the LOD was 0.4 ng mL^−1. Cr^VI was reduced by ascorbic acid	179
Cu	Edible oils (sunflower, olive, flaxseed, corn) and baby oil	FAAS	An LLME procedure, based on 500 μL of triethylamine as the extraction solvent, 750 μL of 7.5 mol L^−1 HNO3 as a hydrophilicity-switching trigger, 30 mL of oil sample and 5 min extraction time, was applied prior to determination of Cu by FAAS. With a pre-concentration factor of 22.7, LOD and LOQ values were 6.9 ng mL^−1 and 23.0 ng mL^−1. RSDs were <4.7% (intra-day) and <9.4% (inter-day) and recoveries ranged from 85% to 100%	193
Cu	Radish	spICP-MS	See Ce, ref. 173	173
Cu	Food samples (grapes, apple, tea, and rice), water samples	FAAS	A sorbent, consisting of modified poly m-phenylenediamine/carbon nanotube electrospun nanofiber, for thin film micro-extraction of Cu from food and water samples, was developed and characterised using FTIR, SEM and TEM. The determination of Cu by FAAS, after micro-extraction, was achieved with an LOD of 0.32 ng mL^−1, linearity ranging from 2.0 to 500 ng mL^−1, and RSDs% of 2.9% (intra-day) and 3.7% (inter-day)	191
Cu	CRMs, diet supplements, tea, water	FAAS	Cu was determined by FAAS, in various types of samples (CRMs, diet supplements, black tea, water) after achieving a pre-concentration factor of 20 using SPME on pyrocatechol violet impregnated magnetic graphene oxide. The LOD and LOQ were 4.0 μg L^−1 and 13.3 μg L^−1, respectively, and the RSD was 4.93%	192
F	CRM (corn starch), plant-based materials	ETV-ICP-AES	A novel procedure for the direct determination of F in solid samples by means of ETV-ICP-AES was developed. 2 mg of sample were placed into the ETV furnace and heated at 250 °C for 15 s followed by a step at 2200 °C for 20 s. Internal standardisation was carried out using the Ar 404.442 nm emission line. Calibration strategies based on either matrix-matched CRMs or standard solutions were assessed and allowed to perform quick screenings of F content in plant based materials	73
Fe	Beer	HR-CS-AAS	See Cl, ref. 69	69
Fe	Cereal flakes	HR-CS-AAS	See Cd, ref. 70	70
Hg	Food, water, soil	ICP-AES	See Co, ref. 189	189
Hg species	Rice	ICP-MS, CVG-AFS, SR-XRF	Total Hg: 200 mg was dissolved in 5 mL HNO3 + 1 mL H2O2 in a stainless-steel bomb at 160 °C for 5 h. Following evaporation at 90 °C to 1 mL, dilution to 5 mL with 2% HNO3 + 0.1% β-mercaptoethanol and centrifugation, the ICP-MS LOD was 0.009 μg L^−1. The method was validated by analysis of SRMs (GBW08508 and NIES 10). MeHg: 50 mg was mixed with 2 mL 25% KOH–CH3OH, shaken for 4 h at 60 °C, diluted to 10 mL and centrifuged (4000 rpm for 30 min). The supernatant was ethylated by NaBEt4. The CV-AFS LOD was 0.002 μg L^−1. Elements (Cu, Fe, Hg, Se and Zn) in grains were mapped by SR-XRF	93
Hg species	Mushrooms	HPLC-ICP-MS	Hg species were solubilised from powdered samples (0.1–0.2 g) by ultrasonic assisted extraction (10 mL 12 g L^−1L-cysteine for 40 min at 40 °C). After centrifugation (9000 rpm for 15 min) and filtration (0.22 μm), analytes (Hg^II, MeHg, EtHg, and PhHg) were separated on a C-8 column by 10 mol L^−1 ammonium acetate + 0.8 g L^−1L-cysteine at pH 4.0 with a MeOH gradient. The LODs ranged from 0.6–4.5 μg kg^−1. Total Hg was determined by a direct mercury analyser	174
I	Food samples (potatoes, pasta and rice)	ICP-MS	The I uptake by potatoes, pasta and rice during cooking with table salt enriched with KIO3 was assessed. Iodine was extracted from cooked food samples with 0.5% NH3 prior to ICP-MS analysis. All products showed an increase in I content	232
I	Cow’s milk	ICP-MS	Iodine was measured by ICP-MS in 96 cow’s milk samples from 24 locations in the US, after alkaline digestion of 2 g milk with KOH, followed by stabilization with NH4OH and sodium thiosulfate, according to an AOAC method. The I content ranged from 31 to 251 μg per serving (240 mL), compared to the recommended daily intake of 150 μg	233
Mg	Infant formula powder	LIBS	A PLSR model was developed and applied to LIBS data to predict the Mg content of infant formula powder. The LOD was 150 mg kg^−1	85
Mn	Coffee and wastewater	FAAS	Mn was extracted from coffee and wastewater samples using 1.50 mL of 0.02% w/v 3-[[(2-hydroxyphenyl)imino]methyl]-2-naphthalenol, as the ligand, 500 μL of a DES (choline chloride–phenol mixture, 1 : 2 molar ratio), 1.0 mL THF as the emulsifier agent and 2.0 mL of buffer at pH 10. An enrichment factor of 92.9 was achieved. The LOD and LOQ values were 0.52 μg L^−1 and 1.72 μg L^−1, respectively, and RSD was 2.7%	68
Ni	Food samples (crab, shellfish and rice samples)	ETAAS	A magnetic nano-adsorbent (silk fibroin–Fe3O4–EDTA) was developed, assessed and applied to the extraction of Ni from various food samples (crab, shellfish, rice) prior to analysis by ETAAS. The extraction procedure was carried out using 4 mg L^−1 of the nano-adsorbent, incubated with the sample at pH 6 and 28 °C, for 21 min, followed by separation of the nano-adsorbent on a magnet bar. An LOD of 0.0017 μg L^−1 was achieved	195
Ni	CRMs (stream sediment, human hair and mussel) environmental water samples	ICP-MS	See Co, ref. 188	188
Ni	Drinking water	FAAS	See Cd, ref. 186	186
Ni	Baby foods (ready meals, fruits, desserts, paps)	ETAAS	The content of Ni in 85 commercial premade baby foods was measured by ETAAS after dry ashing 2 g of samples in a sand bath at 300 °C for 2 h, followed by 3 h in a muffle furnace at 560 °C. The ETAAS program consisted of a drying step from 90 °C to 110 °C, a pyrolysis step at 1150 °C and atomisation at 2550 °C. The LOD was 12.5 μg kg^−1. The observed values ranged from <LOD to 225.7 μg kg^−1. Mean recoveries ranged between 78.0% and 90.1% and RSD were <10.3%. Estimated daily intakes of Ni for children were estimated	234
Ni	Beverages (chamomile tea, coffee)	FAAS	A procedure based on DLLME and slotted quartz tube FAAS achieved an LOD of 4.9 ng mL^−1 and an LOQ of 16.4 ng mL^−1. The pre-concentration factor was 66.4. Recoveries ranged from 88.1% to 120.2%	66
Pb	Milk	FAAS	A micro-extraction procedure for Pb in milk was developed using 8 mL of sample with 1.0 mL of Na2HPO4 buffer (pH 12) and 1.0 mL of 0.10% (w/v) dithizone in EtOH and vortex mixed for 30 s. Then 0.50 mL of DES (1 + 1 phenol and choline chloride) was injected into the solution with a syringe, followed by 1.0 mL THF. After each addition, the solution was vortex mixed for 10 s. Finally, the organic phase was separated by centrifuging the solution for 2.0 min at 6000 rpm and analysed by slotted quartz tube FAAS. Spike recoveries were 102.5% and 103.2% and RSD was 3.1%. The LOD and LOQ (8.7 μg L^−1 and 29 μg L^−1, respectively) were substantially better than those for FAAS, however not low enough for the determination of Pb in milk samples at the current EU limit of 20 μg L^−1	196
Pb	Food CRMs (rice, wheat and tea)	ETAAS	See Cd, ref. 187	187
Pb	Dietary supplements, SRM	HR-CS-ETAAS	Dietary supplement aliquots (0.8 and 1.0 mg) were placed directly onto a Zr-coated platform in the furnace, together with Pd(NO3)2–Mg(NO3)2 as a chemical modifier. The LOD and LOQ were 2.16 pg and 7.2 pg, respectively; RSD was <5% and recovery from a SRM was 116%	71
Pb	Chocolate mussels, red wine, rice	ICP-AES	See As, ref. 49	49
Pb	Drinking water	FAAS	See Cd, ref. 186	186
Pb	Milk	FAAS	Pre-concentration of Pb in milk was achieved by magnetic dispersive SPE using Fe3O4@SiO2@3-chloropropyltriethoxysilane@o-phenylendiamine NPs as the adsorbent and analysed by FAAS. Linearity ranged from 2.0 to 250 ng mL^−1. With a pre-concentration factor of 66, an LOD of 0.11 ng mL^−1 was achieved. RSDs were <3.4% (intra-day) and 5.5% (inter-day) and recoveries varied between 81.6% and 97.2%	197
Pb	Drinking water	XRF, ICP-AES	10 000-fold pre-concentration of Pb was achieved by filtering 2 L of tap water on an activated carbon felt. This enabled fast Pb detection for water samples in the range 0–150 ppb by portable XRF. The procedure was compared with ICP-AES (LOD: 1 ppb) on 15 samples and could be applied to other elements (Ca, Cu, Fe, Mn and Zn)	235
Pd	Water (well, stream, tap and sea water)	FAAS	Selective micro-extraction of Pd from water samples was achieved using a newly synthesised complexing agent (N-(3-chloro-4-fluorophenyl)-N-phenylthiourea). The Pd-complex, formed in highly acidic conditions, was extracted into CHCl3. After centrifugation, the organic phase was diluted with HNO3 and analysed by FAAS. Linearity ranged from 10 to 1000 μg L^−1 with an LOD of 2.28 μg L^−1	198
Sb	Cow’s milk, CRM		Sb speciation (Sb^III, Sb^V, residual, digestible and total Sb) in cow’s milk was achieved using DMSPE with fibrous TiO2@g-C3N4 nanocomposites as adsorbent and ICP-MS after sample pre-treatment with artificial gastric juice. At pH values between 2.0 and 4.0, only Sb^III was adsorbed onto the nanocomposite material, which was subsequently eluted with 1.5 mol L^−1 HNO3. After reduction of Sb^V to Sb^III, total Sb was determined by the same procedure and the concentration of Sb^V was calculated as the difference. The LOD for Sb^III was 0.37 pg mL^−1 and RSD at 1.0 ng mL^−1 was 4.1%	199
Sb	Herbal teas (bergamot tea, mint tea)	FAAS	A pre-concentration procedure was developed to allow the determination of Sb in herbal teas by slotted quartz tube FAAS, based on vortex assisted ligandless dispersive SPE with ZrNPs as the sorbent material. A 180-fold enrichment factor was achieved with an LOD and an LOQ of 8.0 μg L^−1 and 26.8 μg L^−1, respectively. Linearity ranged from 30 to 250 μg L^−1 and recoveries from spiked bergamot tea samples varied between 93 and 102%	112
Sb	Rice	LA-ICP-MS	See As, ref. 124	124
Se	Rice	HG-AAS	See As, ref. 171	171
Se	Dietary supplements	ICP-MS	Five different combinations of isotope and cell operating modes of the second quadrupole were compared: standard mode without a gas cell (^82Se), KED mode with He for kinetic discrimination of the argon-dimer (^78Se, ^82Se) and DRC mode with methane (^80Se, ^78Se). Accurate results for a yeast CRM (SELM-1), after microwave-assisted digestion, for each set of conditions were obtained. But, when analysing 28 mineralised dietary supplements, the results did not agree, although the determination of ^78Se in the KED mode was the most robust. Extractions with a TRIS buffer showed variable efficiencies depending on the form of Se	236
Se	Green tea	FAAS	Brewed green tea was filtered and diluted 1 + 49 with deionized water. Se (oxidation state not specified) was pre-concentrated by dispersive SPE with ZrNPs and measured by slotted quartz tube FAAS. The LOD was 5 μg L^−1	65
Se	Beans	ETAAS, CS-ETAAS	Samples were dissolved by microwave-assisted digestion. The researchers concluded that the direct determination of Se at concentrations <5 ng g^−1 was not possible because of spectral interferences by PO, NO and Fe with the Se line at 196.026 nm. HG with in-atomiser trapping on an Ir modifier gave an acceptable LOD of 30 pg g^−1	98
Se species	Food (mushroom, honey, cinnamon, leek, corn flour, kiwi, grape molasses, rice, ground beef, and egg)	ETAAS	Samples were dissolved by microwave-assisted digestion (6 mL HCl + 2 mL of HNO3) and Se^IV pre-concentrated by DLLME as the dibenzyldithiocarbamate. The CHCl3 phase was diluted to 0.2 mL with 10% HNO3 in EtOH. Total Se was determined after Se^VI was reduced with HCl, and Se^VI was determined by difference. The method was validated by analysis of two SRMs. The LOD was 1 μg L^−1	180
SeMet	Selenized yeast	ICP-MS	Seven different extraction procedures were evaluated. Mild conditions with alkaline SDS resulted in the incomplete hydrolysis of all peptides and proteins necessary to prevent analyte permeation through the asymmetric FFF separation membrane	24
Si	Coffee creamer	ICP-MS	The quantification and size characterization of SiO2NPs in commercial high fat coffee creamer was achieved using asymmetric flow FFF coupled to ICP-MS, after extraction of SiO2NPs with hexane	237
Si	Beer	HR-CS-AAS	See Cl, ref. 69	69
Ti	Food samples (confectionary foods, chewing gums, chocolates and white coloured foods)	ICP-AES, TEM	The levels of TiO2, a food additive, determined in confectionary foods, chewing gums, chocolates and white coloured foods by ICP-AES, ranged from 3 to 2400 mg kg^−1. The size of TiO2 particles in the same foods samples was determined by TEM and varied from 30 to 410 nm	119
V	Beverages (apple juice, beer, red and white wine)	ICP-MS, TXRF	TRXF measurements revealed V concentrations in diatomite filtering aids ranging from 38 to 368 mg kg^−1. To assess the release of V from diatomite used for clarification in beverage production, 1 g of apple juice, beer, red and white wine was digested in a microwave oven with 4 mL HNO3–2 mL ultrapure water up to 180 °C. Digested samples were made up to 40 mL with water, then diluted 1 + 3. ICP-MS analysis was performed in KED mode, using a He and Rh as the internal standard. The results confirmed the release of V from the diatomite filter aids in the respective filtered beverages	238
V	Water (bottled)	TXRF	V was determined in bottled water by TXRF with an improved LOD of 1.4 μg L^−1, due to a new sample preparation method. The procedure required partial dissolution with acetone of the hydrophobic film coating the sample holder, followed by deposition of a 200 μL drop of sample, heating and drying under mechanical vibration	239
Various	Coffee	PIXE	The analysis by PIXE of samples of ground and spent coffee, obtained after drip brewing with mineral water at temperatures varying from 20 °C to 80 °C, revealed different, temperature dependent, extraction patterns for elements such as Ca, Cl, Cu, K, P, Rb and Si	109
Various	Vegetables (onion, radish, fenugreek, French basil, spinach and roselle)	ICP-AES	Ten “microgreens”, including onion, radish, fenugreek, French basil, spinach and roselle, were analysed for protein, dietary fibre profile, elemental profile and vitamin (ascorbic acid, alpha-tocopherol and beta-carotene) content. They provided moderate to good intake of nutrients, excellent for vitamins	240
Various	Olive oil	LIBS	The classification of olive oil, based on its geographical origin, was achieved using elemental data obtained by LIBS on 36 samples, assisted by machine learning algorithms	83
Various	Himalayan salt, table salt	PIXE	A comparison between local table salt and Himalayan salt samples, carried out using PIXE, indicated that the latter contained larger amounts of Fe, K, Mg, Si and Ti	110
Various	Honey, sweetener syrups	LIBS	LIBS spectra were obtained for 6 pure honey samples, 2 sweetener syrups (high fructose corn syrup and sugar cane syrup) and 228 honey samples fortified with these adulterants from 5% to 95% (w/w). PLS discriminant analysis classification models were developed and validated, allowing the detection of the adulterants with 100% accuracy. For the quantification of adulterants, PLSR calibration models were investigated with promising results for samples containing more than 20% (w/w) of the adulterants	81
Various	Rice grains, papaya roots	LIBS, WDXRF	LIBS, WDXRF and FTIR were applied to investigate elemental and molecular pathological changes caused by plant diseases. In two papers, potential changes associated with either false smut disease, affecting rice plants, or root knot nematodes, a type of parasite infesting papaya crops, were studied	86 and 87
Various	Honey	MIP-AES, AMS	Samples of acacia honey, collected over the period from 1958 to 2018, were analysed by MIP-AES, for their elemental composition, and by AMS, for radiocarbon dating. The results indicated elemental changes with time, allowing for the potential use of acacia honey as a biomarker of environmental changes	241
Various	Saffron, radish, and corn silk	LIBS	The potential for rapid discrimination between saffron and other plants using PCA applied to the elemental profile determined by LIBS was evaluated, using three plants (saffron, radish, and corn silk) as target materials	84
Various (4)	Acacia honey, rape honey, high fructose corn syrup	LIBS	The determination of Ca, K, Mg and Na by LIBS, combined with chemometric methods allowed the identification of adulteration of acacia honey with high fructose corn syrup and rape honey. The methods used included univariate analysis, PLS regression and techniques to assess variable importance	82
Various (4)	Chinese herbal foods	CVG-AFS	Samples were dried (80 °C for 12 h), then ground (80 mesh) and 500 mg transferred to a PTFE vessel, followed by 12 mL HNO3. After 24 h, 3 mL of 30% H2O2 was added and the mixture heated at 120 °C for 6 h. HCl was added and the solution heated for 30 min (to reduce Se^VI to Se^IV). After cooling, thiourea was added (to reduce As^V to As^III, and Sb^V to Sb^III). LODs were 0.05 ng mL^−1, 0.006 ng mL^−1, 0.03 ng mL^−1 and 0.05 ng mL^−1 for As, Hg, Sb, and Se, respectively	92
Various (5)	Teas, CRM	ETAAS	Five elements (As, Cd, Cu, Hg, Pb) were determined in teas by ETAAS, following a new vortex-assisted liquid phase micro-extraction procedure based on a solidified DES, consisting of 1-decyl-3-methylimidazolium chloride and n-butanoic acid	177
Various (5)	Wine	ICP-AES	The determination of 5 elements (Ba, Ca, Mg, Mn and Sr) in 125 wines, combined with information regarding soil type, meteorological conditions and wine colour, provided useful insights to explain their variability and potential use for origin classification	181
