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 spectrometry
  • 4.3 Atomic emission spectrometry and laser induced breakdown spectroscopy
  • 4.4 Vapour generation procedures and atomic fluorescence spectrometry
  • 4.5 X-ray spectrometry

• 5 Nanomaterials

• 6 Applications: clinical and biological materials
  • 6.1 Metallomics
  • 6.2 Imaging with LA-ICP-MS and XRF
  • 6.3 Multielement applications
  • 6.4 Progress for individual elements

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

• 8 Applications: foods and beverages
  • 8.1 Progress for individual elements
  • 8.2 Progress in applications

• Abbreviations

• References

---

Abstract

This review covers publications from the second half of 2015 to the middle of 2016. Techniques and applications relevant to clinical and biological materials, foods and beverages are discussed in the text, presenting the key aspects of the work referenced, while the tables provide a summary of the publications considered. An important development seen within the year is the way in which nanomaterials feature in many of the topics considered in this review. With nanoparticles used as antibacterial agents (containing Ag) in medical devices and in food packaging materials, as vehicles for delivering drugs to tissues and cells and in horticulture, methods for analysis of various sample types that may contain nanomaterials have been developed. Properties of nanoparticles and nanotubes were also exploited for analyte preconcentration in measurement procedures. The implications of recent regulations relating to concentrations of trace elements in infant foods, arsenic in rice and methyl mercury in seafoods were evident, with methods specifically developed with these challenges in mind. Emerging topics to look out for in coming years include applications using triple quadrupole ICP-MS and portable XRF spectrometers.

---

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 also from the previous year.^2–6 The same writers also prepared a perspective review to highlight trends in techniques and applications throughout the 30 years of the Journal of Analytical Atomic Spectrometry.⁷

Progress with techniques to show the distribution and concentration of elements in tissues has featured extensively in our recent ASU reviews. In a tutorial review, Hare and colleagues⁸ discussed techniques providing quantitative and element specific data together with specific location within tissues. They emphasised the importance of avoiding disruption to the native chemical environment during sample preparation and analysis whilst concluding that no single technique is ideal to achieve a profile of all aspects of metal biochemistry. Bodzon-Kulakowska and Suder⁹ suggested ICP-MS as a particularly valuable technique, arguing that it uniquely identifies structures and demonstrates molecular changes in clearly defined regions of the sample. While, all would agree with the suggestion that there are advantages in combining with other techniques to provide further information, those familiar with other techniques might differ with the statement about uniqueness. Plant biology does not often feature in these Updates so it is refreshing to see the review Boughton et al.¹⁰ on mass spectrometry imaging as applied to primary metabolism, natural products, plant defence, plant responses to abiotic and biotic stress, plant lipids and the developing field of spatial metabolomics, Limbeck et al.¹¹ reviewed the advantages associated with LA-ICP-MS for the quantification of trace elements in environmental, biological and medical samples. Commenting on the large variety of sample types, e.g. bone, tissues, blood, and problems with sample ablation behaviour and with elemental fractionation, they concluded that general methodological procedures must be adapted to specific applications.

Several reviews addressed analytical issues relating to clinical conditions. It has long been recognised that there is no single marker for Zn status. The most widely used approach, measurement of the concentration in serum or plasma, is subject to many influences other than the available fraction within the body. A group of experts, the Biomarkers of Nutrition for Development; Zinc Expert Panel, reviewed the evidence for a large number of possible markers to assess Zn status in infants and children.¹² Of those considered, there was insufficient information to recommend most with the Panel recommending dietary Zn intake, plasma Zn concentration and height-for-age. Vanhaecke's group in Ghent suggested that high precision isotopic analysis of essential elements using MC-ICP-MS allows investigation of uptake and metabolism by determining the isotopic composition of an element in different compartments.¹³ Examples cited included Cu in Wilson's disease and Fe in haemochromatosis. Larner¹⁴ proposed that there may be a role for such measurements in studies of cancer. Isotopic fractionation in these disorders has been described in the past and, with the limited availability of the necessary instrumentation, progress with this approach will inevitably be quite slow. Ceko et al.¹⁵ reviewed the potential for XRA spectroscopy, XRF imaging and ICP-MS to investigate the possible roles for trace elements in the ovary while Ring et al.¹⁶ referred to both the surgical and analytical involvement in understanding the metallosis that can develop in patients who have received metal-on-metal orthopaedic implants.

Four reviews for measurements of specific elements in clinical specimens should be mentioned. A comprehensive discussion of the analytical approaches to speciation, as relevant to As toxicity, was prepared by Maher and colleagues from Austria and Australia¹⁷ This excellent review addressed how the integrity of species should be retained throughout collection, storage, preparation and analysis. Extraction and separation conditions must be optimised for particular sample types e.g. those for As-lipids being quite different to those required for polar species. The authors emphasised the importance of mass balances and that extraction and column recoveries should be determined for each step of a procedure to ensure accurate results. Methods should also be thoroughly validated using As recovery experiments and analysis of CRMs. Problems encountered when measuring concentrations of halogens by ICP-OES and ICP-MS, such as high ionisation potentials, low wavelength emission lines, and interferences due to matrix effects were addressed by Mesko et al.¹⁸ These authors reviewed the options available for sample preparation, namely dilution, solubilisation, extraction and those involving matrix decomposition. Deficiency of I is widespread throughout the world and can cause neurological disorders to the foetus, therefore assessing I status by measuring concentrations in urine is an important application. Doggui et al.¹⁹ discussed the techniques for the determination of I and associated analytical performance. Grygo-Szymanko et al.²⁰ assessed the sample preparation procedures and analytical methods for speciation and fractionation of Mn in a wide range of specimen types.

Of relevance to topics discussed in detail in later sections of this Update are the reviews of applications of HPLC-ICP-MS to elemental speciation,²¹ AFS²² and methodologies for determining concentrations of elements in foods for infants.²³

2 Metrology, interlaboratory studies, reference materials and reference ranges

Reference ranges for elements are influenced by both biological and environmental factors, with age, sex and genetic patterns playing a key role in the first group and climate, environmental pollution, diet, occupation and other life-style elements as examples of the latter. The variability of background levels of selected elements in biological fluids and tissues was clearly illustrated by Zhang et al.,²⁴ in a survey of 648 subjects, from 12 to 60 years old, living in suburbs southwest of Beijing (China). The concentrations of Cd, Cu, Mn, Pb and Zn were measured by ICP-MS in blood samples collected from March 2009 to February 2010. Besides confirming the findings of other studies of the general populations, e.g. increased Cd levels in smokers and older people and higher Pb levels in men compared to women, the study also concluded that average blood Cu and Zn concentrations for this population were lower than observed in other countries. Reference ranges for the essential elements were established as 541–1475 μg L⁻¹ (Cu); 5.80–25.2 μg L⁻¹ (Mn) and 2349–9492 μg L⁻¹ (Zn). For the distribution of the toxic metals Cd and Pb, the geometric means of 0.68 μg L⁻¹ and 42.6 μg L⁻¹, respectively, and 95^th percentiles, determined as 5.30 μg L⁻¹ and 100 μg L⁻¹, were a matter of concern. For Pb, these findings confirmed those mentioned in last year’s review,¹ from a large study of Chinese children and a national survey of the blood Pb levels in the Korean population. Over the years, improved knowledge of the toxic effects of certain elements on developing organisms led to increased interest in assessing potential exposure to elements in children and their relative reference ranges. Since sample size is a major drawback in performing such studies, the group of Goullè²⁵ focused on exploiting the capabilities of ICP-MS for simultaneous determination. The concentrations of 25 elements (Ag, As, Be, Bi, Cd, Co, Cr, Cu, Hg, Li, Mn, Mo, Ni, Pb, Pt, Rb, Se, Sn, Sr, Te, Tl, U, V, W and Zn) were measured in whole blood and plasma samples of only 0.3 mL. Reference ranges were determined for these elements in a sample of 99 children (48 of them less than 10 years old and 51 between 10 and <18 years old). The results were similar to those reported for adults; however, the levels of Cd, Hg, Pb and Zn appeared to increase with age, whereas Co, Cr, Cu, Mo, Ni and Sn concentrations showed a decreasing trend in older children. In comparison with biological fluids, hair is a much easier type of sample to obtain and store, however, its usefulness in biomonitoring programs is controversial, partly because of the difficulties in establishing reference values for this matrix. In their survey of 4384 women and 1524 men living in Moscow and Moscow region, Skalny and coworkers²⁶ measured the hair levels of 11 elements (Al, As, Be, Bi, Cd, Hg, Li, Ni, Pb, Sn and Sr) using the NexION 300D ICP-MS. All subjects were occupationally non-exposed adults. They concluded that sex-related differences influenced the levels of these elements in hair, with significantly higher levels of Al, As, Cd, Hg, Li and Pb in male hair samples and, conversely, higher concentrations of Bi, Ni, Sn and Sr in those from women.

Nutritional products intended for children require significant attention due to the importance for normal development and potentially higher risks caused by a lower body weight and dependence from limited food types. Over the past two years, the European Union maximum allowable levels of Cd and Pb in infant formulas and other food consumed primarily by young children were reduced. This may have triggered efforts to revise and revalidate robust methods, where possible by collaborative studies, for the simultaneous determination of several elements in infant formulas, adult nutritionals and milk based products. A review of the state-of-the-art of analytical methods approved by AOAC was published by Poitevin et al.,²³ who also addressed current needs, analytical gaps and future trends in this area. The same group²⁷ exploited the features of existing official methods to develop a robust procedure, based on ICP-MS, with or without collision/reaction cell, and acid digestion, for the determination of 10 trace elements (Al, As, Cd, Co, Cr, Hg, Mo, Pb, Se and Sn) in infant formulas, adult nutritionals and milk based products. The figures of merit obtained in a single laboratory validation met the AOAC criteria for official method performance. A different group, led by J. J. Thompson,²⁸ worked on the extension of the scope of an existing AOAC method (2011.19, see below), to determine 12 minerals and trace elements (Ca, Cr, Cu, Fe, K, Mg, Mn, Mo, Na, P, Se and Zn) in infant and nutritional formulas. Following microwave-aided acid digestion, the samples were analysed by ICP-MS, using Ge and Te as internal standards. The in-house validation study, performed on 11 matrices and the NIST SRM 1849a, gave repeatability RSDs less than 5%, recovery of spiked amounts ranging from 90.1% to 109% and relative bias from the certified values of SRM 1849a between 96.2% and 107.7%, thus meeting the AOAC method performance requirements. An initial assessment of reproducibility RSD across 6–7 laboratories and seven matrices indicated RSD values ranging from 2.5% (Cu) to 7.1% (P). The same group²⁹ reported the results of the collaborative study of the AOAC Method 2011.19. The analytical protocol involved microwave-assisted digestion of samples of seven matrices among infant, paediatric and adult nutritional formulas with HNO₃, H₂O₂ and internal standard, followed by measurement of the concentrations of Cr, Mo and Se (only) by ICP-MS with collision/reaction cell. The target requirements for the repeatability and reproducibility RSDs (≤5% and ≤15%, respectively) were met for almost all of the matrices, with repeatability RSDs ranging from 1.0 to 7.0% and values from 2.5 to 13.4% for the reproducibility RSD observed among nine laboratories from four different countries, using seven different models of ICP-MS instruments. The LOQs for ready-to-feed formulas were set as 20 μg kg⁻¹ for Cr and Mo and 10 μg kg⁻¹ for Se. Since I is also important for growth and development, especially in babies and children, Zywicki and Sullivan³⁰ coordinated an additional collaborative study on the same matrices, including the NIST SRM 1849a. The aim was to validate the measurement of I concentrations by ICP-MS according to the AOAC First Action Official Method (SM) 2012.15. This time, sample preparation required alkaline digestion with a KOH solution, carried out in an oven or an open-vessel microwave system. After addition of NH₄OH and Na₂S₂O₃, dilution to a stated volume, filtration and further dilution, the digests were then analysed. The repeatability RSD ranged from 0.77% to 4.78% and values for the reproducibility RSD were between 5.42% and 11.5% which met the AOAC criteria for method performance.

Exposure to MeHg through the diet is a matter of concern, especially for countries where large amounts of seafood of marine origin are traditionally consumed. Although current maximum levels in food of marine origin are based on the total content of Hg, work is in progress to evaluate the feasibility and usefulness to measure Hg species, in particular MeHg, because of its greater neurotoxicity, affecting developing organisms to a larger extent. Valdersnes et al.³¹ reported the results of a collaborative study of a method for the determination of MeHg by GC-ICP-MS, using ID for higher reliability of the measurement results. Sample preparation involved spiking with an appropriate amount of the ²⁰¹Hg enriched MeHg, extraction with TMAH, adjustment to pH 5, addition of hexane and sodium tetraethylborate to derive the sample. After centrifugation, the supernatant organic phase was analysed using GC-ICP-MS. Seven seafood products (mussel tissue, squid muscle, crab claw meat, whale meat, cod muscle, Greenland halibut muscle and the CRM NRCC DOLT-4 Dogfish liver), with MeHg concentrations ranging from 0.035 to 3.58 mg kg⁻¹ (as Hg) on a dw basis, were analysed by the eight participants. The observed repeatability RSD values were between 2.1% and 8.7%, whereas reproducibility RSD ranged from 5.8% to 42%.

In recent years, much effort has been spent in devising procedures to determine As species, and especially iAs in food. In January 2016 maximum limits for iAs in rice were established in the EU and, subsequently, a recommendation was issued inviting EU Member States to assess the levels of iAs in other food types over a period of three years. This work is to be accomplished based on progress made on the standardisation of analytical methods. A standard method (EN 16802) was published by the European Committee for Standardization (CEN),³² to measure concentrations of iAs in food samples. Arsenic species are extracted, in a heated water bath, into a diluted HNO₃–H₂O₂ solution, the As^III is oxidised to As^V, with determination by AE-HPLC-ICP-MS. Another collaborative trial was conducted by Fiamegkos et al.³³ to determine the performance characteristics of an analytical method based on FI-HG-AAS. Sample preparation involved solubilisation of the protein matrix with HCl, to denature proteins and release all As species into solution, followed by extraction of iAs with CCl₃ and back-extraction in an acidic medium. The method was tested on seven RMs, covering a broad range of matrices (mussels, cabbage, hijiki seaweed, fish protein, rice, wheat, mushrooms) with concentrations ranging from 0.074 to 7.55 mg kg⁻¹. The LOQ was 0.010 mg kg⁻¹ of iAs. The repeatability RSD ranged from 4.1 to 10.3%, while the reproducibility RSD ranged from 6.1 to 22.8%. The authors recommended following the SOP available from https://ec.europa.eu/jrc/en/interlaboratory-comparisons, to ensure that the determination of iAs is free of interferences from other known As species, except for MMA^V, which, however, is usually only a minor As species in food.

The importance of the traceability of measurement results to common references is widely recognised but often difficult to achieve, because of the lack of reference methods and materials, to act as steps in the traceability chains from the routine laboratory result to a higher metrological reference. The Joint Committee for Traceability in Laboratory Medicine was set up to evaluate the claim of potential reference methods and materials in this area. Greaves and co-workers³⁴ investigated the performance of ICP-MS, for the determination of Cl in sweat, as a biomarker for the diagnosis of cystic fibrosis, aiming to submit a candidate reference method to the JCTLM. Calibration was based on pure NaCl solutions and linearity was demonstrated over the range from 7.4 mmol L⁻¹ (LOQ) to 225 mmol L⁻¹. The authors compared the results obtained, using either online (Ge and Sc) or off-line (Ga) internal standards, on the EQA samples provided by the Royal College of Pathologists of Australasia Quality Assurance Program for sweat electrolyte testing. Based on the evidence of lower imprecision and bias (off-line IS: RSD 1.9%, bias 1.5 mmol L⁻¹; online IS: RSD 8.0%, bias 3.8 mmol L⁻¹) the first procedure was recommended. Because of the variable effect of the matrix on analytical determinations, the traceability of measurement results often relies on demonstration of negligible bias using CRMs. In laboratory medicine, where samples are often analysed with minimal sample pre-treatment, calibration via matrix CRMs is frequently the choice to assure the traceability of the results. The commutability of such calibrators, i.e. their capability to reproduce the analytical behaviour of the analyte(s) in the test sample(s), is a key issue. A large study, by Ge et al.,³⁵ explored the analytical behaviour of patient samples (48) and processed (CRMs, EQAS samples) materials (24) for the determination of Mg in serum, using 12 routine kits, including calibrators, applied on a Hitachi 7180 automatic analyser. The results were compared with those obtained by ICP-MS, as the reference method. The overall precision was good (RSD < 2.01%), while the overall mean bias ranged from −0.04 (−4.52%) to 0.06 mmol L⁻¹ (7.20%). Whereas human serum pools were commutable with all kits, matrix effects for external quality assessment samples, CRMs and calibrators were observed in some of the assays. Aqueous reference materials could only be used for calibration in ICP-MS. Significant efforts are therefore required for the assessment of materials to be used as reference in analytical work. This year, two reports of such work were presented, supporting RMs for Hg isotopic composition in hair and for Hg species concentrations in tuna fish. Cold vapour MC-ICP-MS was applied by Yamakawa et al.³⁶ to determine Hg isotopic ratios in an existing human hair CRM (NIES CRM No. 13). The results identified mass-dependent and mass-independent fractionation values for Hg isotopes (¹⁹⁹Hg, ²⁰⁰Hg, ²⁰¹Hg, ²⁰⁴Hg). The study was supported by quality assurance measures including determination of the Hg isotopic compositions of secondary RMs (IAEA-085, IAEA-086, and CRPG-RL24H) and a bilateral interlaboratory comparison with the University of Pau. Factors affecting the preparation of a tuna fish candidate RM for total Hg and MeHg mass fractions (i.e. stability during production, homogeneity and analyte segregation) were investigated by Nomura and coworkers.³⁷ Methyl mercury was determined after selective extraction of the samples with toluene and L-cysteine for 15 min at 80 °C. Comparable results were obtained with different analytical techniques for total Hg mass fraction by CV-AAS (3.03 ± 0.22 μg g⁻¹), by solid sampling ETAAS (2.99 ± 0.43 μg g⁻¹) and with a direct mercury analyser (2.95 ± 0.15 μg g⁻¹). The MeHg mass fraction was 2.49 ± 0.13 μg g⁻¹ by CV-AAS and 2.44 ± 0.46 μg g⁻¹ by solid sampling ETAAS. The best procedure for RM preparation was fine grinding in a cryogenic mill. Analyte segregation, studied at different particle size, for both analytes, was not observed. Stability studies, covering a period of 10 months and storage temperatures up to 50 °C, showed no loss of analytes. In order to guarantee a similar level of protection worldwide from potential risks due to the presence of contaminants in the environment, water and food, it is important that the means to assure comparable measurements are widespread to the whole analytical community. It is therefore interesting to mention the report of a proficiency testing scheme for Cd, Cu and Fe in drinking water, carried out in Indonesia using external, rather than consensus, assigned values.³⁸ The organisers justified their choice pointing out, that although consensus values are easier to obtain, they may be significantly biased and suffer from large uncertainties when a large proportion of the participants are relatively inexperienced. Thus the Research Centre for Chemistry of the Indonesian Institute of Sciences organised this proficiency testing scheme using a material already assessed as a candidate RM by two independent techniques: ICP-AES and ETAAS. The results of this exercise confirmed good performance for 94% (Cu) and 71% (Fe) of the 62 participants, whereas, for the trace element Cd, much poorer performance was highlighted using this approach.

3 Sample collection and preparation

3.1 Collection, storage and preliminary preparation

It is important that the conditions under which specimens are stored between collection and analysis are established, so as to avoid either contamination or losses of the analytes. The large number of samples analysed for the US NHANES reports necessarily have to be stored prior to analysis. The laboratory responsible for the trace element analyses reported their results from a 12 month temperature stability study of iHg, EtHg and MeHg in whole blood.³⁹ Two pools of bovine blood were stored at −70, −20, 4, 23 and 37 °C. The concentrations of the three species were determined at 1, 2, 4 and 6 weeks and 2, 4, 6, 8, 10 and 12 months. It was found that the analytes were stable for at least one year when stored at up to room temperature (23 °C) but for no more than 2 weeks when at 37 °C. Palcic et al.⁴⁰ established a method to determine concentrations of the urinary metabolites of (2-chlorovinyl)arsenous acid, otherwise known as the chemical warfare agent, Lewisite. An accelerated Arrhenius matrix-matched stability study demonstrated that the metabolites are stable in urine for more than a year. Readers may wonder why such an assay was thought to be necessary. The authors point to such munitions being recovered from the sea by fishing trawlers and to potential exposures around former production facilities, and to occupational exposure due to the destruction of stockpiled materials.

It is generally understood that a timed collection of urine is the ideal specimen type when assessing excretion of trace elements, due to the variation in water content throughout the day. As it is often considered inconvenient, concentrations of elements in spot samples are usually expressed as a ratio to the concentration of creatinine or, sometimes, the specific gravity. However, there are reports that suggest these corrections may not be reliable and Hoet et al.⁴¹ presented further work that supports such reservations if adopted for environmental studies. Paired 24 h collection and random spot urine samples were obtained from 39 volunteers with no occupational exposure to trace elements, and analysed to determine concentrations of 17 elements. Results from the spot samples were presented as; absolute concentration, ratio to creatinine and with the concentration adjusted to specific gravity. These three values were then compared to those of the 24 h collection samples. There was no single pattern of relationship between urinary concentrations measured in spot and 24 h collections and the authors concluded that, in the context of environmental exposure, creatinine adjustment is not recommended, and that specific gravity adjustment appears to be a more reliable alternative. However, for some elements neither method appeared to be suitable. In contrast to the results of Hoet et al.,⁴¹ Li et al.⁴² reported that for I in urine the ratio to creatinine is an acceptable way to report results. These authors measured I in 24 h collection and spot urine specimens, and in serum, from 222 pregnant women living in an I adequate area of China. They found that I intake, as assessed by spot urine concentration in pregnant women, is inaccurate and increases the apparent prevalence of iodine deficiency. These workers suggested that the I to creatinine ratio better reflects both the 24 h excretion and the circulating I levels during pregnancy and the postpartum period. Clearly the conundrum is unsolved! The experience of this reviewer suggests that factors such the time of collection of spot samples, the level of exposure and the elements being investigated all contribute to the reliability of ‘corrected’ results' and that experimental procedures should be validated in advance of conducting studies involving collection of urine.

3.2 Digestion, extraction and preconcentration

3.2.1 Clinical and biological materials. While some form of acid digestion or extraction remains the most widely used general approach to sample preparation, alkaline dissolution is a favoured alternative, as demonstrated by Mohd-Taufek et al.⁴³ These workers measured concentrations of 21 elements in human milk samples, by ICP-MS, following dissolution with EDTA, NH₄OH, isopropanol and Triton X-100.

Quantitative imaging of elements in tissues and other samples is relevant to understanding their physiological roles, as discussed in further detail in Section 6.2. Perrin et al.⁴⁴ compared preparation procedures to assess which best preserves the site and concentration of elements in single cells, for analysis by μPIXE and SR-μXRF spectrometry. Using a rat primary hippocampal neuron model, losses of K and Mg and increases in Cl, Na, S and Zn were seen in chemically fixed cells compared to cryofixed cells. The K/Na and Cl/K cellular ratios also indicated good preservation of the chemical and structural integrity of cryofixed cells, but not those fixed by either formaldehyde or methanol. From these results their recommended procedure involved rinsing cells with ammonium acetate, rapid freezing by plunging into liquid nitrogen-chilled cryogenic fluid and freeze-drying at low temperatures.

New reagents used to preconcentrate analytes are reported each year in these Updates. Recently, nanomaterials have featured more widely in developments for the analysis of biological and food samples. Oxidized carbon nanotubes covered with layers of bovine serum albumin, described as restricted-access carbon nanotubes were prepared and characterised by Barbosa et al.⁴⁵ This material had a protein exclusion capacity of almost 100% and a maximum Pb²⁺ adsorption capacity of 34.5 mg g⁻¹. It was used in an online SPE system coupled to a thermospray flame-furnace AA spectrometer to determine concentrations of Pb in untreated human blood serum. An LOD of 2.1 μg L⁻¹, an enrichment factor of 5.5, and inter- and intra-day precisions (RSD) of <8.1% were reported. Recoveries from spiked samples were from 89.4 to 107.3%. Mortada et al.⁴⁶ used hydroxyapatite nanorods, prepared from recycled eggshell, for the preconcentration of Cu, Pb and Zn in various sample types including blood and urine. X-ray diffraction, TEM and FTIR spectrometry were used to confirm the structure of the material which, as with the carbon nanotubes, was also used for SPE. Methods were optimised for the adsorption and extraction in either batch or column procedures, followed by FAAS determination. Detection limits for Cu, Pb and Zn were 0.72, 5.12 and 0.55 μg L⁻¹, respectively, with an enrichment factor of 250. Accuracy was verified using biological CRMs. Magnetic Fe₃O₄@ZIF-8 core–shell nanoparticles (NPs) were used by Zou et al.⁴⁷ to preconcentrate iAs. The NPs, together with the adsorbed As, were separated by a magnet and were then dissolved in HCl for detection of the As by HG-AFS. Under the optimized conditions, the sensitivity for iAs was significantly increased through preconcentration, with satisfactory analytical figures of merit obtained. Alonso and colleagues⁴⁸ also employed magnetic NPs, functionalised with 1,5-bis(di-2-pyridil)methylene thiocarbohydrazide, in a procedure developed to measure concentrations of Hg in biological and seawater samples. These workers employed an FI-SPE-CV-ETAAS system and avoided any problems caused by back pressure in the column by applying an external magnetic field to immobilise the NPs. The accuracy of the proposed method was demonstrated by analysing three CRMs and a preconcentration factor of 5.4 was calculated.

3.2.2 Foods and beverages. Microextraction techniques minimise solvent requirements and enhance the ‘green’ aspect of analytical science while coping with complex food matrices. The optimisation of a supramolecular solvent-based extraction and preconcentration scheme was reported⁴⁹ for use in water and food samples. The supramolecular solvent, formed from decan-1-ol (125 μL) and THF (450 μL), was used to transfer the CoII, complexed with DDTC (300 μL, 0.1% w/v) for use in a microsampling FAAS. The preconcentration factor was 30 with an RSD of 1.51% and an LOQ 6.32 μg L⁻¹. The combination of ionic liquid (IL) and multi-walled carbon nanotube (MWCNT) was described by Grijalba et al.⁵⁰ to extract and preconcentrate iAs in a complex food matrix. The IL, tetradecyl(trihexyl)phosphonium chloride, formed an ion pair with As^VMoO₄—after removal of PO₄²⁻ with BuOH–CHCl₃ (1 : 3) mix—which was adsorbed onto the MWCNT. The MWCNTs were then dispersed in tetradecyltrimethylammonium bromide and injected directly into the graphite tube for determination of As by ETAAS using mixed Mg–Pd matrix modifier. No digestion or back-extraction procedures were required as MWCNTs were pyrolyzed before atomisation. An enrichment factor of 70 was obtained with 5 mL of garlic extract. The LOD was 7.1 ng L⁻¹ and RSDs, at 5 ng L⁻¹, were 5.4% and 4.8% for As^III and As^V, respectively. A green and simple low solvent-volume precipitation technique for separation and preconcentration of Cd in food samples followed by microsampling FAAS detection was proposed.⁵¹ An azo dye, normally used for spectrophotometric assay, 1-(2-thiazolylazo)-2-naphthol (200 μL, 1%), was used to precipitate Cd after hot plate mineralisation and adjustment to pH 8. The LOD was 0.25 μg L⁻¹ and the RSD was 5.5% at a concentration of 20 μg L⁻¹. Silver is frequently used as an anti-microbial agent but has other applications and AgNPs were used as a solid sorbent in ultrasound-assisted dispersive SPME to determine concentrations of Hg in water samples using HR-CS ETAAS.⁵² The optimised conditions for effective adsorption of Hg²⁺ from 10 mL water were pH 3.5, and 10 mg of AgNPs. After sonication, the AgNPs, separated by centrifugation, were dissolved in 7 M HNO₃. The preconcentration factor was 15 with an RSD of 7% and an LOD of 0.005 ng mL⁻¹. Some other DLLME-based techniques allowed the detection of 0.8 μg Cr^VI per mL in drinking water, coupled with TXRF,⁵³ and the determination of Pd^II in various food and water matrices using 2-mercaptobenzimidazole as a complexing agent for use with FAAS.⁵⁴

4 Progress with analytical techniques

4.1 Mass spectrometry

Removal of polyatomic interferences by way of collision/reaction cell technology commonly features in this section, particularly in regard to determination of low analyte concentrations in biological matrices. With a focus on blood and serum Co concentrations <5 μg L⁻¹, Kerger et al.⁵⁵ directly compared two Q-ICP-MS platforms with different methods of interference removal; DRC technology employing NH₃ as a reaction gas versus CRC using He for KED. The use of NH₃ reaction gas with Co-spiked blood, and to a lesser extent with serum, resulted in interference (most likely from ⁴²Ca¹⁴NH₃ and ⁴³Ca¹⁴NH₂) adding approximately 3 μg L⁻¹ to the background signal. Under optimised conditions, the CRC system measured consistently lower Co concentrations and achieved lower LODs compared with the DRC methodology (serum: 0.14 versus 0.21 μg L⁻¹; whole blood: 0.62 versus 1.7 μg L⁻¹). These findings may have implications for choice of technology with the growing interest in application of lower blood Co thresholds in clinical practice. In another study,⁵⁶ the gas mixture CH₃F/He was evaluated for quantitation of trace levels of As in blood, serum and urine achieving an excellent instrumental LOD of 0.8 ng L⁻¹. The CH₃F reacts with As with high efficiency to form AsCH₂⁺, which can be monitored at an interference free m/z ratio of 89, avoiding Cl-based interferences on m/z 75. Addition of He to the gas mixture was required to obtain good reaction efficiency and led to only a modest decrease in sensitivity compared with using He in KED mode. An optimal S/N ratio was achieved with a gas flow rate of 1.6 mL min⁻¹ and a rejection parameter, q, of 0.55 to 0.65. In the final method, the workers used a triple cone interface to reduce dispersion of the ion beam and several strategies to reduce non spectral matrix effects, including IS correction with Te, adequate sample dilution (at least 1 in 50 for blood) and addition of 3% EtOH to all samples and standards. Taking account of the dilution factor, the functional LOD was 20 ng L⁻¹ and the LOQ was 70 ng L⁻¹. Results obtained from biological RMs were in good agreement with certified values and average recoveries from spiked whole blood were 96.8 ± 3.9%. Peng et al.⁵⁷ proposed a method for ultra-trace detection of the carcinogen BrO₃⁻ in drinking water, which following a 6 minute isocratic IC separation using a high salt mobile phase (KOH), employed N₂O reaction gas to convert ⁷⁹Br⁺ to ⁷⁹Br¹⁶O⁺ with off-mass detection at m/z 95. The S/N ratio was enhanced by two orders of magnitude with respect to standard mode. Additionally, an online aerosol dilution technique successfully alleviated the signal suppressing effects of the high-salt eluent. The LOD and LOQ were 0.013 and 0.04 μg L⁻¹ respectively.

The use of ICP-QQQ-MS for interference-free analysis of As and halides was highlighted in two publications^40,58 this year. Palcic et al.⁴⁰ described development of a rapid, high throughput LC-ICP-MS/MS method for quantitative analysis of the major Lewisite metabolite (2-chlorovinyl)arsenic acid with qualitative identification of minor metabolites in urine. Prior to MS detection, the metabolites were oxidised using H₂O₂ and then separated by RP-LC with isocratic elution using ion-pairing reagent, tetrabutylammonium hydroxide, resulting in a 5.73 minute injection to injection cycle. Thereafter Cl-based interferences were removed using O₂ gas in the octopole reaction cell with detection of the 75 → 91 m/z transition, corresponding to the reaction product AsO⁺. The LOD was 2.2 μg L⁻¹ and the LOQ, 7.4 μg L⁻¹, while a good between day RSD was achieved of 3.1–3.3%. Triple quadrupole ICP-MS was also applied to determination of drug related Cl and Br compounds in blood plasma as an alternative to radiolabelling.⁵⁸ This technology permitted the use of H₂ gas, which has a better reaction efficiency with Cl than O₂, to determine interference-prone Cl by formation of ClH₂⁺ and detection of m/z ratio transition 35 → 37. The Br was detected “on mass” at an m/z of 79, which together with its higher IP, resulted in 5 to 6-fold improvement in sensitivity. The LOQs for Br and Cl were 0.01 mg L⁻¹ and 0.05 mg L⁻¹, respectively. Prior HPLC separation of the drug-related compounds from inorganic Cl was essential and measures were taken to reduce Cl contamination. This involved spontaneous nebulisation to avoid leaching of Cl from peristaltic pump tubing and disconnection of the HPLC system during the initial stage of the chromatographic run to avoid introduction of inorganic salts into the ICP. Recoveries from plasma spiked with diclofenac and 4-bromobenzoic acid were within the range 95–105% and precision was typically <10%. The necessity of contamination control for Cl, to take advantage of the superior sensitivity capabilities of ICP-QQQ-MS was illustrated in a study⁵⁹ describing speciation of chlorate and chlorite in blood, serum and plasma. Despite prior separation of chlorite from excess chloride and use of ICP-QQQMS with H₂ reaction gas, a chlorine background persisted and no improvement in instrumental LODs was achieved compared with single quadrupole ICP-MS (∼0.1 mg L⁻¹ for chlorite and ∼0.2 mg L⁻¹ for chlorate).

Cordeau et al.⁶⁰ explored a novel application of macrobore LC-ICP-MS for the quantitation of Se-labelled peptides as a suitable alternative to conventional radiolabelling methods. Inclusion of LC separation to isolate the less polar Se peptides from salts present in the biological matrix compared with direct FI analysis afforded a 4-fold increase in signal intensity and an apparent pre-concentration of 183% of the Se-peptides. The final LC method consisted of a C8 RP stationary phase with gradient elution using MeOH : H₂O with a 100 μL injection volume and a total run time of 11 minutes. Isotope ⁷⁸Se was measured in CRC mode using He (9.6 mL min⁻¹) to suppress Ar polyatomic interferences and a linear response was obtained (R² = 0.9991) from 25 to 1000 ng L⁻¹ Se. The LOD and LOQ were 17 and 57 ng L⁻¹ Se, equivalent to approximately 20 and 72 fmol of injected Se respectively. Bias calculated from spiked samples was <9%. When the method was applied to determine the dissociation constant of a vasopressin ligand containing two selenocysteine residues with its receptor system, results were comparable to those obtained using radiolabelling.

This year there have been two reports addressing isotopic analysis in biological samples by MC-ICP-MS, an area highlighted in last year's Review. The first study⁶¹ evaluated a method for S isotopic analysis in serum and its application to patients with cancer or rheumatoid arthritis. The serum was acid digested and S purified using strong AEC, which attained a S : P ratio of ≥20 and removed the matrix effects of phosphate. High resolution mode (R ∼ 9000) in combination with aerosol desolvation was employed to resolve spectral interferences with the MC-ICP-MS. Correction for instrumental mass bias towards preferential detection of heavier ions was achieved via the standard-sample-standard bracketing technique. Addition of stoichiometric amounts of NH₄ improved SO₄ transmission and stability through the desolvation process while minimising clogging of the cones and membrane. For an 8 mg L⁻¹ standard solution, good reproducibility was observed for δ³³S and δ³⁴S (±0.15‰ and ±0.10‰ respectively, n = 25) although this was slightly poorer in the biological matrix. Similarly, agreement of results with accepted values for inorganic RMs was within 0.3‰ but a positive bias (0.75–1.0‰) was apparent when results from biological samples were compared with those by the reference method, EA-IRMS. Anoshkina et al.⁶² focused on direct isotopic Fe analysis in acid digested whole blood compared with prior chromatographic separation by AEC. Here, the authors used a standard sample introduction system and reported that an Fe concentration of 1 mg L⁻¹ was required to minimise spectral interference. Instrumental mass discrimination correction was found to be most effective employing IS correction with 1 mg L⁻¹ Ni in combination with external correction by standard–sample–standard bracketing. Evaluation of the influence of matrix factors on δ⁵⁶Fe values using synthetic solutions demonstrated a negative bias of up to −0.17‰ in the presence of 1 mg L⁻¹ Na and K while 1% (w/v) glucose resulted in a positive bias of 0.08‰. No significant differences in isotopic values measured in blood samples (n = 10) were observed when compared with prior Fe isolation by AEC and reproducibility was highly comparable (between day RSD (n = 8): 0.03 and 0.06‰ for δ⁵⁶Fe and δ⁵⁷Fe respectively).

Use of ICP-MS alongside ESI-MS/MS (dual analysis) enables concurrent detection of elements and their pertaining molecular species following isolation of the compounds of interest by a common separation procedure. Bierla et al.⁶³ performed speciation of lipophilic Se compounds in sunflower oil by RP-HPLC followed by simultaneous determination of elemental Se and specific Se-species by ICP-MS and high mass accuracy (<1 ppm) HR-ESI-MS/MS respectively. While identical chromatographic conditions served for sample introduction into both MS systems, the use of non-aqueous mobile phase in the gradient elution resulted in constant changes in the ICP-MS plasma conditions affecting detection sensitivity, which had to be corrected for mathematically post run. A total of 11 Se-containing triglycerol derivatives were identified for the first time using the final method. Another interesting example of dual analysis⁶⁴ involved CF-GE coupled online to ICP-MS with splitting of the chromatographic eluents for corresponding peptide mass fingerprinting by MALDI-TOF/TOF-MS to allow monitoring of intracellular metals and their associated proteins. The procedure was demonstrated through a study of Ni and Zn associated with two H. pylori metallochaperones, HypA and HspA, overexpressed in E. coli cells. The GE was performed using a three-step voltage program with a tris-glycine buffer where the metal-binding proteins were eluted at the last step and transferred to ICP-MS in 50 mmol L⁻¹ NH₄NO₂. When monitoring the metals by ICP-MS, use of the ion ratio with ³⁴S was found to be essential to correct for protein loss during the GE separation. Calibration curves formed from a series of Zn,Ni-HpHypA samples with loaded metal concentrations ranging from 15 to 90 μmol L⁻¹ demonstrated good linearity for both metals (R² > 0.99) corroborating the well-preserved integrity of the metal–protein complexes and applicability for determining metal–protein stoichiometries.

In the arena of nanoparticle analysis, several reports utilising single particle ICP-MS in food related matrices are of relevance to this section although the lack of reports in biological matrices is of note. Single particle ICP-MS methodology enables direct quantification of single particle size, concentration and size distribution in low concentration solutions and can distinguish individual particles from dissolved metal ions. In a study assessing internalisation of 10 nm AgNPs into Arabidopsis plant roots and leaves,⁶⁵ a macerating enzyme digestion method as an alternative to acid digestion was demonstrated to release NPs from plant tissue without affecting the NP size distribution. After optimisation of the single particle ICP-MS method using a 30 nm Ag particle standard, the LOD, defined as the minimum detectable size of a single NP, was 10 nm. The most frequent NP size detected in the plant tissues was 26–27 nm, suggesting aggregation of the 10 nm particles. Two other studies applied single particle-ICP-MS to investigate AgNPs released from plastic food containers into food simulants.^66,67 The approach described by Ramos et al.⁶⁶ included a flow rate of 50 μL min⁻¹ and a short dwell time of 10 ms to prevent multiple particles being counted as one. Using calibration curves prepared from 40 and 60 nm particle solutions, the LODs were 0.5 and 10 ng L⁻¹ for 40 and 60 nm AgNPs respectively. The prepared food extracts were diluted with H₂O to yield a total Ag concentration at the ng L⁻¹ level (particle number concentrations < 10⁸ L⁻¹). When applied to food simulants, single particle-ICP-MS analysis confirmed the presence of AgNPs (18–30 nm) in all extracts investigated and AgNP concentration number varied from 4.0 to 1510 × 10⁶ L⁻¹. Witzler et al.⁶⁸ applied single particle ICP-MS for the quantification of Au and AgNPs in a fruit juice matrix with no sample digestion and compared use of dissolved calibration with NP size calibration. However, accuracy and precision of the measurements were confounded by inadequate signal to noise ratios and NP instability, particularly for diameters <30 nm.

4.2 Atomic absorption spectrometry

A novel approach to the measurement of Ag²⁺and AgNPs was reported by Leopold and her colleagues.⁶⁹ Silver has well-recognised antibacterial properties and AgNPs may be used in food packaging. The extent to which transfer into food samples may then occur can be determined by separate measurement of the AgNPs and the Ag²⁺. This was shown to be possible using solid sampling CS-ETAAS, due to a difference in atomisation time between the two species, of approximately 2 seconds. The strategy was applied to analyses of eight foods spiked with Ag²⁺ or AgNPs, fresh pepper after being in silver-impregnated polypropylene food storage boxes and a mussel sample used in an Ag-exposure study. The differences in the appearance times of the atoms from AgNPs and Ag²⁺ was matrix dependent but were significant with all sample types. Preliminary experiments pointed to the variation being related to components such as carbohydrate and protein.

Differences in atomisation temperatures was exploited by Almeida et al.,⁷⁰ to sequentially determine Cr and Mn in vegetable oil and biodiesel, produced from soybean, canola, sunflower and corn, by CS-ETAAS. Atomisation temperatures of 1900 and 2500 °C for Mn and Cr respectively, produced two distinct signals. Samples were prepared by dilution with MIBK. The LOQs were 3.57 ng g⁻¹ for Cr and 4.48 ng g⁻¹ for Mn. Recoveries from spiked samples were 79 to 115% for Cr and 79 to 108% for Mn. External calibration was used and results were comparable with those obtained following acid digestion.

A high resolution CS-ETAAS procedure was also used by Ozbek and Akman⁷¹ to measure concentrations of F in baby foods. The F was determined as the CaF molecule at 606.440 nm, using solid sampling. The limit of detection was 0.20 ng and the F concentrations of samples ranged from <LOQ to 3.75 μg g⁻¹. Micro-scale distribution of F in samples was also determined.

Sample preparation using functionalised magnetic NPs was discussed above.⁴⁸ The analysis, for Hg in biological samples, was completed via an FI manifold that included SPE, CV generation and ETAAS. The optimized procedure had an LOD of 7.8 ng L⁻¹ with an RSD of 1.7% at a concentration of 1.0 μg L⁻¹ Hg.

4.3 Atomic emission spectrometry and laser induced breakdown spectroscopy

Two impressive reports describe work with novel systems for plasma-based AES. Butaciu et al.⁷² developed apparatus for ETV-capacitively coupled plasma microtorch AES. The relatively inexpensive miniaturised instrumentation included a low power (15 W) and low Ar consumption (150 mL min⁻¹) microplasma, connected to a small-sized Rh ETV device and a low resolution microspectrometer. The system was used to provide a procedure to measure concentrations of Cd in dietary products that would meet the requirements of EU Commission Regulations. Samples were microwave digested with HNO₃ and H₂O₂, and concentrations determined by standard additions calibration. The LOD in the original sample was 0.001–0.009 mg kg⁻¹, recovery of added Cd was 103 ± 15%. These and other figures of merit were similar to those given by an ETAAS method. A procedure to measure concentrations of SeMet and Se^IV in hydrolysed biofortified yeast was reported by Barrientos et al.⁷³ These workers used RP-HPLC with post-column HG and Se quantification by AES with a nitrogen-MIP. Freeze-dried sample was hydrolysed with methanesulfonic acid and HPLC separation performed with a mobile phase containing 0.08% v/v heptafluorobutyric and methanol (92 : 8), the column effluent was online mixed with K₂S₂O₈ 6% m/v, NaOH 3% m/v, passed through a reaction coil in a water bath at 60 °C, and then 10 M HCl was added prior to HG using NaBH₄. On-column LOQs were 59 ng Se mL⁻¹ for Se^IV and 0.52 μg mL⁻¹ for SeMet.

The use of LIBS has shown some increase in publications in comparison to previous years^1,74 and with an increase in scope of developments and applications. Zhong et al.⁷⁵ described the analysis of aqueous samples by LIBS through interfacing with ultrasonic nebulisation. The concept was to improve ablation efficiency by applying LIBS to the dense cloud of droplets formed from the nebulisation process. The apparatus was tested with Cu, Fe, Mg, Mn, Na, Pb and Zn and achieved good linearity and low LODs even with laser pulse energy as low as 30 mJ. The methodology certainly showed potential. Gaubeur et al.⁷⁶ reported the combination of DLLME and LIBS to detect Cr species in drinking water samples. This approach provided significant improvements in sensitivity and detection limits compared to LIBS alone. The complexation and extraction of Cr^VI was achieved with diethyldithiocarbamate in an acidic medium. The total Cr was also determined following complete oxidation, with the mass balance accounting for Cr^III. Whilst the work achieved an LOD and LOQ at relevant ranges for Cr species (3.1 and 10 μg kg⁻¹ respectively), the authors noted the limitations in precision with the current approach due to the drying of the extractant on an aluminium substrate prior to LIBS detection.

The potential of fsLIBS as a surgical diagnostic tool was described by Gill et al.⁷⁷ as a real time method to ensure complete removal of bone tumours. The study compared primary and secondary bone tumours with matched control/normal bone. Although no difference was observable between control bone and secondary tumours, a variation in the Mg/Ca ratio was found with primary bone tumours in comparison to normal bone. The authors suggested the technique could be applied in surgery to ensure total tumour removal thus minimising the need for follow up surgery. Jaswal and co-workers⁷⁸ presented the analysis of cholesterol gallstones and pigment gallstones using LIBS in comparison to WDXRF. The WDXRF enabled collection of both qualitative and quantitative data, but the LIBS approach was limited to qualitative only due to the challenges of calibration with such a heterogeneous sample matrix. However, the elemental ratios showed good correlations between LIBS and WDXRF. Some of the authors from this study also published a method implementing an optimised LIBS system to quantify Br in bread.⁷⁹ The excitation–acquisition delay, gate width and excitation source were investigated to maximise the sensitivity for Br, using the atomic transition line Br I at 827.2 nm. Spiked dried bread pellets were prepared as calibrants and four samples from a local market were assessed. The calibration standards and samples were also measured by ICP-MS to provide confirmation of the LIBS data with good correlations observed. An LOD of 5.09 ppm was achieved with an RSD of 2.8% at 20 ppm, indicating the potential of LIBS for Br detection with minimal sample preparation.

4.4 Vapour generation procedures and atomic fluorescence spectrometry

A review of advances in AFS reported during 2014 and 2015 was prepared by Butcher²² which referred to photochemical vaporisation, and novel atom cells. Applications of diffusive gradients in thin films for speciation were also mentioned but in the context of environmental analysis rather than foods and biological materials.

The interesting developments involving vapour generation of Hg and the elaborate system to generate Se hydrides from yeast were discussed in the preceding sections on AAS and AES.^48,73All other work of interest considered developments in the determination of As. Magnetic Fe₃O₄@ZIF-8 core–shell NPs used for the adsorption and preconcentration of iAs prior to HG-AFS⁴⁷ was described in the sample collection and preparation section. Variation in the efficiency of generating hydrides of different As species were shown to be abolished with post-column HG. This was the conclusion of Marschner et al.⁸⁰ from their work to measure concentrations of As^III, As^V, MMA and DMA in human urine. An FI manifold was prepared with a new design of gas–liquid separator that reduced fluctuations in the flow of hydride to the atomiser. With 100% HG of all species, it was possible to calibrate the assay with single species standardisation using DMA. It was also shown that ion exchange afforded better separation than did ion-pair chromatography and that this allowed analysis of undiluted urine. Analysis of the NIST urine SRM gave acceptable values. The LODs for As^III, As^V, MMA and DMA were reported to be 40, 97, 57 and 55 pg mL⁻¹, respectively.

In our recent ASU reviews we referred to regulations on the As concentrations in water, animal feed and now, in the EU, in rice. The implications of these regulations on analytical techniques were elaborated upon by Petursdottir et al.⁸¹ This is discussed in the section on As in foods (8.1.1).

4.5 X-ray spectrometry

A comprehensive review of recent advances in X-ray spectrometry⁴ accompanies this Update and includes applications with clinical and biological materials, foods and beverages. Imaging applications of XRF spectrometry are described separately in Section 6.2 below.

The use of X-ray spectrometry for biological studies was well represented over the past year. Analysis related to metal implants is also a regular feature in this review and two papers described the application of μXRF. Bilo et al.⁸² examined the effect of atomic layer deposition with TiO₂ as a coating mechanism for Co–Cr implants in comparison to bare metal. In the bare bone models, Co and Cr particles were found in the surrounding tissues after one month, with increasing levels after two months. Yet, with TiO₂ coating, a concentration gradient from the alloy surface to the tissue was observed but did not change over time. Golasik et al.⁸³ applied SR-μXRF to evaluate the release of ionic Ti from implants. This was emulated with rat models administered with ionic Ti (6 mg Ti per kg body weight). The liver, spleen and kidneys were quantitatively analysed for Ti and also Ca, Cu, Fe, K and Zn using NIST SRM 1831 and single element standards as the calibrants. Results were expressed as mass deposits per unit area (μg cm⁻²) and showed Ti mainly accumulated in the liver and was correlated to Ca levels, with the image maps showing the distribution was as expected for the administration route and time of exposure. The work provided some insight into the behaviour of ionic Ti in relation to metallic implants. Another example of SR-μXRF for biological samples was presented by Wandzilak et al.⁸⁴ who determined various elements including Ca, Cl, Cu, Fe, K, P, S and Zn in tumours in brain tissues in comparison to healthy tissues. The external calibration method was applied using NIST SRM 1832 and SRM 1833 standards with results expressed in μg cm⁻². It was found that there was a decrease in the concentrations of Ca, Fe, P and S whilst an increase in Zn in tumours with a high malignancy grade in comparison to control samples. High resolution distribution maps were also generated in which features such as blood vessels could be identified. With the aid of multivariate discriminant analysis, it was possible to model the elemental profiles and correctly classify the histopathological types in 99.93% of cases. The work added much needed information regarding the elemental changes that occur due to tumour growth in the brain. The issue of calibration for SR-XRF was discussed by Trunova et al.⁸⁵ and tested by comparing a biological sample (NCSZC 85005 Beef liver) using a geological material (IAEA Soil-7) as the reference sample. It was found that with fluorescence radiation absorption correction, for the different matrices, it was possible to achieve good recoveries for Br, Ca, Cu, Mn, Fe, K, Rb, Sr and Zn in the beef liver CRM. Typically, Se analysis in biological samples with XRF is challenging due to the low sensitivity and low occurrence. However, Li and Yu⁸⁶ have addressed this problem by using a high energy beam (100 kV, 600 W) for EDXRF with a 3D optics design. The LOD was found to be 0.1 μg g⁻¹ and samples up to 5.1 μg g⁻¹ were determined after fully optimising the system, offering a non-destructive, sensitive technique to investigate Se levels in biological matrices.

The role of portable XRF technology featured more heavily in comparison to previous years, highlighting the development and improvement of the technique. Of interest to those engaged with monitoring occupational exposure to lead is the onsite XRF analysis described by Gorce and Roff.⁸⁷ Inhalable dust and fumes were assessed through air sampler filters whilst surface wipes evaluated the removable Pb residues. The portable XRF data was compared against a laboratory based WDXRF system with good correlations observed between the techniques: R² of >0.99 for airborne measurements and 0.88 for the wipes. This demonstrated the effectiveness of the technique to rapidly determine occupational exposure of Pb onsite with fit for purpose accuracy and includes the potential to extend the scope of analysis to other elements. Fleming et al.⁸⁸ focussed the use of portable XRF on the analysis of As and Se for in vivo detection in soft tissue. Hamsters were used as models for animal tissue and were dosed with either As-only, Se-only, As and Se combined or control. The LODs were calculated using doped phantom calibrants, achieving 1.00 ± 0.05 ppm for As and 0.83 ± 0.02 ppm for Se. It was found that the liver, gall bladder and intestines had consistently higher levels of both As and Se with an As : Se ratio of approximately 2.5 : 1 observed regardless of the organ tested. The authors also discussed the merits of extending the analysis to in vivo human tissue analysis. Shehab and co-workers⁸⁹ also investigated As and Se using a portable device and K-shell XRF benchtop system to develop an in vivo method for the analysis of human skin. Nylon-backed polyester resins spiked with As and Se were used as skin phantoms (concentration levels confirmed by NAA). Once optimised, LODs were calculated for both systems; the portable XRF system obtained 0.59 ± 0.03 ppm As and 0.75 ± 0.02 ppm Se with a 1 minute measurement time whilst the benchtop system achieved 0.35 ± 0.01 ppm and 0.670 ± 0.004 ppm respectively with a 30 minute measurement time. When the effective does was also considered, the benchtop system was found to be superior for As but approximately equivalent (within the error) to the portable instrument for Se. This could enable in field in vivo determination of As and Se, but as the authors indicated, the relevance of this in a biomonitoring setting could have limited scope given the LODs and lack of knowledge regarding the distribution/heterogeneity of As and Se within the skin. McIver et al.⁹⁰ used portable XRF to analyse As in human finger nail samples as a rapid screening tool. The data were compared against ICP-MS and the agreement between the methods was acceptable when the sample mass was above 30 mg, however it was not often possible to obtain that quantity. Therefore the authors concluded that the portable XRF system had potential but was currently not a suitable replacement for ICP-MS analysis. The determination of As using portable XRF remained the focus in a study by Hagiwara et al.⁹¹ but in order to quantify As levels in drinking water. The workers developed a method to pre-concentrate 50 mL of water using a pre-treated SPE disc, which could then be measured directly by the portable XRF system. The approach was tested with different As species, with no effect observed, and compared against AAS, with good correlation found. The LOD was 2 μg L⁻¹ and had a linear range of 10 to 100 μg L⁻¹. With no power or gas requirements, it offered an opportunity to provide rapid onsite As detection for remote testing activities in rural locations.

Forensic applications of XRF spectrometry were highlighted in two historical cases based on the same 18^th century burial site in Portugal. The first instance⁹² applied μXRF to determine Pb concentrations in human remains of a male, approximately 30 years old. Bone and tooth samples were examined with both bulk levels and spatial profiles obtained. Certified reference materials of powdered bone, pressed into pellets, were used to determine accuracy with good recoveries obtained. Elemental maps elucidated the major locations of Pb accumulation: 120 ± 10 μg g⁻¹ in cortical bone, tibia and fibula; 250 ± 20 μg g⁻¹ in the trabecular region; highest levels presented in the ribs, 560 ± 30 μg g⁻¹. Within the tooth, Pb was found in the pulp and root at 130 ± 50 μg g⁻¹. However, the surrounding soil was found to contain around 50 μg g⁻¹, indicating ante mortem exposure. An additional component of the study compared the results obtained from the benchtop μXRF system with a second system equipped with a high powered X-ray tube with a water cooled tungsten anode and a changeable molybdenum secondary target. The X-ray tube, the secondary target and the sample were in a triaxial geometry. This setup provided lower LODs and improved precision over the benchtop equipment. The authors of this work published a second paper⁹³ focussed on the Hg content from a female approximately 50 years old as preliminary hair analysis indicated Hg was present at over 5% w/w. A similar instrument set up was reported in addition to the comparison between benchtop μXRF system and triaxial geometry XRF spectrometer. The workers measured hair, a molar tooth, cranium, ribs and a hand phalange, concluding that the elevated Hg exposure was likely through use of Hg-based compounds for medical purposes as the localised soil concentration was low.