Various (5)	Rice	ICP-MS, CVG-AFS, SR-XRF	See Hg ref. 93	93
Various (6)	Instant soups, CRMs	MIP-AES	Six elements (Cu, K, Mg, Mn, P and Zn) were simultaneously determined in instant soups by MIP-AES. 500 mg of the grounded and homogenised sample underwent microwave-aided digestion with 1.0 mL conc. HNO3–1.0 mL 30% H2O2–6.0 mL ultrapure water, for 40 min, up to a maximum temperature of 200 °C and 1000 W power. The digested samples were made up to 15 mL with ultrapure water. Precision was <8.4%. Spiking experiments and analysis of 3 CRMs (NIST SRM 1515 apple leaves, NIST SRM 1568b rice flour and NIST SRM 3234 soy flour) yielded recoveries ranging from 83.3% to 107.5%. LODs were: 0.09 mg kg^−1 (Cu), 4.9 mg kg^−1 (K), 1.0 mg kg^−1 (Mg), 0.04 mg kg^−1 (Mn), 5.4 mg kg^−1 (P) and 0.88 mg kg^−1 (Zn)	75
Various (7)	Canned food (tomatoes, sardines)	ICP-AES, SEM/EDX	An investigation of potential migration of chemical elements from cans into simulated media or food during shelf storage was carried out measuring 7 elements (Al, Cd, Cu, Fe, Pb, Sn and Zn) by ICP-AES and by examining the can inner surface by SEM/EDX. Acidic media and or the presence of tomato aided the migration of Fe and Zn, both in simulants and in food and caused the corrosion of the inner sides of the cans	42
Various (8)	Rice wholemeal flour, sourdough, muffins	FAAS	Sample (3 g) was ashed for 10 h at 550 °C then dissolved in 20 mL of 20% HCl the LODs were 0.2 mg kg^−1 for K and Mg; 0.6 mg kg^−1 for Fe; 0.3 mg kg^−1 for Ca, Cr, Cu, Mn and Zn. The researchers concluded that fermentation of rice flour with lactic acid bacteria could lead to new gluten-free products with improved nutritional values	242
Various (8)	Sugarcane products	MIP-AES	To analyse samples of sugars (refined, crystal, demerara and brown) and derived products (molasses and brown sugar candy) by MIP-AES, 1000 mg sample were acid digested with 5.0 mL conc. HNO3–1.0 mL H2O2 for 4 h at 150 °C, using a reflux system. The LODs were 0.03 mg kg^−1 (Ba); 0.03 mg kg^−1 (Ca); 0.04 mg kg^−1 (Cu); 0.08 mg kg^−1 (Fe); 0.18 mg kg^−1 (K); 0.03 mg kg^−1 (Mg); 0.02 mg kg^−1 (Mn) and 0.26 (Na) mg kg^−1, respectively. Spike recoveries ranged from 80% to 119%	243
Various (9)	Milk, milk products, infant formula, and adult nutrition products	ICP-AES	Microwave-aided digestion and ICP-AES was used for the determination of 9 elements (Ca, Cu, Fe, K, Mg, Mn, Na, P, and Zn) in several matrices. This was tested in a collaborative study with 14 participants, involving 25 products and the NIST SRM 1849a. Recoveries of certified values ranged from 98% to 101% for the nine elements. Reproducibility RSD was <10% for Ca, K, Mg, Na and P, in all cases, and in the fortified nutritional product samples for the trace elements, with the exception of Mn. The reproducibility RSD for Cu, Fe, Mn and Zn ranged from 13.2 to 82.8% in these unfortified samples and from 11.4% to 55.0% in the dairy samples	26
Various (9)	Rice, SRM	ICP-AES, HG-AAS	Microwave-aided digestion of 500 mg of rice samples was carried out using 5 mL 1 mol L^−1 HNO3–2.5 mL 30% H2O2, with a heating program of 400 W for 4 min and 800 W for 10 min. The digests were diluted up to 25 mL prior to analysis by ICP-AES (14 elements) or HG-AAS (As). The destruction of organic matter, monitored via spectrophotometric determination of the residual C, was >89%. Acceptable performances, in terms of recoveries of CRMs and precision, were achieved for Al, Ca, Cr, Cu, Fe, K, Mn, Mo and Ni. For these elements, LODs ranged from 0.0087 mg kg^−1 (Mn) to 1.6 mg kg^−1 (Ca)	44
Various (9)	Edible honeysuckle (Lonicera japonica Thunb.)	ICP-MS	The concentrations of 9 elements (As, Cd, Cr, Cu, Hg, Mn, Ni, Pb and Zn) were determined by ICP-MS in both honeysuckle flowers and after soaking or boiling for 30 min. The study showed no evidence of risk from the consumption of these flowers, except for a potential risk associated with the presence of Cd	244
Various (9)	Infant formulas, CRMs	ICP-AES	A simple preparation method, based on treatment of infant formula samples in an ultrasonic bath, was developed for the simultaneous determination of 9 elements (Ca, Cu, Fe, K, Mg, Mn, Na, P and Zn) by ICP-AES. Recoveries and RSD ranged from 91% to 105% and from 1.1% to 5.2%, respectively. In addition, the best conditions to conduct in vitro studies of the dialyzability of these elements from infant formulas were investigated	45
Various (10)	Edible flowers	ICP-AES	10 elements (Cd, Co, Cu, Fe, Mn, Ni, Pb, Sr, V, and Zn) were determined in edible flowers. The study showed higher concentrations of Mn, in the flowers of Acmella oleracea, Cu and Sr in those of basil (Ocimum basilicum) and pumpkins (Cucurbita moschata and C. pepo) and Ni in those of orange daylily (Hemerocallis fulva)	245
Various (10)	Tofu	ICP-AES	The content of 10 elements (Ca, Co, Cu, Fe, K, Mg, Mn, Mo, Na, Zn) were determined in 130 samples of tofu by ICP-AES to assess the contribution of tofu consumption to the dietary intake of these elements	246
Various (10)	Teas (loose leaf and bags)	ICP-AES	To aid classification of black and green teas, the total content of chemical elements (Al, Ba, Ca, Cu, Fe, Mg, Mn, Ni, Sr and Zn) and their fractions were determined in tea infusions by ICP-AES. Fractions were operationally defined on the basis of their separation, due to hydrophobicity and charge, carried out by tandem column SPE, using reverse-phase and strong cation-exchange extraction tubes connected in series	247
Various (10)	Food supplements	EDXRF	For all samples, the powder was pressed for 2 min under 10 tons to make a cylindrical pellet of diameter 20 mm and a thickness 1 mm. Spectra were acquired with a custom tri-axial setup that lowered the LODs by polarization of the incident beam at a secondary target. Values for As, Ca, Cl, Cu, Fe, Mn, K, Pb, Sr and Zn ranged from 1 to 60 μg g^−1. The method was validated by the analysis of six CRMs	167
Various (10)	Coconut sugar	EDXRF	In the attempt to distinguish coconut sugar from cheaper alternatives, such as cane and beetroot sugar, elemental profiles, consisting of 10 elements (Br, Ca, Cl, Cu, Fe, K, P, Rb, S and sr) were determined in 11 coconut, 10 cane and 1 beetroot sugar samples by EDXRF	103
Various (11)	Wine (red)	MIP-AES	Eleven elements were determined in red wine samples by MIP-AES. Dilution ratios were established as 1 + 99 (K); 1 + 49 (Ca, Mg, Na); 1 + 9 (Al, Fe, Mn, Rb, Sr) and 1 + 3 (Cr, Cu). Three matrix-matched quantification methods were evaluated (standard additions; modified standard dilution and multi-energy calibration) of which, the first allowed quantification of all analytes, the second only for the seven minor elements and the third only for the four major ones. Chemometrics (PCA) was applied to evaluate the association between element distribution and wine origin	76
Various (11)	Food supplements	WDXRF	The content of 11 elements (Cr, Cu, Ir, Mo, Mn, Ni, Os, Pb, Pt, Rh, Ru) was determined by WDXRF in 25 food supplements for weight loss to assess potential risks associated with the consumption of these products	169
Various (13)	Human milk	ICP-MS	25 mL human milk samples were collected on three occasions, at 2 week intervals, from 20 breast-feeding mothers, aged 20–40 y. 0.5 mL of sample was digested in a microwave oven at 180 °C for 25 min with 7 mL 65% HNO3–1 mL 30% H2O2, then cooled to 50 °C for 30 min. The content of 13 elements (Ca, Cr, Cu, Fe, K, I, Mg, Mn, Mo, Na, P, Se and Zn) was determined by ICP-MS. Inter-assay RSD ranged from 0.5% to 9.9%. On the same occasions, the food consumption of the subjects, over the previous 24 h, was recorded by interviews and the respective dietary intake calculated according to the available information. The results suggested low maternal intake of oligo-elements and decreasing concentrations of essential minerals with lactation stage	222
Various (14)	Edible animal tissues (snails, woodcock, pheasant, and hare)	ICP-MS	The intake of essential (Co, Cu, Fe, Mn, Mo, Ni and Zn) and non-essential (Al, As, Cd, Cr, Hg and Pb) trace elements from home-processed food obtained from uncommon animal food sources (snails, woodcock, pheasant, and hare) were evaluated	248
Various (14)	Rice	ICP-MS	See As ref. 88	88
Various (14)	Food CRM (Bovine liver, tea leaves, mussel)	TXRF	“Suspension sampling” was investigated as a faster and more convenient sample preparation method for low power TXRF. Samples were ground and suspended in Milli-Q water, which proved better than Triton X-100 or HNO3 as a dispersant. Better results were achieved with smaller average particle size and lower sample mass (20 mg). Ga was used as the internal standard. Recoveries for most elements ranged from 80% to 120% and the RSD was <15%	102
Various (14)	Quinoa (Chenopodium quinoa Willd)	ICP-AES	The content of chemical elements (B, Ca, Co, Cu, Fe, K, Mg, Mn, Mo, Na, Ni, P, S and Zn) in quinoa (Chenopodium quinoa Willd.) and the effect of processing (polishing and milling) as well as cooking (boiling and steaming) was investigated by ICP-AES analysis, after digestion of 500 mg of polished quinoa with 7.5 mL conc. HNO3 in a microwave oven for 2 h at 90 °C	172
Various (16)	Food samples, plant-based, CRMs	ICP-MS	Food sample slurries were introduced directly into ICP-MS for the determination of 16 elements (As, Ba, Ca, Cr, Cu, Fe, Ge, K, Li, Mg, Pb, Rb, Sr, Ti, V and Zn). In the slurry, 90% of the particles were <0.8 μm in diameter. Efficient transportation and complete ionization were achieved for particle size of 8.5 μm and 3–4 μm, respectively. Calibration was performed with aqueous standards, after demonstrating differences <5% with calibration using CRM slurries. Recovery and precision (RSD), assessed on food CRMs, ranged from 93% to 106% and from 0.3% to 8.6%. LODs determined on procedural blanks 500-fold diluted ranged from 2 to 224 ng g^−1	175
Various (16)	Herbal infusions and teas	TXRF	Boiling tri-distilled water was added to 1 g of herbal infusion/tea material (i.e. leaves, bark, flowers, fruits) (50 mL) or tea bags (150 mL), then left for 10 min. Oolong and Pu-erh teas required a multiple-steepings procedure. Thirteen elements (Br, Ca, Cl, Cu, Fe, K, Mn, Ni, P, S, Sr, Rb and Zn) were found in all extracts (29 samples) whereas another three elements (Ba, Co, Sc) only occasionally. PCA was applied to these data to identify characteristic profiles	249
Various REEs (16)	Chinese tea (black tea, green tea, oolong tea and dark tea)	ICP-MS	The contents of 16 REEs in 3011 tea samples, from various areas of China, was investigated by ICP-MS to assess their role in enhancing the yield and quality of tea. Five elements (Ce, La, Nd, Sc and Y) represented >80% of the REES	250
Various (17)	Milk (cow, goat, buffalo, yak, and camel), CRM	ICP-MS	Milks from several animals (cow, goat, buffalo, yak and camel) were analysed by ICP-MS. Samples were thawed overnight at 4 °C, then 2 g (0.3 g CRM) were treated overnight with 6 mL 65% HNO3–2 mL 30% H2O2, prior to microwave-aided digestion at a maximum temperature of 190 °C. The digests were diluted to 25 mL with ultrapure water prior to ICP-MS analysis. LOQs ranged from 0.0003 mg kg^−1 (Cd) to 2.67 mg kg^−1 (Ca) and recoveries of certified values varied from 88.2% to 114.3%. Chemometric analysis was applied to evaluate the correlation between the distribution of Al, Ca, Cd, Cr, Cu, Fe, K, Mg, Mn, Na, Ni, Pb, Se, Sn, Sr and Zn and milk type	170
Various REEs (17)	Rice	ICP-MS	See As ref. 88	88
Various (20)	Wine	LA-ICP-MS, ICP-MS	A procedure was reported for the determination of 20 elements (Al, B, Ba, Ce, Co, Cs, Cu, La, Li, Mg, Mn, Nd, Pb, Pr, Rb, Sr, U, V, Y and Zn) in wine samples by LA-ICP-MS, without any sample pre-treatment and calibration based on either the NIST SRM 610 glass or aqueous standard solutions with In as internal standard. A comparison with ICP-MS was carried out. LODs for the LA-ICP-MS approach ranged from 0.01 ng mL^−1 to 1 ng mL^−1	183
Various (20)	Foodstuffs (Brazil nut, golden berries, heart of palm and dietary supplements containing acai fruit)	ICP-MS	The content of major, minor and trace elements in selected foodstuffs from the Amazon region and dietary supplements was assessed by ICP-MS after microwave-aided acid digestion, to estimate the contribution of the consumption of such items to dietary intakes	251
Various (21)	Wine	ICP-AES	The elemental profile of non-commercial traditional wines from rural areas was investigated using data for 61 elements measured in 62 homemade white and red wines and multivariate pattern recognition analysis. The information provided by Fe, Mg and Sr allowed wines to be differentiated by geographic regions	252
Various (5) and REEs (16)	Teas, CRMs	ICP-MS	The characteristics of tea slurries (particle size, stabilization and homogenization) as well as methods to achieve them, were thoroughly examined. The proposed procedure involved milling of 0.10 g of a dried or freeze-dried tea or CRM sample in a mixture of 1.5 mL 0.5% (m/v) polyethylene imine–0.5 g ZrO beads for 120 s, to achieve average particle sizes in the range of 0.9–1.0 μm, with 90% of the particles <1.5 μm. The slurries were then diluted to 50 mL with 0.5% (m/v) polyethylene imine and introduced directly into ICP-MS. Aqueous standards were used for calibration, since particles with an average size of 1 μm showed a behaviour comparable to that of aqueous solutions. The LODs ranged from 0.03 μg kg^−1 (for Tm) to 1.2 μg kg^−1 (for Cr). Analysis of tea CRMs gave recoveries between 90% and 105%. The method was applied to the screening of As, Cd, Cr, Hg and Pb as well as 16 REEs in 20 Pu’er teas	178
Various (30)	Total diet study	ICP-MS	In a multi-regional total diet study in Sub-Saharan Africa, 2700 food samples, grouped into 225 food composite samples, corresponding to 13 major food groups, were analysed for the content of 30 elements according to a validated method based on ICP-MS. Compliance of Al, As, Cd, Hg and Pb levels with limits set by the Codex Alimentarius and the impact of artisanal cookware were discussed	253
Various (41)	Dairy milk, plant-based milk alternatives, CRMs	ICP-MS, CVAAS	The levels of 41 elements were assessed in samples of milk (cow, goat, and donkey) and plant-based milk alternatives (from soy, rice, oat, spelt, almond, coconut, hazelnut, walnut, cashew, hemp and quinoa)	254
Various (41)	Seafood (clams and oysters)	spICP-MS, ICP-MS	41 elements were detected by ICP-MS in acid digested shellfish seafood (clams and oysters), of which 20 were selected for further analysis by spICP-MS to characterise NP concentrations and particle size distributions. 2 mL of 20% v/v TMAH was added to 0.1 g of seafood tissue, homogenised and sonicated for 60 min at 37 °C in an ultrasonic bath, then placed on a shaker for 24 h at 80 rpm. The samples were then filtered and diluted with 0.1% Triton X-100 to at least 1% TMAH concentration before analysis. With this procedure, recovery, determined using AuNPs, was 95.9%, as compared to 103.6% by means of enzymatic digestion. NPs of Ce, Gd, La, Pr, and Y were detected in clams at concentrations ranging from 0.6 ng g^−1 to 37.7 ng g^−1. In oysters, NPs (Ce, La, Nb, Pr, and Y) concentrations varied between 4.2 and 19.7 ng g^−1	176
Various (42)	Saffron	ICP-MS	The determination of 42 elements and 5 stable isotopes (^13C, ^15N, ^34S, ^2H and ^18O) by ICP-MS in 67 samples of saffron from different geographical areas allowed to discriminate the spice according to origin with a specificity between 83% and 100%	185
Various (61)	Wines	ICP-MS	Elemental profiles and stable isotope ratios (^13C : ^12C, ^18O : ^16O) were determined in 142 wine samples to derive a method to classify by geographical origin. For ICP-MS analysis, 5 mL of sample was placed on an electric heating plate at 80–90 °C until the volume was reduced to 2 mL. After cooling to room temperature, 5 mL of conc. HNO3 and 2 mL of H2O2 were added and the sample evaporated to 2 mL at 110–120 °C, cooled to room temperature and diluted to 25 mL with ultrapure water. Carbon and oxygen isotope ratios were determined by IRMS, after equilibrating 200 mL of wine with CO2 at 50 °C for 12 h. Multivariate analysis was applied to these sets of data to develop models for the correct origin classification	184
Zn	Mineral water, isotonic sports drinks	HG-AFS	Zn was determined in selected beverages using a new online pre-concentration system coupled with HG-AFS. Samples of mineral water (10 mL) or isotonic sports drinks (5 mL) were adjusted to pH 8.00 with 0.1 mol L^−1 NaOH then diluted to a final volume of 25 mL with 6.5 mmol L^−1 Na2CO3 buffer. A mini-column, filled with 100 mg of polyurethane foam was coupled to the FI system and impregnated with 0.10% 4-(2-pyridylazo) resorcinol (w/v) at a flow rate of 0.50 mL min^−1 for 6 h, followed by 10% (w/v) NaOH, 5% (v/v) HCl and ultrapure water. After loading the sample onto the column, the Zn complex was eluted into the HG-AFS, where ZnH2 was formed by reaction with NaBH4. The method was optimised using a full two-level factorial design. With an enrichment factor of 88.92, the LOD and LOQ were 0.03 μg L^−1 and 0.09 μg L^−1, respectively. Zn concentration in mineral water and isotonic sports drinks ranged from 29.00 to 78.59 μg L^−1	95

---

8.1 Progress for individual elements

8.1.1 Arsenic. Arsenic in rice is still a common topic for study with many papers furthering studies previously reported. Gu et al.⁹¹looked at the levels of As in rice products with a particular focus on infant foods and regulatory levels in Australia. Whilst European regulations limit levels to <0.1 mg kg⁻¹ of iAs in infant foods Australia has only a general limit of <1 mg kg⁻¹ total As, for all populations. This study measured total As and iAs in a variety of rice based infant foods. The concentration of total As was determined by ICP-MS following digestion in HNO₃, whereas that of iAs was measured by CV-ICP-MS following digestion with HClO₄ and reaction with HBr and hydrazine sulphate. The level of iAs was typically found to be between 60% and 85% of the total As content, meaning that, in more than 75% of the samples analysed, the level of iAs was >0.1 mg kg⁻¹. The authors concluded that the current regulatory limit may be insufficient to protect the infant population from unacceptable exposure to iAs.

Organic arsenic species are frequently studied, often as oxo-arsenic species, less frequently studied are the analogue thio-arsenic species. Kaňa et al.⁴⁸ developed a method for separating thio and oxo arsenic species in the same samples, particularly to establish the levels of thio species naturally present in foods. A successful ion pair HPLC-ICP-MS method for separation and identification of thio and oxo species was reported. Known amounts of arsenic-containing compounds were spiked into fish, seafood and mushroom samples prior to a simple extraction into water by use of an ultrasonic bath. Spike recoveries of all species were between 96 and 121%. Unsurprisingly, oxo-arsenic species were common in all sample types, thio-arsenic species were only found to be naturally present in two of the shellfish samples and at around 20 ng g⁻¹.

8.1.2 Mercury. The number of papers focusing on Hg were fewer in this review period than reported in the previous ASU.¹ The main paper of note⁹³ looked at total Hg and MeHg distribution in rice crops and how this was affected by an increase in Se supplementation of crops grown in the field. The addition of Se at the rate of 0.15 mg kg⁻¹ soil led to reduction of MeHg by between 23 to 86%. The distribution of MeHg was shown to be decreasing from the roots, straw, grains and hulls. Increasing the amounts of Se was shown to reduce the levels of MeHg and increase the levels of Cu, Fe and Zn as shown by XRF mapping of the grains, thus showing a potential improvement in the nutritional quality of the grains.

8.2 Single and multielement applications in food and beverages

8.2.1 Dietary intake. A less common technique, MIP-AES, was used to study a variety of instant soups for nutritional content.⁷⁵ Samples were digested in dilute HNO₃–H₂O₂ by microwave-assisted digestion and analysed by MIP-AES for Cu, K, Mg, Mn, P and Zn. The method was verified by analysing CRMs, giving recoveries between 83% and 108% for all elements, and spiked samples giving recoveries of 86% and 106%, thus Carvalho et al.⁷⁵ demonstrated the application to be suitable for analysis of this matrix.

8.2.2 Human milk and infant formula. The use of LIBS to measure Mg in infant milk formula was explored by Markiewicz-Keszycka et al.⁸⁵ Total content and within sample variability were investigated with good correlation (R² = 0.93) seen between samples measured by a reference method (FAAS) and LIBS. The LOD was established to be 150 mg kg⁻¹, which, although relatively high compared to other method such as AAS, is appropriate for Mg levels in formulas. Good homogeneity was demonstrated within sample aliquots. The method offers potential for rapid, semi-destructive measurement of Mg (and other nutritional elements) by LIBS, potentially in real-time process control situations.

With an increase in meat-free diets, it is perhaps unsurprising that there would be interest in vegan options for infant nutrition. Gu et al.⁹¹ investigated a range of rice based infant formulas and foods for iAs and total As levels. Formula was shown to have a lower percentage of iAs (37.4%) than rice pasta and crackers (84.8% and 74.3% respectively). In all cases levels were above the EU limit of 0.1 mg kg⁻¹ (See Section 8.1.1).

8.2.3 Dairy products. In a novel study, 17 elements were measured in milk from 5 different mammalian species, namely, cow, goat, buffalo, yak and camel.¹⁷⁰ Microwave-assisted digestion followed by ICP-MS was used to measure nutritional, trace and toxic elements. Results were categorised into major elements Ca (516–888 mg g⁻¹), K (698–1377 mg g⁻¹) and Na (253–428 mg g⁻¹), minor elements Al (0.277–0.493 mg g⁻¹), Cu (0.165–0.522 mg g⁻¹), Fe (1.01–2.54 mg g⁻¹), Mg (58.5–96.7 mg g⁻¹), Mn (0.156–0.256 mg g⁻¹), Sr (0.695–3.05 mg g⁻¹) and Zn (3.11–5.81 mg g⁻¹) and trace elements which were found in decreasing concentrations in the order (Ni > Sn > Se > Pb > Cr > As > Cd). The data obtained underwent principal component and linear discriminant (LDA) statistical analysis with significance set at p < 0.05. From this test, As, Ca, Cr, Cu, Fe, K, Mg, Na, Ni, Pb, Se, Sr and Zn were selected to generate the discriminant model. It was found that here was a strong correlation between elemental composition and species, with the LDA model successfully predicting species in most cases. The technique could be useful in discriminating milk types and could be a useful tool for demonstrating authenticity and origins of milks, particularly those with higher market values.

8.2.4 Cereals. Arsenic and Sb are often found together as soil pollutants and therefore end up in plant structures. As these elements share the same s²p³ outer electron configuration, some of the behaviour in biological systems is similar. In a study by Wu et al.,¹²⁴ HG-AFS was used to measure total As and Sb in paired rice and soil samples, speciation was carried out by HPLC-ICP-MS in whole rice samples, μXRF was used to examine location of As, Fe, Mn and Zn within the grain and due to spectral interferences, LA-ICP-MS was used to measure As and Sb. Having located soil As hotspots, μXANES was used for further investigation. Although Sb was more prevalent in the soils than As (Sb 5.91–322.35 mg kg⁻¹ and As 0.01–57.21 mg kg⁻¹), more As was found in the plant due to its greater bioavailability, whereas, for Sb, a bioavailability as low as 2% was previously reported. Both elements followed the same reduction in concentration in plant structures following root > shoot > husk > grain with As species accumulating in embryo and aleurone layer, while Sb was mostly found in husk, bran and aleurone layer. The μXRF analysis showed a close correlation between As and Fe concentrations within the grains and μXANES showed that DMA was the prevalent As species with some iAs being present at between 19–63% of the total As content. The authors concluded that As is more significant as a pollutant in Sb/As contaminated soils, due to its higher bioavailability, and that Fe plays an important role in remediation of soils contaminated with As.

In another study, 10 different varieties of Brazilian rice were tested for REEs, nutritional and toxic elements, and As species.⁸⁸ The speciation was carried out by HPLC-ICP-MS coupled with HG, which reduced LODs to 0.004 mg L⁻¹ for DMA, 0.003 mg L⁻¹ for MMA and As^III and 0.010 mg L⁻¹ for As^V. Mean values across the 10 varieties of rice studied showed that iAs was the major species but all samples analysed had levels of As within the acceptable limits for infant consumption. There was strong evidence for increased concentrations of nutritional elements causing a reduction in As and Cd, demonstrating that biofortification can both improve nutritional values of rice and reduce the levels of toxic elements, thus concluding that a daily intake of 70 g of rice could provide the RDIs for Mn, a significant percentage of the RDI for Zn and Mg and 1300% of the RDI of Fe.

Elik et al.¹⁷¹ developed a novel, green and rapid sample preparation procedure, based on a natural deep eutectic solvent (NADES) and ultrasonic-assisted microextraction, for the determination of iAs and iSe in rice by GF-AAS. Samples were pre-digested using conventional microwave-assisted digestion, the extraction method utilised ultrasonic extraction into a proline/malic acid 1 + 1 NADES mixture with THF. To validate the method, CRMs were spiked and excellent recoveries were seen for both iAs (97.2–98.8%) and iSe (97.9–98.9%).

In India, rice is consumed by two thirds of the population and is a key crop with a large impact on the economy. The second largest rice growing region in India is Uttar Pradesh, but with considerably lower yields than other regions, attributed to disease, among which false smut. Sharma et al.⁸⁶ used FTIR, LIBS and WDXRF to better understand the effect of false smut on the rice grains. It was found that decreasing concentrations of Ca, Cu, Fe, Mg, Si and Zn, and increasing concentrations of Cr, K, P, and S were good indicators of the presence of the false smut disease. This team then applied the same approach to nematode infected papaya roots.⁸⁷ Again there were differences seen between healthy and diseased plants, with an increase in Al, Ca, Fe, Mg and Si concentrations and a decrease in Cu, K, Mn, Na, P and Zn levels in the infested plants. It was also observed that some elements, such as Mo and Zr, were entirely absent in healthy plants but present in plants with the infection, possibly due to weakening of the plants natural defence systems by the nematodes.

A less commonly consumed grain, quinoa, was the focus of a study by Mhada et al.¹⁷² The team looked at the processing of the grains and the effect on saponins, minerals and total phenols. Macro-element (Ca, K, Mg, Na, P and S) and micro-element (Fe, Zn, Mn, B, Cu, Ni, Co and Mo) levels were measured by ICP-OES, at different stages of grain processing and cooking. During processing, levels of undesirable bitter saponins were reduced by a process of polishing, but this process also reduced the mineral content significantly. Milling was found to reduce the mineral content further and 50% of mineral content was lost due to elimination of teguments and embryos where these minerals are contained. Cooking was also found to reduce mineral content, particularly elements such a B and K which saw a 40% drop. This effect was reduced if the grains were steamed. To maximise the nutritional benefit of this grain, the authors proposed a mixture of wet and dry processing to reduce saponins, retaining whole grains and cooking by steaming.

8.2.5 Vegetables, fruits, mushrooms and nuts. The fate of nanoparticles of Ce and Cu in radish plants and simulated digestion was studied by Hayder et al.¹⁷³ Cerium or CuNPs were dosed into hydroponic solutions in which radish plants were grown under controlled conditions. After the growing period, total Ce and Cu concentrations were measured in the roots and leaves by ICP-MS following microwave-assisted digestion. A higher level of both elements was seen in the roots (with Ce more prevalent than Cu: CeO NP: 1156.3 ± 801.10 μg g⁻¹, CuO NP: 862.52 ± 14.96 μg g⁻¹) than in leaves (CeO NP: 4.0947 ± 2.0782 μg g⁻¹, CuO NP: 16.7434 ± 1.0640 μg g⁻¹). Samples also underwent enzymic gastric and gastrointestinal digestion and it was found that CeO NPs were stable throughout the digestion process, leading to the inference that these NPs could be transferred to the bloodstream, whereas Cu could be present as both ionic and NP forms.