A number of reports highlighted some interesting and different applications of XRF analysis. Bahadir et al.⁵³ described a method for Cr speciation by DLLME with TXRF. With the sample preparation achieving species separation and an optimised TXRF system, the workers were able to achieve an LOD of 0.8 μg L⁻¹ and good linearity across 3 orders of magnitude (5 to 4000 μg L⁻¹). Romero et al.⁹⁴ presented a method for Hg detection in waters utilising photochemical vapour generation and trapping with quartz substrates coated in PdNPs prior to TXRF detection. The optimised approach was shown to achieve good recoveries ranging from 96 to 100%, good precision (3% RSD at 1 μg L⁻¹, and an LOD of 54 ng L⁻¹). Borgese and colleagues⁹⁵ published a review on TXRF as a screening tool for food analysis. They examined food quality monitoring applications which included beverages, cereals, fruits, vegetables, animal derivatives and dietary supplements. The review also examined the outlook on the transfer of these research methods to routine testing laboratories. This year has seen the implementation of the ICH Q3D Guidelines for Elemental Impurities for new drug products which requires manufacturers to demonstrate that regulated elements are below set limits. Following this, existing products must meet compliance by December 2017, leaving a large amount of testing to be completed. The work by Figueiredo et al.⁹⁶ suggested that WDXRF could potentially fit this role. Following the validation requirements from the European Medical Agency and the United States Pharmacopeia, the LODs, precision and accuracy was assessed for As, Cd, Cr, Cu, Hg, Ir, Mn, Mo, Ni, Os, Pb, Pd, Pt, Rh, Ru and V. However, the acceptance criteria were only met for Cu, Cr, Ir, Mn, Mo, Ni, Os and Pt, perhaps highlighting the remaining challenges to use WDXRF as a compliance tool for all regulated elements.

5 Nanomaterials

From the detailed discussion in other sections of this Update it is evident that nanomaterials are involved in many of the developments associated with the topics considered in this review. Throughout this article we have described work in which a variety of nanomaterials are employed for a diverse range of applications involving the use of analytical atomic spectrometry. Reports of analyte pre-concentration included oxidised carbon nanotubes covered with layers of bovine serum albumin,⁴⁵ hydroxyapatite nanorods prepared from eggshell,⁴⁶ NPs with a magnetic core,^47,48 a quartz substrate coated with PdNPs⁹⁴ and AgNPs.⁵² Silver has antibacterial properties that are exploited in many situations including incorporation of AgNPs in food packaging. Transfer of ionic Ag and AgNPs from wrapping materials, cartons and beverage containers, including babies plastic bottles, into contents was shown to readily occur.^66–69 Various medical devices, such as urinary catheters, have also been prepared with incorporation of Ag into the material or with AgNPs as a surface coating, to prevent infection. The amount of Ag released was shown to vary considerably, but enough to raise safety concerns in some situations.⁹⁷ Nanoparticles as vehicles for drug delivery is a further clinical area of interest. Recent work with AgNPs and AuNPs reported on procedures to characterise pharmaceutical NPs and the distribution to different tissues and excretion.^97–99 A reversible inflammatory response, with accumulation of AgNPs in macrophages, was observed in another study.¹⁰⁰ Nanoparticles are also important in agriculture. Accumulation of AgNPs, predominantly at the middle lamella and cell walls in root tissue of Arapidopsis thaliana was seen, with some translocated toward the leaves. This finding may be of significance in plants where the leaves are consumed.⁶⁵

6 Applications: clinical and biological materials

Table 1 summarises developments associated with clinical and biological specimens and medicines.

Table 1 Clinical and biological materials, pharmaceuticals, traditional medicines and supplements

Element	Matrix	Technique	Sample treatment/comments	Reference
Ag	Tissues (liver, kidneys, spleen, stomach, small intestine)	AAS	Measurement of Ag in rat tissues following acute (2000 mg kg^−1) and sub-acute acute (200 mg kg^−1 for 30 days) exposure via the GI tract to 12 nm AgNPs. The highest Ag concentrations were detected in the liver (0.87 ± 0.37 μg g^−1) and kidneys (0.24 ± 0.02 μg g^−1)	99
Ag	Water; saline; plasma	ICP-MS	The total Ag and AgNPs released from medical devices (n = 5) into aqueous device extracts under various conditions (37 to 50 °C over 1 hour to 7 days) were evaluated using ICP-MS in conjunction with other NP characterisation techniques. Total Ag extracted from the devices ranged from 1 × 10^−1 to 1 × 10^6 ng cm^−2	97
Ag	Murine primary macrophages	ICP-MS; PIXE	A study to assess the effect of AgNP exposure (single high non-lethal dose or repeated low doses) on a model cell system. Using ICP-MS, PIXE and imaging techniques, Ag intracellular content and distribution were shown to be similar upon acute and repeated low dose exposure although acute exposure led to a stronger toxicological macrophage response	100
Al	EBC; urine	ICP-MS	Concentrations of Al and Be in EBC were evaluated as markers of occupational exposure. Both EBC Be and Al were higher in workers exposed to the respective elements compared with controls, however, EBC Be was more closely correlated with a cumulative Be exposure index than urine Be concentrations	134
As	Biological fluids (serum, blood and urine)	ICP-MS	Use of gas mixture CH3F/He enabled interference-free detection of As in samples with high Cl content. Reaction product, AsCH2^+, was monitored at m/z 89. The use of Te as an IS and addition of 3% v/v EtOH to all samples and calibration standards mitigated matrix effects. An LOD of 0.8 ng L^−1 was achieved	56
As	Urine; As species in frozen human urine (NIST 2669)	AEC-ICP-MS	The As species: As(III); As(V); MMA(V); DMA(V); AB and arsenocholine in diluted urine were separated in 9 minutes using a 2% MeOH-modified H2O : HNO3 gradient elution prior to ICP-MS detection with Ge as the IS. Species LOQs ranged from 0.3 to 0.9 μg L^−1 and RSDs (n = 11) from 8 to 11%. The method was validated using spiked urine (recoveries 94 to 103%) and a CRM before application to volunteers not exposed to As through diet	104
As	Cancer cells	HPLC-ICP-MS	A method to identify 13 common As metabolites with special attention to thiolated or thiol conjugated arsenicals was applied to the analysis of cancer cells, incubated with novel chemotherapeutic agent, darinaparsin. The As metabolite profiles were compared with results obtained using a previously developed method	105
As	Urine	LC-ICP-MS/MS	A high throughput method for quantitative determination of major Lewisite metabolites ((2-chlorovinyl)arsonic acid (CVAOA) and (2-chlorovinyl)arsenous acid (CVAA)) and qualitative determination of minor bis-metabolites. Following sample treatment with H2O2 to oxidise all CVAA to CVAOA, and LC separation, As was detected specifically using m/z transition 75 → 91. The method was linear (R^2 = 0.998) over concentration range 50 to 3500 μg L^−1, precision was 3.1 to 3.3% and the LOD was 2.2 μg L^−1 for 0.1 mL of urine. Chromatographic selectivity was verified against other As species and accuracy of the method confirmed through participation in PT schemes (biases −7.1 to +4.3%)	40
As	Soft tissue	XRF	Portable XRF was applied to in vivo tissue measurements of As and Se following dosing of hamsters (n = 7) with either As only, Se only or As and Se. The LOD for As, determined from phantom calibration trials, was 1.00 ± 0.05 ppm. Tissue concentrations ranged up to 26.4 ± 1.4 ppm for As and 11.8 ± 0.8 ppm for Se with highest concentrations detected in the liver, gall bladder and small intestine	88
As	Skin	XRF	Performance of an upgraded portable XRF system was compared with that of an advanced benchtop XRF system for in vivo measurement of As and Se in human skin. The LODs for As were 0.59 ± 0.03 and 0.35 ± 0.01 ppm for the portable device and benchtop analyser respectively. The performance of the portable device was concluded to be inferior for As analysis but equivalent for Se	89
Au	Tissues	ICP-MS	The biodistribution of 10 nm AuNPs in rats was compared after intervaginal space injection (ISI) in the tarsal tunnel and intravenous injection (IVI). Quantitative analysis of Au demonstrated ISI to result in similar AuNP content of the lung, heart and intestine and higher AuNP content of skin and muscle with respect to IVI	138
Au	Organs	ICP-OES	The clearance of 80 nm polymeric micelles loaded with 0.9 or 0.5 nm AuNPs, administered intravenously to mice, was studied by determination of organ Au content. Use of micelles containing the larger AuNPs resulted in a greater reduction in liver and spleen Au concentrations at 1 and 3 months post injection (72 and 67% versus 38 and 35% respectively) and increased faecal and urine Au concentrations (7.5 and 100-fold) one day post injection	98
Au	Organs (kidney, spleen and liver)	LA-ICP-MS	A quantitative imaging study in mice demonstrating that surface charge dictates the sub organ distributions of AuNPs in the kidney, liver and spleen following intravenous injection of functionalised AuNPs	114
B	Serum	ICP-MS	Serum B concentrations evaluated weekly after implantation of different ratios of nano-hexagonal boron nitride–hydroxy apatite composites in rat femurs (n = 120) were not found to be statistically higher than those of a control group (n = 6), supporting suitability of the composite for future biomaterial use	135
Be	EBC; urine	ICP-MS	See Al, ref. 134	134
Br	Plasma	HPLC-ICP-MS/MS	A method for separation and determination of drug-related Br and Cl in plasma. Following chromatographic separation of drug-related entities from inorganic chloride, ^35Cl was detected by ICP-QQQMS with H2 as a reaction gas, at m/z transition 35 → 37, while Br was monitored “on mass”. Repeatability was <10% RSD and recoveries were between 95 and 105% with a linear dynamic range up to 1 and 5 mg L^−1 for Br and Cl respectively. The LOQ for Br was 0.01 mg L^−1	58
Cl	Plasma	HPLC-ICP-MS/MS	The LOQ for Cl was 0.05 mg L^−1	58
Cl	Blood	IC-ICP-MS, IC-ICP-QQQMS	In a method for determining chlorate and chlorite in blood, the Cl species were separated by IC before detection by ICP-QMS or ICP-QQQMS with H2 gas. The LODs obtained with ICP-QMS (0.5 and 1.0 mg L^−1 for chlorite and chlorate respectively) showed no improvement with use of ICP-QQQMS due to a persistent Cl background	59
Cl	Sweat	ICP-MS	Development of a potential candidate reference method for sweat Cl for use in diagnosis of cystic fibrosis. Calibration standards were prepared from NaCl. Comparison of off-line (Ga) with online (Ge and Sc) internal standard methods by way of a PT scheme assessment indicated superior performance of the off-line IS (RSD 1.9%, bias 1.5 mmol L^−1versus RSD 8.0% and bias 3.8 mmol L^−1) although the difference in bias did not reach statistical significance. Linearity was determined up to 225 mmol L^−1 with an LOQ of 7.4 mmol L^−1	34
Co	Serum	AAS	A study to assess serum Cr, Co and Mo concentrations in patients with ceramic on metal hip arthroplasty. Serum Cr and Co were significantly higher at 3 years follow up than pre-implant surgery while Mo was not significantly different. When compared with MoM total hip arthroplasty patients, serum Cr levels were significantly lower while Co levels were lower but didn't reach statistical significance	125
Co	Simulated body fluids	ICP-MS	Metal ion release of a CoCrMo alloy subjected to different electrochemical and tribocorrosion conditions in NaCl, PBS and, PBS containing albumin, was investigated. Cobalt was detected in the highest concentrations under most conditions tested	123
Co	Biological fluids (blood, serum and synovial fluid)	ICP-MS	A study evaluating factors affecting clinical Co measurement: interference removal system (NH3 reaction gas versus He collision cell); sample preparation method (acid digestion versus alkaline dilution); matrix and storage conditions. The LODs observed ranged from 0.2 to 0.5 μg L^−1 in serum and synovial fluid and from 0.6 to 1.7 μg L^−1 in whole blood. While Co measurements were very comparable at Co concentrations >5 μg L^−1, at lower concentrations, alkaline dilution and He collision cell technology allowed superior polyatomic interference removal and increased sensitivity	55
Co	Bone tissue	μXRF	Metal migration into animal bone tissue from inserted bare and TiO2-coated Co–Cr alloys was evaluated using μXRF mapping. The use of a TiO2 coating was demonstrated to reduce diffusion of Co and Cr particles into biological tissues	82
Cr	Simulated body fluids	ICP-MS	See Co, ref. 123	123
Cr	Serum	AAS	See Co, ref. 125	125
Cr	Gingival crevicular fluid (GCF)	AAS	Elements Cr and Ni in GCF were assessed in 24 subjects at baseline, 1 month and 6 months post fitting of fixed orthodontic appliances. Mean Cr and Ni concentrations at baseline were reported as 1.978 ± 0.721 and 3.894 ± 1.442 μg g^−1 with Cr increasing by 200 and 700% at 1 and 6 months respectively while Ni increased by 150 and 510%	130
Cr	Blood	ET-AAS	A sample preparation method for speciation of Cr (III and VI) using DLLME before detection with ET-AAS was validated and applied to in vitro studies. The dispersant solvent, ionic liquid 1-hexyl-3-methylimidazolium hexafluorophosphate dissolved in acetone was rapidly injected by a syringe into blood samples containing Cr(III), which had previously been complexed by N-acetyl cysteine. The method provided a preconcentration factor of 10. Under optimal conditions, the linear range was 0.03 to 4.4 μg L^−1, RSDs were typically <5% and the LOD was 0.005 μg L^−1. The method was validated using an SRM	173
Cu	Liver	ICP-MS	Liver samples from cull cattle (n = 150) from a single abattoir were wet ashed and analysed for Cu. Approximately 40% of dairy cattle had liver Cu content greater than the Animal Health and Veterinary Laboratories Agency reference range (<8000 μmol kg^−1 dry matter) in contrast to 16.9% of beef cattle	174
Cu	Liver biopsy	LA-ICP-MS	Imaging analysis with a spatial resolution of 10 pm of paraffin-embedded liver needle biopsies for Cu and Fe from two patients with Wilson's disease and a negative control. Quantification used an external calibration based on matrix-matched gelatine standards and achieved an LOD for Cu of 1 μg g^−1. Good differentiation of the diseased samples from the control was observed through the inhomogeneous distribution of Cu and high Cu concentrations (up to 1200 μg L^−1)	112
Cu	Brain	SR-XRF	Elemental content of Cu, Fe and Zn in brain (substantia nigra pars compacta region) was examined in a Parkinson's disease mouse model and found to be high with respect to a control group. Rapamycin treatment reduced Cu, Fe and Zn content, which was associated with functional improvement	118
Cu	Urine	FAAS	A CPE sample preparation method for determination of Cu in urine by FAAS. Following reduction of Cu(II) to Cu(I) by ascorbic acid, reaction of 2,9-dimethyl-1,10-phenanthroline (neocuproine) with Cu(I) produced a hydrophobic chelate, which was extracted and preconcentrated using non-ionic TritonX-114. After optimisation of pH, reactant concentrations, equilibrium temperature and incubation time, an LOD for Cu of 0.01 μg L^−1 in 10 mL urine was achieved. The RSD was <3.5% for Cu concentrations from 0.5 to 100.0 μg L^−1. Analysis of an CRM yielded satisfactory results	175
Cu	Biological and food samples: whole blood (seronorm trace elements whole blood L-1 and L-3); urine (seronorm trace elements urine); tomato (NCS ZC85006)	FAAS	A novel SPE method using hydroxyapatite nanorods (HAPNRs) prepared from eggshell for subsequent determination of Cu, Zn and Pb by FAAS. Following optimisation, a preconcentration factor of 250 was achieved. Accuracy of the method was verified using the CRMs listed opposite. For Cu, the LOD and LOQ were 0.72 and 2.40 μg L^−1 respectively and linearity was verified up to 250 μg L^−1	46
Eu	Tissues (nose, extrapulmonary airway and lung)	ICP-MS	Inhaled particle deposition in developing rats was studied following aerosol exposure to 380 μg m^−3 of 30 nm Eu-doped Gd2O3 NPs. The deposited dose in the listed tissues was determined by ICP-MS and observed to be significantly greater in juvenile compared with adult and neonatal rats (2.22 versus 1.47 and 0.097 ng g^−1 body weight respectively). Toxicity data suggested the lanthanide particles were not acutely toxic or inflammatory	176
Fe	Liver biopsy	LA-ICP-MS	See Cu, ref. 112. The LOD for Fe was 5 μg g^−1	112
Fe	Brain	XRF	See Cu, ref. 118	118
Fe	Bovine ovaries	XRF	A baseline study of healthy ovaries (n = 19) in which histological sections (n = 40) were examined by XRF, following minimal sample reprocessing, to determine high resolution spatial distribution of elements Fe, Se and Zn and their localisation to specific ovarian structures	120
Fe	Blood	MC-ICP-MS	Direct analysis of acid-digested blood was compared with the standard protocol of prior chromatographic separation for isotopic Fe measurements by MC-ICP-MS. Influence of biological matrix elements was evaluated using synthetic solutions and bias of the Fe isotopic composition in the presence of Na and K, and high glucose concentrations, was limited by dilution of the whole blood digest to 0.75–1.5 mg L^−1 Fe. Instrumental mass discrimination was corrected for using a combination of internal (admixed Ni) and external standard correction. Good agreement with the reference procedure was demonstrated	62
Gd	Tissues (nose, extrapulmonary airway and lung)	ICP-MS	See Eu, ref. 176	176
Gd	Skin	ICP-MS; LA-ICP-MS; HILIC-ICP-MS	Assessment of Gd deposition in the skin of a patient with normal renal function after exposure to large cumulative doses of Gd-based contrast agents. Analysis by ICP-MS demonstrated high levels of Gd (14.5 ± 0.4 μg g^−1), while LA-ICP-MS analysis showed deposition to be in the deep layers of the skin. The presence of intact Gd-chelate species was confirmed by speciation analysis by HILIC-ICP-MS although most of the Gd present could not be further characterised	115
Hg	Human tissue and hair	CV-AAS	Analysis of post mortem samples from a corpse exhumed 6 years after death, revealed high tissue Hg concentrations (small intestine, 1516 ng g^−1; large intestine, 487 ng g^−1; liver, 1201 ng g^−1; heart, 1023 ng g^−1; and scalp hair, 227 ng g^−1) consistent with death due to Hg poisoning	141
Hg	Biological samples; sea water	CV-ETAAS	A new chelating sorbent, consisting of NPs functionalised with 1,5-bis(di-2-pyridil)methylene thiocarbohydrazide contained within a magnetic-based reactor, was used online for SPE prior to CV-ET-AAS. Following optimisation, the LOD was 7.8 ng L^−1 and precision was 1.7% RSD at 1 μg L^−1 Hg. A preconcentration factor of 5.4 was achieved. Results obtained from CRMs were in good agreement with certified values	48
Mg	Serum	ICP-MS	Validation of a candidate reference method for Mg and Na. Prior to ICP-MS analysis, samples spiked with Al internal standard were digested with 69% HNO3 at 150 °C. Measurements were calibrated by bracketing calibrators and serum Mg and Na concentrations calculated by comparing ion ratios of the serum samples to calibrators. The measurement RSDs were 0.07 to 0.14% for Na and 0.21 to 0.36% for Mg while analytical recoveries were 100.11 and 100.25% for Mg and Na respectively. Method accuracy was verified through analysis of CRMs and participation in inter-laboratory comparisons	177
Mo	Simulated body fluids	ICP-MS	See Co, ref. 123	123
Mo	Serum	AAS	See Co, ref. 125	125
Na	Serum	ICP-MS	See Mg, ref. 177	177
Ni	GCF	AAS	See Cr, ref. 130	130
Ni	Intracellular proteins	CF-GE-ICP-MS	An online CF-GE-ICP-MS method was used in conjunction with peptide mass fingerprinting to monitor intracellular metals, Ni and Zn, and their associated proteins. Correction for protein loss during the GE separation was achieved by simultaneous monitoring of, and correction, with ^34S	64
P	Brain and prostate tissue	LA-ICP-QQQMS	A LA-ICP-QQQMS method was evaluated in three tune modes (no gas, H2 and O2/H2 mixture) for application to elemental bio-imaging. When the method was applied to a prostate cancer biopsy sample using H2/O2 mode, high levels of Zn were identified in the area correlating with cancer. The O2 mass shifts, ^80Se → ^80Se^16O; ^31P → ^31P^16O, were used to image P and Se in mouse brain	111
Pb	Biological and food samples	FAAS	See Cu,^46 the sorption capacity of HAPNRs for Pb was 2.53 mmol g^−1, the LOD and LOQ were 5.12 and 17.06 μg L^−1 respectively, while linearity was verified up to 400 μg L^−1	46
Pb	Serum	TS-FF-AAS	Restricted access nanotubes (RACNTs) formed from oxidised carbon nanotubes covered with layers of bovine serum albumin, which extract Pb^2+ ions directly from untreated serum while excluding almost all serum proteins, were used for online SPE coupled to TS-FF-AAS. Under optimised conditions, the LOD was 2.1 μg L^−1, an enrichment factor of 5.5 was achieved, intra- and inter-day precision were <8% and recoveries of Pb from untreated whole blood ranged from 89.4 to 107.3%	45
Pb	Human remains: bone and teeth	μXRF; triaxial geometry	Quantitative evaluation of Pb in human remains by non-destructive μXRF was compared with a triaxial geometry set up. Accuracy was verified through analysis of bone RMs. Quantitative calculations were performed by fundamental parameter and compare mode methods. The LOD for μXRF was 3 μg g^−1 compared with 2 μg g^−1 for triaxial geometry. The method was applied to detect unusually high concentrations of Pb in ribs at 560 ± 30 μg g^−1 and trabecular bone at 250 ± 20 μg g^−1. Pulp and root in teeth contained up to 130 ± 50 μg g^−1 of Pb	92
Pb	Teeth (North East England)	LA-ICP-MS	Isotopic composition and concentration determination of Pb in deciduous teeth was used to provide a high resolution time frame of exposure to Pb during foetal development and early childhood in children (n = 10) born in 1997 and 2000 in North East England. Isotope measurements were acquired by LA-ICP-MS at contiguous 100 μm intervals across thin sections of teeth. Isotope data was integrated with histological analysis to assign estimated ages to each ablation point of the teeth samples	116
Pt	Cancer cells	ICP-MS	An ICP-MS method was used to study NP-mediated delivery of multinuclear Pt(IV) drugs into cancer cells	142
Pt	Cell lines; tissues	ICP-MS; LA-ICP-MS	The deposition behaviour of two chemically distinct Pt complexes was investigated in healthy and cancer cell lines, and in mice, and compared with cis-platin. Post intraperitoneal injection of 4 mg kg^−1 of the complexes, analysis of mouse organs showed that highest Pt accumulation was in the liver, kidney and lungs. Bio-imaging by LA-ICP-MS of paraffin-embedded liver and kidney sections with a spatial resolution of 50 μm showed differential distribution of Pt in kidney sections with the two complexes	113
S	Blood and serum	AEC-MC-ICP-MS	A method for S isotopic analysis involved AEC, with attention to removal of PO4, followed by MC-ICP-MS detection. Stoichiometric amounts of NH4 were added to improve SO4 transmission and stability through the desolvation process. Use of high resolution (R ∼ 9000) allowed the main spectral interferences to be resolved. On analysis of standard solutions, isotopic abundances were within errors of accepted values and reproducibility was ±0.15 and ±0.10‰ for δ^33S and δ^34S respectively. The final method was applied to patients with cancer and rheumatoid arthritis	61
Se	Brain and prostate tissue	LA-ICP-QQQMS	See Se, ref. 111	111
Se	Soft tissue	XRF	See As, ref. 88. The LOD for Se was 0.83 ± 0.02 ppm	88
Se	Bovine ovaries	XRF	See Fe, ref. 120	120
Se	Skin	XRF	See As, ref. 89. The reported LODs for Se were 0.75 ± 0.02 and 0.670 ± 0.004 ppm respectively for the portable and benchtop XRF analysers	89
Se	Biological samples	EDXRF	A high-energy linearly polarised beam EDXRF spectrometry system in tandem with a 3D optics design was applied to determination of Se concentrations (range 0.1 to 5.1 μg g^−1) in biological samples. Effects of applied voltage, acquisition time, secondary target and various filters were investigated. The LOD for Se was 0.1 μg g^−1	86
Se	Hepatoma HepG2 cells	HPLC-ICP-MS; ESI-MS; ESI-Q-TOF-MS	An unknown Se metabolite was detected in the cytosolic fraction by HPLC-ICP-MS when human hepatoma HepG2 cells were exposed to sodium selenite. This was identified as selenocyanate by ESI-MS and ESI-Q-TOF-MS. The selenocyanate was less toxic to HepG2 cells than selenite or cyanide, suggesting it was formed to reduce the toxicity of selenite	102
Se	Plasma and urine	LC-ICP-MS; ICP-MS	Changes in blood plasma Se concentration and 24 hour urinary Se metabolite excretion in healthy subjects (3 TMSe eliminators and 4 non-TMSe eliminators) after oral administration of sodium selenate (50 μg Se) were assessed using ICP-MS and LC-ICP-MS. While the main elimination product in urine of all subjects was un-metabolised selenate, TMSe eliminators showed lower selenate excretion with a corresponding increase in TMSe excretion	101
Se	Plasma	LC-ICP-MS	A method for detection of low molecular weight Se metabolites in plasma consisted of initial ultrafiltration using a 3000 Da cut off and then post column isotope dilution-based quantitation. The method was validated using standards of Se species in aqueous solution and plasma before application to cancer patients receiving selenite treatment. Of three low molecular weight Se compounds observed, two were verified by other methods	178
Ti	Organs (liver, spleen and kidney)	SR-μXRF	A study assessing Ti distribution in the liver, spleen and kidneys of rats following administration of ionic Ti (6 mg kg^−1) either as a single intravenous dose or via 30 days oral administration. The Ti was mainly retained in the kidneys, while a non-uniform distribution of Ti in the liver was observed, with formation of Ca enriched single aggregates	83
Ti	Blood and tissues	ICP-MS	An assessment of Ti release into blood and surrounding tissue from implanted Ti alloy growth guidance sliding devices and its clinical consequences in growing children (25 with implants, 13 without). Measurement of Ti and V by ICP-MS showed higher blood concentrations (2.8 and 4 fold for Ti and V respectively) in 90% of study subjects with respect to controls, while median content of Ti in soft tissues adjacent to the implant was increased by more than 1500 fold. Metallosis-associated complications occurred in 20% of the study subjects	128
Tl	Urine; hair; tap water (Tuscany, Italy)	ICP-MS	Quantitation of Tl in urine (n = 150) and hair (n = 318) in residents of an area of Tuscany in which Tl contamination was identified in tap water (2 to 10 μg L^−1). Concentrations of up to 498 ng g^−1 in hair and up to 5.44 μg L^−1 in urine were detected	144
U	Bone	SR-μXRF	A study of the initial distribution of U within cortical and trabecular bone in contaminated rat femurs. The U was found to be preferentially located in calcifying zones	121
V	Blood and tissues	ICP-MS	See Ti, ref. 128	128
Zn	Brain and prostate tissue	LA-ICP-QQQMS	See Se, ref. 111	111
Zn	Bovine ovaries	XRF	See Fe, ref. 120	120
Zn	Brain	XRF	See Cu, ref. 118	118
Zn	Biological and food samples	FAAS	See Cu, ref. 46 sorption capacity of HAPNRs for Zn was 2.37 mmol g^−1, the LOD and LOQ were 0.55 and 1.83 μg L^−1 respectively, while linearity was established up to 300 μg L^−1	46
Zn	Intracellular proteins	CF-GE-ICP-MS	See Ni, ref. 64	64
Various (25)	Blood and plasma (France)	ICP-MS	A study to elucidate reference ranges in healthy children (n = 99) in France. An ICP-MS method used 0.3 mL of sample to quantitate 25 elements in whole blood and 23 in plasma	25
Various (17)	Urine	ICP-MS	Evaluation of the use of systematic creatinine adjustment in random urine spot samples to predict 24 hour urine excretion for 17 elements in non-occupationally exposed subjects (n = 39). Upon comparison of four models, the ratio with specific gravity was found to be a more reliable alternative to creatinine correction with no over adjustment except for V. For Cd, Cr, Cu, Ni and Te none of the models accurately reflected 24 hour urine excretion	41
Various (16)	Pharmaceuticals; dietary supplements	WDXRF	An assessment of the feasibility of measuring 16 elements by a WDXRF method in final drug products. Figures of merit were: linearity correlation coefficients >0.9; LODs and LOQs of 0.6 to 5.4 and 1.7 to 16.4 μg L^−1 respectively; spiked recoveries of 70 to 150% and RSDs <20%. European Pharmacopeia requirements were met for Cu, Cr, Ir, Mn, Mo, Ni, Os and Pt	96
Various (8)	Brain tissue	μXRF	Concentrations and spatial resolution of Ca, Cl, Cu, Fe, K, P, S and Zn obtained by μXRF were compared in neoplastic and healthy brain tissue and correlated with malignancy grade. A decrease in the average concentration of Ca, Fe, P and S but an increase in Zn was observed in tissues with a high grade of malignancy. In new malignancy cases, histopathological types were correctly classified in 99.16% of cases	84
Various (8)	Cells	SR-XRF	Imaging of elements Ca, Cd, Cl, Co, Cu, K, Ni and Zn in single organelles and other subcellular features using SR-μXRF and nano-SR-XRF with both micro and nanoprobes in combination with TEM	122
Various (6)	Single cells; brain tissue	PIXE; SR-μXRF	Four sample preparation methods for single cells for subsequent imaging of Cl, K, Mg, Na, S and Zn were compared. The protocols used a combination of rinsing solutions (PBS or ammonium acetate) and fixation methods (cryofixation or chemical fixation with 3% paraformaldehyde or MeOH). Chemically fixed cells showed a disturbance of elemental distributions and cellular ion ratios. The optimal sample preparation method used ammonium acetate rinsing then rapid freezing in chilled cryogenic fluid followed by freeze-drying at low temperatures. The protocol was validated on rat primary hippocampal neurons	44
Various (5)	Skin	SR-μXRF	Histological skin sections from a metal wire worker with asymptomatic grey pigmentation were investigated via 2D SR-μXRF. The deposits were identified as Cu, Ni, Sr and Zn, which were strongly associated with maximal S concentrations	133

---

6.1 Metallomics

A comprehensive review of recent advances in speciation and related applications⁶ accompanies this Update and includes applications with clinical and biological materials, foods and beverages. A detailed review of As speciation was mentioned in Section 1.