Speciation of toxic elements in mushrooms was reported in two papers. Zou et al.¹⁷⁴ looked at Hg species whilst Komorowicz et al.⁴⁶ focussed on As speciation. For Hg speciation HPLC-ICP-MS was employed to determine iHg as well as MeHg, EtHg and PhHg species in wild mushrooms. Organic Hg species are of interest as these are deemed to be highly toxic and generally more bioavailable than iHg. For this study, 30 samples of wild mushroom were harvested and 65 samples were purchased. All samples were dried to constant weight and then ground to <0.15 mm. Sample aliquots of 0.1–0.2 g underwent ultrasonic extraction in 10 mL of 12 g L⁻¹L-cysteine, at temperatures maintained below 40 °C, to minimise volatilization and degradation of Hg species. The resulting extracts were centrifuged and filtered, to obtain 1 mL solution for analysis. Speciation was carried out using a C8 column, with a mobile phase of MeOH–10 mmol L⁻¹ NH₄CH₃COO–0.8 g L⁻¹L-cysteine (pH = 4.0), followed by ICP-MS analysis in no gas mode at mass ²⁰²Hg. Total Hg was measured using a direct mercury analyser. Optimised conditions were established and linearities, LOD and LOQ established for each species. The LOQs were found to be between 2.0 μg kg⁻¹ and 15 μg kg⁻¹, comparable to literature values. Analysis of the samples found that the predominant species was iHg. Quantifiable concentrations of MeHg were reported for 7 of the 16 samples analysed, while EtHg and PhHg were only detected in 1 sample each. All samples gave acceptable extraction efficiencies when compared to the total Hg measurements (83.8% to 97.5%). It was concluded that the method was a rapid technique for Hg speciation, with a low MeOH uptake and reduced carbon deposition on the cones. Although typically low levels of Hg were seen, typically <3 mg kg⁻¹, the presence of Hg, particularly MeHg, in mushrooms should not be ignored.

In the second study by Komorowicz et al.,⁴⁶ IC-ICP-MS and SEC-UV-Vis were used to study As speciation and the behaviour of these species in model digestive systems – as an assessment of bioavailability. Samples were prepared using ultrasonic assisted extraction (0.3 g sample into 8 mL extraction solution, for 5 min at 23 °C); microwave-assisted extraction (0.3 g, 8 mL ultrapure H₂O, 15 min ramp to 80 °C, hold for 30 min at 80 °C); microwave-assisted extraction and enzymic assisted extraction with 3 steps by the Unified BARGE Method (saliva (S1), saliva and gastric (S2), saliva, gastric and intestinal (S3)). For the determination of total As, samples were prepared by microwave-assisted digestion. Total As content was highly variable, with concentrations ranging from 0.78 mg g⁻¹ to 68.3 mg g⁻¹. In all cases, iAs was shown to be a minor component (<2%), whereas AB or MMA were the major organic As species found. The microwave- and ultrasound-assisted extraction methods yielded similar results and recoveries. The bioavailability of As from the S1, S2 and S3 samples was found to be in the ranges: 73–102%, 74–115% and 18–87%, respectively, of the enzymic assisted extraction. The extracts obtained after enzymic assisted extraction were subjected to SEC-UV-Vis followed by ICP-DRC-MS. The main signal was obtained in the area of a Mr of ∼5 kDa, for all mushroom extracts, and the signal was confirmed as being an As bound protein, with a secondary unknown signal. Further work is proposed to understand these signals and establish a quantification method for the enzymic assisted extraction products.

Cui et al.¹⁷⁵determined the concentrations of 16 elements (As, Ba, Ca, Cr, Cu, Fe, Ge, Li, Mg, Pb, Rb, Sr, Ti, V and Zn) in slurried food samples by ICP-MS. Rather than solubilising samples, the team ground dried food samples to a fine size to allow for direct nebulisation. It was found that a particle size of <8.5 μm provided for efficient transportation to the plasma, though particles of 3–4 μm were required for complete ionisation. A variety of CRMs were tested, giving recoveries between 93% and 106%, thus demonstrating the potential for a simplified preparation of food samples for nutritional studies.

8.2.6 Fish and seafood. Zhou et al.¹⁷⁶ examined elemental composition, by ICP-MS, and the presence of NPs in clams and oysters by spICP-MS. The total concentrations of 41 elements were measured after acid digestion of the samples. From these results, 20 elements were identified for further analysis by spICP-MS, following alkaline digestion, which was utilised to maintain integrity of the particles. Within the 20 elements, 6 REEs were detected in NP form. These were present as particles, ranging from 30 to 65 nm in oysters and from 35 to 55 nm in clams, with Ce, La, Pr and Y common in both species, while Gd and Nd were only observed in either clams or oysters, respectively. The NPs accounted for between 3.4% and 50% of the total metal content.

Metal migration was studied ⁴² in standard and easy-open can types, using simulated fillings (oil, CH₃COOH and EtOH) and canned tomatoes, tomato puree and sardines in oil and tomato. Cans were left for long periods, to simulate products on the market, then their levels of Al, Cd, Cu, Fe, Pb, Sn and Zn were determined by ICP-OES, to assess metal migration from the packaging. In the model systems, CH₃COOH was shown to be the most corrosive agent, as indeed these samples were analysed after a mere 265 days due to corrosion being observed on the cans. An increase in corrosion was also observed in the samples of canned sardines in oil and tomato. Regarding the simulated fillings, oils and EtOH showed a far smaller level of corrosion. In these samples, Fe and Zn were the elements showing the highest migration, with only minimal amounts of Al, Cu, Pb and Sn measured. Cadmium was seen in the sardine samples, but this was concluded to be the endogenous Cd content of the fish. Scanning electron microscopy and EDXRF were used to examine the areas of corrosion. The results of these assessments indicated that the observed migration of Fe and Zn ions originated from the high corrosion zones, particularly badly tinned, uncoated junction joints and easy-open lids.

8.2.7 Drinking water and non-alcoholic beverages. A procedure based on a Vortex assisted LPME was used to investigate the levels of potentially toxic elements (As, Cd, Cu, Hg and Pb) in a range of teas by ETAAS. Ahmadi-Jouibari et al.¹⁷⁷ developed a method for a ‘close to room temperature’ extraction using 1-decyl-3-methylimidazolium chloride–n-butanoic acid as the extraction mixture. The sample was digested with HNO₃–H₂O₂ in a microwave oven and an aliquot of the filtered digestate underwent vortex assisted LPME with diethyl dithiophosphoric acid, as a complexing agent, and the extraction mixture, at 50 °C. The mixture was cooled on ice to solidify the enriched extraction mixture, which was then removed, melted and dissolved in ACN for analysis. This procedure gave an enrichment factor between 164 and 235, allowing for very low LODs to be achieved (Cu & Pb – 0.03 μg kg⁻¹, As & Hg – 0.1 μg kg⁻¹ and Cd – 0.005 μg kg⁻¹). Spiked samples and CRMs were analysed to assess the method’s recovery, which was found to be between 88% and 110% for all targeted elements. On the basis of these data, the method was considered a cost effective, sensitive and suitable option to determine As, Cd, Cu, Hg and Pb levels in teas.

da Costa et al.⁹⁵utilised SPE and HG-AFS to determine Zn in mineral waters and isotonic drinks. Samples were pH adjusted to 8.0 with NaOH and buffered with Na₂CO₃ solutions, giving a sample dilution of 2.5-fold for waters and 5-fold for isotonic drinks. Columns were prepared using a 0.1% 4-(2-pyridylazo) solution and used in an on-line pre-concentration system. Samples were loaded onto the column, then eluted with 3% HCl directly into the HG-AFS. The pre-concentration conditions were optimised using factorial and Box–Behnken design. The method, proved to be effective by evaluation of samples with known Zn content, giving an enrichment factor of 89, recoveries between 84% and 113%, excellent LODs (0.03 μg L⁻¹) and a linear range between 0.09 and 25 μg L⁻¹.

Guo et al.¹⁷⁸developed a direct particle nebulisation technique for analysis of toxic elements and REEs in teas by ICP-MS. These researchers studied the effect of particle size on the analytical results and concluded that a maximum size of 3 μm was necessary to achieve complete (>95%) ionization. Responses from slurried samples were compared with those from the same, acid digested, samples, showing good correlation (R² = 0.993) and demonstrating the reliability of a simplified preparation for tea samples. The method was further verified by analysis of CRM tea samples, returning 89–109% recovery for toxic elements and REEs.

Micro-extraction techniques for the determination of trace elements in water were reported in several papers. Determination of Cd by MIP-AES was investigated with an enrichment factor between 42 and 46 and a reported LOQ of 1 mg L⁻¹.⁷⁴ Chromium speciation was carried out by complexation of Cr^III with 1-(2-pyridylazo)-2-naphthol followed by FAAS. The method gave an LOD of 0.4 ng mL⁻¹ and a pre-concentration factor of 39. The same method was also applied and found suitable for rice and sausage following extraction with HCl.¹⁷⁹

Selenium speciation by ETAAS was also studied.¹⁸⁰ Selenium (Se^IV) in water samples (tap, drinking, river, sea, spring and mineral) was preferentially complexed with sodium dibenzyldithiocarbamate and extracted into CHCl₃ and THF. A series of spiked samples showed recoveries between 97–100%. Total Se was also measured and Se^VI calculated by difference. Food samples which had previously undergone microwave-assisted acid digestion were also subject to the same treatment and good recoveries from RMs were obtained.

8.2.8 Alcoholic beverages. Several studies have looked at the elemental compositions of wine. In their work, Blotevogel et al.¹⁸¹used ICP-OES to evaluate the effect of weather and soil chemistry on elemental content in wines. A variety of wines from Europe were diluted with H₂O and analysed for Ba, Ca, Mg, Mn and Sr. The results, combined with meteorological and soil composition data, underwent multi-variate statistical analysis. Soil chemistry, unsurprisingly, had an impact on the elemental profile of the wines, particularly with respect to Ca. Climate also influenced the elemental content of wines, with higher rainfall giving greater Ba and Mn contents while higher Sr concentrations were associated with higher summer temperatures. The Ca : Mg ratios were also linked to wine colour, and therefore winemaking style, likely due to precipitation of Ca during fermentation.

In another study a team from the University of Guanajuato⁷⁶ utilised MIP-AES to look at major and trace elements in Mexican wines. A successful simple dilute and analyse method was established, the results obtained also indicated that the regionality of the wines could be identified by the elevated presence of certain elements, with the alkali and alkali earth elements being key indicator elements. Concentrations of Al 0.28–0.78 mg L⁻¹, Ca 63.9–121 mg L⁻¹, Cu 0.10–0.42 mg L⁻¹, Fe 0.21–2.11 mg L⁻¹, K 866–4896 mg L⁻¹, Mg 56.6–164 mg L⁻¹, Mn 0.69–1.72 mg L⁻¹, Na 32.7–326 mg L⁻¹, Rb 0.35–6.31 mg L⁻¹, and Sr 0.56–1.90 mg L⁻¹ were observed, but Cr was not detected.

Aluminium speciation in wine was investigated using LC-ICP-MS with a PSDVB–sulfonate exchanger analytical column for separation. The method developed by Karas et al.¹⁸² was found to successfully separate several organic and inorganic species, including inorganic Al³⁺, deemed to be the species of most concern. In general, white wines were found to contain higher levels of aluminium, as a comparison, the average contents were: Al¹⁺ 0.51 mg L⁻¹; red 0.34 mg L⁻¹. Al²⁺ white 0.11 mg L⁻¹; red 0.06 mg L⁻¹. Al³⁺ white 0.53 mg L⁻¹; red 0.09 mg L⁻¹. The dilution approach was shown to be suitable and torch clogging did not occur.

An unusual application of direct LA-ICP-MS on liquid samples was utilised by Liao et al.¹⁸³ As the internal standard In was added to 1 mL of each wine sample at the ratio of 10%, which was then placed into wells in a Teflon sheet to separate the samples. After LA, the resulting vapour was swept into the ICP-MS using He. A suite of 20 elements was determined by this method and, on the same set of samples, also by conventional solution nebulisation ICP-MS following acid digestion. With the optimised conditions, results showed good correlation with conventional ICP-MS. The use of LA-ICP-MS allowed for rapid analysis of wine samples, with minimal impact of effects from varying dissolved solids and EtOH levels. By virtue of direct analysis and elimination of dilution steps in the preparation, LODs were less than 1 ng mL⁻¹, for the majority of the elements of interest, and around 0.01 ng mL⁻¹ for Ce, La, Pr and U. The authors suggest that the technique could easily be employed for other beverages as well.

Su et al.¹⁸⁴ carried out a study combining data from ICP-MS and IRMS to verify the origin of Chinese wines (see Section 8.2.8).

Beer, as the world most widely consumed alcoholic beverage, has also been studied. One such study⁶⁹ looked at Cl⁻, Fe and Si in beers using HR-CS-GF-AAS following an unusual preparation technique, utilising UV light in the presence of peroxide. The levels of Cl⁻ and Fe are both important in the manufacture of beer, affecting taste, clarity and stability. Silicon is studied as a nutritional component, with bioavailable Si believed to be important in reducing the toxicity of ingested Al. By use of a xenon discharge lamp, simultaneous AAS measurements of Fe and Si levels could be carried out, whilst with molecular absorption it was possible to determine Cl⁻ stabilised by In. The obtained LOD and LOQ were 0.05 mg L⁻¹ and 0.17 mg L⁻¹ for Cl⁻, 0.08 mg L⁻¹ and 0.26 mg L⁻¹ for Fe and 2.0 μg L⁻¹ and 6.7 μg L⁻¹ for Si. The method was applied for the analysis of 10 beer samples, with concentrations ranging between 106 mg L⁻¹ and 277 mg L⁻¹ for Cl, between 15 mg L⁻¹ and 37 mg L⁻¹ for Si, and between <20 μg L⁻¹ and 73 μg L⁻¹ for Fe. Four spiked samples, covering at least 2 orders of magnitude in concentration, were prepared for each element and returned satisfactory recoveries, within ±20% of the spiked amounts, thus indicating that the proposed method may be a useful rapid single technique for the determination of these elements in beer.

8.2.9 Authenticity. Su et al.¹⁸⁴ measured the concentrations of a large suite of elements by ICP-MS and the carbon and oxygen isotope ratios by IRMS in wine. The authors noted that the information on the oxygen isotope ratio was particularly useful for geographical classification, as δ¹⁸O ratios are affected by higher latitude and altitude. Multivariate analysis showed that Ag, In, Mn, Li, Re, Ta and Th concentrations combined with δ¹⁸O ratio data allowed to identify the regional origin of a wine with an 87.3% success rate.

For a sample type not previously covered by these ASUs, Perini et al.¹⁸⁵ measured 42 elements by ICP-MS and stable isotope analysis (δ¹³C, δ¹⁵N, δ³⁴S, δ²H and δ¹⁸O) in 76 saffron samples from Morocco, Iran and Italy, produced between 2016 and 2019. Saffron has historically been a highly prized spice and, as such, production has expanded to many countries across the planet, away from the traditional growing areas of Turkey, Greece, Iran and Italy. The origin of saffron has an impact on its market value and, as such, it is prone to mislabelling and fraud. It was found that modelling with Cr, Cs, Eu, K, Mn, Mo, Nd, Ni, Pb, Rb, Sr, Zn concentrations together with δ¹³C, δ¹⁵N, δ³⁴S and δ²H, allowed for geographical discrimination, with 100% discrimination by country and an additional 83–84% discrimination of regional harvests within Italy. Individual components were identified as key markers for specific countries or growing seasons.

The use of coconut products has grown in popularity in recent years. Coconut sugar is one of these products, demanding a high price, often 100 times the price of cane sugar, making it attractive to fraudulent enterprises. Authentication is frequently carried out by determination of C isotopic ratios, which requires expensive equipment and experienced analysts. Zdiniakova et al.¹⁰³ used EDXRF to develop elemental profiles of coconut, cane and beetroot sugars. Bromine, Ca, Cl, Cu, Fe, K, P, Rb, S and Sr were used for the study, and it was shown that coconut sugar was often mineral enriched, whereas only Ca and Fe appeared at levels above the LOQ in cane sugars. The authors proposed that with handheld XRF devices and multivariate analysis it would be possible to establish methods for field assessment of sugar. Samples with levels below the LOQ for Br, Cl, Cu, K, P, Rb, S and Sr could be considered for further testing to confirm authenticity.