Much of the biochemistry of essential trace elements is directed or controlled by binding to proteins and methods to determine the proteins involved and their binding properties are required to understand mechanisms of how trace elements are biologically active within cells. Wang et al.⁶⁴ used GE-ICP-MS to investigate metal binding to two proteins, Helicobacter pylori metallochaperones HypA and HspA, in E. coli cells. Metals and S measured at the same time showed binding of Co, Cu, Ni and Zn to these proteins and that the properties were perturbed by Bi. Of interest was the observation that binding in vivo was not necessarily the same as found in in vitro experiments.

Two aspects of Se metabolism to be mentioned were reported during this review period. A study reported by Jager et al.¹⁰¹ used HPLC-ICP-MS to identify and measure concentrations of Se metabolites in human urine. Following intake of Na₂SeO₄, approximately 30% of the dose was excreted in the urine, largely as unmetabolised SeO₄²⁻, with most eliminated during the first 2 h. Of the seven healthy subjects who participated, three also excreted TMSe⁺ and one excreted two different seleno-sugars. The results suggest considerable variation in the normal metabolism of inorganic Se compounds. Anan et al.¹⁰² found a previously unidentified metabolite in the cytosolic fraction of human hepatoma cells exposed to Se as Na₂SeO₃. Using ESI-MS and ESI-Q-TOF-MS the metabolite was shown to be selenocyanate. This species was also formed in vitro by mixing cell homogenate, Na₂SeO₃ and GSH suggesting a reaction between Se²⁻ and endogenous cyanide. When administered to rats, the selenocyanate became incorporated into selenoproteins and selenometabolites. The authors suggested that formation of selenocyanate may act to reduce the toxicity of selenite.

Methods to separate the main As species of biological interest are now well established. In our 2015 ASU we reported the work of Leese et al.¹⁰³ who measured the concentrations of As species in urine of subjects without occupational exposure. A similar study has now been carried out in Spain by Moreda-Piñeiro et al.¹⁰⁴ with urine samples from individuals not chronically exposed to As via water and food. Stice et al.¹⁰⁵ developed an HPLC-ICP-MS method that extends the number of identifiable As metabolites to 13. These included several thiolated and thiol-conjugated arsenicals. The procedure was applied to analysis of cancer cells incubated with dimethylarsino-glutathione which, as darinaparsin, has been developed to treated refractory malignancies.

Concentrations of Cl are regularly determined in serum or plasma in clinical chemistry departments using ion specific electrodes. Schwan et al.⁵⁹ established an assay for Cl speciation in blood samples with IC separation and detection by ICP-MS. Using a 20 μL injection volume the LODs were 0.5 and 1.0 mg Cl per L for chlorite and chlorate, respectively. The assay was validated and applied to the monitoring of concentrations of chlorite and chlorate in blood samples from farm animals following use of disinfectants with chlorate salts.

Sunflower oil enriched with selenite produces a mixture of selenitriglycerides which have anti-oxidant and anti-cancer properties. Using HPLC-ICP-MS and HPLC with HR, high mass accuracy-ESI-MS, Bierla and colleagues⁶³ developed procedures for separation and identification of these selenospecies. In an unusual approach, a non-aqueous mobile phase gradient was successfully employed. Samples were prepared by dissolving sunflower oil in isopropanol or as a methanolic extract. The ICP-MS showed 14 selenospecies and the ESI-MS information indicated 11 Se-triglycerol derivatives suggested to be formed from the oxidation of linoleic acid and mixed lineleates/oleinates.

Multielement speciation was developed by Marcinkowska et al.¹⁰⁶ to determine concentrations of the toxic agents As^III, As^V, Cr^VI, Sb^III and Sb^V in drinking water. Good separation was achieved by anion-exchange HPLC and detection by ICP-MS with O₂ as the reaction gas, all within in 15 min. The optimised, validated method was precise (1–7 to 2.4% RSD) and sensitive with LODs of 0.067, 0.068, 0.098, 0.083 and 0.038 μg L⁻¹ for As^III, As^V, Cr^VI, Sb^III and Sb^V, respectively. Recoveries of added species were between 91 and 110%.

6.2 Imaging with LA-ICP-MS and XRF

A number of review articles were published summarising the benefits and challenges of MS imaging capability for biological and environmental research. Boughton et al.¹⁰ provided a comprehensive tutorial and review of MS-based imaging covering both elemental and organic techniques for plant biology. The authors described the advances and evolution of instrumentation over the last decade and the benefits it has brought to this diverse area of research. Hare et al.⁸ published a tutorial paper on the state of the art imaging technology for biological applications which covered MS, XFM and fluorescent probes. Numerous examples were used to explain the approaches with advantages and limitations also covered. The authors concluded that sensitivity, selectivity and spatial resolution were the key properties to visualise the role of elements in biological samples. Van Malderen et al.¹⁰⁷ produced a thorough evaluation and summary of recent developments with LA-ICP-MS, highlighting the technical improvements that have arisen through deeper understanding of the ablation process. As transport efficiency, speed and resolution have improved, LA-ICP-MS has opened new opportunities to investigate applications such as nanoparticle analysis, single cell detection, cell biomarker assays using multiplexed metal-conjugated antibody tagging and 3D imaging. The paper concluded that significant gains have been made but yet further work is still required. In particular, the mass spectrometer can be a limiting factor when considering multielemental detection for these fast ablation events. In another tutorial review,¹⁰⁸ the practical use of MatLab software for LA-ICP-MS imaging applications was presented, using biological tissue samples as an example. It provided a useful guide for users on the important parameters for data treatment and included a significant electronic supplement with step by step guides. Limbeck et al.¹¹ discussed the role of LA-ICP-MS in the life sciences and environmental chemistry. The review focussed on the challenges of calibration and critically evaluated the approaches currently reported in a variety of matrices ranging from feathers, fish scales, tissues and protein gels, to powdered samples. Although it mainly considered quantification techniques for bulk samples, many of the same issues exist for imaging. Continuing on this theme, Thieleke and Vogt¹⁰⁹ described a novel calibration strategy for LA-ICP-MS utilising ID for the characterisation of solid reference materials. The approach involved the preparation of thin polymeric films containing the enriched isotope spike which was placed under the sample, thus enabling the sample and spike to be simultaneously ablated. Lead was used as an example. A polymeric film doped with ²⁰⁴Pb was produced for the enriched spike layer whilst an in-house prepared sample containing natural Pb and the CRM BAM-H010 were used for development and validation of the methodology. The authors also applied both single and double ID calculations, concluding an equal performance in terms of accuracy (obtaining recoveries within 1% of the CRM value) and expanded uncertainties of 11% and 8% respectively. However, extra analysis time was required to ensure complete ablation of all layers and the ID equations relied upon a uniform layer thickness which can be variable in real biological tissue samples.

A novel development for LA-ICP-MS describing a new two-volume laser ablation cell and integrated ICP torch was reported by Douglas et al.¹¹⁰ which was specifically designed to improve sample transport efficiency and minimise washout times. The paper provided the technical design details and performance characteristics but in brief, using NIST SRM 612, it represented a 300 to 500-fold increase in peak maximum and an 8 to 14-fold increase in peak area in comparison to the standard single volume cell, without affecting the background levels. Washout times were <5 ms which was a considerable improvement over regular systems which are typically in the range of 0.1 to 1.5 s. Furthermore, an application of the new design was demonstrated with Gd-labelled human T cells for single cell analysis. The new LA cell generated peak widths of <10 ms at ‘full width 10% maximum area’ and improved sensitivity up to 6-fold compared to other ablation cells. The prototype has since been further refined and developed into a commercial product.

The application of LA-ICP-MS for biological imaging was predominantly featured during the past year. Bishop et al.¹¹¹ reported the development of an ICP-QQQ-MS method with LA to determine multiple elements with challenging interferences. Three tune modes were first optimised in solution to compare the performance of Cu, Fe, Mn, P, Se and Zn in either no-gas mode, H₂ mode or a mixed O₂/H₂ mode. Elements such as P and Se are known to readily react with O₂ and can be mass shifted by +16, effectively moving the analyte away from the interference and detected off mass. Yet other elements such as Fe and Zn can be determined on mass within the same tune mode. The ICP-MS parameters were refined further with LA using NIST SRM 612 and in-house prepared tissue calibration standards. Examples of bioimages were then presented using the mixed O₂/H₂ mode for a prostate cancer biopsy (quantitative Zn analysis, on mass) and mouse brain tissue slice (Se and P qualitative images only, off mass). Whilst the work presented the first use of LA-ICP-QQQ-MS for on and off mass detection for biologically relevant elements with spectral interferences, the improvements in LODs were actually minimal in comparison to H₂ alone. Wilson's disease was the focus of a paper by Hachmöller and co-workers¹¹² who applied LA-ICP-MS to liver biopsy samples to investigate the distribution of Cu and Fe in patients with the condition. Spiked gelatine standards were used for calibration, achieving LODs of 1 μg g⁻¹ for Cu and 5 μg g⁻¹ for Fe with a spatial resolution of 10 μm which enabled small needle biopsy samples to be analysed. The results showed an inhomogeneous distribution of Cu with concentrations reaching 1200 μg g⁻¹ in the Wilson's disease cases. It was also found that high levels of Fe were inversely correlated to Cu. Lum et al.¹¹³ described the development of two potential Pt-based anti-cancer drugs and combined LA-ICP-MS analysis with an in vitro cell line study and quantitative ICP-MS measurements. The new compounds were tested in comparison to cis-platin as the reference, using mouse models to investigate bioavailability. The ICP-MS data showed the highest levels accumulated in the liver and kidney, with the least in the heart and brain. Imaging with LA-ICP-MS was performed on paraffin embedded sections of liver and kidney to determine the Pt distribution within the tissues. Based on the results, it was possible to differentiate between the two new Pt-containing drugs despite the structural similarities of the complexes.

A study on the effect of functionalised Au nanoparticles was reported by Elci et al.¹¹⁴ to investigate the influence of surface chemistry and biodistribution. The Au nanoparticles had a 2 nm core and were treated with pentanethiol and tetra(ethylene glycol) with four different end groups: two were positively charged but with different hydrophobicity, a negatively charged group and a neutral group. The nanoparticles were then administered to four groups of mice plus a control. The organs were analysed for total Au content by ICP-MS and overall it was found that kidney, liver, lung and spleen had the highest Au levels, with the positively charged particles accumulating to a much greater degree compared to the neutral or negative particles. Quantitative images were produced for sections of these tissues with calibration achieved via Au nanoparticles spiked homogenised beef liver and chicken breast (to represent spleen). The LA method used a 50 μm spot size which enabled the researchers to locate the specific regions of accumulation within the tissues and to further understand the in vivo biochemistry for different nanoparticle surface functionalisation.

An unusual clinical investigation was published by Roberts et al.¹¹⁵ following a patient with significant exposure to gadolinium-based contrast agents due to 61 MRI brain scans over an 11 year period. A combination of analytical techniques was used including ICP-MS for total Gd determination in skin biopsies and LA-ICP-MS to investigate the distribution of Gd within the skin layers. A high concentration of Gd was found in the skin (14.5 ± 0.4 μg g⁻¹) whilst the imaging data revealed Gd deposits within the deep layers of the subcutis that was contradictory to previously reported observations. Another interesting application of LA-ICP-MS, although not strictly imaging, was a pilot study to evaluate the potential of using Pb isotope ratios in deciduous teeth as a method to track childhood Pb exposure.¹¹⁶ Tooth samples from children aged 5–8 years old (n = 10) living in northern England were collected, thinly sectioned then isotope patterns from across the whole tooth (from enamel to pulp) were measured. In combination with environmental isotope data (e.g. airborne Pb sources such as PM10, PM2.5), it was possible to observe differences in children born after the year 2000, following the ban of leaded petrol, and those born 1997.

The complementary use of LA-ICP-MS and LIBS as a tool to achieve complete elemental mapping capability in biological samples was highlighted by Bonta and co-workers.¹¹⁷ A tandem LIBS/LA system was used to enable simultaneous LIBS analysis for C, H, K, Na and O, during LA. With online coupling to ICP-MS, trace elements (namely Au, Cu, Fe, P, Pt and Zn) could be detected simultaneously to complement the LIBS data. The instrumental parameters were optimised and elemental maps were generated for a human tumour slice, previously exposed to cis-platin. The approach showed promise to determine the location of major, minor and trace elements within one measurement session.

During the past year, the interest in XRF for biological imaging applications continued and received significant attention, covering a number of important topics ranging from the role of essential elements in the pathogenesis of diseases to single cell analysis. Tian et al.¹¹⁸ studied the distribution of Cu, Fe and Zn in mouse brain models induced with Parkinson's disease and treatment with Rapymycin, a neuron protecting drug. The substantia nigra pars compacta region of the brain was quantitatively imaged, with NIST SRM 1577a bovine liver and NIST SRM 612 glass used as calibrants, and included Compton scattering normalisation. Combining this data with behavioural tests and microscopy analysis provided evidence that the drug treatment reduced the levels of Cu, Fe and Zn whilst dopamine neurons and motor functions recovered. The researchers concluded that the oxidative properties these elements have within the body could play a role in Parkinson's onset and progression and that Rapymycin could have protective activity through metal homeostasis. The use of μXRF in combination with XANES was described by Hachmöller and colleagues¹¹⁹ to investigate elemental distributions in liver samples of patients with Wilson's disease, which was a complementary study to the LA-ICP-MS method cited above.¹¹² In this work, Cu, Fe, Mn and Zn were analysed in a needle biopsy sample of liver from a Wilson's disease sufferer. A benchtop μXRF system acted as a screening method to identify the regions of interest for further investigation by SR-μXRF which enabled higher resolution images to be generated (4 μm). The Cu and Fe trends matched that of the previously reported LA-ICP-MS results¹¹² where Cu was inversely proportional to Fe; in this work, Zn distribution followed the pattern of Fe whereas Mn was evenly dispersed. As a further development, the workers also used XANES for Cu speciation as currently little is known about Cu oxidation states in Wilson's disease. The data showed the presence of both Cu(I) and Cu(II) however the authors concluded this could only be considered indicative and that further work on this area to elucidate protein and redox behaviour should be undertaken. Ceko et al.¹²⁰ reported the determination of Fe, Se and Zn in developing bovine ovaries through imaging with SR-μXRF. The high resolution of the technique enabled spatial distributions of the elements to be clearly identified within specific organ components. It was also observed that Fe and Zn were co-localised whilst Se was found infrequently around the antrum of large follicles. The study was based on a healthy population but does add baseline data for the elemental distribution within developing ovaries. Bourgeois et al.¹²¹ applied SR-μXRF to bones from rats exposed to U, to further understand the accumulation in vivo. The high spatial resolution of this technique enabled the researchers to evaluate the distribution within specific regions. It was found that intravenous U was located in calcifying zones and rapidly collected in femoral metaphyses, in calcifying cartilage and in recently formed bone tissue along trabecular bone. The researchers concluded that the results show the effect of bio-mineralisation rather than passive absorption.

As developments in XR beams continue, the ability to analyse single cells offered an opportunity to understand elemental profiles within the cell. To achieve this, the cell must be well preserved with minimal changes from the sample preparation process. Perrin et al.⁴⁴ investigated four different approaches widely used as fixation techniques. Both μPIXE (spot size 0.8 μm) and SR-μXRF (beam size 21 × 38 nm²) were used, enabling a very high degree of spatial resolution. The optimal method was cell rinsing with 150 mM ammonium acetate at pH 7.4 followed by cryofixation then freeze dried at low temperature. This approach minimised losses from diffusible elements such as Mg and K whilst maintaining structural integrity. Kashiv and co-workers¹²² described their efforts to image single cells and subcellular features by combining TEM and SR-XRF equipped with both a microprobe and nanoprobe. Insulin producing mouse pancreatic beta cell lines were used as the models and were cryofixed and freeze substituted. The high spatial resolution of TEM for structural analysis provided a means to select suitable cells for SR-XRF analysis. The microprobe had a beam size of 150 × 150 nm² with sufficient sensitivity to detect Ca, Cd, Cl, Co, Cu, K, Ni and Zn (Cd was intentionally added to the cells to investigate the impact of a non-essential element). This provided a relatively quick way to establish regions of interest which were then analysed with the nanoprobe (40 × 40 nm² beam spot size). The combination of data highlighted that areas rich in protein, such as the nucleus and mitochondria, were predominantly concentrated in Cd, Co, Cu, Zn, and potentially Ni, reflecting their affinity to protein binding. Conversely, Cl and K were mainly present in the free cell solution. The analytical capability of nanoprobe SR-XRF to determine elemental distributions at the subcellular level was clearly demonstrated and has the potential to provide deeper insight into the role of elements within the cell.

6.3 Multielement applications

6.3.1 Specimens analysed to investigate metallic implants and biomaterials. Kerger et al.⁵⁵ investigated variables that may influence the measurement of Co in serum, blood and synovial fluids. Beginning with the instrumentation it was found that Agilent and Perkin Elmer ICP-mass spectrometers provided similar LODs (<0.2 μg L⁻¹) but that there were more polyatomic interferences with Perkin Elmer equipment when concentrations were near the detection limit. Equivalent results were obtained when samples were prepared by either digestion or dilution. It was concluded that ICP-MS methods are appropriate for analysis of specimens for clinical purposes but with some slight reservations associated with low concentration samples.

Release of metals from metallic hip prostheses continues to be investigated. Mechanisms involved in the degradation of implants are complex, as is evident from the range of clinical responses with some patients having no complications and others requiring revision of the implant. Corrosion of the alloy and wear of the implant are thought to be involved in degradation and Espallargas et al.¹²³ conducted an extremely thorough investigation of metal ion release from a low carbon Co/Cr/Mo alloy widely used in biomedical implants. Highly polished discs of the metal were incubated in electrolyte solutions and human plasma simulant at 37 °C and subjected to a range of electrochemical and mechanical (tribo)corrosion conditions. Metal ion release was determined using ICP-MS. Cobalt was the principal metal released under both electrochemical and tribo-corrosion conditions. Addition of albumin dramatically increased Mo release under electrochemical conditions and increased both Cr and Mo release under tribo-corrosion conditions. It was hypothesised that, under tribo-corrosion conditions, metal degradation was due to depassivation (primarily Co dissolution) and mechanical detachment of wear particles. They concluded that in vitro testing of implants with a combination of electrochemical and tribo-corrosion conditions, with measurement of metal release using ICP-MS, is a powerful tool for assessing the bio-corrosion properties of metals used in biomedical appliances. Using XRF spectrometry, it was shown by Bilo et al.⁸² that a layer of TiO₂ on the surface of Co–Cr alloy implants, inserted into animal bone, limited the migration of metal particles into surrounding tissues. Any possible negative effect from the inclusion of Ti was, however, not considered!

The causes and mechanisms of formation of pseudotumours in patients with metal-on-metal hip replacements, remain ambiguous. Catelas et al.¹²⁴ proposed that blood lymphocyte subpopulations may help to identify reactions occurring in different patient groups. The lymphocyte populations were compared in patients with failed hip implants, with or without a pseudotumour, and patients with well-functioning hip implants. Serum concentrations of Co and Cr were determined using ICP-MS and peripheral blood populations of T and B lymphocytes analysed by flow cytometry after immuno-staining. Of the various T and B cells monitored, the percentage of memory T helper, memory cytotoxic T, memory B and type 1 T helper cells were all significantly reduced in the blood of patients with failed implants and pseudotumours. At the same time, the serum Co concentrations of these patients (6.8 μg L⁻¹) were higher than in those without pseudotumours (0.9 μg L⁻¹). These results imply a type IV hypersensitivity reaction in patients with pseudotumours. The results do not, however, demonstrate whether Co has a causal role in the generation of pseudotumours.

Concentrations of Co, Cr and Mo were measured in the serum of patients with ceramic-on-metal and metal-on-metal arthroplasty 3 years following the operation, by Cadossi et al.¹²⁵ The Cr and Co concentrations increased compared with pre-operative levels but were significantly lower in the ceramic, compared with the metal, group of patients. There were no changes in the Mo levels. Functional outcomes were excellent in the ceramic-on-metal group of patients. Sakamoto et al.¹²⁶ described the experience of a woman given a polyethylene-on-metal total hip arthroplasty. In the course of the operation, the greater trochanter was broken which was fixed using a trochanteric cable grip reattachment. Six years later she had severe pain in the hip and investigations showed a mass, presumed to be a pseudotumour, severe loss of bone and a broken cable. Serum concentrations of Co and Cr were normal but the Ti was increased. The mass was removed during surgery, as was the broken cable. Analysis of the pseudotumour identified Ti-particles with the likely source being, the titanium femoral stem as the cable was made from Co–Cr alloy. This observation is consistent with other studies where increased serum Ti concentrations were seen from patients with total hip replacement using Ti femoral stems.¹²⁷

Metallic devices are used for various other clinical conditions and investigations of metal concentrations in tissues or biological fluids are occasionally reported. Lukina¹²⁸ and colleagues from three countries, collaborated to assess the release of metals from growth guidance sliding devices. These implants are used to treat children with scoliosis as they allow rods to slide during growth. The study involved 25 children, 13 of whom received the growth guidance sliding device made from a Ti₆Al₄V alloy. Approximately 6 years later, blood Ti and V concentrations were increased in the study group (by 2.8 and 4 times, respectively). Five patients had metallosis-related complications and revision operations were carried out in two subjects. Analysis of the removed soft tissue showed extremely high concentrations of Ti.

Dental devices are among the most widely used items likely to release metals to surrounding tissues. Adverse effects have been reported and attributed to Ni from orthodontic appliances – Ni brackets, molar bands and Ni–Ti alloy wires. Release of Ni from self-ligating orthodontic appliances into saliva was assessed by Golz et al.¹²⁹ Thirty young patients (10–13 years) fitted with self-ligating brackets, bands or arch wires were monitored. Saliva samples were collected at defined time points before and after fitting of the appliance with Ni concentrations determined using ICP-MS. The baseline median Ni concentration prior to treatment was 21.85 μg L⁻¹ and levels between 13.7 and 86.3 μg L⁻¹ were recorded at follow up time points. The authors reported significant increases in salivary Ni in the initial month after fitting of all appliance types, but noted that levels subsequently returned to baseline values by 8 weeks. They concluded that self-fixing appliances released Ni in the short term at levels similar to levels previously documented for conventional orthodontic appliances. A similar investigation was conducted by Amini et al.¹³⁰ but these workers analysed gingival crevicular fluid, a sample type not previously used and for which trace element concentrations are unknown. Specimens were collected before and at 1 and 6 months after fixing orthodontic devices to 24 subjects. The gingival index increased with time, indicating inflammation of the gums while Cr and Ni concentrations were also increased at 1 and 6 months by 200 and 700% (Cr) and by 150 and 510% (Ni). Simsek et al.¹³¹ implanted metal-containing endodontic repair materials into subcutaneous tissue of 15 rats. After 45 days, tissues were removed from these and control animals. Of the seven metals determined, Cr was increased in the brain and kidney and Mg was increased in kidney and liver, of the test animals. However, the concentrations were deemed not to be at harmful levels and the three materials tested were inferred to be non-toxic.

6.3.2 Biological fluids. There is little to report from the period covered by this review. The novel work involving gingival crevicular fluid was described in the previous section.¹³⁰ Basal concentrations of Cr and Ni were 1.98 ± 0.72 and 3.84 ± 1.44 μg g⁻¹, respectively.¹³⁰ Isotopic ratio measurements for Ca, Cu, Fe and Zn in various biological fluids¹³ were cited in the earlier section on reviews. A technique to measure concentrations of Br and Cl in plasma, for the purposes of drug studies,⁵⁸ was mentioned in the section on ICP-MS (4.1).

Whether urine collected over 24 h is preferable to a single spot-sample was discussed in the section on Sample collection. Wang et al.¹³² considered whether or not any urine collection may be indicative of exposure. These workers noted that there can be large variations in the concentrations of metals urine samples due to inconsistent exposure. For this study 11 men collected urine samples on eight occasions during a period of 3 months. From measurements of the concentrations of As, Cd, Co, Cu, Mo, Ni and Pb, the 3 month averages were calculated and compared with individual spot-sample concentrations. The data were used to estimate the subject-specific mean, taken to represent exposure during that period. They computed that the number of spot-samples required to reliably estimate the mean (to within 20%) was from 3 for Cd to 27 for Zn.

6.3.3 Tissues. Compared to our previous Updates, all the work of interest discussed in this section involves analyses using XRF spectrometry and most refer to developments with the instrumentation rather than reports of the concentrations of elements in specific tissues. Concentrations of As and Se were measured in various soft tissues. This work by Fleming et al.⁸⁸ was essentially aimed at testing whether or not a portable XRF spectrometer could be used for in vivo analysis. Details of this work were described above (Section 4.5).