Milk authenticity from a variety of species was studied by Chen et al.,¹⁷⁰ as discussed in Section 8.2.3.

Abbreviations

AAS	atomic absorption spectrometry
AB	arsenobetaine
AC	arsenocholine
ac	alternating current
ACN	acetonitrile
AD	Alzheimer’s disease
AE	atomic emission
AEC	anion exchange chromatography
AEC-ICP-MS/MS	anion exchange chromatography-ICP-MS/MS
AES	atomic emission spectrometry
AF	atomic fluorescence
AFS	atomic fluorescence spectrometry
AMS	accelerator mass spectrometry
AOAC	Association of Official Analytical Chemists
APDC	ammonium pyrrolidinedithiocarbamate
ASU	Atomic Spectrometry Update
BARGE	Bioaccessibility Research Group of Europe
BMI	body mass index
CE	capillary electrophoresis
CEC	cation-exchange chromatography
CE-ICP-MS	capillary electrophoresis-ICP-MS
CRM	certified reference material
CS	continuum source
CS-AAS	continuum source-AAS
CSF	cerebrospinal fluid
CV	cold vapour
CV-AES	cold vapour-AES
CV-AFS	cold vapour-AFS
CVD	chemical vapour deposition
CVG	chemical vapour generation
CVG-AFS	chemical vapour generation-AFS
CV-ICP-MS	cold vapour-ICP-MS
DBD	dielectric barrier discharge
dc	direct current
DDTP	o,o′-diethyldithiophosphate
DES	deep eutectic solvent
DLLME	dispersive liquid–liquid microextraction
DMA	dimethylarsenic
DRC	dynamic reaction cell
DSPE	dispersive solid phase extraction
DTPA	diethylenetriaminepentaacetic acid
dw	dry weight
EAE	experimental autoimmune encephalomyelitis
EAR	estimated average requirement
EC50	half maximal effective concentration
EDXRD	energy dispersive X-ray diffraction
ELISA	enzyme-linked immunosorbent assay
EPA	Environmental Protection Agency
ERM	European Reference Material
ESI	electrospray ionization
ESI-MS	electrospray ionization-MS
ETAAS	electrothermal atomic absorption spectrometry
EtHg	ethylmercury
EtOH	ethanol
ETV	electrothermal vaporisation
ETV-GD-AES	electrothermal vaporisation-GD-AES
ETV-ICP-AES	electrothermal vaporisation-ICP-AES
EU	European Union
FAAS	flame atomic absorption spectrometry
FAO	UN Food and Agricultural Organization
FFF	field flow fractionation
FI	flow injection
FSIQ	Full Scale Intelligence Quotient
FTIR	Fourier transform infrared
GBCA	Gd-based contrast agent
GC	gas chromatography
GC-AFS	gas chromatography-AFS
GC-CV-ASF	gas chromatography-CV-ASF
GD	glow discharge
GDM	gestational diabetes mellitus
GF-AAS	graphite furnace atomic absorption spectrometry
HAS	human serum albumin
HCL	hollow cathode lamp
HDP	hypertensive disorders of pregnancy
HDPE	high density polyethylene
HG	hydride generation
HG-AAS	hydride generation-AAS
HG-AFS	hydride generation-AFS
HG-ICP-MS	hydride generation-ICP-MS
HPLC	high performance liquid chromatography
HPLC (AEC)-ICP-DRC-MS	high performance liquid chromatography (AEC)-ICP-DRC-MS
HPLC-ESI-MS	high performance liquid chromatography-ESI-MS
HPLC-ESI-MS/MS	high performance liquid chromatography-ESI-MS/MS
HPLC-HG-AFS	high performance liquid chromatography-HG-AFS
HPLC-ICP-MS	high performance liquid chromatography-ICP-MS
HPLC-ICP-QQQ-MS	high performance liquid chromatography-ICP-QQQMS
HR	high resolution
HR-CS-ETAAS	high resolution-continuum source-ETAAS
HR-CS-GF-AAS	high resolution-continuum source-GF-AAS
HR-ICP-MS	high resolution-ICP-MS
HR-TEM	high resolution-TEM
IAEA	International Atomic Energy Agency
iAs	inorganic arsenic
IC	ion chromatography
IC-ICP-MS	ion chromatography-ICP-MS
ICP	inductively coupled plasma
ICP-AES	inductively coupled plasma atomic emission spectrometry
ICP-DRC-MS	inductively coupled plasma–dynamic reaction cell-mass spectrometry
ICP-MS	inductively coupled plasma mass spectrometry
ICP-MS/MS	inductively coupled plasma tandem mass spectrometry
ICP-OES	inductively coupled plasma optical emission spectrometry
ICP-QMS	ICP-quadrupole MS
ICP-QQQMS	ICP-triple quadrupole MS
ID	isotope dilution
id	internal diameter
ID-SF-ICP-MS	isotope dilution- SF-ICP-MS
iHg	inorganic mercury
IRMS	isotope ratio mass spectrometry
iSe	inorganic Se
KED	kinetic energy distribution
KXRF	K-line XRF
LA	laser ablation
LA-ICP-MS	laser ablation ICP-MS
LC	liquid chromatography
LC-ESI-MS/MS	liquid chromatography- ESI-MS/MS
LC-ESI-Q/TOFMS	liquid chromatography-ESI-quadrupole/time of flight MS
LC-HG-AFS	liquid chromatography-HG-AFS
LC-ICP-MS	liquid chromatography-ICP-MS
LC-MS	liquid chromatography-MS
LC-MS/MS	liquid chromatography-MS/MS
LDA	linear discriminant analysis
LIBS	laser induced breakdown spectroscopy
LLE	liquid liquid extraction
LLME	liquid liquid microextraction
LOD	limit of detection
LOQ	limit of quantification
LPE	liquid phase extraction
LPME	liquid phase microextraction
μXANES	micro XANES
μXRF	micro XRF
MALDI	matrix-assisted laser desorption ionization
MC	multicollector
MC-ICP-MS	multicollector-ICP-MS
MDL	method detection limit
MEC	multienergy calibration
MeHg	methyl mercury
MeOH	methanol
MeSeCys	methylselenocysteine
MIP-AES	microwave induced plasma-AES
MIP-OES	microwave induced plasma-OES
MMA	monomethylarsenic
MoM	metal on metal
M r	molecular weight
MRI	magnetic resonance imaging
MS	mass spectrometry
MS/MS	mandem MS
NAA	neutron activation analysis
nanoSIMS	nano secondary ion mass spectrometry
nanoSR-XRF	nano synchrotron radiation XRF
NIST	National Institute of Standards and Technology
NP	nanoparticle
NRCC	National Research Council of Canada
NTD	neural tube defect
PAR	4-(2-pyridylazo)resorcinol
PCA	principal component analysis
PDMS	polydimethilsiloxane
PhHg	phenyl mercury
PIXE	particle-induced X-ray emission
PLS	partial least squares
PLS-DA	partial least squares discriminant analysis
PLSR	partial linear squares regression
PMD	post-mortem delay
PTFE	poly(tetrafluoroethylene)
QC	quality control
QMS	quadrupole MS
QQQ	triple quadrupole
RBC	red blood cells
RDI	recommended dietary intake
REE	rare earth element
RM	reference material
RNI	recommended intake (RNI)
RP	reverse phase
RSD	relative standard deviation
RSDr	repeatability relative standard deviation
RSDR	reproducibility relative standard deviation
RT-PCR	real time-polymerase chain reaction
SD	standard deviation
SDA	standard dilution analysis
SDS	sodium dodecylsulfate
SEC	size exclusion chromatography
SEC-ICP-MS	size exclusion chromatography-ICP-MS
SEC-MC-ICP-MS	size exclusion chromatography-MC-ICP-MS
SeCN	selenocyanate
SEC-UV-Vis	size exclusion chromatography-UV-Vis
SeCys	selenocysteine
SeCys2	selenocystine
SEM	scanning electron microscopy
SeMet	selenomethionine
SeSug1	methyl-2-acetamido-2-deoxy-1-seleno-b-D-galactopyranoside
SeSug2	methyl-2-acetamido-2-deoxy-1-seleno-b-D-glucopyranoside
SeSug3	methyl-2-amino-2-deoxy-1-seleno-b-D-galactopyranoside
SF	sector field
SF-ICP-MS	sector field-ICP-MS
SIMCA	soft independent modelling of class analogy
SIMS	secondary ion mass spectrometry
sp	single particle
SPE	solid phase extraction
spICP-MS	single particle ICP-MS
SPME	solid phase microextraction
SR-μXRF	synchrotron radiation-μXRF
SRM	standard reference material
SR-XAS	synchrotron radiation XAS
SR-XRF	synchrotron radiation XRF
TEM	transmission electron microscopy
THF	tetrahydrofuran
TMA	tetramethylarsenic
TMAH	tetramethylammonium hydroxide
TMAO	trimethylarsine oxide
TMSe	trimethylselenium ion
TOF	time-of-flight
TOF-SIMS	time-of-flight-SIMS
Tris	tris(hydroxymethyl)aminomethane
TXRF	total reflexion XRF
US	United States
UV	ultraviolet
UV-Vis	ultraviolet-visible
WDXRF	wavelength dispersive XRF
WHO	World Health Organisation
XANES	X-ray absorption spectroscopy near-edge structure
XAS	X-ray absorption spectroscopy
XRD	X-ray diffraction
XRF	X-ray fluorescence

Conflicts of interest

There are no conflicts to declare.

References

1. A. Taylor, A. Catchpole, M. P. Day, S. Hill, N. Martin and M. Patriarca, J. Anal. At. Spectrom., 2020, 35(3), 426–454.
2. E. H. Evans, J. Pisonero, C. M. Smith and R. N. Taylor, J. Anal. At. Spectrom., 2020, 35(5), 830–851.
3. C. Vanhoof, J. R. Bacon, A. T. Ellis, U. E. A. Fittschen and L. Vincze, J. Anal. At. Spectrom., 2019, 34(9), 1750–1767.
4. R. Clough, C. F. Harrington, S. J. Hill, Y. Madrid and J. F. Tyson, J. Anal. At. Spectrom., 2020, 35(7), 1236–1278.
5. S. Carter, R. Clough, A. Fisher, B. Gibson, B. Russell and J. Waack, J. Anal. At. Spectrom., 2019, 34(11), 2159–2216.
6. J. R. Bacon, O. T. Butler, W. R. L. Cairns, J. M. Cook, C. M. Davidson, O. Cavoura and R. Mertz-Kraus, J. Anal. At. Spectrom., 2020, 35(1), 9–53.
7. F. Gyngard and M. L. Steinhauser, J. Anal. At. Spectrom., 2019, 34(8), 1534–1545.
8. P. Massonnet and R. M. A. Heeren, J. Anal. At. Spectrom., 2019, 34(11), 2217–2228.
9. S. K. H. Shah, J. Iqbal, P. Ahmad, M. U. Khandaker, S. Haq and M. Naeem, Radiat. Phys. Chem., 2020, 170,  DOI:10.1016/j.radphyschem.2019.108666.
10. Q. Q. Wang, W. T. Xiangli, G. Teng, X. T. Cui and K. Wei, Appl. Spectrosc. Rev., 2020 DOI:10.1080/05704928.2020.1791151.
11. S. Matsuyama, K. Maeshima and M. Shimura, J. Anal. At. Spectrom., 2020, 35(7), 1279–1294.
12. D. G. Filatova, V. V. Es’kina, V. B. Baranovskaya and Y. A. Karpov, J. Anal. Chem., 2020, 75(5), 563–568.
13. M. Martinez and M. Baudelet, Anal. Bioanal. Chem., 2020, 412, 27–36.
14. X. Feng, H. H. Zhang and P. Q. Yu, Crit. Rev. Food Sci. Nutr., 2020 DOI:10.1080/10408398.2020.1776677.
15. B. Hashemi and S. Rezania, Microchim. Acta, 2019, 186(8), 578,  DOI:10.1007/s00604-019-3668-2.
16. E. O. Er, G. D. Bozyigit, C. Buyukpinar and S. Bakirdere, Crit. Rev. Anal. Chem., 2020 DOI:10.1080/10408347.2020.1797465.
17. M. Locateni, A. Tartaglia, S. Piccolantonio, L. A. Di Iorio, E. Sperandio, H. I. Ulusoy, K. G. Furton and A. Kabir, Curr. Anal. Chem., 2019, 15(7), 731–744.
18. C. Luvonga, C. A. Rimmer, L. L. Yu and S. B. Lee, J. Agric. Food Chem., 2020, 68(7), 1910–1934.
19. J. R. Camurati and V. N. Salomone, J. Toxicol. Environ. Health, Part B, 2019, 23(1), 1–12,  DOI:10.1080/10937404.2019.1672364.
20. J. Zhang, D. Baralkiewicz, Y. Z. Wang, J. Falandysz and C. T. Cai, Chemosphere, 2020, 246,  DOI:10.1016/j.chemosphere.2019.125685.
21. R. Milačič and J. Ščančar, TrAC, Trends Anal. Chem., 2020, 127, 115888,  DOI:10.1016/j.trac.2020.115888.
22. K. Katerinopoulou, A. Kontogeorgos, C. E. Salmas, A. Patakas and A. Ladavos, Foods, 2020, 9(4), 489,  DOI:10.3390/foods9040489.
23. X. Wei, Y. Lu, X. Zhang, M. L. Chen and J. H. Wang, TrAC, Trends Anal. Chem., 2020, 127, 115886,  DOI:10.1016/j.trac.2020.115886.
24. D. B. Alcantara, R. F. Nascimento, G. S. Lopes and P. Grinberg, Anal. Methods, 2020, 12(26), 3351–3360.
25. I. Swiatkowska, N. Martin and A. J. Hart, J. Trace Elem. Med. Biol., 2019, 53, 120–128.
26. H. Cruijsen, E. Poitevin and S. L. Brunelle, J. AOAC Int., 2019, 102(6), 1845–1863.
27. X. Y. Fan, Q. Li, Z. G. Jin, X. X. Yu, M. L. Ding and Y. Ju, Clin. Chim. Acta, 2020, 508, 249–253.
28. M. Garcia, E. Vassileva, S. Azemard and A. Canals, Food Anal. Methods, 2020, 13, 390–402,  DOI:10.1007/s12161-019-01661-x.
29. O. Mihai, M. S. Kawamoto, K. L. LeBlanc, P. Grinberg, A. R. D. Nogueira and Z. Mester, J. Food Compos. Anal., 2020, 87,  DOI:10.1016/j.jfca.2020.103421.
30. J. Baudry, J. F. Kopp, H. Boeing, A. P. Kipp, T. Schwerdtle and M. B. Schulze, Eur. J. Nutr., 2020, 59, 3045–3058,  DOI:10.1007/s00394-019-02143-w.
31. J. B. Judith, B. Annabelle, A. Rafael, H. Max, Z. Ursina, B. Michel, P. Fred, B. Murielle and D. Vincent, Swiss Med. Wkly., 2020, 150,  DOI:10.4414/smw.2020.20207.
32. B. C. Son, C. K. Lee, C. H. Suh, K. H. Kim, J. H. Kim, S. U. Jeong, D. H. Kim, J. Y. Ryu, S. W. Lee, S. J. Kim, Y. M. Kwon and Y. B. Park, J. Occup. Environ. Hyg., 2019, 16(12), 763–774,  DOI:10.1080/15459624.2019.1668000.
33. M. T. L. Ballesteros, B. G. Barrado, I. N. Serrano, S. I. Alvarez, M. D. G. Anaya and M. J. G. Munoz, J. Trace Elem. Med. Biol., 2020, 57, 60–67.
34. R. L. Sato, G. G. Li and S. Shaha, Am. J. Obstet. Gynecol., 2006, 194(6), 1683–1688.
35. K. A. Layne, D. M. Wood, M. Dixon-Zegeye, J. R. H. Archer, K. Raja and P. I. Dargan, Invest. Radiol., 2020, 55(7), 405–411.
36. R. Soler-Blasco, M. Murcia, M. Lozano, X. Aguinagalde, G. Iriarte, M. J. Lopez-Espinosa, J. Vioque, C. Iniguez, F. Ballester and S. Llop, Environ. Int., 2019, 130, 104835,  DOI:10.1016/j.envint.2019.05.029.
37. B. B. Gump, B. Hruska, P. J. Parsons, C. D. Palmer, J. A. MacKenzie, K. Bendinskas and L. Brann, Environ. Res., 2020, 185,  DOI:10.1016/j.envres.2020.109308.
38. K. Sera, S. Goto, T. Hosokawa and Y. Saitoh, Nucl. Instrum. Methods Phys. Res., Sect. B, 2020, 477, 144–149.
39. A. F. Robson, P. Lockett, L. Tetlow and C. Chaloner, Ann. Clin. Biochem., 2020, 57(3), 246–248,  DOI:10.1177/0004563220915949.
40. D. Verrey, S. Durand, O. Thomas, V. Lelevrier, P. Quenel and B. Le Bot, J. Exposure Sci. Environ. Epidemiol., 2019, 29(5), 706–717.
41. A. Taylor, Ann. Clin. Biochem., 1986, 23, 364–378.
42. M. Safta, I. Limam, R. Jebali, M. Marzouki, M. R. Driss and R. Kalfat, Int. J. Environ. Anal. Chem., 2020 DOI:10.1080/03067319.2020.1786550.
43. J. Hildebrand and T. Goen, J. Trace Elem. Med. Biol., 2020, 61, 126538,  DOI:10.1016/j.jtemb.2020.126538.
44. I. J. S. da Silva, A. F. Lavorante, A. P. S. Paim and M. J. da Silva, Food Chem., 2020, 319,  DOI:10.1016/j.foodchem.2020.126435.
45. M. I. A. Fioravanti, R. F. Milani, E. L. de Paiva and M. A. Morgano, Anal. Methods, 2020, 12(25), 3225–3234.
46. I. Komorowicz, A. Hanc, W. Lorenc, D. Baralkiewicz, J. Falandysz and Y. Z. Wang, Chemosphere, 2019, 233, 223–233.
47. L. Cheng, X. A. Yang, M. T. Shi and W. B. Zhang, J. Chromatogr. A, 2020, 1619,  DOI:10.1016/j.chroma.2020.460915.
48. A. Kaňa, M. Sadowska, J. Kvicala and O. Mestek, J. Food Compos. Anal., 2020, 92,  DOI:10.1016/j.jfca.2020.103562.
49. D. Martinez-Rubio, G. Grindlay, M. Llaver, R. G. Wuilloud and J. Mora, J. Anal. At. Spectrom., 2020, 35(5), 933–942.
50. M. T. Liu, X. F. Mao, J. X. Liu, L. Ding, X. Na, G. Y. Chen and Y. Z. Qian, At. Spectrosc., 2019, 40(3), 83–90.
51. Y. Zhou, Z. Q. Chen, J. X. Zeng, J. X. Zhang, D. X. Yu, B. Zhang, X. W. Yan, L. M. Yang and Q. Q. Wang, Anal. Chem., 2020, 92(7), 5286–5293.
52. Y. K. Tanaka, R. Iida, S. Takada, T. Kubota, M. Yamanaka, N. Sugiyama, Y. Abdelnour and Y. Ogra, ChemBioChem, 2020, 21(22), 3266–3272,  DOI:10.1002/cbic.202000358.
53. F. Aureli, M. Ciprotti, M. D’Amato, E. D. da Silva, S. Nisi, D. Passeri, A. Sorbo, A. Raggi, M. Rossi and F. Cubadda, Nanomaterials, 2020, 10(5), 888,  DOI:10.3390/nano10050888.
54. A. F. R. Garcia, M. Corte-Rodriguez, M. Macke, K. L. LeBlanc, Z. Mester, M. Montes-Bayon and J. B. Bettmer, Analyst, 2020, 145(4), 1457–1465.
55. B. Callejon-Leblic, G. Rodriguez-Moro, A. Arias-Borrego, A. Pereira-Vega, J. L. Gomez-Ariza and T. Garcia-Barrera, J. Chromatogr. A, 2020, 1619,  DOI:10.1016/j.chroma.2020.460919.
56. L. Fu, H. L. Xie, J. H. Huang and L. Chen, Anal. Chim. Acta, 2020, 1112, 1–7.
57. N. Solovyev, A. Ala, M. Schilsky, C. Mills, K. Willis and C. F. Harrington, Anal. Chim. Acta, 2020, 1098, 27–36.
58. J. Y. Hu, Z. Y. Li, H. Zhang, R. Liu and Y. Lv, Anal. Chem., 2020, 92(12), 8523–8529.
59. S. Wilschefski, M. Baxter and G. Pool, Ann. Clin. Biochem., 2020, 57(3), 234–241.
60. L. N. Zheng, Y. B. Sang, R. P. Luo, B. Wang, F. T. Yi, M. Wang and W. Y. Feng, J. Anal. At. Spectrom., 2019, 34(5), 915–921.
61. M. Aliste, S. Arranz, A. Sánchez-Ortega, M. C. Sampedro, N. Unceta, A. Gómez-Caballero, A. Vallejo, M. A. Goicolea and R. J. Barrio, J. Forensic Sci., 2020, 65(4), 1094–1101.
62. M. Aranaz, M. Costas-Rodriguez, L. Lobo, H. Gonzalez-Iglesias, F. Vanhaecke and R. Pereiro, J. Pharm. Biomed. Anal., 2020, 177,  DOI:10.1016/j.jpba.2019.112857.
63. E. Paredes, V. Malard, C. Vidaud, E. Avazeri, R. Ortega, A. Nonell, H. Isnard, F. Chartier and C. Bresson, Analyst, 2019, 144(20), 5928–5933.
64. N. Atsever, T. Borahan, E. G. Bakirdere and S. Bakirdere, J. Pharm. Biomed. Anal., 2020, 186,  DOI:10.1016/j.jpba.2020.113274.
65. N. E. Karlidag, E. S. Kocoglu, M. Toprak, O. Yilmaz and S. Bakirdere, Food Chem., 2020, 329,  DOI:10.1016/j.foodchem.2020.127210.
66. M. Saylan, B. T. Zaman, E. G. Bakirdere and S. Bakirdere, J. Food Compos. Anal., 2020, 88,  DOI:10.1016/j.jfca.2020.103454.
67. F. Uzcan and M. Soylak, Int. J. Environ. Anal. Chem., 2020 DOI:10.1080/03067319.2020.1776865.
68. B. Tişli, T. U. Gösterişli, B. T. Zaman, E. G. Bakirdere and S. Bakirdere, Anal. Lett., 2020 DOI:10.1080/00032719.2020.1789871.
69. L. N. Pires, J. S. Almeida, F. D. Dias and L. S. G. Teixeira, Food Anal. Methods, 2020, 13, 1746–1754,  DOI:10.1007/s12161-020-01787-3.
70. O. Zverina, J. Kuta, P. Coufalik, P. Koseckova and J. Komarek, Food Chem., 2019, 298,  DOI:10.1016/j.foodchem.2019.125084.
71. G. C. Leal, P. Mattiazzi, F. Rovasi, T. Dal Molin, D. Bohrer, P. C. do Nascimento, L. M. de Carvalho and C. Viana, J. Food Compos. Anal., 2020, 86,  DOI:10.1016/j.jfca.2019.103360.
72. Q. S. Deng, C. Yang, H. T. Zheng, J. X. Liu, X. F. Mao, S. H. Hu and Z. L. Zhu, J. Anal. At. Spectrom., 2019, 34(9), 1786–1793.
73. P. Maung and D. Beauchemin, J. Anal. At. Spectrom., 2020, 35(6), 1097–1102.
74. R. Serrano, G. Grindlay, P. Niedzielski, L. Gras and J. Mora, J. Anal. At. Spectrom., 2020, 35(7), 1351–1359.
75. L. S. B. Carvalho, C. S. Silva, J. A. Nobrega, E. S. B. Morte, D. Santos and M. G. A. Korn, J. Food Compos. Anal., 2020, 86,  DOI:10.1016/j.jfca.2019.103376.
76. T. L. E. Cruz, M. G. Esperanza, K. Wrobel, E. Y. Barrientos, F. J. A. Aguilar and K. Wrobel, Spectrochim. Acta, Part B, 2020, 164,  DOI:10.1016/j.sab.2019.105754.
77. P. Modlitbová, Z. Farka, M. Pastucha, P. Pořízka, K. Novotný, P. Skládal and J. Kaiser, Microchim. Acta, 2019, 186(9), 629,  DOI:10.1007/s00604-019-3742-9.
78. Y. W. Chu, Z. J. Zhang, Q. Y. He, F. Chen, Z. Q. Sheng, D. Zhang, H. L. Jin, F. G. Jiang and L. B. Guo, J. Adv. Res., 2020, 24, 353–361.
79. M. A. Gondal, R. K. Aldakheel, M. A. Almessiere, M. M. Nasr, J. A. Almusairii and B. Gondal, J. Pharm. Biomed. Anal., 2020, 183,  DOI:10.1016/j.jpba.2020.113153.
80. A. Taylor, N. Barlow, M. P. Day, S. Hill, N. Martin and M. Patriarca, J. Anal. At. Spectrom., 2019, 34, 426–459.
81. M. G. Nespeca, A. L. Vieira, D. Santos, J. A. G. Neto and E. C. Ferreira, Food Chem., 2020, 311,  DOI:10.1016/j.foodchem.2019.125886.
82. J. Peng, W. Xie, J. Jiang, Z. Zhao, F. Zhou and F. Liu, Foods, 2020, 9(3), 341,  DOI:10.3390/foods9030341.
83. N. Gyftokostas, D. Stefas and S. Couris, Appl. Sci., 2020, 10(10), 3462,  DOI:10.3390/app10103462.
84. Y. Z. Zhangcheng, Y. Z. Liu, S. Saleem, Q. H. Zhang, Y. Chen, Y. F. Qu and X. Lu, J. Laser Appl., 2020, 32(3) DOI:10.2351/7.0000137.
85. M. Markiewicz-Keszycka, M. Zhao, X. Cama-Moncunill, T. El Arnaout, D. Becker, C. O’Donnell, P. J. Cullen, C. Sullivan and M. P. Casado-Gavalda, Int. Dairy J., 2019, 97, 57–64.
86. N. Sharma, Kamni, V. K. Singh, S. Kumar, Y. Lee, P. K. Rai and V. K. Singh, Appl. Phys. B: Lasers Opt., 2020, 126(7) DOI:10.1007/s00340-020-07475-8.
87. N. Sharma, Y. Khajuria, V. K. Singh, S. Kumar, Y. Lee, P. K. Rai and V. K. Singh, At. Spectrosc., 2020, 41(3), 110–118.
88. B. M. Freire, V. D. Santos, P. D. F. Neves, J. Reis, S. S. de Souza, F. Barbosa and B. L. Batista, Anal. Methods, 2020, 12(16), 2102–2113.
89. A. I. G. de las Torres, I. Giraldez, F. Martinez, P. Palencia, W. T. Corns and D. Sanchez-Rodas, Food Chem., 2020, 315,  DOI:10.1016/j.foodchem.2020.126215.
90. X. P. Yu, W. J. Cui, Q. Wang, Y. F. Guo and T. L. Deng, Anal. Bioanal. Chem., 2019, 411, 7251–7260,  DOI:10.1007/s00216-019-02093-6.
91. Z. Y. Gu, S. de Silva and S. M. Reichman, Int. J. Environ. Res. Public Health, 2020, 17(2), 415,  DOI:10.3390/ijerph17020415.
92. X. Deng, R. S. Li and S. P. Deng, J. Fluoresc., 2020, 30(4), 949–954.
93. Y. Y. Li, W. J. Hu, J. T. Zhao, Q. M. Chen, W. Wang, B. Li and Y. F. Li, Ecotoxicol. Environ. Saf., 2019, 182,  DOI:10.1016/j.ecoenv.2019.109447.