Normal, primary and secondary bone cancer specimens were analysed using fsLIBS by Gill et al.⁷⁷ The objective was to determine whether discrimination of normal from cancerous tissue was possible, ideally in the course of an operation. The work showed that Ca : Mg ratios were decreased in primary cancer, compared with normal bone tissue. In secondary cancer samples, where there are both normal and cancerous cells, there was no significant difference compared to normal tissue. The results indicated the prospect of using fs lasers for identification and ablation of primary bone tumours.

Two studies were reported where XRF spectrometry was used to determine concentrations of elements in brain tissues. Wandzilak et al.⁸⁴ prepared high resolution maps of the spatial distribution and concentrations of Ca, Cl, Cu, Fe, K, P, S and Zn in brain tumour tissue and surrounding structures such as blood vessels and areas of calcification. The results obtained using μXRF spectrometry showed decreased concentrations of Ca, Fe, P and S but increased Zn, in cancerous tissue and that the changes correlated with the tumour malignancy grade. The specificity of this approach was demonstrated by comparison of results against the histological classification. Agreement was found for 99.93% of the cases used to develop the model and 99.16% when applied to new cases. While not suggested by these workers it would appear that the approach could be used to give an analytical diagnosis while only the difficult cases would have to be seen by an experienced histopathologist. Tian et al.¹¹⁸ considered concentrations of Cu, Fe and Zn in neurodegenerative disease. Using SR-XRF spectrometry, these workers found higher concentrations in the substantia nigra pars compacta of a Parkinson's disease mouse model compared to control tissues. Treatment with the drug Rapamycin reduced the concentrations of these elements and also reversed changes to the dopamine neurons and motor nerve functionality.

The review by Ceko et al.,¹⁵ on analysis of trace elements in ovarian tissue, was mentioned in Section 1. The same authors also report some of their own work using XRF spectrometry.¹²⁰ The research focussed on Fe, Se and Zn which appeared to be localised to specific structures. Co-localisation of Fe and Zn was shown while Se, surrounding the antrum of large follicles, appeared infrequently. The authors opined that the high resolution over large areas could have a significant impact on understanding the mechanisms of ovarian development.

Work from the group led by Chettle features regularly in our Updates. In addition to the report of Fleming et al.⁸⁸ described earlier in this section, this team investigated the feasibility of using in vivo XRF spectrometry to measure concentrations of As and Se in human skin. In the project reported by Shehab et al.⁸⁹ the performances of an upgraded portable XRF system and an advanced version of a benchtop XRF system were evaluated and compared (see Section 4.5 for further details of the Experimental conditions). An interesting case report by Breier et al.¹³³ described the investigation of a 75 year-old man who had unusual grey pigmented skin lesions on his face and fingers. He had previously worked for many years cutting high voltage cables and never used a face shield or gloves. Biopsies removed for histological examination showed inclusions within macrophages. Unstained sections analysed by SR-XRF spectrometry showed deposits of Cu, Ni, Sr and Zn, all of which were the main components of the cable with which he had worked. The metals were found in association with S, suggestive of binding to protein thiol groups.

6.4 Progress for individual elements

6.4.1 Arsenic. Work regularly continues to be published on the speciation of As in biological matrices, particularly as environmental and public health regulation starts to focus on health risks from iAs and organo-As species rather than total As concentrations. An essential aspect of As speciation measurement is the maintenance of species integrity during sample collection, preparation and analysis using combined chromatographic and atomic spectrometric methodologies. At any stage of the analytical procedure there is the risk of altering the original As species. Maher et al.¹⁷ discussed this in a comprehensive review of As speciation measurement. They emphasised that As extraction is matrix dependent and cannot be achieved for all sample types using a single set of conditions. They also stressed that mass balance, extraction and chromatographic column recoveries need to be quantified at each step. Methods must be validated by spike recovery studies and wherever possible analysis of CRMs. Stice et al.¹⁰⁵ developed an HPLC-ICP-MS method for the separation and quantitative determination of 13 As metabolite species previously identified in biological matrices. The method focused on the separation and determination of thiolated and thiol-conjugated As species. The method was used to determine As species in human cancer cells incubated with darinaparsin, an organic arsenical chemotherapy agent. They considered the method to be a valuable tool in identifying the As species formed during metabolism of environmental or occupational exposure and As-based therapeutic treatments. There is always a feeling of déjà vu when reading through some papers in preparation for writing this review. Methods for the quantitative determination of As species in human urine have been developed over the past 30 years. Papers continue to be published on this topic, although in many cases it is not clear what advantages the new approach offers over already published methods. A Spanish group¹⁰⁴ presented a method for the determination of As^III, As^V, DMA, MMA, AB and AC in urine using anion exchange chromatography coupled with ICP-MS. Satisfactory chromatographic separation within 9 minutes was achieved using gradient elution with two mobile phases, H₂O and 20 mM HNO₃ both containing 2% MeOH. Urine samples were simply filtered and diluted 1 : 4 v/v with H₂O. Reported LODs were between 0.3 and 0.9 μg L⁻¹ and the method was validated by analysing NIST SRM 2669, which has certified values for As species. The authors used the method to determine concentrations of As species in volunteers neither environmentally or occupationally exposed to As. Marschner et al.⁸⁰ optimised experimental conditions for the quantitative determination of As species in human urine using HG-AFS. They reported that the conditions employed produced 100% efficient generation of AsH₃ from all As species by reaction with 2 mol L⁻¹ HCl and 2.5% NaBH₄. Use of a novel gas–liquid separator design reduced signal noise from fluctuations in AsH₃ flow to the atomiser. Reported LODs were 40 ng L⁻¹, 97 ng L⁻¹, 57 ng L⁻¹ and 55 ng L⁻¹ for As^III, As^V, MMA and DMA respectively. The authors considered the method offered better results than other methods without the need for sample pre-treatment.

Two papers by a Canadian group^88,89 describing excellent work on the determination of As and Se in skin and organs using XRF are discussed in detail in Section 4.5.

It was somewhat disconcerting, though possibly a reflection of the times we are in, to find a paper⁴⁰ describing a rapid emergency response method for the detection of Lewisite exposure biomarkers in human urine, using LC-ICP-MS/MS. Chromatography conditions were optimised for the separation and quantitative determination of the primary and secondary Lewisite metabolites, (2-chlorovinyl)arsonic acid, bis-(2-chlorovinyl)arsonic acid and (1-chlorovinyl)arsonic acid. Urine samples were diluted with water and reacted with 6% H₂O₂ to oxidise any (2-chlorovinyl)arsenous acid to the pentavalent metabolite. The authors showed the method to be accurate, robust, specific and sensitive, with a reported LOD of 2.2 μg L⁻¹. Stability studies of Lewisite metabolites showed them to be extremely stable with an in vivo oxidation half-life for (2-chlorovinyl)arsenous acid of 6.2 days at 37 °C. One hopes that the methodology is never needed.

6.4.2 Beryllium. Chronic occupational exposure to Be can lead to severe lung diseases including cancer and beryllosis, a pulmonary granulomatosis. Hulo et al.¹³⁴ commented on the limitations of urinary Be measurements as a biomarker. As inhalation is the primary route of exposure, they examined the merit of determining Be in exhaled breath as a biomarker of exposure. Exhaled breath condensate (EBC) and spot urine samples were collected from aluminium production plant workers with a history of Be exposure as assessed by air Be measurements. Concentrations of Al and Be in EBC were determined using ICP-MS, with external calibration. Reported LODs were 0.001 μg L⁻¹ and 0.1 μg L⁻¹ for Be and Al respectively. Urine Al and Be concentrations were determined using ICP-DRC-MS with He as the collision cell gas. Calibration was by standard additions. Reported LODs were 0.0064 μg L⁻¹ and 4.1 μg L⁻¹ for Be and Al respectively. The researchers reported that EBC Be levels were elevated in samples from exposed workers compared with controls and that the highest concentrations were found in EBC samples from pot room workers. A significant positive correlation between Be and Al in EBC, but no correlation was observed between concentrations of either element in EBC and the corresponding urine samples. They also observed a positive correlation between EBC Be and cumulative Be exposure in pot room workers but not welders. The researchers concluded that, despite low atmospheric concentrations, Be could be quantitatively determined in EBC from exposed workers. They considered that measurement of Be in EBC could provide additional toxicokinetic information complimentary to that provided by urine Be measurements and that EBC could be used to improve the biomonitoring of exposed workers.

6.4.3 Boron. In past reviews we have highlighted the growing interest in the release of metals, as soluble ions or particulates, from surgical implants. Most of the research has concentrated on implants containing Co, Cr, Ni, Ti and V. More recently, attention has focused on the potential of boron nitride–hydroxyapatite (BN–HA) nanocomposites in orthopaedic and dental implants due to their biocompatibility with bone tissue. However, to date there appears to have been no research published on the in vivo stability of BN–HA implants. Atila and colleagues¹³⁵ used ICP-MS to monitor serum B levels in rats following implantation of BN–HA composites into femur bones. Serum sample were diluted 20-fold with NH₄OH/EDTA/Triton X-100/C₄H₉OH and B quantitatively determined using matrix matched calibration standards. The reported LOD was 0.2 μg mL⁻¹. The researchers measured serum B levels over the first four weeks following implantation and reported that serum B levels did not change significantly at any time point after implantation. They concluded that implanting BN–HA nanocomposite did not result in increased serum B levels in the short or long term, therefore it could be considered a promising new material for implantation technologies.

6.4.4 Chlorine. In recent years the review team have noted the increasing use of ICP-MS for quantitative measurement of non-metallic elements in biological matrices. The potential use of chlorate salts as in vivo disinfectant for animals bred for food led Schwan et al.⁵⁹ to develop a method for the separation and quantitative determination of Cl species in whole blood using Q-ICP-MS coupled with IC. With an injection volume of 20 μL, reported LODs for Cl species were 0.5 mg L⁻¹ for chlorite and 1.0 mg L⁻¹ for chlorate. The researchers investigated the use of ICP-QQQ-MS in MS/MS mode with H₂ as reaction gas to overcome polyatomic interferences on ³⁷Cl. However, S/N ratios could not be improved. The measurement of sweat Cl⁻ is a recognised method for the diagnosis of cystic fibrosis. Recommended clinical chemistry methods for the determination of Cl⁻ in biological fluids are generally based on electrochemistry (ion-specific electrodes or coulometry). Collie and colleagues³⁴ described work to develop an approach for the measurement of Cl⁻ in sweat using ICP-MS, to present to the Joint Committee for Traceability in Laboratory Medicine as a candidate reference method for CF diagnosis. External calibration standards covering the expected concentration range were prepared from NaCl. Two variants of the method were compared, using either online (instrument based) internal standardisation with Ge and Sc or off-line (sample based) internal standardisation with Ga. Both methods gave highly reproducible results and excellent agreement with target values for proficiency test samples from the Royal College of Pathologists Australasian QA programme, although the off-line method showed increased accuracy. The authors considered both methods suitable for recognition as reference methods for diagnosis and monitoring of CF.

Klencsar et al.⁵⁸ used HPLC coupled with ICP-MS/MS to separate and quantitatively determine total drug-related Br and Cl in human plasma. Chromatography was used to separate the drug based Br and Cl from free inorganic Cl⁻. Interference-free determination of ³⁵Cl was achieved using H₂ as reaction gas and measuring the ³⁵ClH₂ reaction product at m/z 37. Bromine was measured ‘on-mass’ at m/z 79. Reported LODs were 0.05 μg L⁻¹ and 0.01 μg L⁻¹ for Cl and Br respectively. The method was used to analyse plasma samples from patients participating in a clinical study of a newly-developed Cl- and Br-containing pharmaceutical agent. The authors considered the method to be a suitable alternative to standard methodologies using radio-labelled drugs.

6.4.5 Chromium and cobalt. Much of the work of interest involving Co and Cr relates to metal implants and is discussed in Section 6.3.1. Kerger et al.⁵⁵ undertook a comprehensive study on factors affecting the quantitative determination of Co in blood, serum and synovial fluid using two commercial ICP-MS systems with CRC technology. Method LODs of 0.2–0.5 μg L⁻¹ were reported for synovial fluid and serum, and between 0.6 μg L⁻¹ and 1.7 μg L⁻¹ for whole blood. Both acid digestion and detergent dilution sample pre-treatments gave comparable results. The authors considered that the ICP-MS technology used gave reliable results for Co concentrations above 5 μg L⁻¹ but that polyatomic interferences may affect accurate quantitative determination at very low Co concentrations in complex matrices.

6.4.6 Copper and zinc. Green (or sustainable) chemistry was a very popular topic in our last review¹ and green methods continue to be published. Mortada et al.⁴⁶ described a precipitation method for the preparation of hydroxyapatite nanorods from recycled egg shells. The nanorods were used as a novel solid phase extraction medium for the pre-concentration of Cu, Pb and Zn from biological matrices for quantitative determination using FAAS. Maximum sorption capacities were between 2.37 and 2.53 mmol L⁻¹ for all 3 ion species and the pre-concentration factor for the method was 250, giving reported LODs of 0.72 μg L⁻¹, 6.12 μg L⁻¹ and 0.55 μg L⁻¹ for Cu, Pb and Zn respectively. The method was validated by analysing whole blood and urine CRMs.

Hachmöller et al.¹¹² presented interesting work on the distribution of Cu and Zn in biological tissues. These workers used LA-ICP-MS to analyse liver needle biopsy samples from patients with Wilson's disease, a rare disorder of Cu metabolism. Concentrations of the two elements were also quantitatively determined using external calibration with matrix matched gelatine standards. The authors readily differentiated two samples from patients with the disease from a control sample. Copper concentrations up to 1200 μg g⁻¹ were measured and Cu distribution was not homogeneous. The researchers also reported an inverse relationship between the two elements in regions of high Cu and high Fe levels.

Burton and colleagues¹³⁶ claimed to present a ‘novel omics’ approach to facilitate personalised cancer screening. They used ICP-MS to quantitatively determine 22 elements in urine samples from 136 female patients newly diagnosed with breast cancer. Only two metals, Cu and Pb were significantly increased in the cancer patients compared with healthy controls. Using multivariate statistical methods the researchers considered the approach compared favourably with other breast cancer diagnostic methods. These workers proposed that this ‘omics’ approach offered a non-invasive approach to early cancer detection and hypothesised that urinary Cu and Pb might be a potential breast cancer biomarker. It will be interesting to see if this ‘omics’ approach is taken up and developed further by other research groups.

Two papers reported elevated concentrations of Zn in tumour tissue. Bishop et al.¹¹¹ used ICP-QQQ-MS operated in no-gas, H₂ or H₂/O₂ modes to overcome polyatomic interferences on the quantitative determination of trace elements in tissue sections. Using the H₂/O₂ mixed gas mode, they observed high concentrations of Zn in cancerous tissue from a prostate biopsy. They considered that ICP-QQQ-MS technology with mixed cell gases provides a sensitive method for elemental bio-imaging. Wandzilak et al.⁸⁴ used μXRF to map elemental distributions in human brain tumour tissue. The researchers found that increased concentrations of Zn and decreased concentrations of Fe were correlated with tumour malignancy.

6.4.7 Gadolinium. Methods developed to monitor Gd levels in patients undergoing MRI scans, where the element is used as a contrast agent have been regularly reported. Roberts et al.¹¹⁵ reported the case of an individual having normal renal function but presenting with high concentrations of Gd in skin as a result of exposure to large cumulative doses of Gd-based contrast agent over an 11 year period involving 61 brain MRI scans. The researchers determined Gd in skin biopsies from the forearm and lower extremities using LA-ICP-MS and ICP-MS coupled with hydrophilic interactive LC. Skin Gd levels of up to 14.5 μg g⁻¹ were determined which the authors noted were similar to levels previously observed in MRI scan patients with nephrogenic systemic fibrosis. Laser ablation ICP-MS showed the Gd to be deposited within the deep layers of the skin and speciation analysis using LC-ICP-MS revealed the presence of intact Gd-chelate species. It was concluded that Gd skin deposition can occur in patients with normal renal function who are exposed to high cumulative doses of contrast agents. They considered further studies to address the potential consequences of deposition of Gd in skin and other tissues and organs.

6.4.8 Gold. The interest in nanomaterials, highlighted in last year's review¹ has increased considerably in this review period, becoming a dominant feature of the review. Gold NPs are receiving substantial interest as both imaging and therapeutic agents. Alzaki and colleagues⁹⁸ noted, however, that rather than being novel, biological applications of AuNPs began in the 1970's with the development of gold based immunostaining. The researchers investigated the biodistribution and clearance of two formulations of AuNP-loaded micelles, containing either 0.9 nm or 5 nm particles, injected i.v. into mice. Concentrations of Au in body organs and tissues, faeces and urine were quantitatively determined using ICP-OES at post injection times between one day and three months. Tissues, organs and whole blood were digested with HNO₃/HCl (aqua regia). Highest levels of Au were determined in the liver and spleen. A marked reduction in all concentrations was observed between day 1 and 3 months post injection. The primary clearance mechanism was hypothesised to be through hepatobiliary excretion. The presence of Au in the urine of mice administered the 0.9 nm AuNPs indicated some degree of renal excretion. The observation that smaller NPs were excreted more efficiently led the authors to propose that excretion of administered NPs might be improved if clusters of smaller particles were used instead of larger single NPs. This leads very neatly onto the work by Tsai et al.¹³⁷ who emphasised the importance of traceable and verifiable characterisation methods for functional AuNPs in imaging or drug delivery biomedical applications. They presented a comprehensive orthogonal characterisation study using physical (electrospray-differential mobility analysis), microscopic (AFM, TEM, SEM) and spectrometric (SAXS, ICP-MS) techniques to obtain the key material characteristics of commercially-available surface-modified AuNP colloids.

Elci et al.¹¹⁴ used ICP-MS to quantify the sub-organ distribution of AuNPs in kidney, liver and spleen of mice injected i.v. with functionalised AuNPs. They reported that sub-organ distribution was strongly influenced by the surface charge of the NPs. In the kidney, positively charged particles accumulated in the glomeruli, which led the researchers to hypothesise that positive NPs may be filtered at a different rate to negatively or uncharged particles. In the spleen, both negatively and positively charged particles accumulated in the red pulp whilst uncharged particles were found in the white pulp. In the liver, positively charged particles accumulated in the hepatocytes whilst negatively charged particles were more heterogeneously distributed. The authors concluded that neutral NPs may interact with the immune system to a greater extent than charged particles. Shi et al.¹³⁸ presented an important study on the biodistribution of AuNPs following intervaginal space injection (ISI) in the tarsal tunnel of rats, which the authors considered a novel anatomical structure for fluid transport. The authors compared the biodistribution of NPs injected via this route with conventional i.v. injection (IVI). Concentrations of Au in blood and target organs were quantitatively determined using ICP-MS. Significantly different distribution patterns were identified for the two injection routes. Unsurprisingly, concentrations in blood were significantly lower following ISI compared with IVI but Au levels were higher in skin and muscle following ISI. The authors concluded that the results demonstrated a direct circulation-independent AuNP transportation pathway which may help in improving the biodistribution of nanomaterials.

6.4.9 Iron. Methods for the accurate isotopic analysis of trace elements in biological matrices normally involve some form of sample treatment to remove the element of interest from the complex matrix. Anoshkina et al.⁶² reported studies to directly determine Fe isotopes in acid digests of whole blood, without any form of Fe separation, using MC-ICP-MS. Preliminary work with synthetic solutions was undertaken to investigate the potential interferences from other matrix elements and remaining organic material. They noted that, in the presence of high concentrations of Na and K, Fe isotope composition showed a low bias, whereas glucose concentrations >1.0% led to a high bias for isotope measurements. The authors reported that by diluting blood samples to give a total Fe concentration of 0.75–1.5 mg L⁻¹ and correcting for mass bias with an internal standard of Ni, accurate and precise isotope ratio measurements could be made. This was confirmed by comparing results with those obtained by a reference method employing prior chromatographic separation of Fe from the blood matrix.

Two groups described studies on the intracellular localisation of Fe and other essential elements. An Australian group¹²⁰ presented the findings of a comprehensive study to examine the distribution of Fe, Se and Zn in bovine ovaries in relation to anatomical structures. They used XRF spectrometry to analyse 40 ovary cross-sections focusing on antral follicles, corpora lutea and blood vessels. It was argued that minimal sample processing enabled high resolution information on the spatial distribution of the three elements. Co-localisation of Fe and Zn in the structures examined was noted. The group considered the data provided a valuable baseline study of healthy ovaries which could be extended to disease states in order to gain a better understanding of ovarian health. In order to gain better insight into the relationship between metal elements in the brain and neurodegenerative diseases, such as Parkinsonism, a Chinese group¹¹⁸ used SR-XRF spectrometry to study the localisation of Cu, Fe and Zn, in brain sections of mice with drug-induced Parkinsonism. They observed that drug-treated mice had higher concentrations of all three elements in the substantia nigra pars compacta. Treatment with rapamycin reduced the elemental concentrations and restored dopamine function. The researchers considered the method to be a valuable tool for monitoring pathological changes in Parkinsonism and in assessing efficacy of drug therapies.

6.4.10 Lead. Over the past 30 years it has been very reassuring to report on ever declining reference levels for blood Pb in general populations in the industrialised world. In this review period, Tagne-Fotso et al.¹³⁹ reported the results of a large study on blood Pb levels in the general population of northern France. Two thousand individuals aged between 20 and 59 were recruited. Blood Pb was determined using ICP-MS. A geometric mean blood Pb of 18.8 μg L⁻¹ was reported with a 95% confidence interval of 18.3–19.3 μg L⁻¹. Multivariate regression models identified smoking, consumption of common beverages (such as tea, coffee and wine) and raw vegetables, some DIY and leisure activities and living in pre 1950s housing, as the major factors influencing blood Pb.

Although environmental levels of Pb exposure have declined dramatically in recent years, there is still concern over exposure to Pb in early childhood. Shepherd et al.¹¹⁶ studied the concentration and isotopic composition of Pb in dentine and enamel from deciduous teeth in order to assess foetal and early childhood Pb exposure. Total Pb concentrations and ²⁰⁸Pb : ²⁰⁶Pb, ²⁰⁷Pb : ²⁰⁶Pb ratios were determined at 100 μm intervals across thin sections of teeth using LA-ICP-MS. Integrating the isotope data with histological analysis enabled the researchers to assign estimated ages to each ablation point. They reported significant differences in both the concentration and isotope ratios of Pb between individuals reflecting differences in exposure during early childhood, as the lowest dentine Pb levels were found in children born in 2000 or later, following the withdrawal of Pb from petrol. The Pb isotope ratios correlated closely with those found in modern day industrial particulate matter aerosols (PM10s and PM2.5) indicating airborne pollution as the primary exposure source. The authors considered this pilot study demonstrated that the method had the potential to assess Pb exposure in large study cohorts.

With ever decreasing environmental Pb levels, attention has focused on dietary factors contributing to Pb exposure. Regular consumption of game meat is rare in the UK but apparently much more common in Canada, to such an extent that research was funded by the Quebec Institute of Public Health into the potential health risks from consumption of lead-shot big game meat. A study¹⁴⁰ was performed to assess both actual and perceived health risks associated with the consumption of white-tailed deer and moose killed by lead ammunition. Concentrations of Pb in meat samples were determined using ICP-MS. Lead was detected in 90% of deer samples and 70% of moose samples analysed. Levels exceeding 0.1 mg kg⁻¹ were recorded in 33% of deer samples and 13% of moose samples. Questionnaire returns from hunters were analysed using Monte-Carlo simulation methods and enabled the researchers to conclude that a small proportion of hunters and their families might be at risk from increased Pb exposure from game meat consumption. They considered it important to inform the relevant population of this hazard and how it might be avoided.

Methods still continue to be developed for the determination of Pb in serum. Barbosa et al.⁴⁵ described a method for the direct determination of Pb from untreated blood serum using carbon nanotubes coated with BSA. The nanotubes had a reported adsorption capacity for Pb of 34.5 mg g⁻¹. They were packed in a mini-column in an online solid-phase extraction system coupled to thermospray-FFAAS. An enrichment factor of 5 was reported giving an LOD of 2.1 μg L⁻¹. Unfortunately in our opinion, this LOD is still insufficiently sensitive to be of value for ‘real’ applications other than for monitoring individuals with very high blood lead levels.

Bench-top μXRF systems offer the capability to provide quantitative multielement measurements and mapping with the advantages of being non-destructive and requiring small sample sizes, properties that are particularly relevant for archaeological studies. Dias et al.⁹² used a commercial bench-top polycapillary μXRF system and an XRF system with tri-axial geometry to study the distribution of Pb in bones and teeth of an 18^th century male skeleton. The two systems gave comparable quantitative data as assessed by the analysis of pressed pellets of SRM bone material and an LOD of 3 μg g⁻¹ was reported. Mapping of Pb in cross sections of teeth and bone revealed enrichment in the tooth pulp and root and in inner spongy bone. Levels of Pb determined in different bones from the skeleton ranged from 120 μg g⁻¹ in tibia to 560 μg g⁻¹ in ribs and showed a similar distribution relationship to in vivo concentrations found in long-term exposed lead workers. Levels of Pb in the soil surrounding the skeleton were very low, leading the researchers to conclude that skeletal Pb was due to ante mortem exposure.

6.4.11 Manganese. Grygo-Szymanko et al.²⁰ undertook a comprehensive review of analytical methods for the speciation and quantitative determination of Mn in environmental and biological matrices. The review encompassed simple sample preparation techniques, such as extraction, through to complex hyphenated techniques for separation of Mn species.

6.4.12 Mercury. In a paper that could well make a good crime novel, Lech¹⁴¹ described the quantitative determination of Hg in organs and hair samples from an exhumed female cadaver, autopsied six years after death on the order of the public prosecutor's office. At the time of death, no toxicological tests were undertaken and death was attributed to myocardial infarction. Subsequent CVAAS analysis revealed elevated Hg and elevated concentrations ranging from 227 ng g⁻¹ in scalp hair to 1516 ng g⁻¹ in small intestine, 1201 ng g⁻¹ in liver and 1023 ng g⁻¹ in heart muscle. Analysis of soil around the burial site gave Hg concentrations of 0.5–1.5 ng g⁻¹. These results strongly supported new evidence which indicated the woman may have been poisoned. A Portuguese group⁹³ identified extremely high concentrations of Hg in hair samples taken from a human cadaver from the 18^th or 19^th century. Analyses were made using two EDXRF spectrometry systems, one using tri-axial geometry the other using a micro-beam. Exceptionally high hair Hg concentrations, over 5% w/w were reported. Much lower levels were reported in the cranium and there was no evidence of elevated Hg in the soil surrounding the burial site. The researchers hypothesised that the high hair concentrations were due to the use of a mercury-based compound on the scalp.

6.4.13 Platinum. The merit of NP mediated drug delivery was demonstrated by Gao et al.¹⁴² They used ICP-MS to quantitatively determine the intracellular delivery of therapeutic Pt into cancer cells. Use of multinuclear Pt, as polymer-di-cis Pt^IV NPs led to both increased intracellular Pt concentrations and increased Pt-DNA adduct formation. Mechanistic studies showed the uptake of Pt to be due to endocytosis rather than passive diffusion or metal-ion transporter mediated uptake. Another Chinese group¹¹³ used LA-ICP-MS to visualise the tissue distribution of two novel Pt drug complexes in mice following intraperitoneal injection. A commercial cis-platin drug was used as a reference comparator. Quantitative determination of Pt in six major organs showed highest concentrations in liver followed by kidneys, lung, testes heart and brain. Deposition trends, monitored using LA-ICP-MS matched the findings of the total elemental analysis. It was particularly noted that the kidney distribution patterns observed for the different Pt complexes were significant to enable discrimination between the two novel complexes. The distribution maps were sufficiently different to be able to distinguish the two novel complexes. They considered LA-ICP-MS to be a valuable tool for therapeutic drug development.