94. A. Santa-Rios, B. D. Barst and N. Basu, Anal. Chem., 2020, 92(5), 3605–3612.
95. M. A. P. da Costa, D. L. F. da Silva, F. D. Dias, A. D. Santos and W. N. L. dos Santos, Anal. Methods, 2020, 12(13), 1711–1719.
96. S. Liu, X. C. Duan and J. S. Sun, Anal. Sci., 2020, 36(4), 419–423.
97. X. Yuan, K. J. Li, Y. T. Zhang, Y. T. Miao, Y. Xiang, Y. Sha, M. Zhang and K. Huang, Microchem. J., 2020, 155, 104695,  DOI:10.1016/j.microc.2020.104695.
98. A. A. Shaltout, J. Bouslimi and H. Besbes, Food Chem., 2020, 328,  DOI:10.1016/j.foodchem.2020.127124.
99. D. R. Chettle and F. E. McNeill, Physiol. Meas., 2019, 40(12), TR01,  DOI:10.1088/1361-6579/ab6019.
100. A. Gruber, R. Müller, A. Wagner, S. Colucci, M. Vujiç Spasiç and K. Leopold, Anal. Bioanal. Chem., 2020, 412, 6419–6429,  DOI:10.1007/s00216-020-02614-8.
101. D. S. Almeida, M. M. Brigido, M. J. Anjos, S. T. Ferreira, A. S. Souza and R. T. Lopes, X-Ray Spectrom., 2019, 48(5), 452–464.
102. J. Gao, C. Sheng, Y. Q. Zhu, M. Dong, R. Qian and S. J. Zhuo, Spectrosc. Spectral Anal., 2020, 40(3), 945–949.
103. T. Zdiniakova and M. B. de la Calle, Eur. Food Res. Technol., 2020, 246, 2101–2109,  DOI:10.1007/s00217-020-03559-z.
104. E. K. Sichangi, H. K. Angeyo and A. Dehayem-Kamadjeu, Spectrochim. Acta, Part B, 2019, 158, 105635,  DOI:10.1016/j.sab.2019.105635.
105. J. Nguyen, D. Crawford, D. Howarth, B. Sukhu, A. Pejović-Milić and J. L. Gräfe, Physiol. Meas., 2019, 40(8) DOI:10.1088/1361-6579/ab3c0e.
106. Y. Izumoto, K. Fukutsu, K. Takamura, Y. Sakai, Y. Oguri and H. Yoshii, J. Radiol. Prot., 2020, 40(3), 692–703.
107. S. Q. Jiang, X. Y. Wu, J. L. Sun, G. Chen, R. Tang, Z. Li, R. Y. Wei, L. Liang, X. J. Zhou, D. L. Chen, J. Li, H. Gao, J. Zhang and Z. T. Zhao, Chin. Med. J., 2019, 132(16), 1959–1964.
108. K. O. Lorentz, W. de Nolf, M. Cotte, G. Ioannou, F. Foruzanfar, M. R. Zaruri and S. M. S. Sajjadi, J. Archaeol. Sci., 2020, 120,  DOI:10.1016/j.jas.2020.105193.
109. R. Debastiani, C. E. I. dos Santos, M. M. Ramos, V. S. Souza, L. Amaral and J. F. Dias, Nucl. Instrum. Methods Phys. Res., Sect. B, 2020, 477, 154–158.
110. T. Kuhn, P. Chytry, G. M. S. Souza, D. V. Bauer, L. Amaral and J. F. Dias, Nucl. Instrum. Methods Phys. Res., Sect. B, 2020, 477, 150–153.
111. R. K. Virk, M. Shafique, K. Vaid, S. Verma, M. P. Bansal, M. L. Garg and B. P. Mohanty, Nucl. Instrum. Methods Phys. Res., Sect. B, 2020, 477, 91–97.
112. N. E. Karlidag, M. Toprak, Z. Tekin and S. Bakirdere, J. Food Compos. Anal., 2020, 92,  DOI:10.1016/j.jfca.2020.103583.
113. J. Kruszewska, D. Kulpinska, I. G. Grabowska-Jadach and M. Matczuk, Metallomics, 2020, 12(3), 408–415.
114. H. M. Fahmy, O. A. Ali, A. A. Hassan and M. A. Mohamed, J. Trace Elem. Med. Biol., 2020, 61, 126505,  DOI:10.1016/j.jtemb.2020.126505.
115. Z. M. Chang, H. Zhou, C. Yang, R. Zhang, Q. N. You, R. H. Yan, L. Li, M. F. Ge, Y. G. Tang, W. F. Dong and Z. Wang, J. Mater. Chem. B, 2020, 8(23), 5019–5025.
116. M. T. Matter, J. H. Li, I. Lese, C. Schreiner, L. Bernard, O. Scholder, J. Hubeli, K. Keevend, E. Tsolaki, E. Bertero, S. Bertazzo, R. Zboray, R. Olariu, M. A. Constantinescu, R. Figi and I. K. Herrmann, Adv. Sci., 2020, 7(15), 2000912,  DOI:10.1002/advs.202000912.
117. N. D. Donahue, E. R. Francek, E. Kiyotake, E. E. Thomas, W. Yang, L. Wang, M. S. Detamore and S. Wilhelm, Anal. Bioanal. Chem., 2020, 412, 5205–5216,  DOI:10.1007/s00216-020-02783-6.
118. J. Santelli, S. Lechevallier, D. Calise, D. Marsal, A. Siegfried, M. Vincent, C. Martinez, D. Cussac, R. Mauricot and M. Verelst, Acta Biomater., 2020, 108, 261–272.
119. S. Sungur, P. Kaya and M. Koroglu, Food Addit. Contam., Part B, 2020, 13(4), 260–267,  DOI:10.1080/19393210.2020.1769193.
120. B. Katarzyna, R. M. Taylor, J. Szpunar, R. Lobinski and R. A. Sunde, Metallomics, 2020, 12(5), 758–766.
121. K. Takahashi and Y. Ogra, Metallomics, 2020, 12(2), 241–248.
122. M. A. Harvey, P. D. Erskine, H. H. Harris, G. K. Brown, E. A. H. Pilon-Smits, L. W. Casey, G. Echevarria and A. van der Ent, Metallomics, 2020, 12(4), 514–527.
123. C. J. Greenhalgh, E. Karekla, G. J. Miles, I. R. Powley, C. Costa, J. de Jesus, M. J. Bailey, C. Pritchard, M. MacFarlane, J. H. Pringle and A. J. Managh, Anal. Chem., 2020, 92(14), 9847–9855.
124. T. L. Wu, X. D. Cui, P. X. Cui, S. T. Ata-Ul-Karim, Q. Sun, C. Liu, T. T. Fan, H. Gong, D. M. Zhou and Y. J. Wang, Environ. Pollut., 2019, 252, 1439–1447.
125. A. Jagielska, A. Ruszczynska, E. Bulska and B. Wagner, J. Anal. At. Spectrom., 2020, 35(7), 1340–1350.
126. M. D. Pour, L. Ren, E. Jennische, S. Lange, A. G. Ewing and P. Malmberg, J. Anal. At. Spectrom., 2019, 34(8), 1581–1587.
127. M. K. Andersen, S. Krossa, T. S. Hoiem, R. Buchholz, B. S. R. Claes, B. Balluff, S. R. Ellis, E. Richardsen, H. Bertilsson, R. M. A. Heeren, T. F. Bathen, U. Karst, G. F. Giskeodegard and M. B. Tessem, Anal. Chem., 2020, 92(4), 3171–3179.
128. E. Ender, M. A. Subirana, A. Raab, E. M. Krupp, D. Schaumloffel and J. Feldmann, J. Anal. At. Spectrom., 2019, 34(11), 2295–2302.
129. C. W. He, W. X. Song, T. A. Weston, C. Tran, I. Kurtz, J. E. Zuckerman, P. Guagliardo, J. H. Miner, S. V. Ivanov, J. Bougoure, B. G. Hudson, S. Colon, P. A. Voziyan, G. Bhave, L. G. Fong, S. G. Young and H. B. Jiang, Proc. Natl. Acad. Sci. U. S. A., 2020, 117(27), 15827–15836.
130. L. Lemelle, A. Simionovici, P. Colin, G. Knott, S. Bohic, P. Cloetens and B. L. Schneider, Commun. Biol., 2020, 3(1), 364,  DOI:10.1038/s42003-020-1084-0.
131. M. R. Gherase and D. E. B. Fleming, Crystals, 2020, 10(1), 12,  DOI:10.3390/cryst10010012.
132. S. J. Yin, C. R. Wang, J. Wei, D. Wang, L. Jin, J. F. Liu, L. L. Wang, Z. W. Li, A. G. Ren and C. H. Yin, Environ. Int., 2020, 139,  DOI:10.1016/j.envint.2020.105688.
133. A. Ovayolu, G. Ovayolu, E. Karaman, T. Yuce, M. A. Ozek and V. A. Turksoy, Congenital Anomalies, 2020, 60(5), 136–141,  DOI:10.1111/cga.12363.
134. W. P. Zhang, W. Y. Liu, S. S. Bao, H. X. Liu, Y. Z. Zhang, B. Zhang, A. F. Zhou, J. Chen, K. Hao, W. Xia, Y. Y. Li, X. Sheng and S. Q. Xu, Environ. Int., 2020, 135, 105391,  DOI:10.1016/j.envint.2019.105391.
135. C. G. Howe, S. F. Farzan, E. Garcia, T. Jursa, R. Iyer, K. Berhane, T. A. Chavez, T. L. Hodes, B. H. Grubbs, W. E. Funk, D. R. Smith, T. M. Bastain and C. V. Breton, Environ. Res., 2020, 184,  DOI:10.1016/j.envres.2020.109294.
136. N. S. de Morais, D. A. Saraiva, C. Corcino, T. Berbara, A. Schtscherbyna, K. Moreira, M. Vaisman, E. K. Alexander and P. Teixeira, Thyroid, 2020, 30(12), 1792–1801,  DOI:10.1089/thy.2019.0462.
137. X. Y. Wang, D. Gao, G. F. Zhang, X. Zhang, Q. Li, Q. Gao, R. J. Chen, S. Z. Xu, L. Huang, Y. Zhang, L. X. Lin, C. R. Zhong, X. Chen, G. Q. Sun, Y. Song, X. F. Yang, L. P. Hao, H. Y. Yang, L. Yang and N. H. Yang, Environ. Int., 2020, 135, 105370,  DOI:10.1016/j.envint.2019.105370.
138. Q. Q. Zhang, X. N. Li, X. Liu, M. R. Dong, J. P. Xiao, J. Wang, M. Y. Zhou, Y. D. Wang, D. Ning, W. J. Ma, W. Zhu, T. Liu and B. Zhang, Chemosphere, 2020, 246,  DOI:10.1016/j.chemosphere.2019.125732.
139. W. C. Jiang, G. W. Sun, X. Y. Wen, S. S. Men, W. B. Cui, M. Jing, X. W. Jia, Z. A. Hu, D. N. Pu, S. C. Zhang, X. Z. Yuan, X. T. Liu, X. R. Zhang and C. B. Wang, Clin. Chem. Lab. Med., 2020, 58(6), 873–882.
140. I. Swiatkowska, N. G. Martin, J. Henckel, H. Apthorp, J. Hamshere and A. J. Hart, J. Trace Elem. Med. Biol., 2020, 57, 9–17.
141. X. L. He, F. X. Reichl, S. Milz, B. Michalke, X. Wu, C. M. Sprecher, Y. Yang, M. Gahlert, S. Rohling, H. Kniha, R. Hickel and C. Hogg, Dent. Mater., 2020, 36(3), 402–412.
142. C. Freire, P. Vrhovnik, Z. Fiket, I. Salcedo-Bellido, R. Echeverria, P. Martin-Olmedo, G. Kniewald, M. F. Fernandez and J. P. Arrebola, Sci. Total Environ., 2020, 719,  DOI:10.1016/j.scitotenv.2020.137458.
143. M. Scholefield, S. J. Church, J. S. Xu, S. Kassab, N. J. Gardiner, F. Roncaroli, N. M. Hooper, R. D. Unwin and G. J. S. Cooper, Metallomics, 2020, 12(6), 952–962.
144. K. Midander, L. Schenk and A. Julander, Regul. Toxicol. Pharmacol., 2020, 115, 104693,  DOI:10.1016/j.yrtph.2020.104693.
145. A. Stojsavljevic, B. Rovcanin, J. Jagodic, D. D. Radojkovic, I. Paunovic, M. Gavrovic-Jankulovic and D. Manojlovic, Environ. Res., 2020, 186,  DOI:10.1016/j.envres.2020.109538.
146. X. M. You, Y. Xiao, K. Liu, Y. Q. Yu, Y. Y. Liu, P. P. Long, H. Wang, L. E. Zhou, Q. F. Deng, Y. H. Lin, X. M. Zhang, M. A. He, T. C. Wu and Y. Yuan, Environ. Chem., 2020, 17(4), 304–313.
147. A. Marvelli, B. Campi, G. Mergni, M. E. Di Cicco, P. Turini, P. Scardina, R. Zucchi, M. Pifferi, G. Taccetti, A. Paolicchi, G. la Marca and A. Saba, Anal. Bioanal. Chem., 2020, 412, 6909–6916,  DOI:10.1007/s00216-020-02821-3.
148. S. Q. Wang, B. Hesse, M. Roman, D. Stier, H. Castillo-Michel, M. Cotte, J. P. Suuronen, A. Lagrange, H. Radbruch, F. Paul, M. Taupitz, E. Schellenberger, I. Sack and C. Infante-Duarte, Invest. Radiol., 2019, 54(10), 617–626.
149. A. L. Stanescu, D. W. Shaw, N. Murata, K. Murata, J. C. Rutledge, E. Maloney and K. R. Maravilla, Pediatr. Radiol., 2020, 50(3), 388–396.
150. P. Day, S. Erdahl, S. Eckdahl, J. Bornhorst and P. J. Jannetto, Ann. Clin. Biochem., 2019, 56(6), 638–645.
151. H. L. Li, Z. F. Chen, J. X. Li, R. Liu, F. Zhao and R. Liu, J. Appl. Toxicol., 2020, 40(12), 1636–1646,  DOI:10.1002/jat.4023.
152. Y. T. Zou, D. C. Wang, S. L. Yu, X. Q. Cheng, L. Y. Xia, Y. C. Yin, S. W. Xie, Q. Cheng, L. Qiu and X. L. Lian, Clin. Biochem., 2020, 82, 99–104.
153. E. J. McAllum, D. J. Hare, I. Volitakis, C. A. McLean, A. I. Bush, D. I. Finkelstein and B. R. Roberts, Prog. Neurobiol., 2020, 186,  DOI:10.1016/j.pneurobio.2019.101744.
154. B. Michalke, D. Willkommen and V. Venkataramani, J. Visualized Exp., 2020, 159, e61055,  DOI:10.3791/61055.
155. J. X. Lu, J. Cai, T. X. Ren, J. H. Wu, D. Q. Mao, W. D. Li, Y. Zhang, J. H. Piao, J. Wang, L. C. Yang, X. G. Yang and Y. X. Ma, Nutr. Metab., 2019, 16(1), 55,  DOI:10.1186/s12986-019-0384-1.
156. S. R. Scott, M. M. Shafer, K. E. Smith, J. T. Overdier, B. Cunliffe, T. W. Stafford and P. M. Farrell, Archaeometry, 2020, 62(1), 109–129,  DOI:10.1111/arcm.12513.
157. J. Dantzer, P. Ryan, K. Yolton, P. J. Parsons, C. D. Palmer, K. Cecil and J. M. Unrine, Sci. Total Environ., 2020, 700,  DOI:10.1016/j.scitotenv.2019.134519.
158. B. Lelievre, S. Triau, P. Codron, Y. Mariau, F. Papin-Lefebvre, A. Collin, G. Drevin, G. Le Roux, M. Briet, D. Boels and F. Letournel, Forensic Toxicol., 2020, 38(2), 505–510.
159. B. Bortot, M. Mongiat, E. Valencic, S. Dal Monego, D. Licastro, M. Crosera, G. Adami, E. Rampazzo, G. Ricci, F. Romano, G. M. Severini and S. Biffi, Int. J. Nanomed., 2020, 15, 4793–4810.
160. A. Moghaddam, R. A. Heller, Q. Sun, J. Seelig, A. Cherkezov, L. Seibert, J. Hackler, P. Seemann, J. Diegmann, M. Pilz, M. Bachmann, W. B. Minich and L. Schomburg, Nutrients, 2020, 12(7), 2098,  DOI:10.3390/nu12072098.
161. F. Maass, B. Michalke, D. Willkommen, C. Schulte, L. Tonges, M. Boerger, I. Zerr, M. Bahr and P. Lingor, J. Trace Elem. Med. Biol., 2020, 57, 110–115.
162. F. P. Phiri, E. L. Ander, R. M. Lark, E. H. Bailey, B. Chilima, J. Gondwe, E. J. M. Joy, A. A. Kalimbira, J. C. Phuka, P. S. Suchdev, D. R. S. Middleton, E. M. Hamilton, M. J. Watts, S. D. Young and M. R. Broadley, Environ. Int., 2020, 134,  DOI:10.1016/j.envint.2019.105218.
163. A. Gawor, A. Ruszczynska, M. Czauderna and E. Bulska, Animals, 2020, 10(5), 808,  DOI:10.3390/ani10050808.
164. J. Hildebrand, A. Greiner, H. Drexler and T. Goen, J. Trace Elem. Med. Biol., 2020, 61, 126519,  DOI:10.1016/j.jtemb.2020.126519.
165. C. Comera, C. Cartier, E. Gaultier, O. Catrice, Q. Panouille, S. El Hamdi, K. Tirez, I. Nelissen, V. Theodorou and E. Houdeau, Part. Fibre Toxicol., 2020, 17(1), 26,  DOI:10.1186/s12989-020-00357-z.
166. P. Albendea, I. Sierra, C. Hernandez, A. I. Barrado and A. Yllera, Radiat. Phys. Chem., 2020, 174, 108929,  DOI:10.1016/j.radphyschem.2020.108929.
167. F. H. Reboredo, A. Barbosa, M. M. Silva, M. L. Carvalho, J. P. Santos, M. F. Pessoa, F. Lidon, J. C. Ramalho and M. Guerra, Exposure Health, 2020, 12, 917–927,  DOI:10.1007/s12403-020-00354-9.