6.4.14 Selenium. In our review series we have regularly reported on the ‘discovery’ of new Se species and metabolites in biological matrices, many of which have no known function. However, the results of an interesting piece of work by Anan et al.¹⁰² demonstrated a selenite detoxification mechanism in mammalian cells. Incubating human hepatoma cells with Na₂SeO₃ led to the formation of an ‘unknown metabolite which the authors detected in the cell cytosol using HPLC-ICP-MS’. The researchers identified the metabolite as SeCN⁻ using ESI-MS and ESI-Q-TOF-MS. They hypothesised that SeCN⁻ was formed by intracellular reaction between selenide (or other products of selenite reduction) and endogenous cyanide. They considered that, as SeCN⁻ was less toxic than either of the reactants, it was formed to reduce the intracellular toxicity of selenite. A German team¹⁴³ sought to characterise Se species in the CSF of sheep administered a diet supplemented with Na₂SeO₃ (0.2 mg kg⁻¹ dry weight) over a two year period. Control animals were fed a diet with a marginal Se concentration of <0.05 mg kg⁻¹ dry weight. At the conclusion of the feeding regime, paired CSF and serum samples were collected and Se measured and speciated using IEC coupled with ICP-DRC-MS. A mean serum Se level of 96.5 ± 18 μg L⁻¹ was determined in the Se supplemented animals compared with a mean level of 33 ± 5 μg L⁻¹ in the control group. Corresponding concentrations of Se in CSF were 6.1 ± 1.6 μg L⁻¹ in the supplemented animals and 4.38 ± 1 μg L⁻¹ in the controls. Selenoprotein P and GPx were the principal Se species in serum. The authors noted that the CSF : serum Se ratio was higher than the CSF : serum albumin ratio, which led them to hypothesise that there was local generation of Se species in the brain. A good positive correlation between albumin and albumin bound Se led them also to hypothesise a role for albumin in Se transportation across the CSF barrier.

Another German group¹⁰¹ studied the in vivo metabolism of Se in a volunteer group of healthy subjects administered a dietary supplement of either Na₂SeO₄ or selenised yeast. The group was categorised into TMSe⁺ eliminators and non-eliminators. Total Se concentrations in plasma and urine were determined using ICP-MS. Further speciation analysis of Se in urine was performed using LC-ICP-MS. Plasma Se increased similarly for both supplement types within 2–3 h of administration, from 84.5 ± 13.2 μg L⁻¹ to 97.4 ± 13.2 μg L⁻¹ for selenite versus 89.5 ± 12.9 μg L⁻¹ to 92.1 ± 13.9 μg L⁻¹ for the yeast. The major urinary species was unmetabolised selenate (30 ± 6.9% of the administered dose). As expected, TMSe⁺ was only found in the urine of the TMSe⁺ eliminators and this group eliminated a significantly lower amount of selenate. Only one individual metabolised Se to Se-containing carbohydrates, which were also detected in urine. The authors hypothesised that different metabolism routes might imply different dispositions to pharmacological or toxic effects from exposure to inorganic Se compounds.

6.4.15 Silver. As with Au in this review period, the focus of attention for Ag has been on the quantitative determination of Ag nanoparticles (AgNPs) in biological matrices. Several papers on this topic merit comment, some of which are discussed in the other sections of this review. The interest has been stimulated by the use of Ag nanomaterials as potent antimicrobial agents in food packaging and containers. Feichtmeier et al.⁶⁹ described a fast, sensitive screening method for the determination of AgNPs in complex biological matrices, using solid-sampling CS-ETAAS. Taking advantage of the recently observed phenomenon of a delay in the atomisation signal for AgNPs compared with ionic Ag⁺, the researchers developed a method that enabled differential analysis of AgNP and Ag⁺ without elaborate pre-treatment. They reported that atomisation delays between AgNPs and Ag⁺ varied in different matrices between −2 s and +2 s, but were always significant thereby allowing discrimination between the two Ag species. A Russian team⁹⁹ examined the biodistribution of AgNPs following acute single dose (2000 mg kg⁻¹ b.w.) or sub-acute repeated dose (250 mg kg⁻¹ b.w.) intragastric administration of 12 nm non-coated AgNPs. Tissue Ag concentrations were determined using AAS. For both administration regimes, highest levels of Ag were determined in the liver (0.87 ± 0.37 μg g⁻¹) and kidneys (0.24 ± 0.02 μg g⁻¹). Silver was also found in spleen, stomach and small intestine but at much lower concentrations. The very low tissue levels determined, compared with the high doses administered, led the authors to conclude that excretion of Ag was very efficient. No post mortem patho-morphological abnormalities were observed and haematological and biochemical parameters were unaffected compared with control animals. The authors concluded that AgNPs were readily absorbed from the GI tract but did not cause significant adverse effects. Aude-Garcia et al.¹⁰⁰ studied the intracellular concentrations and distribution of Ag in murine macrophages exposed to either a single high dose or repeated low doses of AgNPs, using a battery of analytical techniques including PIXE, TEM, EDXRF spectrometry and ICP-MS. They noted that accumulation in macrophages was similar in both exposure scenarios. Also, that partial Ag expulsion occurred once macrophages were removed from the exposure. An assessment of function revealed that acute exposure led to a strong macrophage response, including reduced phagocytic activity and increased release of inflammatory cytokines. The responses were reversible and were much less marked in the cells repeatedly exposed to low doses of AgNPs.

Sussman et al.⁹⁷ investigated the release of AgNPs from medical devices. The researchers used SEM to confirm the presence of AgNP coatings on a number of marketed medical devices. Release of Ag from these devices incubated in either H₂O, saline or human plasma was quantitatively determined using ICP-MS. Over incubation periods of one to seven days, Ag leaching ranged from 10⁻¹ to 10⁶ ng cm⁻². One particular device released significantly more Ag than the others. The authors considered that potential release of Ag from some medical devices needs to be incorporated into the risk assessments for these products.

6.4.16 Thallium. Whilst chronic environmental exposure to Tl is extremely rare, a geological survey of the Valdicastello region of Tuscany identified alarmingly high levels of Tl in the local groundwater, due to acid leaching from pyrite ores in abandoned mines. Analysis of tap water in the area identified levels of Tl between 2 and 10 μg L⁻¹ and led the health authorities to impose a ‘do not drink’ order. Campanella et al.¹⁴⁴ monitored Tl exposure of the local population by measuring Tl levels in urine and hair samples. Thallium was quantitatively determined using ICP-MS. Hair concentrations, representing longer term exposure, ranged from 1 to 498 ng g⁻¹ whilst urine concentrations, reflecting more recent contact, ranged from 0.046 to 5.44 μg L⁻¹. The authors observed a clear relationship between elevated Tl levels in biological samples and residency in zones with elevated tap water Tl.

Although current medical opinion considers the use of Prussian blue as the appropriate therapy for acute Tl intoxication, Saljooghi et al.¹⁴⁵ suggested that a combined treatment with the chelators deferasirox and desferrioxamine, two recognised chelating agents for treating iron overload, might be an efficient chelation therapy for Tl intoxication. Groups of rats were injected with a single dose of TlCl and given the chelating agents either individually or as a combined treatment at two dosing levels, 75 mg kg⁻¹ or 150 mg kg⁻¹. Serum Tl and both urine Tl and Fe were quantitatively determined using ETAAS. The researchers reported that the chelators were only effective at the higher dose. Desferrioxamine was more effective than deferasirox in increasing urinary Tl excretion but no improvement was observed with the combined treatment.

6.4.17 Titanium. There has been a renewed interest in Ti over the current review period. A Polish research team^83,146 investigated the biodistribution and localisation of ionic Ti following either a single i.v. dose or repeated oral doses of Ti (6 mg kg⁻¹ b.w.). Concentrations of Ti in serum and major organs were quantitatively determined using ETAAS. Organ distribution studies were undertaken using micro-SR-XRF spectrometry. It was reported that Ti was not uniformly distributed but formed aggregates, some of which were also enriched with Ca. The group identified kidney as the principal target organ followed by liver and spleen. The authors claimed the studies contributed to the understanding of the toxicokinetics of ionic Ti and the potential health risks from exposure to Ti. However, it should be questioned how relevant the data might be, given the very high doses of Ti administered.

A study reported by Sakamoto et al.¹²⁶ described a case of the development of a pseudotumour in an elderly patient who had undergone metal-on-metal total hip arthoplasty. Magnetic resonance imaging indicated the presence of the pseudotumour which was surgically removed. Quantitative determination of metal elements in the tumour tissue revealed the main source of metal to be debris from the implant which was constructed of a titanium alloy. The authors concluded that release of particulate Ti from metal implants has the potential to cause the formation of symptomatic pseudotumours.

6.4.18 Uranium. Bone is a primary target organ for the accumulation of U following occupational or environmental exposure. However, the mechanisms of U uptake and accumulation are poorly understood. In an effort to gain further insight into bone U mobilisation, Bourgeois et al.¹²¹ used ICP-MS and SR-μXRF spectrometry to study the distribution of U in bones of rats administered an i.v. injection of UO₂(NO₃)₂. Total U, measured in HNO₃ digests of ashed bone using ICP-MS, was 26 ± 3 ppm. Femur samples were fixed in EtOH and embedded in methyl methacrylate. Sections (150 μm depth) were cut and 2 mm areas analysed for U using full-field micro-beam XRF spectrometry. High U concentrations were observed in traebecular and cortical bone sections and specifically located at the edges of the bone sections, close to the cartilage growth plates. This accumulation at actively growing regions of bone led the authors to hypothesise that it arose from active transport and deposition via metal-binding proteins rather than simple chemical precipitation with phosphate. They considered more investigations were needed to further clarify the mechanisms of U deposition in the skeleton.

7 Applications: drugs and pharmaceuticals, traditional medicines and supplements

Traditional Chinese medicines are widely used throughout the world. A number of investigations have shown that some preparations may contain undue amounts of elements such as As and Hg that can be harmful to health. Niuhuang Jiedu is a traditional Chinese medicine that contains realgar (arsenic sulfide mineral). Jin et al.¹⁴⁷ used fully validated ICP-MS and HPLC-ICP-MS procedures to measure the concentrations of As^III, As^V, MMA, DMA, AB, ACh and 20 other trace elements, in preparations of this medicine. Their results indicated that some manufacturers formulated lower amount of realgar than required in the Chinese Pharmacopoeia in their preparations. In experiments to determine the extraction into water and gastrointestinal fluid almost identical profiles for total As and iAs were found indicating that most of the As must be as the toxic species. For the other trace elements, extraction rates were higher into gastrointestinal fluid. The authors concluded that potential toxicity of Cd, Cu, Hg and Pb is low but that the true As toxicity should be further investigated. A method for alkaline extraction of Cr from traditional Chinese medicines to minimise inter-conversion of Cr^VI to Cr^III was reported by Li et al.¹⁴⁸ The Cr^VI was separated by anion exchange chromatography and concentrations measured by ICP-MS. The procedure was applied to analysis of Chinese medicines.

A number of prescription, non-prescription medicines and dietary supplements available in Poland were analysed by Kowalski and Frankowski¹⁴⁹ to determine concentrations of Hg. Those with the greatest levels were found in the prescription drugs ranging from, 0.9–476.1 ng g⁻¹. Non-prescription medicines and dietary supplements were generally lower in Hg content, at 2–45.8 and 0.9–16.7 ng g⁻¹, respectively. Taking account of the recommended daily doses, the total intakes suggested that measurement of Hg in pharmaceuticals is warranted to avoid potentially harmful exposure. Figueiredo et al.,⁹⁶ like Kowalski and Frankowski,¹⁴⁹ referred to national and international regulations for metal impurities in pharmaceuticals and dietary supplements. They however, chose to investigate whether WDXRF spectrometry might be a useful alternative analytical technique to the more usual ICP-MS. To this end, a procedure to measure concentrations of 16 elements was evaluated. With reported LODs of 1.7–16.4 μg g⁻¹, recoveries of 70–150% and RSDs of less than 20%, the method was said to meet the requirements of the European Pharmacopeia for Cu, Cr, Ir, Mn, Mo, Ni, Os, and Pt, and of the United States Pharmacopeia for Ir, Ni, Os and Pt. Kauffman and colleagues at the US FDA¹⁵⁰ considered the contribution of pharmaceutical excipients to the concentrations of elemental impurities in medicines regulated by the US Pharmacopoeia Chapter 232 and the International Conference on Harmonization's Q3D Guideline on Elemental Impurities. A method involving microwave digestion and ICP-MS was used to survey 190 samples from 31 different excipients and 15 samples of 8 drugs. Their results indicated that low levels of elemental impurities were present in the materials tested.

Cordeau et al.⁶⁰ developed a method using HPLC-ICP-MS for pharmacology studies. This approach was adopted to avoid use of radio-labelling and to provide low LODs. The approach was successfully tested by determining the dissociation constant of Se-labelled vasopressin with the V1A receptor system and comparing results with conventional radioactivity assays. Interest in the medical applications of NP technology is developing rapidly. To overcome resistance to multinuclear Pt-prodrugs, Gao et al.¹⁴² tested NP-mediated delivery of polymer-di-cisPt(IV) to A2780DDP cells. Measurements of Pt by ICP-MS showed that more drug was present in cancer cells and greater number of Pt-DNA adducts were formed. Additional experiments proved that uptake of the NPs into cells involved a mechanism of endocytosis rather than by passive diffusion or the Cu transporter-1 mediated active transportation, thus providing an explanation for how the polymer-di-cisPt(IV) micelles overcome cisplatin resistance. Consideration of possible neurotoxicity from widely used AgNPs was addressed by Strickland et al.¹⁵¹ An experimental in vitro neural network model using rat primary cortical cells was established with recording of spontaneous and pharmacologically-induced (bicuclline, a GABA antagonist) action potential spikes and active electrodes. This model was then tested against AgNO₃ and AgNPs. A greater number of active electrodes were recorded following exposure to citrate- and PVP-coated 10 nm AgNPs. The number of action potential spikes increased following bicuclline treatment when exposed to 75 nm AgNPs. Application of AgNO₃ produced fewer spontaneous or pharmacologically-induced action potential spikes and active electrodes indicating that Ag^II does not mediate these effects. These results demonstrate that, in vitro, non-cytotoxic concentrations of 10 nm citrate- and 75 nm PVP-coated Ag neuroparticles alter neural network function.

8 Applications: foods and beverages

Table 2 summarises developments associated with foods and beverages.

Table 2 Foods and beverages

Element	Matrix	Technique(s)	Sample preparation/comments	Ref.
Ag	Plastic food containers	ICP-MS; TEM	The release of Ag nanoparticles from four commercially available plastic food containers was investigated using different food simulant solutions (Milli-Q grade water, 3% acetic acid and 10% EtOH) and measuring total amount of Ag by ICP-MS and particle size and number concentration by single-particle ICP-MS and TEM energy dispersive spectrometry	67
Ag	Plastic food containers	ICP-MS; SEM-EDX	After confirming the presence of Ag nanoparticles by SEM-EDX, the release of Ag nanoparticles from a plastic baby feeding bottle and a food box was evaluated using food simulant solutions (water, 3% v/v acetic acid, and 10% and 90% v/v EtOH) between 20 and 70 °C for up to 10 days using single-particle ICP-MS	66
Ag	Radish (Raphanus sativus) sprouts	ICP-AES	The effects of AgNPs on crop plants was investigated studying radish (Raphanus sativus) sprouts. The uptake of Ag and mineral nutrients, quantified by ICP-AES, was compared with other parameters. Ag uptake reduced the water content, the root and shoot lengths (above 500 mg L^−1 Ag) and the levels of B, Ca, Cu, Mg, Mn and Zn, compared with the control plants	179
Ag	Water and fruit (apple and orange) juices	ICP-MS	A method for the determination of Ag and Au nanoparticles with single-particle ICP-MS in water and fruit juices, after dilution with a digestion reagent, was validated. Particle size were determined with bias and precision <15%	68
As	Drinking water	XRF	To achieve the on-site quantitation of total As in drinking water, a 50 mL sample, adjusted to pH 3 with diluted HCl, was passed through a Ti and Zr-loaded carbon disk, subsequently examined by mobile XRF spectrometry. The LOD was 2.0 μg L^−1	91
As	Drinking water	HPLC-ICP-MS	A procedure is proposed for the multielemental speciation of As^III, As^V, Cr^VI, Sb^III and Sb^V in drinking water using anion-exchange HPLC-ICP-MS, with a DRC and O2 as the reaction gas. Five species (H3AsO3, H2AsO4^−, SbO2^−, Sb(OH)6^−, CrO4^2−) were fully separated within a 15 min run, with gradient elution using 3 mM EDTANa2-36 mM NH4NO3 as mobile phase. The reported LODs ranged from 0.038 (Sb^V) to 0.098 μg L^−1 (Cr^VI) and recoveries were between 91% and 110%	106
As	Environmental samples, biological fluids and food samples	HPLC-ICP-MS; HPLC-HG-AFS	A review addressing conditions to preserve the integrity of inorganic and organic As species during collection, storage, sample preparation and measurement. In most cases optimal conditions depend on the As species and the sample type	17
As	Fish, rice, CRMs	ICP-MS; HPLC-ICP-MS	Enzymatic methods (based on trypsin or cellulase) for the extraction of As and As species from food were evaluated by comparison with microwave-assisted digestion with HNO3–H2O2, using CRMs and test samples (fish and rice). Measurements were carried out by ICP-MS and HPLC-ICP-MS	180
As	Garlic	ETAAS	Arsenomolybdate complexes (As^V) forming an ion pair with trihexyl(tetradecil)phosphonium chloride were extracted on 1.0 mg of multi-walled carbon nanotubes dispersed in the same ionic liquid. After centrifugation and re-dispersion with tetradecyltrimethylammonium bromide as a surfactant, an aliquot was injected into the graphite furnace. The method allowed the separation of As^III and As^V. The LOD was 7.1 ng L^−1 and RSDs (n = 6) at 5 μg L^−1 were 5.4% and 4.8% for As^III and As^V, respectively	50
As	Rice	HG-AFS; HPLC-ICP-MS	After extraction of total As from rice using HNO3, As^V was reduced to As^III using thiourea. Inorganic As was then separated from organic As using polystyrene resin cartridges and quantified by HG-AFS. An LOD of 1.1 μg kg^−1 and RSDs <6% were achieved. The method was validated by analysis of CRMs and comparison with HPLC-ICP-MS	181
As	Rice	HPLC-ICP-MS	The content of iAs in 30 rice samples was determined using a fast screening method based on the reaction of AsH3 with HgBr2 (Gutzeit method) and compared with the results obtained by HPLC-ICP-MS. No bias was observed. The LOQ was 50 μg kg^−1 and RSD was 12%	182
As	Rice	—	A review of methods for the determination of As in rice, discussing their advantages and weaknesses and reporting recent developments	152
Au	Water and fruit (apple, orange) juices	ICP-MS	See Ag, ref. 68	68
Br	Bread	LIBS; ICP-MS	Br levels from 157 to 451 ppm were observed in loaf bread samples using UV pulsed LIBS. The results were confirmed by ICP-MS	79
Br	Drinking water, SRMs	IC-ICP-MS	BrO3^− and Br^− in drinking water samples were separated by IC using isocratic elution with 40 mM KOH. Enhanced sensitivity was achieved by measuring BrO^+, formed after reaction with N2O in a DRC, at the interference free m/z of 95. For BrO3^−, an LOD of 0.013 μg L^−1, and RSD (n = 5, 0.5 μg L^−1 BrO3^−) ranging from 2.1 to 4.5% were achieved	57
Br	Milk whey proteins	ICP-MS	500 mg of whey protein concentrate, hydrolysate or isolate were digested with 25 mM NH4OH using microwave-induced combustion. The determination of the content of Br and I in NIST SRM 8435 “Milk powder” gave recoveries of 102% and 105%, respectively. The method provided LOQs seven (Br) and 281 (I) times lower than IC	183
Cd	Bovine liver CRM, food samples	FAAS	Pre-concentration of Cd from food samples was achieved by precipitation of Cd at trace levels with 1-(2-thiazolylazo)-2-naphthol at pH 8, followed by phase separation, dissolution of the remaining solid particles in conc. HNO3 and microinjection into the FAA spectrometer. A LOD and a LOQ of 0.25 μg L^−1 and 0.83 μg L^−1, respectively, were achieved. An RSD of 5.5% was reported	51
Cd	Dietary products CRMs	ETV-CCP-AES; ETAAS	An ETV-capacitively coupled plasma-AE microspectrometer was tested by comparison with ETAAS for the determination of Cd in food samples after microwave-assisted digestion with an HNO3–H2O2 mixture. Analytical performance complied with EU regulation for food control whereas lower power (15 W) and lower Ar consumption (150 mL min^−1) were necessary to operate this instrument	72
Cd	Food CRMs	ETV-ICP-MS; ICP-MS	Solid sampling of food for Cd determination by means of ETV-ICP-MS was achieved using a tungsten coil trap, a porous carbon vaporiser and an online Ni–Cr ashing furnace. An LOQ of 0.5 pg was reported. The RSD was <10.0% (n = 10) and average recovery for spiked food samples ranged from 99.4% to 105.9% (n = 6). No significant difference was observed for CRMs in comparison with conventional microwave digestion	184
Cd	White and brown meat of warty crab (Eriphia verrucosa)	AAS	Brown meat of warty crab (Eriphia verrucosa) showed significantly high Cd concentrations (up to 5.629 mg kg^−1 wet weight; mean value, 1.465 mg kg^−1 wet weight). Most people only consume the white meat, where Cd concentrations were found to be below LOD. However, in areas where whole crustaceans are traditionally eaten, substantially increased Cd intake may occur	164
Co	Water, cereal, powdered beverage and fruit samples, CRMs	FAAS	125 μL of 1-decanol, 450 μL of THF and 300 mL of 0.1% w/v DDTC were used for the extraction of the DDTC–Co^II complex into the nano and molecular size micelles formed by the mixture 1-decanol–THF at pH 6, followed by centrifugation for 8 min. With a pre-concentration factor of 30, an LOD of 1.89 μg L^−1 and an LOQ of 6.32 μg L^−1 were achieved. The RSD was 1.51% (n = 8)	49
Cr	Drinking water	HPLC-ICP-MS	See As, ref. 106	37/05518
Cr	Drinking water CRM	DLLME-LIBS	Multivariate analysis was applied to the optimisation of experimental factors (pH, DDTC concentration, extractant and volume of disperser solvents) affecting the speciation of Cr^VI by means of DLLME-LIBS	76
Cr	Infant, paediatric, and adult nutritional formulas	ICP-MS	As part of AOAC activities, a method for the determination of Cr, Mo and Se in infant, paediatric, and adult nutritional formulas by ICP-MS with a collision/reaction cell was validated by a collaborative study involving 9 participants, sample preparation involved microwave digestion with HNO3, H2O2 and internal standard. The repeatability RSD ranged from 1.0 to 7.0% and the reproducibility RSD from 2.5 to 13.4%	29
Cr	Water (drinking water and still bottled water)	ICP-MS; HPLC-ICP-MS	Two IDMS procedures were developed and validated for the determination of the concentration of total Cr, by ICP-MS, and Cr^VI, by HPLC-ICP-MS. A DRC was applied to remove interferences	185
Cu	Beverages and foods, SRMs	FAAS	The pre-concentration of Cu by means of ultrasonic-assisted CPE was achieved using Safranin T as the ion-pairing reagent and the non-ionic surfactant, poly(ethyleneglycol-mono-p-nonylphenylether) and NH4Cl as the extracting and salting out agents. LODs of 6.10 ng L^−1 (Cu^II at pH 10) and 24.4 ng L^−1 (Cu^I at pH 8.5) were achieved. RSD was in the range of 2.15–4.80% (n = 5)	186
Cu	Drinking water and walnuts	FAAS	An enrichment factor of 640 was achieved with pre-concentration of Cu^II on a column filled with a resin, followed by elution with 5 mL of 3 M HNO3, and the use of a slotted quartz tube. The LOQ was 0.23 μg L^−1. Analysis of NIST SRM 1570a (spinach leaves) provided acceptable results	187
Cu	Sea water, biological and food samples, CRMs	FAAS	A novel sorbent media (hydroxyapatite nanorods) was prepared from recycled eggshell by precipitation and used for SPE. Cu^2+, Pb^2+ and Zn^2+ were quantitatively sorbed on this matrix between pH 6 and 9 with a pre-concentration factor of 250. The LOQs (10s) for Cu^2+, Pb^2+ and Zn^2+ were 2.40, 17.06 and 1.83 μg L^−1, respectively. Blood and urine CRMs were analysed	46
F	Baby food, SRMs	HR-CS-ETAAS	With an unusual procedure, the concentration of F in various baby foods was determined, via solid sampling, measuring the molecular absorption at 606.440 nm, due to CaF generated in the graphite furnace using 1000 °C (pyrolysis) and 2200 °C (CaF forming step). The LOD was 0.20 ng F and the characteristic mass was 0.17 ng F	71
Gd	Drinking water	HPLC-ICP-MS	Improved sensitivity was reported for the speciation of Gd from iatrogenic sources in drinking water. Sample introduction involved USN combined with HILIC with a diol-based stationary phase. For frequently applied Gd-based contrast agents, LODs <20 pmol L^−1 were achieved	168
Hg	Fish and seafood	GC-ICP-IDMS	A method for the determination of MeHg in foodstuffs of marine origin by GC-ICP-IDMS after dissolution, derivatisation and extraction of the samples was validated in a collaborative inter-laboratory study. The 8 participants analysed mussel tissue, squid muscle, crab claw meat, whale meat, cod muscle, Greenland halibut muscle and a CRM based on dogfish liver (NRCC DOLT-4), over a range of MeHg concentrations ranging from 0.035 to 3.58 mg kg^−1 (as Hg) dry weight. Repeatability RSD ranged from 2.1% to 8.7% and reproducibility RSD from 5.8% to 42%	31
Hg	Seafood	CVAAS	A novel Fe3O4@SiO2@polythiophene magnetic nanocomposite was applied to the preconcentration of Hg^2+ from seafood samples prior to determination by CVAAS. An LOD of 0.02 ng mL^−1 and RSD <9.2% were observed. The sorption capacity was assessed as 59 mg g^−1	155
Hg	Seafood	CV-AFS	Using a home-made electrochemical flow cell, Hg atoms were generated by electrolytic reduction on the surface of a glassy carbon cathode and then measured by AFS. The method was applied to the determination of MeHg in seafood with an LOD of 1.88 × 10^−3 ng mL^−1 (MeHg). Recoveries from spiked samples ranged from 87.6% to 103.6% and the RSD was <5%. MeHg levels in local seafood samples were between 3.7 and 45.8 ng g^−1	154
Hg	Sushi, CRMs	AAS	The total Hg content was determined on 60 mg of fresh sushi samples by thermal decomposition amalgamation AAS using a direct mercury analyser. The recoveries of certified values from various CRMs ranged between 94% and 112%. LOD and LOQ were 0.4 and 1.4 μg kg^−1, respectively	188
Hg	Tuna fish CRM	CV-AAS; ETAAS; AAS	A candidate tuna fish CRM for Hg and MeHg mass fractions was characterised using different analytical techniques (CV-AAS, solid sampling ETAAS and direct mercury analyser)	37
Hg	Water	XRF	An LOD of 54 ng L^−1 Hg was achieved, with an enrichment factor of 127, when measuring Hg by total reflection XRF. The sample treatment involved generation of Hg vapour by online photo-UV reduction in the presence of 2 M acetic acid and its trapping, as a result of amalgamation, onto quartz substrates coated with nanostructured palladium. Repeatability RSD at 1 μg L^−1 was 3%	94
Hg	Water, CRMs	HR-CS-ETAAS	A procedure for the pre-concentration, by a factor of 15, of Hg from water samples, based on ultrasound-assisted dispersive micro SPE with 10 mg silver nanoparticles as a solid sorbent, was reported. An LOD of 0.005 ng mL^−1 was achieved. The RSD for real water samples ranged from 6 to 11%	52
I	Food samples, CRMs	ICP-MS	I concentrations were measured in 7 SRMs and 21 fortified food samples ICP-MS following alkaline extraction and addition of 3% 2-propanol. The reported LOQ in solid food samples (unusually calculated as 30 s of the method) was 36 ng g^−1. RSDs ranged from 2% to 7% and the observed bias was between 81% and 119%	158
I	Infant and adult/paediatric formulas	ICP-MS	As part of the AOAC activities, a method for the determination of total I in nutritional formulas was validated in a collaborative study. 13 laboratories analysed infant and adult nutritional products including low-fat, soy-based, and milk-based formulas as well as the NIST SRM 1849a. The method involves microwave-assisted digestion of the samples with a KOH solution, followed by addition of NH4OH and sodium thiosulfate, filtration, dilution and analysis by ICP-MS. The repeatability RSD ranged from 0.77% to 4.78% and the reproducibility RSD from 5.42% to 11.5%	30
I	Milk whey proteins	ICP-MS	See Br, ref. 183	183
I	Tomato fruits	ICP-AES	The possibility of I supplementation via coating of tomato fruits with an edible film of I-doped chitosan was assessed by characterisation of the edible film by FTIR and SEM, I content and I stability studies, assessment of tomato antioxidant activity, freshness and storability	161
Mn	Drinking water, components of drinking water distribution systems	XANES; XRF	The presence and possible resuspension of Mn from mineral deposits on brass and lead components of drinking water distribution systems was assessed in situ using synchrotron-based micro XANES and μXRF mapping	166
Mo	Infant, paediatric, and adult nutritional formulas	ICP-MS	See Cr, ref. 29	29
Mo	Milk, vegetables, foodstuffs, SRMs	FAAS	Ultrasonic-thermostatic-assisted CPE allowed faster and simpler pre-concentration of Mo and V from milk, vegetables and foodstuffs, as compared to traditional CPE. Stable anionic oxalate complexes of Mo^VI and V^V were associated with [^189phenoxazin-5-ylidene]azanium; sulfate (Nile blue A) at pH 4.5, and, as such, extracted into the micellar phase of polyoxyethylene(7.5)nonylphenyl ether (PONPE 7.5). Enhancement factors of 115 (Mo^VI) and 145 (V^V) were reported, as well as RSDs ≤3.5% and recoveries of spiked amounts ranging from 95.7% and 102.3%	189
Na	Tomato ketchup sauce	FAAS; FAES; ICP-AES	Salt content has to be clearly indicated on food labels in the EU. Three sample preparation methods (dispersion in water, wet digestion and dry ashing) and four analytical techniques (FAAS, FAES, ICP-AES and thermometric endpoint titrimetry) were compared using tomato ketchup as a sample	190
Pb	Sea water, biological and food samples	FAAS	See Cu, ref. 46	46
Pb	White-tailed deer, moose	ICP-MS	The content of Pb in 72 samples of meat from animals killed by lead ammunition was determined as part of an investigation of potential Pb exposure for consumers in Canada. The mean Pb levels of 0.28 and 0.17 mg kg^−1 in white-tailed deer (n = 35) and moose (n = 37), respectively, were associated with potential health risk for adults and children consuming such meat frequently	140
Pb	Wine	ICP-MS	Pb concentration and the ratios ^204Pb : ^206Pb, ^207Pb : ^206Pb and ^208Pb : ^206Pb, determined in 94 red wines produced in various regions of Brazil, were reported. Wine samples (1 g) were reacted with 1 g conc. HNO3 and then diluted to 10 g with ultrapure water	170
Pd	Water, soil, and food samples	FAAS	The determination of Pd was achieved with an enhancement factor of 20.7 by DLLME followed by FAAS. Pd^II complexes with 2-mercaptobenzimidazole were extracted in CHCl3 and subsequently dispersed in acetone. The LOD (3s) was 8 μg L^−1 and an RSD (n = 6) <4.8% was reported	54
Rh	Food	ETAAS	A preconcentration procedure, based on adsorption of the Rh 2-(5-bromo-2-pyridylazo)-5-diethylaminophenol/tetraphenylborate ion associated complex at the surface of an alumina column, was applied to the determination of Rh levels in food	191
Sb	Drinking water	HPLC-ICP-MS	See As, ref. 106	106
Se	Biofortified yeast	HPLC-HG-MIP-AES	A method was developed to determine Se species (SeMet and Se^IV) in hydrolyzed yeast, based on ion-pair RP-HPLC, post-column HG and N2 sustained MIP-AES. Freeze-dried biomass was hydrolyzed with methanesulfonic acid. A 0.08% v/v heptafluorobutyric acid in MeOH was used as the mobile phase. After addition of 6% m/v K2S2O8–3% m/v NaOH and heating at 60 °C in a water bath, HG was achieved in the presence of 10 M HCl, NaBH4 2% m/v and NaOH 0.3% m/v	73
Se	Feed, milk	AAS	Supplementing dairy cow feed with Se yeast (0.3 mg of Se kg^−1 of dry matter) increased Se levels in milk and cheese without affecting their quality	156
Se	Food (fish, shellfish, meat, dried seaweeds), CRMs	ICP-MS	Satisfactory recovery was obtained for Se from food samples, dried to powder and subsequently mixed with an equivalent amount of water and heated in a sealed vessel in an electric oven at 220 °C for 120 min	157
Se	Infant, paediatric, and adult nutritional formulas	ICP-MS	See Cr, ref. 29	29
Se	Seafood	HPLC-IC-HG-AFS	Se-Cys, Se-Met and Se^IV were separated in 10 min on an anion-exchange column (PRP-X100), using a mixture (39 : 1, v/v) 30 mM NH4H2PO4–MeOH at the flow rate of 1.0 mL min^−1. Analytical conditions for HG-AFS detection were chosen as: 2.0% KBH4, 5.0% HCl; lamp current, 90 mA; photomultiplier tube voltage, 280 V; flow rates of carrier and shielding gas 300 mL min^−1 and 800 mL min^−1. LODs for Se-Cys, Se-Met and Se^IV were 1.66, 0.990, 1.10 μg L^−1, respectively. The repeatability RSD was <5.0% (n = 10) and recoveries from spiked seafood samples ranged from 87.3% to 103%	165
Se	Se-enriched sunflower oil	HPLC-ICP-MS	A combination of analytical techniques (HPLC-ICP-MS, using ^78Se for detection, and ES Orbitrap MS) was applied to the speciation of Se in Se-enriched sunflower oil. A non-aqueous mobile phase gradient was used for both RP-HPLC-ICP-MS and EI-MS	63
Si	Food additive (synthetic amorphous silica)	ICP-MS; TEM	A combined approach, using various analytical techniques (dynamic light scattering, multiangle light scattering, asymmetric flow-FFF, ICP-MS and TEM) was applied to the detection and characterisation of Si nanoparticles in synthetic amorphous silica used as food additive E551	192
V	Milk, vegetables and foodstuffs	FAAS	See Mo, ref. 189	189
Zn	Sea water, biological and food samples	FAAS	See Cu, ref. 46	46
Various (4)	Candy products	ICP-MS	The content of Al, Cd, Ni and Pb, assessed in 263 samples of 12 types of candy products, by ICP-MS after acid mineralization, were in the ranges 21.28–62.91 μg g^−1 (Al), 0.12–1.01 μg g^−1 (Cd), 0.40–1.27 μg g^−1 (Ni) and 1.03–7.14 μg g^−1, (Pb), respectively	159
Various (5)	Milk powder	ICP-AES	Three acid digestion procedures (open system with a cold finger); conventional open system (4 h at 120 °C) and microwave digestion (heating ramp to 180 °C for 45 min) were evaluated for the preparation of milk powder samples prior to determination of Ca, Cu, Fe, K and Na by ICP-AES	193
Various (5)	Vegetable sprouts of garden cress (Lepidium sativum), mung beans (Vigna radiata), soybeans (Glycine max)	HR-CS-AAS	The possibility to increase the mineral content of vegetable sprouts was investigated using hydroponic nutrient solutions enriched with Ca^2+, Cr^3+, Mg^2+, Se^2+and Zn^2+. Elemental concentrations were determined in dried plant material by HR-CS-AAS after microwave-assisted extraction. Instrumental LODs were reported as 0.010, 0.029, 0.013, 0.151 and 0.030 mg L^−1 for Ca^2+, Cr^3+, Mg^2+, Se^2+ and Zn^2+, respectively, with recoveries >98% and RSDs <5%. Garden cress showed the highest bioconcentration factor	162
Various (7)	Water, meat products, and baby food, CRMs	FAAS	The pre-concentration, by a factor of 100, of Cd^II, Co^II, Cu^II, Fe^III, Cr^III, Pb^II and Zn^II by means of SPE on a poly[N-(3-methyl-1H-indole-1-yl)]-2-methacrylamide-co-2-acrylamido-2-methyl-1-propane sulfonic acid-co-divinylbenzene resin as sorbent at pH 8 was reported. LODs between 0.12 and 1.6 μg L^−1, RSD (n = 10) ≤1.8% and recoveries between 95% and 102% were achieved	194
Various (10)	Gluten-free foods	ICP-MS	A survey of the content of 10 elements (As, Cd, Co, Cr, Cu, Fe, Mn, Ni, Pb and Zn) in gluten-free foods was carried out and the results compared to those of other studies	160
Various (10)	Infant formula, adult nutritionals, and milk products	ICP-MS	As part of AOAC activities a method for the simultaneous determination of Al, As, Cd, Co, Cr, Hg, Mo, Pb, Se and Sn in infant formulas, adult nutritionals, and milk based products was developed combining features of existing standard methods	27
Various (11)	Food (milk powder, chocolate powder, and soluble coffee samples), milk CRMs	HR-CS-FAAS; ICP-MS	A closed-vessel conductively heated digestion system (CHDS) was evaluated for the digestion of milk powder, chocolate powder and soluble coffee samples. The concentrations of Ca, Cu, Fe, K, Mg, Mn, Na, and Zn were measured by HR-CS-FAAS and those of Cd, Mo and Se by ICP-MS/MS. When compared with microwave oven digestion, CHDS gave comparable performances on both CRMs and test samples	195
Various (12)	Infant, paediatric and adult nutritional formulas	ICP-MS	As part of AOAC activities, a method for the determination of 12 elements (Ca, Cr, Cu, Fe, K, Mg, Mn, Na, Mo, P, Se and Zn) in infant, paediatric and adult nutritional formulas by ICP-MS after microwave-aided digestion was developed and assessed. The repeatability RSD was <5% for all elements and intermediate precision RSD was <5% for 9 of them. Recoveries for spiked matrixes and CRMs varied from 90.1% to 109%	28
Various (19)	Organic and conventionally grown rice	ICP-MS	An approach combining information on the levels of concentration of 19 trace elements and a data mining technique (support vector machine) allowed to predict the authenticity of organic rice samples with an accuracy of 98% when using the whole dataset and of 96% when using only the elements Ca and Cd	171
Various (halogens)	Food and biological materials	ICP-AES; ICP-MS	A review, over the last 20 years, of sample preparation techniques best suited for the determination of halogens in food and biological materials by ICP-AES and ICP-MS	18
Various	Clinical and biological materials, foods and beverages	—	An ASU review of advances in the field	1
Various	Food	XRF	A comprehensive overview of the applications of TXRF in the field of food analysis	95
Various	Food	AFS	A review of advances in instrumentation and novel applications of AFS, including food analysis	22
Various	Food simulants	—	An investigation of metal release from 8 coffee machines and 11 electric kettles identified increased release after decalcification, especially for Cr, Mn, Ni, Pb and Zn, exceeding, in some cases, the limits proposed by the European Council	196

---

8.1 Progress for individual elements

8.1.1 Arsenic. It is recognised that regulations for maximum allowable concentrations of As in foods, drinking water and animal feed are unsatisfactory in that they refer to total As rather than to the toxic species. Regulation for As in rice has recently been introduced by the EU. In a most timely contribution to this challenging topic, Petursdottir et al.⁸¹ discussed the toxicity of As species, regulations and the implications for analytical techniques. Reference is made to a strategy, previously suggested by these authors, whereby analysis is directed to the determination of (i) toxic iAs, (ii) non-toxic arsenobetaine and (iii) a potentially toxic fraction (all other organoarsenicals).

Access to reliable testing equipment in the field can be very beneficial in providing regulatory decision making information rapidly to a local community. Discussed in previous years' review as a laboratory technique, Hagiwara et al.⁹¹ described an effective on-site technique for quantitation of As in drinking water using a Ti–Zr loaded carbon disc SPE followed by mobile XRF spectrometry. Both iAs and organic As species were adsorbed and an LOD of 2 μg L⁻¹ was reported.

Though development of these field kits and low-cost laboratory methodologies are still prevalent, we now appear to be in a time for reflection on As analysis on water and food. Costa et al.¹⁵² presented a general review (74 references) on analytical strategies while Maher et al.¹⁷ provided a significant appraisal of (479 references) measurement by HPLC-ICP-MS and HPLC-HG-AFS and also cautioned that the As species determined should not be artefacts of the preservation or extraction procedures. Yogarajah and Tsai¹⁵³ reviewed (172 references) options for the detection of As in drinking water in terms of novel sensors and concluded microfluidic devices still have a bright future.

8.1.2 Mercury. Using EC reduction on a glassy-carbon electrode in CV-AFS, a notable improvement in the LOD for MeHg determination in seafood was obtained by Zu and Wang.¹⁵⁴ The EC cell (70 mm × 4 mm × 0.5 mm; 0.14 mL optimised volume) comprised a bar of glassy carbon and a slice of Pt embedded in PTFE blocks. The optimum sample volume was 0.5 mL injected into an electrolytic flow and the total analysis time was 60 s. The authors noted a greater sensitivity towards MeHg (compared to iHg) with an LOD of 1.88 × 10⁻³ ng mL⁻¹ and the RSD was 2.0% at 2 ng mL⁻¹ MeHg. The improvement in selectivity towards Hg^II by a novel Fe₃O₄–SiO₂ magnetic nanocomposite was described by Abolhasani et al.¹⁵⁵ Tetraethyl orthosilicate was anchored to Fe₃O₄-NPs which were then modified through polymerisation of thiophene in the presence of KMnO₄. The particles were characterised by FTIR, TEM and XRD and the capacity of this new nanosorbent was 59 mg g⁻¹. Optimised conditions were pH 5.2; 6.8 min extraction time; 26 mg nanosorbent. Elution was achieved using 3.0 mL 0.3 M thiourea in 0.34 M HCl over 8.4 min. Under these conditions the LOD was 0.02 ng mL⁻¹ and the RSD was <9.2% using CV-AAS.

8.1.3 Selenium. Eleven new species of Se ^IV -triglycerides in sunflower oil that have been shown to confer antioxidant and anticancer properties, were elucidated by Bierla et al.⁶³ using a new non-aqueous gradient with RP-HPLC-ICP-MS. The polar fraction of selenised sunflower oil was obtained by partitioning the sample between hexane and a MeOH–H₂O mixture (9 : 1). The MeOH fraction recovered for analysis was dissolved in ACN-butan-2-ol mobile phase. The effect of Se supplementation to cattle feed on the Se distribution in cow milk and the quality of mozzarella cheese, was studied by Liu et al.¹⁵⁶ Addition of 0.3 mg Se kg⁻¹ dw resulted in a Se concentration in milk of 35.81 μg L⁻¹ and the distribution in different milk fractions of fat, casein, and whey protein were 9.82, 45.56, and 44.62%, respectively. Whereas the composition and texture of Se-containing mozzarella did not change compared to the control, physical properties such as meltability, flowability and stretchability were improved following eight weeks of storage, with resultant lower pH and water activity. After storage, most of the Se species detected by HPLC-ICP-MS were SeMet and SeCys. A simple extraction method based on sub-critical water, for the determination of Se in food samples by ICP-MS, was described by Ohki et al.¹⁵⁷ The powdered seafood, meat or seaweed samples (0.1 g) were mixed with H₂O and carbon fibre in a sealed PTFE vessel and heated to 220 °C for 2 h. After filtration, the supernatant was directly analysed by ICP-MS yielding an LOD of 0.01 μg g⁻¹. Dissolution of insoluble Se after 2 h was confirmed by SEC-ICP-MS.

8.1.4 Other elements. Analysis of halogens in foodstuffs, particularly F in baby food, is of great importance but can often involve long sample preparation. An improved approach has been developed with solid sampling HR-CS-ETAAS using the molecular absorption of CaF.⁷¹ Digestion of the sample was avoided by drying at 100 °C then introduction into the graphite tube along with 10 μL 4000 μg L⁻¹, Ca solution. Pyrolysis was carried out at 1000 °C and molecular vaporisation was set at 2200 °C during which the molecular absorbance for CaF was measured at 606.440 nm. The LOQ of the method was 3.75 μg g⁻¹ and the measured F content of a CRM was within its 95% confidence interval. Analysis of food samples for I can be compromised by incomplete alkaline digestion or from adsorption of I onto plastic- or glassware. Todorov and Gray,¹⁵⁸ therefore, revisited and modified hotplate digestion with TMAH by adding 3% propan-2-ol as C modifier and surfactant to enhance signal and minimise surface adsorption. The LOD of 36 ng g⁻¹ was sufficient for the determination of I in food samples. Bias, measured using NIST SRMs, ranged from 85% to 102% and reproducibility was 1–7% RSD.

The migration of antimicrobial AgNPs from plastic packaging into food matrices continues to be of concern. Mackevica et al.⁶⁷ studied the migration from four brands of container into H₂O, 10% EtOH or 3% acetic acid, migration solutions. Determination of total Ag was by ICP-MS and AgNP size was by single particle-ICP-MS. The total Ag ranged from 13 to 42 μg g⁻¹ with acetic acid leaching the highest amount, releasing 3.1 ng cm⁻². A very similar study,⁶⁶ also using ICP-MS, was carried out to compare plastic material from a baby bottle and a food box, using similar food simulants to the above study, but conducted at 20, 40 and 70 °C. With the increase in use of AgNPs in agriculture, an appropriate detection and quantification is required to ensure crop health and food safety. Bao et al.⁶⁵ used an enzyme to release AgNPs from plant material (modelled using Arapidopsis thaliana) followed by single particle-ICP-MS to elucidate deposition patterns and possible translocation towards aerial parts of the plant. Macerozyme R-10, a complex enzyme from Rhizopus sp., containing cellulase, hemicellulase, and pectinase, was used at 37 °C for 24 h, to liberate NPs from roots and shoots. The authors found that in treated plants, NPs accumulate predominantly at the middle lamella and cell walls in root tissue and that some AgNPs can be translocated toward the leaves. This will obviously be relevant to plants where the leaves are consumed.

8.2 Progress in applications

8.2.1 Dietary intake studies. Sugar candy is not usually considered as a component of a staple diet, requiring the dietary intake of heavy metals (Al, Cd, Ni, Pb) to be studied. However, Marin-Martinez et al.¹⁵⁹ analysed 263 samples of 12 different types of candy available in Spain. Determined by ICP-MS following digestion with HNO₃–HClO₄ with V₂O₅ as a catalyst, concentration ranges were 21.28–62.91 μg g⁻¹ Al, 0.12–1.01 μg g⁻¹ Cd, 0.40–1.27 μg g⁻¹ Ni and 1.03–8.14 μg g⁻¹ Pb. The candy consumption of children was assessed, through questionnaires, (alarmingly showing children in the age range 1–2 years eat the largest amounts) but the authors did not calculate intake per unit body mass.

8.2.2 Human milk and infant formula. Several multielement methods have been developed with determination by ICP-MS. The procedure developed by Mohd-Taufek et al.⁴³ analysed low volumes (0.2 mL) of human milk, for studies of pre-term mothers whose milk production is low. The method required digestion with EDTA, NH₄OH solution, isopropanol and Triton X-100 followed by simultaneous determination of eight trace elements by ICP-MS. Helium (4.5 mL min⁻¹) was used as the collision gas. Method validation, using pooled human milk samples, showed good repeatability and reproducibility values of <7 and <6%, respectively, and accuracy using NIST 1549 was within 80–120%. Concentrations, (mean 20SD) from 12 samples, for Br, Cu, Fe, I, Mn, Mo, Se, were 812.6 ± 128.7, 220.8 ± 32.9, 47.3 ± 99.9, 113.5 ± 17.1, 1.37 ± 0.14, 0.37 ± 0.12, 14.3 ± 5.8, 1390.6 ± 211.5 μg L⁻¹, respectively. The authors claim this is the first indication of high Br in human milk.

A similar ICP-MS method in infant formula, adult nutritional and milk products with prior high pressure acid digestion was reported by the Nestlé Research Centre.²⁷ This single-laboratory validation met all AOAC criteria for method performance. An AOAC First Action Method (2011.19) for the determination by ICP-MS of Cr, Mo and Se in infant formula and adult nutritional products was recorded²⁹ and involved sample preparation by microwave-assisted digestion with HNO₃ and H₂O₂. Repeatability and reproducibility ranged from 1.0 to 7.0% and 2.5 to 13.4% across all elements.

8.2.3 Cereals, flour and rice. As gluten-free cereal products become more readily available for patients with Coeliac disease these items require monitoring. Mehder et al.¹⁶⁰ measured concentrations of As, Cd, Co, Cr, Cu, Fe, Mn, Ni, Pb and Zn in commercial goods available in Saudi Arabia, using ICP-MS after acid digestion. It was found that Fe and Mn were at higher concentrations than has been reported in previous studies. Given the use of KBrO₃ as an improver and bleaching agent in white flour, the same authors used LIBS (266 nm pulsed laser as the excitation source and a gated intensified charge-coupled device camera as detector) to determine residual Br in white bread.⁷⁹ Calibration was carried out by the method of standard addition using the Br I (827.2 nm) atomic transition line. The LOD was determined to be 5.09 ppm and the RSD was 2.8% (at 20 ppm Br) and the results were within 5% of those obtained using an ICP-MS method. Levels of Br in white bread from the author's local Saudi Arabian market ranged from 157 to 352 ppm, although only four samples were analysed.

8.2.4 Vegetables, fruit, mushrooms and nuts. An approach proposed to prevent thyroid disease in regions where there is I deficiency is to coat mature cherry tomatoes with a thin film (1.5 μm) of I-doped chitosan (CT-I).¹⁶¹ The CT-I film was created by adding chitosan powder (0.75 g) to acetic acid (50 mL, 1% v/v) under vigorous stirring into which KI (100 μg mL⁻¹) was added slowly. Washed and dried tomatoes were dipped into solutions of 0.5% or 1.5% CT-I yielding 0.31 ± 0.04 and 0.38 ± 0.07 μg g⁻¹, respectively, per an average weight of the fruit, as analysed by ICP-AES. At this level only about 540 cherry tomatoes would have to be consumed a day! Another option for trace element supplementation is hydroponic enrichment as demonstrated by Blicharska et al.¹⁶² Nutrient solutions containing Ca, Cr, Mg, Se and Zn were used in the cultivation of garden cress (Lepidium sativum), mung beans (Vigna radiata), and soybeans (Glycine max), either directly or after germination in nutrient solution. Dried plant material following the experiment was subjected to microwave-assisted extraction (10 mL 65% HNO₃ per 0.2 g sample), then measurement of the metal concentrations by HR-AAS. Garden cress had the highest bioconcentration and Zn showed the greatest biofortification and lowest toxicity across all plant types. As the presence of Cr and Mn can have a significant impact on the stability and useful life of vegetable oils, a method for their determination in soybean, canola, sunflower and corn oils after dilution with MIBK was described by Almeida et al.⁷⁰ Two sequential atomisation temperatures (1900 and 2550 °C) were used with a rapid wavelength change (279.4817 and 357.8687 nm) was made possible using HR-CS-ETAAS. The LODs were 1.07 and 1.34 ng g⁻¹, with RSDs being 12 and 8% at 10.0 ng g⁻¹, respectively and recoveries ranged from 79 to 115%.

8.2.5 Fish & seafood. The distribution of protein-bound As between shrimp meat, shell and visceral samples was studied by HG-AFS after alkali extraction-isoelectric precipitation and microwave digestion.¹⁶³ The average concentrations of total As were found to be 0.123, 0.234, and 0.183 μg g⁻¹ while that of As bound to proteins was 0.626, 4.01, and 5.91 μg g⁻¹, respectively; the discrepancy is not explained. Although only the meat is typically consumed, this information is relevant given the financial importance of the industrial recovery of protein from shrimp heads, which makes up 50% of dried head weight, and where As contamination may occur. This trend of visceral organs preferentially bioaccumulating heavy metals was mirrored by Ariano et al.¹⁶⁴ who carried out a complete assessment of the health risk associated with consumption of the Warty Crab (Eriphia verrucosa). European Union regulations for Cd in crab are only set for white meat, ignoring the potentially dangerous contribution from brown meat. Samples of white meat (muscle of claws and appendages) and brown meat (contents of the cephalothorax), collected from the southern Tyrrhenian Sea (Italy), were microwave digested and Cd determined by ETAAS with Mg(NO₃)₂ as matrix modifier. In 40 white meat samples the Cd concentration was below the LOQ (0.55 μg kg⁻¹) but for brown meat the mean Cd level was 1.465 mg kg⁻¹ (SD 1.790 mg kg⁻¹). This Cd concentration correlated positively with carapace width, (cw) (p < 0.05), where mean Cd was 0.322 mg kg⁻¹ when cw < 6.0 cm and 2.729 mg kg⁻¹ where cw > 6.0 cm. Therefore for a 100 g weekly consumption (62% white meat, 38% brown meat) the upper mean Cd concentration for cw > 6.0 cm was 103.70 μg per week equating to 55% of the PTWI set by the European Food Standards Agency. The simultaneous determination of both iSe^IV and organic-Se species (SeCys and SeMet) in seafood using ion exchange HPLC coupled to low cost HG-AFS detection was elaborated by Xie et al.¹⁶⁵ The method comprised the extraction of Se species from freeze-dried samples, in KOH–MeOH, then separation using AEC with 30 mM NH₄H₂PO₄ and MeOH (39 : 1, v/v). Analysis of samples of eight different fish species (Yellow croaker, Grass carp, Water ripples, Catfish, Squid, Hairtail, ‘shellfish’ and Crucian carp) purchased at a market in Whenzhou, Zhejiang Province, China, indicated that iSe was not detected (LOD 1.10 μg L⁻¹) and the mean (SD) concentrations of SeCys and SeMet were 2.67 μg g⁻¹ (0.21) and 1.37 μg g⁻¹ (0.11).

8.2.6 Meat and meat products. In previous ASU reviews, we have seen Pb poisoning highlighted in Scottish grouse or from cows licking Pb-painted feed hoppers; this year Fachehoun et al.¹⁴⁰ described the risk to the 545 938 permit-holding Canadian hunters who collectively harvested 61 067 white-tailed deer and 28 141 moose. The Pb concentrations in meat of volunteered specimens, from cervids shot with Pb ammunition, were determined by ICP-MS. The mean and median levels (deer: 0.28 vs. 0.004 mg kg⁻¹ and moose: 0.17 vs. 0.003 mg kg⁻¹) demonstrated that a normal distribution was not observed due to the fact that concentrations in 40% and 50% of white-tailed deer and moose samples were below the LOQ. Nevertheless, a simulated individual exposure dose of 0.118 μg kg⁻¹ d⁻¹ (95^th percentile of 0.305 μg kg⁻¹ d⁻¹) was calculated using dietary information from self-administered questionnaires.