168. A. M. D. Lopes, P. R. Chellini and R. A. de Sousa, Anal. Lett., 2020, 53(13), 2096–2110,  DOI:10.1080/00032719.2020.1729169.
169. A. Figueiredo, I. M. Costa, T. A. Fernandes, L. L. Goncalves and J. Brito, Nutrients, 2020, 12(4), 954,  DOI:10.3390/nu12040954.
170. L. Chen, X. Li, Z. M. Li and L. G. Deng, RSC Adv., 2020, 10(12), 6736–6742.
171. A. Elik, A. Demirbas and N. Altunay, Anal. Methods, 2019, 11(27), 3429–3438.
172. M. Mhada, M. L. Metougui, K. El Hazzam, K. El Kacimi and A. Yasri, Foods, 2020, 9(5), 660,  DOI:10.3390/foods9050660.
173. M. Hayder, J. Wojcieszek, M. Asztemborska, Y. Zhou and L. Ruzik, J. Sci. Food Agric., 2020, 100(13), 4950–4958,  DOI:10.1002/jsfa.10558.
174. H. M. Zou, C. Zhou, Y. X. Li, X. S. Yang, J. Wen, C. X. Li, S. J. Song and C. J. Sun, Anal. Bioanal. Chem., 2020, 412, 2829–2840,  DOI:10.1007/s00216-020-02515-w.
175. H. Cui, W. Guo, L. L. Jin, Y. E. Peng and S. H. Hu, J. Anal. At. Spectrom., 2020, 35(3), 592–599.
176. Q. F. Zhou, L. H. Liu, N. A. Liu, B. He, L. G. Hu and L. N. Wang, Ecotoxicol. Environ. Saf., 2020, 198,  DOI:10.1016/j.ecoenv.2020.110670.
177. T. Ahmadi-Jouibari, N. Noori and N. Fattahi, Toxin Rev., 2019 DOI:10.1080/15569543.2019.1633543.
178. W. Guo, R. Wang, W. X. Wang and Y. E. Peng, J. Food Compos. Anal., 2020, 91,  DOI:10.1016/j.jfca.2020.103517.
179. M. Fasihi, M. Rajabi, B. Barfi and S. M. Sajjadi, Int. J. Environ. Anal. Chem., 2020 DOI:10.1080/03067319.2020.1784408.
180. O. D. Uluozlu, At. Spectrosc., 2019, 40(4), 133–138.
181. S. Blotevogel, E. Schreck, C. Laplanche, P. Besson, N. Saurin, S. Audry, J. Viers and P. Oliva, Food Chem., 2019, 298,  DOI:10.1016/j.foodchem.2019.125033.
182. K. Karas, A. Ziola-Frankowska and M. Frankowski, Molecules, 2020, 25(5), 1069,  DOI:10.3390/molecules25051069.
183. X. H. Liao, T. Luo, S. H. Zhang, W. Zhang, K. Q. Zong, Y. S. Liu and Z. C. Hu, J. Anal. At. Spectrom., 2020, 35(6), 1071–1079.
184. Y. Y. Su, J. Gao, Y. F. Zhao, H. S. Wen, J. J. Zhang, A. Zhang and C. L. Yuan, J. Food Prot., 2020, 83(8), 1323–1334.
185. M. Perini, S. Pianezze, L. Ziller, M. Ferrante, F. Ferella, S. Nisi, M. Foschi and A. A. D’Archivio, J. Mass Spectrom., 2020, 55(11), e4595,  DOI:10.1002/jms.4595.
186. C. Ozcan and S. Akozcan, Desalin. Water Treat., 2020, 175, 219–228.
187. L. J. Huang, W. Huang, R. J. Shen and Q. Shuai, Food Chem., 2020, 330,  DOI:10.1016/j.foodchem.2020.127212.
188. S. Z. Chen, Y. X. Liu, J. T. Yan, C. L. Wang and D. B. Lu, At. Spectrosc., 2020, 41(4), 169–174.
189. S. Ozdemir, E. Kilinç and S. Fatih, Sci. Rep., 2020, 10(1), 455,  DOI:10.1038/s41598-019-57401-y.
190. E. F. Fiorentini, M. N. Oviedo and R. G. Wuilloud, Spectrochim. Acta, Part B, 2020, 169, 105879,  DOI:10.1016/j.sab.2020.105879.
191. Z. Mehrani, H. Ebrahimzadeh, A. A. Asgharinezhad and E. Moradi, Microchem. J., 2019, 149, 103975,  DOI:10.1016/j.microc.2019.103975.
192. N. Ozkantar, E. Yilmaz, M. Soylak and M. Tuzen, Food Chem., 2020, 321,  DOI:10.1016/j.foodchem.2020.126737.
193. M. Hassan, Z. Erbas, U. Alshana and M. Soylak, Microchem. J., 2020, 156, 104868,  DOI:10.1016/j.microc.2020.104868.
194. Y. Zhang, P. Wu, H. Li, H. J. Luo, W. H. Luo and Z. X. Lin, Spectrosc. Spectral Anal., 2020, 40(2), 632–636.
195. A. Benvidi, M. Yekrangi, S. Jahanbani and H. R. Zare, Food Chem., 2020, 319,  DOI:10.1016/j.foodchem.2020.126432.
196. T. Borahan, T. Unutkan, N. B. Turan, F. Turak and S. Bakirdere, Food Chem., 2019, 299,  DOI:10.1016/j.foodchem.2019.125065.
197. M. Ramezanpour, S. N. Raeisi, S. A. Shahidi and S. Ramezanpour, Micro Nano Lett., 2020, 15(6), 390–395.
198. Z. Bahadir and A. Mermer, Anal. Lett., 2021, 54(3), 558–571,  DOI:10.1080/00032719.2020.1785484.
199. S. Z. Chen, Y. X. Liu, C. L. Wang, J. T. Yan and D. B. Lu, Food Anal. Methods, 2020, 13, 2102–2110,  DOI:10.1007/s12161-020-01818-z.
200. C. Xiong, M. Stiboller, R. A. Glabonjat, J. Rieger, L. Paton and K. A. Francesconi, J. Trace Elem. Med. Biol., 2020, 61, 126502,  DOI:10.1016/j.jtemb.2020.126502.
201. J. He, A. F. Li, H. Y. Han, S. Su, P. Zhang and E. P. Zhang, Rapid Commun. Mass Spectrom., 2020, 34(4), e8591,  DOI:10.1002/rcm.8591.
202. D. L. Novo, A. S. Henn, E. M. M. Flores and M. F. Mesko, Rapid Commun. Mass Spectrom., 2020, 34(S3), e8675,  DOI:10.1002/rcm.8675.
203. M. Edinborough, M. Pilgrim, S. Fearn, A. Cvetkovic, B. Mihailovic, R. Grbic and K. Edinborough, J. Anal. At. Spectrom., 2020, 35(6), 1199–1206.
204. T. Filippini, K. Upson, G. Adani, C. Malagoli, C. Baraldi, B. Michalke and M. Vinceti, Int. J. Environ. Res. Public Health, 2020, 17(7), 2264,  DOI:10.3390/ijerph17072264.
205. J. Jelsma, M. G. M. Schotanus, R. Senden, I. C. Heyligers and B. Grimm, Hip Int., 2019, 29(6), 638–646.
206. Y. Zhang, Y. J. Li, T. X. Ren, X. B. Liu, L. C. Yang, J. H. Piao, J. Wang and X. G. Yang, Biomed. Environ. Sci., 2020, 33(6), 414–420.
207. C. H. Lin, C. K. Su and Y. C. Sun, Microchem. J., 2020, 154,  DOI:10.1016/j.microc.2019.104569.
208. I. Y. Pavlov, R. L. Parker, E. Lazar-Molnar, F. G. Strathmann and J. C. Delgado, J. Immunol. Methods, 2019, 470, 33–39.
209. F. J. Gil, E. Espinar-Escalona, N. Clusellas, J. Fernandez-Bozal, M. Artes-Ribas and A. Puigdollers, Metals, 2020, 10(6), 702,  DOI:10.3390/met10060702.
210. B. K. Gersten, T. S. Fitzgerald, K. A. Fernandez and L. L. Cunningham, J. Assoc. Res. Otolaryngol., 2020, 21, 303–321,  DOI:10.1007/s10162-020-00759-y.
211. M. Prodana, C. E. Nistor, A. B. Stoian, D. Ionita and C. Burnei, Coatings, 2020, 10(6), 526,  DOI:10.3390/coatings10060526.
212. D. Di Candia, E. Muccino, A. Battistini, M. Boracchi, G. Gentile and R. Zoja, Leg. Med., 2020, 42, 101661,  DOI:10.1016/j.legalmed.2019.101661.
213. Y. Zhang, X. Huo, X. L. Lu, Z. J. Zeng, M. M. Faas and X. J. Xu, Chemosphere, 2020, 257,  DOI:10.1016/j.chemosphere.2020.127257.
214. D. R. McKeating, J. J. Fisher, P. Zhang, W. W. Bennett and A. V. Perkins, J. Trace Elem. Med. Biol., 2020, 59, 126419,  DOI:10.1016/j.jtemb.2019.126419.
215. L. A. Henriquez-Hernandez, D. Romero, A. Gonzalez-Antuna, B. Gonzalez-Alzaga, M. Zumbado, L. D. Boada, A. F. Hernandez, I. Lopez-Flores, O. P. Luzardo and M. Lacasana, Sci. Total Environ., 2020, 706,  DOI:10.1016/j.scitotenv.2019.135750.
216. T. Konz, C. Monnard, M. R. Restrepo, J. Laval, F. Sizzano, M. Girotra, G. Dammone, A. Palini, G. Coukos, S. Rezzi, J. P. Godin and N. Vannini, Anal. Chem., 2020, 92(13), 8750–8758.
217. A. I. Barros, F. C. Pinheiro and J. A. Nobrega, Anal. Methods, 2019, 11(27), 3401–3409.
218. A. Bena, M. Orengia, M. Gandini, B. Bocca, F. Ruggieri, A. Pino, A. Alimonti, F. Ghione and E. Farina, Int. J. Hyg. Environ. Health, 2020, 225,  DOI:10.1016/j.ijheh.2020.113454.
219. E. Caron-Beaudoin, M. Bouchard, G. Wendling, A. Barroso, M. F. Bouchard, P. Ayotte, K. L. Frohlich and M. A. Verner, J. Exposure Sci. Environ. Epidemiol., 2019, 29(5), 613–623.
220. Q. L. Yu, X. H. Sun, J. L. Zhao, L. Zhao, Y. F. Chen, L. N. Fan, Z. X. Li, Y. Z. Sun, M. Q. Wang and F. Wang, Nutr. Metab., 2019, 16, 73,  DOI:10.1186/s12986-019-0395-y.
221. A. Foroutan, C. Fitzsimmons, R. Mandal, H. Piri-Moghadam, J. M. Zheng, A. C. Guo, C. Li, L. L. Guan and D. S. Wishart, Metabolites, 2020, 10(6), 233,  DOI:10.3390/metabo10060233.
222. S. S. Tan, G. L. Khor, E. Stoutjesdijk, K. W. T. Ng, I. Khouw, M. Bragt, A. Schaafsma, D. A. J. Dijck-Brouwer and F. A. J. Muskiet, J. Food Compos. Anal., 2020, 89,  DOI:10.1016/j.jfca.2020.103468.
223. K. Ido, R. Matsushita, S. Fujii, S. Miyashita, T. Umemura, A. Hokura and K. Inagaki, Anal. Sci., 2020, 36(6), 717–722.
224. K. L. Rasmussen, T. Delbey, P. d’Imporzano, L. Skytte, S. Schiavone, M. Torino, P. Tarp and P. O. Thomsen, Heritage Sci., 2020, 8(1), 65,  DOI:10.1186/s40494-020-00407-x.
225. A. C. Fernandez-Escudero, I. Legaz, G. Prieto-Bonete, M. Lopez-Nicolas, A. Maurandi-Lopez and M. D. Perez-Carceles, Sci. Rep., 2020, 10(1), 9964,  DOI:10.1038/s41598-020-66472-1.
226. J. Vysloužil, P. Kulich, T. Zeman, T. Vaculovič, M. Tvrdoňová, P. Mikuška, Z. Večeřa, J. Stráská, P. Moravec, V. J. Balcar and O. Šerý, J. Trace Elem. Med. Biol., 2020, 61, 126511,  DOI:10.1016/j.jtemb.2020.126511.
227. N. J. Clark, R. Clough, D. Boyle and R. D. Handy, Environ. Sci.: Nano, 2019, 6(11), 3388–3400.
228. X. T. Jia, X. Y. Yang, W. Zhao, Y. N. Hu and H. F. Cheng, J. Sep. Sci., 2019, 42(18), 2957–2967,  DOI:10.1002/jssc.201900282.
229. M. Menon, B. Sarkar, J. Hufton, C. Reynolds, S. V. Reina and S. Young, Ecotoxicol. Environ. Saf., 2020, 197,  DOI:10.1016/j.ecoenv.2020.110601.
230. M. Zhao, M. Markiewicz-Keszycka, R. J. Beattie, M. P. Casado-Gavalda, X. Cama-Moncunill, C. P. O’Donnell, P. J. Cullen and C. Sullivan, Food Chem., 2020, 320,  DOI:10.1016/j.foodchem.2020.126639.
231. H. W. Tian, L. Z. Jiao and D. M. Dong, Sci. Rep., 2019, 9(1), 10443,  DOI:10.1038/s41598-019-46924-z.
232. A. K. Meinhardt, A. Muller, A. Burcza and R. Greiner, Food Chem., 2019, 301,  DOI:10.1016/j.foodchem.2019.125293.
233. J. M. Roseland, K. M. Phillips, K. Y. Patterson, P. R. Pehrsson, R. Bahadur, A. G. Ershow and M. Somanchi, Nutrients, 2020, 12(5), 1246,  DOI:10.3390/nu12051246.
234. A. Pereira, L. J. G. Silva, B. D. F. Simoes, C. Lino and A. Pena, J. Food Compos. Anal., 2020, 92,  DOI:10.1016/j.jfca.2020.103541.
235. M. Tighe, M. Bielski, M. Wilson, G. Ruscio-Atkinson, G. F. Peaslee and M. Lieberman, Anal. Chem., 2020, 92(7), 4949–4953.
236. A. Hirtz and K. Gunther, Anal. Lett., 2020, 53(16), 2529–2544,  DOI:10.1080/00032719.2020.1746328.
237. B. Li, S. L. Chua, A. L. Ch’ng, D. Y. Yu, S. P. Koh, H. Phang and P. K. T. Chiew, Anal. Bioanal. Chem., 2020, 412(22), 5499–5512,  DOI:10.1007/s00216-020-02770-x.
238. B. May, T. Dreifke, C. D. Patz, C. L. Schutz, R. Schweiggert and H. Dietrich, Food Chem., 2019, 300,  DOI:10.1016/j.foodchem.2019.125168.
239. T. Sugioka, H. Umeda and S. Kunimura, Anal. Sci., 2020, 36(4), 465–469.
240. M. D. Ghoora, D. R. Babu and N. Srividya, J. Food Compos. Anal., 2020, 91,  DOI:10.1016/j.jfca.2020.103495.
241. T. Varga, Z. Sajtos, Z. Gajdos, A. J. T. Jull, M. Molnar and E. Baranyai, Sci. Total Environ., 2020, 736,  DOI:10.1016/j.scitotenv.2020.139686.
242. M. S. Chis, A. Paucean, S. M. Man, V. Bonta, A. Pop, L. Stan, B. V. Beldean, C. R. Pop, V. Muresan and S. Muste, Foods, 2020, 9(6), 822,  DOI:10.3390/foods9060822.
243. M. R. F. Sampaio, M. T. Lisboa, J. G. Timm, A. S. Ribeiro, M. A. Vieira, D. M. Otero and R. C. Zambiazi, Anal. Methods, 2020, 12(10), 1360–1367.
244. L. X. Fan, L. Chen, W. Y. Cui, Y. J. Dong, X. X. Yuan, L. Wang, J. Y. Liang and S. C. Zhao, J. Environ. Sci. Health, Part B, 2020, 55(10), 921–928,  DOI:10.1080/03601234.2020.1797426.
245. G. Drava, V. Iobbi, R. Govaerts, V. Minganti, A. Copetta, B. Ruffoni and A. Bisio, Molecules, 2020, 25(12), 2891,  DOI:10.3390/molecules25122891.
246. S. Paz, C. Rubio, A. J. Gutierrez, D. Gonzalez-Weller and A. Hardisson, Biol. Trace Elem. Res., 2021, 199(1), 382–388,  DOI:10.1007/s12011-020-02151-6.
247. P. Pohl, A. Szymczycha-Madeja and M. Welna, Arabian J. Chem., 2020, 13(1), 1955–1965.
248. E. Carpene, G. Andreani, E. Ferlizza, S. Menotta, G. Fedrizzi and G. Isani, Life, 2020, 10(5), 75,  DOI:10.3390/life10050075.
249. A. Winkler, M. Rauwolf, J. H. Sterba, P. Wobrauschek, C. Streli and A. Turyanskaya, J. Sci. Food Agric., 2020, 100(11), 4226–4236,  DOI:10.1002/jsfa.10463.
250. H. B. Wang, X. T. Chen, J. H. Ye, X. L. Jia, Q. Zhang and H. B. He, J. Sci. Food Agric., 2020, 100(8), 3360–3369,  DOI:10.1002/jsfa.10369.
251. J. Moreda-Piňeiro, J. Sánchez-Piňero, E. Alonso-Rodrìguez, I. Turnes-Carou, P. López-Mahía and S. Muniategui-Lorenzo, J. Food Meas. Charact., 2020, 14(3), 1314–1324.
252. X. O. Dembroszky, Z. May, T. Hartel and A. R. Zsigmond, Environ. Eng. Manage. J., 2020, 19(4), 625–632.
253. P. Jitaru, L. Ingenbleek, N. Marchond, C. Laurent, A. Adegboye, S. E. Hossou, A. Z. Koné, A. D. Oyedele, C. Kisito, Y. K. Dembélé, S. Eyangoh, P. Verger, B. Le Bizec, J. C. Leblanc and T. Guérin, Environ. Int., 2019, 133, 105197,  DOI:10.1016/j.envint.2019.105197.
254. M. L. Astolfi, E. Marconi, C. Protano and S. Canepari, Food Control, 2020, 116, 107327,  DOI:10.1016/j.foodcont.2020.107327.

---