8.2.7 Drinking water and non-alcoholic beverages. Gerke et al.¹⁶⁶ used μXRF and μXANES to characterise Mn deposits on the lead and brass fittings from two drinking water distribution systems (DWDS) in Ohio, USA, maintained with different disinfectant agents. The ranges of Mn content of the deposits were 1.93–4.45% and 0.21–2.12% w/w, and were in an oxide, an oxyhydroxide and a silicate form. Differences in water treatment (Cl₂vs. NH₂Cl) did not influence the formation of the deposits but in all cases potentially toxic metal ions such as Cr, Cu, Pb and Sr were associated with the deposits. Another disinfectant used in DWDS, O₃, can potentially increase BrO₃⁻ in drinking water, so in order to improve the LOD when using IC, Peng et al.⁵⁷ employed DRC-ICP-MS (with N₂O as reaction gas) with detection of ⁷⁹Br¹⁶O⁺ to avoid background interference from ³⁸Ar⁴⁰ArH⁺ and ³⁹K⁴⁰Ar⁺ ions. The achieved LOD and LOQ were 0.013 and 0.04 μg L⁻¹ for BrO₃⁻ with the RSD (at 0.5 μg L⁻¹) in the range 2.1–4.5%. Analysis of 15 bottled and tap water samples purchased in Wuhan, PRC, revealed that all but three were below the US EPA and Chinese maximum allowable concentrations of 10 μg L⁻¹, with more than half being below the LOD. Total Br ranged from 1 to 47.1 μg L⁻¹. Drinking water contamination due to Tl-bearing pyrites ore from abandoned mines near the spring water source feeding the Tuscan town of Valdicastello Carducci was identified by Campanella et al.¹⁴⁴ The spring water Tl concentration measured by ICP-MS ranged from 5 to 37 μg L⁻¹ but levels in urine collected from inhabitants over a month following the contamination ranged from 0.046 to 5.44 μg L⁻¹ (arithmetic mean 0.74 μg L⁻¹ with SD 0.67 μg L⁻¹; geometric mean was 0.55 μg L⁻¹ and 95^th percentile was 1.88 μg L⁻¹). Urinary Tl levels decreased exponentially with a decrease to 30% of initial value during the first two months but remained twice the Italian population reference value. Hair collected from inhabitants ranged from 1 to 498 ng g⁻¹ (arithmetic mean 41 ng g⁻¹ with SD 68 ng g⁻¹; geometric mean 15 ng g⁻¹ and 95^th percentile 157 ng g⁻¹) and the data reveal that adults (arithmetic/geometric mean 58 ng g⁻¹/23 ng g⁻¹) accumulated more Tl than children (20 ng g⁻¹/8 ng g⁻¹). In the UK, a survey of As in private drinking water supplies in Cornwall, indicated that, in 5% of the wells, the WHO guidance level of 10 μg As L⁻¹ was exceeded.¹⁶⁷ Following the necessary adjustments of urinary As concentrations for AB intake and urinary fluid balance, As profiles revealed the degree of exposure to iAs. The extensive use of chelated-Gd MRI contrast agents—in which typically 1 g Gd per scan is used—has led to their detection in drinking water. A sensitive HPLC-ICP-MS method was developed by Birka et al.¹⁶⁸ who used a diol HILIC column and ultrasonic nebulisation. The LODs were considerably lower than in previous work: 0.6 pmol L⁻¹ for total Gd and for individual contrast agents they were 8 pmol L⁻¹ for Gd-BT-DO3A, 11 pmol L⁻¹ for Gd-DOTA and 14 pmol L⁻¹ for Gd-DTPA. A survey of different waterworks along a 50 km stretch of the Ruhr River in Germany indicated that these compounds were stable to all processing steps across the myriad water purification technologies used. The lack of an integrated multielement speciation analysis for toxic species in drinking water (As^III, As^V, Cr^VI, Sb^III and Sb^V) samples prompted Marcinkowska et al.¹⁰⁶ to propose a HPLC-ICP-DRC-MS method. The approach used an AE column with 3 mM EDTANa₂–NH₄NO₃ as mobile phase/complexing agent and O₂ as reaction gas, allowing samples to be analysed which contain mineral salts up to 650 mg L⁻¹. The LODs were 0.067, 0.068, 0.098, 0.083 and 0.038 μg L⁻¹, respectively. Analysis of only four samples of bottled water showed only the presence of As^V, with only one sample containing Cr^VI.

8.2.8 Alcoholic beverages. Although this section normally contains details of trace metals in wine, as well as beer and spirits, have a care for the workers in those primary industries who may be exposed to heavy metals when preparing wine such as Rio Grande do Sul. Rocha et al.¹⁶⁹ analysed the blood serum of 54 vineyard farmers and 104 controls, in Brazil, to determine As, Cu, Mn, Ni, Pb and Zn by ICP-DRC-MS. All concentrations were two- to four-fold greater than in the control samples (P < 0.05) and correlated with increased pesticide use when seasonal disease pressure required.

8.2.9 Food & wine authenticity. Metal isotope ratios have been used extensively to determine the geographical origin of food and wine with Pb and Sr being the most common. For the first time Almeida et al.¹⁷⁰ have demonstrated that using the ratio ²⁰⁴Pb/²⁰⁶Pb, determined by ICP-MS, it was possible to differentiate between Brazilian red wines (n = 94) produced in the important southern region of Rio Grande do Sul from the north-eastern region of Vale do São Francisco. Furthermore, wines produced in different counties of this latter region can also be differentiated.

The authenticity of organic vs. conventionally-grown produce is frequently discussed in this review. This year, Barbosa et al.¹⁷¹ analysed certified organic (n = 17) and conventional (n = 33) rice grown in Brazil, using ICP-MS. Concentrations of As, B, Ba, Ca, Cd, Ce, Co, Cr, Cu, Fe, La, Mg, Mn, Mo, P, Pb, Rb, Se and Zn were determined following microwave-assisted digestion with HNO₃. Both samples types had high As concentrations (0.164 and 0.198 mg L⁻¹) while the Cd was lower in the organic samples (0.004 vs. 0.011 mg L⁻¹) and Ca concentrations were relatively higher in the organic samples (110.7 vs. 34.8 mg L⁻¹). The data mining technique known as support vector machine was used to classify the samples, using 19 variables, with 98% accuracy and 94% sensitivity while just Ca and Cd concentrations gave an accuracy of 96% and a sensitivity of 88%. Although not conceived as an authenticity application, Marriott et al.¹⁷² demonstrated that by determining trace elements (Ba, Ca, K, Li, Mg, Mn, Na, Rb, Sn, Sr and Zn) within sagittal otoliths (the largest of the CaCO₃-containing structures used for sound detection) of European plaice (Pleuronectes platessa), juveniles were classified back to their nursery ground of origin in the south-eastern Irish Sea. Sagittal otoliths were removed from juvenile plaice (6–15 cm long, n = 15) collected from eight sites, dissolved in a solution of HNO₃ (50%)–HCl (25%) and diluted 1 + 49. Concentrations were determined by ICP-MS. Significant (p < 0.05) differences were seen between the eight sites for all elements with post hoc Bonferroni (sic) tests for Mn, Rb, Sn and Zn being significant at p < 0.001. Canonical variate linear discriminant analysis allowed 71% of samples to be classified to their correct nursery site. This study is an interesting indication of how bony fish can be potentially authenticated to their origin.

Abbreviations

2D	Two dimensional
3D	Three dimensional
AA	Atomic absorption
AAS	Atomic absorption spectrometry
AB	Arsenobetaine
AC	Arsenocholine
ACN	Acetonitrile
AE	Atomic emission
AES	Atomic emission spectrometry
AEC	Anion exchange chromatography
AFS	Atomic fluorescence spectrometry
AOAC	Association of Official Analytical Chemists
ASU	Atomic Spectrometry Update
CC	Collision cell
CF	Continuous flow
CRM	Certified reference material
CS	Continuum source
CRC	Collision reaction cell
CSF	Cerebrospinal fluid
CV	Cold vapour
DDTC	Diethyldithiocarbamate
DLLME	Dispersive liquid–liquid microextraction
DMA	Dimethylarsenic
DNA	Deoxyribonucleic acid
DRC	Dynamic reaction cell
dw	Dry weight
EA	Elemental analysis
EBC	Exhaled breath condensate
EC	Electrochemical
EDTA	Ethylenediamine tetraacetic acid
EDXRF	Energy dispersive X-ray fluorescence
EPA	Environmental Protection Agency
EQAS	External quality assessment scheme
ESI	Electrospray ionisation
ETAAS	Electrothermal atomic absorption spectrometry
ETV	Electrothermal vapour
EU	European Union
FAAS	Flame atomic absorption spectrometry
FAES	Flame atomic emission spectrometry
FI	Flow injection
fs	Femtosecond
FTIR	Fourier transfer infra-red
GABA	Gamma aminobutyric acid
GE	Gel electrophoresis
GPx	Glutathione peroxidise
HG	Hydride generation
HILIC	Hydrophilic interaction liquid chromatography
HPLC	High performance liquid chromatography
HR	High resolution
IAEA	International Atomic Energy Authority
iAs	Inorganic arsenic
IC	Ion chromatography
ICH Q3D	International conference on harmonization Q3D
ICP-AES	Inductively coupled plasma atomic emission spectrometry
ICP-MS	Inductively coupled plasma mass spectrometry
ICP-OES	Inductively coupled plasma optical emission spectrometry
ID	Isotope dilution
IEC	Ion-exchange chromatography
IP	Ionisation potential
IR	Infra red
iSe	Inorganic selenium
JCTLM	Joint Committee for Traceability in Laboratory Medicine
KED	Kinetic energy discrimination
LA	Laser ablation
LC	Liquid chromatography
LIBS	Laser induced breakdown spectroscopy
LOD	Limit of detection
LOQ	Limit of quantification
MALDI	Matrix-assisted laser desorption ionisation
MC	Multicollector
MeHg	Methyl mercury
MIBK	Methyl isobutyl ketone
MIP	Microwave induced plasma
MMA	Monomethylarsonic acid
MRI	Magnetic resonance imaging
MS	Mass spectrometry
MWCNT	Multi-walled carbon nanotubes
NAA	Neutron activation analysis
NIST	National Institute of Standards and Technology
NP	Nanoparticle
PBS	Phosphate buffered saline
PIXE	Particle-induced X-ray emission
PRC	People' Republic of China
PTFE	Poly(tetrafluoroethylene)
PTWI	Provisional tolerable weekly intake
Q-ICP-MS	Quadrupole-ICP-MS
QQQ	Triple quadrupole
RM	Reference material
RP	Reversed phase
RSD	Relative standard deviation
SAXS	Small-angle X-ray scattering
SEC	Size exclusion chromatography
SeCys	Selenocysteine
SEM	Scanning electron microscopy
SeMet	Selenomethioneine
S/N	Signal to noise
SP	Solid phase
SPE	Solid phase extraction
SPME	Solid phase microextraction
SRM	Standard reference material
SR-XRF	Synchrotron radiation X-ray fluorescence
TEM	Transmission electron microscopy
THF	Tetrahydrofuran
TMAH	Tetramethyl ammonium hydroxide
TMSe^+	Trimethylselenium
TOF	Time-of-flight
TS-FF	Thermospray flame furnace
TXRF	Total reflexion XRF
USNHANES	United States National Health and Nutrition Examination Survey
WDXRF	Wavelength dispersive X-ray fluorescence
XANES	X-ray absorption near-edge structure
XRA	X-ray absorption
XRD	X-ray diffraction
XRF	X-ray fluorescence

References

1. A. Taylor, N. Barlow, M. P. Day, S. Hill, M. Patriarca and M. White, J. Anal. At. Spectrom., 2016, 31(3), 554–596.
2. O. T. Butler, W. R. L. Cairns, J. M. Cook and C. M. Davidson, J. Anal. At. Spectrom., 2016, 31(1), 35–89.
3. E. H. Evans, J. Pisonero, C. M. M. Smith and R. N. Taylor, J. Anal. At. Spectrom., 2016, 31(5), 1057–1077.
4. M. West, A. T. Ellis, P. J. Potts, C. Streli, C. Vanhoof and P. Wobrauschek, J. Anal. At. Spectrom., 2015, 30(9), 1839–1889.
5. S. Carter, A. Fisher, R. Garcia, B. Gibson, S. Lancaster, J. Marshall and I. Whiteside, J. Anal. At. Spectrom., 2015, 30(11), 2249–2294.
6. R. Clough, C. F. Harrington, S. J. Hill, Y. Madrid and J. F. Tyson, J. Anal. At. Spectrom., 2016, 31(7), 1330–1373.
7. O. T. Butler, R. Clough, J. M. Cook, E. H. Evans, A. S. Fisher, S. J. Hill, A. Taylor and M. West, J. Anal. At. Spectrom., 2016, 31(1), 32–34.
8. D. J. Hare, E. J. New, M. D. de Jonge and G. McColl, Chem. Soc. Rev., 2015, 44(17), 5941–5958.
9. A. Bodzon-Kulakowska and P. Suder, Mass Spectrom. Rev., 2015, 35(1), 147–169.
10. B. A. Boughton, D. Thinagaran, D. Sarabia, A. Bacic and U. Roessner, Phytochem. Rev., 2015, 15(3), 445–488.
11. A. Limbeck, P. Galler, M. Bonta, G. Bauer, W. Nischkauer and F. Vanhaecke, Anal. Bioanal. Chem., 2015, 407(22), 6593–6617.
12. J. C. King, K. H. Brown, R. S. Gibson, N. F. Krebs, N. M. Lowe, J. H. Siekmann and D. J. Raiten, J. Nutr., 2016, 146(4), 858S–885S.
13. M. Costas-Rodriguez, J. Delanghe and F. Vanhaecke, TrAC, Trends Anal. Chem., 2016, 76, 182–193.
14. F. Larner, Anal. Bioanal. Chem., 2016, 408(2), 345–349.
15. M. J. Ceko, S. O'Leary, H. H. Harris, K. Hummitzsch and R. J. Rodgers, Biol. Reprod., 2016, 94(4), 14.
16. G. Ring, J. O'Mullane, A. O'Riordan and A. Furey, Clin. Biochem., 2016, 49(7–8), 617–635.
17. W. A. Maher, M. J. Ellwood, F. Krikowa, G. Raber and S. Foster, J. Anal. At. Spectrom., 2015, 30(10), 2129–2183.
18. M. F. Mesko, V. C. Costa, R. S. Picoloto, C. A. Bizzi and P. A. Mello, J. Anal. At. Spectrom., 2016, 31(6), 1243–1261.
19. R. Doggui, M. El Ati-Hellal and J. El Ati, Ann. Biol. Clin., 2016, 74(2), 184–195.
20. E. Grygo-Szymanko, A. Tobiasz and S. Walas, TrAC, Trends Anal. Chem., 2016, 80, 112–124.
21. J. Delafiori, G. Ring and A. Furey, Talanta, 2016, 153, 306–331.
22. D. J. Butcher, Appl. Spectrosc. Rev., 2016, 51(5), 397–416.
23. E. Poitevin, J. AOAC Int., 2016, 99(1), 42–52.
24. L. L. Zhang, L. Lu, Y. J. Pan, C. G. Ding, D. Y. Xu, C. F. Huang, X. F. Pan and W. Zheng, Environ. Res., 2016, 140, 10–17.
25. J. P. Goullè, P. Le Roux, M. Castanet, L. Mahieu, S. Guyet-Job and M. Guerbet, J. Anal. Toxicol., 2015, 39(9), 707–713.
26. A. V. Skalny, M. G. Skalnaya, A. A. Tinkov, E. P. Serebryansky, V. A. Demidov, Y. N. Lobanova, A. R. Grabeklis, E. S. Berezkina, I. V. Gryazeva, A. A. Skalny and A. A. Nikonorov, Environ. Toxicol. Pharmacol., 2016, 40(1), 18–21.
27. S. Dubascoux, M. Nicolas, C. F. Rime, J. R. Payot and E. Poitevin, J. AOAC Int., 2015, 98(4), 953–961.
28. J. J. Thompson, L. Pacquette and S. L. Brunelle, J. AOAC Int., 2015, 98(6), 1711–1720.
29. L. H. Pacquette and J. J. Thompson, J. AOAC Int., 2015, 98(6), 1702–1710.
30. R. S. Zywicki and D. M. Sullivan, J. AOAC Int., 2015, 98(5), 1407–1416.
31. S. Valdersnes, P. Fecher, A. Maage and K. Julshamn, Food Chem., 2016, 194, 424–431.
32. EN 16802:2016 Foodstuffs. Determination of elements and their chemical species. Determination of inorganic arsenic in foodstuffs of marine and plant origin by anion-exchange HPLC-ICP-MS.
33. I. Fiamegkos, F. Cordeiro, P. Robouch, D. Vélez, V. Devesa, G. Raber, J. J. Sloth, R. R. Rasmussen, T. Llorente-Mirandes, J. F. Lopez-Sanchez, R. Rubio, F. Cubadda, M. D'Amato, J. Feldmann, A. Raab, H. Emteborg and M. B. de la Calle, Accuracy of a method based on atomic absorption spectrometry to determine inorganic arsenic in food: Outcome of the collaborative trial IMEP-41, Food Chem., 2016, 213, 169–179,  DOI:10.1016/j.foodchem.2016.06.033.
34. J. T. Collie, R. J. Massie, O. A. H. Jones, P. D. Morrison and R. F. Greaves, Clin. Chem. Lab. Med., 2016, 54(4), 561–567.
35. M. L. Ge, H. J. Zhao, Y. Yan, T. J. Zhang, J. Zeng, Y. F. Wang, Q. H. Meng and C. B. Zhang, Clin. Lab., 2016, 62(5), 921–930.
36. A. Yamakawa, A. Takeuchi, Y. Shibata, S. Berail and O. F. X. Donard, Accredit. Qual. Assur., 2016, 21(3), 197–202.
37. R. Chelegao, V. M. O. Carioni, J. Naozuka and C. S. Nomura, J. Braz. Chem. Soc., 2016, 27(4), 712–718.
38. R. Ketrin, E. M. Handayani, I. Komalasari and C. Elishian, Accredit. Qual. Assur., 2015, 20(4), 281–285.
39. Y. L. Sommer, C. D. Ward, Y. Pan, K. L. Caldwell and R. L. Jones, J. Anal. Toxicol., 2016, 40(3), 222–228.
40. J. D. Palcic, J. S. Jones, E. L. Flagg and S. F. Donovan, J. Anal. At. Spectrom., 2015, 30(8), 1799–1808.
41. P. Hoet, G. Deumer, A. Bernard, D. Lison and V. Haufroid, J. Exposure Sci. Environ. Epidemiol., 2016, 26(3), 296–302.
42. C. Y. Li, S. Q. Peng, X. M. Zhang, X. C. Xie, D. Y. Wang, J. Y. Mao, X. C. Teng, Z. Y. Shan and W. P. Teng, J. Clin. Endocrinol. Metab., 2016, 101(3), 1290–1298.
43. N. Mohd-Taufek, D. Cartwright, M. Davies, A. K. Hewavitharana, P. Koorts, P. N. Shaw, R. Sumner, E. Lee and K. Whitfield, 2016, 9(7), 2068–2075.
44. L. Perrin, A. Carmona, S. Roudeau and R. Ortega, J. Anal. At. Spectrom., 2015, 30(12), 2525–2532.
45. V. M. P. Barbosa, A. F. Barbosa, J. Bettini, P. O. Luccas and E. C. Figueiredo, Talanta, 2016, 147, 478–484.
46. W. I. Mortada, I. M. M. Kenawy, A. M. Abdelghany, A. M. Ismal, A. F. Donia and K. A. Nabieh, Mater. Sci. Eng., C, 2015, 52, 288–296.
47. Z. R. Zou, S. L. Wang, J. Jia, F. J. Xu, Z. Long and X. D. Hou, Microchem. J., 2016, 124, 578–583.
48. E. V. Alonso, M. D. M. Guerrero, P. Cueto, J. B. Benitez, J. M. C. Pavon and A. G. de Torres, Talanta, 2016, 153, 228–239.
49. F. Aydin, E. Yilmaz and M. Soylak, RSC Adv., 2015, 5(115), 94879–94886.
50. A. C. Grijalba, L. B. Escudero and R. G. Wuilloud, Spectrochim. Acta, Part B, 2015, 110, 118–123.
51. Z. A. Alothman, M. A. Habila, S. M. Alfadul, E. Yilmaz and M. Soylak, Anal. Methods, 2015, 8(17), 3545–3549.
52. M. Krawczyk and E. Stanisz, J. Anal. At. Spectrom., 2015, 30(11), 2353–2358.
53. Z. Bahadir, V. N. Bulut, M. Hidalgo, M. Soylak and E. Margui, Spectrochim. Acta, Part B, 2016, 115, 46–51.
54. M. Pouyan, G. Bagherian and N. Goudarzi, Microchem. J., 2016, 127, 46–51.
55. B. D. Kerger, R. Gerads, H. Gurleyuk, A. Urban and D. J. Paustenbach, Toxicol. Environ. Chem., 2015, 97(9), 1145–1163.
56. M. R. Florez, E. Garcia-Ruiz, E. Bolea-Fernandez, F. Vanhaecke and M. Resano, J. Anal. At. Spectrom., 2016, 31(1), 245–251.
57. Y. E. Peng, W. Guo, J. Zhang, Q. H. Guo, L. L. Jin and S. H. Hu, Microchem. J., 2016, 124, 127–131.
58. B. Klencsar, E. Bolea-Fernandez, M. R. Florez, L. Balcaen, F. Cuyckens, F. Lynen and F. Vanhaecke, J. Pharm. Biomed. Anal., 2016, 124, 112–119.
59. A. M. Schwan, R. Martin and W. Goessler, Anal. Methods, 2015, 7(21), 9198–9205.
60. E. Cordeau, C. Arnaudguilhem, B. Bouyssiere, A. Hagege, J. Martinez, G. Subra, S. Cantel and C. Enjalbal, PLoS One, 2016, 11(6), 18.
61. E. Albalat, P. Telouk, V. Balter, T. Fujii, V. P. Bondanese, M. L. Plissonnier, V. Vlaeminck-Guillem, J. Baccheta, N. Thiam, P. Miossec, F. Zoulim, A. Puisieux and F. Albarede, J. Anal. At. Spectrom., 2016, 31(4), 1047.
62. Y. Anoshkina, M. Costas-Rodriguez and F. Vanhaecke, J. Anal. At. Spectrom., 2015, 30(8), 1816–1821.
63. K. Bierla, A. Flis-Borsuk, P. Suchocki, J. Szpunar and R. Lobinski, J. Agric. Food Chem., 2015, 64(24), 4975–4981.
64. Y. C. Wang, L. G. Hu, X. M. Yang, Y. Y. Chang, X. Q. Hu, H. Y. Li and H. Z. Sun, Metallomics, 2015, 7(10), 1399–1406.
65. D. P. Bao, Z. G. Oh and Z. Chen, Front. Plant. Sci., 2016, 7, 8.
66. K. Ramos, M. M. Gomez-Gomez, C. Camara and L. Ramos, Talanta, 2016, 151, 83–90.
67. A. Mackevica, M. E. Olsson and S. F. Hansen, J. Nanopart. Res., 2106, 18(1), 11.
68. M. Witzler, F. Kullmer, A. Hirtz and K. Gunther, J. Agric. Food Chem., 2015, 64(20), 4165–4170.
69. N. S. Feichtmeier, N. Ruchter, S. Zimmermann, B. Sures and K. Leopold, Anal. Bioanal. Chem., 2016, 408(1), 295–305.
70. J. S. Almeida, G. C. Brandao, G. L. dos Santos and L. S. G. Teixeira, Anal. Methods, 2015, 8(15), 3249–3254.
71. N. Ozbek and S. Akman, Food Chem., 2016, 211, 180–184.
72. S. Butaciu, T. Frentiu, M. Senila, E. Darvasi, S. Cadar, M. Ponta, D. Petreus, R. Etz and M. Frentiu, Food Control, 2016, 61, 227–234.
73. E. Y. Barrientos, K. Wrobel, J. C. T. Guzman and A. R. C. Escobosa, J. Anal. At. Spectrom., 2016, 31(1), 203–211.
74. A. Taylor, M. P. Day, S. Hill, J. Marshall, M. Patriarca and M. White, J. Anal. At. Spectrom., 2015, 30(3), 542–579.
75. S. L. Zhong, R. E. Zheng, Y. Lu, K. Cheng and J. S. Xiu, Plasma Sci. Technol., 2015, 17(11), 979–984.
76. I. Gaubeur, M. A. Aguirre, N. Kovachev, M. Hidalgo and A. Canals, J. Anal. At. Spectrom., 2015, 30(12), 2541–2547.
77. R. K. Gill, Z. J. Smith, R. R. Panchal, J. W. Bishop, R. Gandour-Edwards and S. Wachsmann-Hogiu, Biomed. Opt. Express, 2015, 6(23), 4850–4858.
78. B. B. S. Jaswal, V. Kumar, J. Sharma, P. K. Rai, M. A. Gondal, B. Gondal and V. K. Singh, Lasers. Med. Sci., 2016, 31(3), 573–579.
79. A. O. Mehder, M. A. Gondal, M. A. Dastageer, Y. B. Habibullah, M. A. Iqbal, L. E. Oloore and B. Gondal, J. Environ. Sci. Health, Part B, 2016, 51(6), 358–365.
80. K. Marschner, S. Musil and J. Dedina, Anal. Chem., 2016, 88(7), 4041–4047.
81. A. H. Petursdottir, J. J. Sloth and J. Feldmann, Anal. Bioanal. Chem., 2015, 407(28), 8385–8396.
82. F. Bilo, L. Borgese, J. Prost, M. Rauwolf, A. Turyanskaya, P. Wobrauschek, P. Kregsamer, C. Streli, U. Pazzaglia and L. E. Depero, Appl. Surf. Sci., 2015, 359, 215–220.
83. M. Golasik, P. Wrobel, M. Olbert, B. Nowak, M. Czyzycki, T. Librowski, M. Lankosz and W. Piekoszewski, BioMetals, 2016, 29(3), 487–494.
84. A. Wandzilak, M. Czyzycki, E. Radwanska, C. D. Adamek, K. Geraki and M. Lankosz, Spectrochim. Acta, Part B, 2105, 114, 52–57.
85. V. A. Trunova, A. V. Sidorina and K. V. Zolotarev, X-Ray Spectrom., 2015, 44(4), 226–229.
86. X. L. Li and Z. S. Yu, Appl. Radiat. Isot., 2016, 111, 45–49.
87. J. P. Gorce and M. Roff, J. Occup. Environ. Hyg., 2016, 13(2), 102–111.
88. D. E. B. Fleming, J. W. Groves, M. R. Gherase, G. N. George, I. J. Pickering, O. Ponomarenko, G. Langan, J. E. Spallholz, M. Alauddin, H. Ahsan, S. Ahmed and P. F. La Porte, Radiat. Phys. Chem., 2015, 116, 241–247.
89. H. Shehab, E. D. Desouza, J. O'Meara, A. Pejovic-Milic, D. R. Chettle, D. E. B. Fleming and F. E. McNeill, Physiol. Meas., 2016, 37(1), 145–161.
90. D. J. McIver, J. A. VanLeeuwen, A. L. Knafla, J. A. Campbell, K. M. Alexander, M. R. Gherase, J. R. Guernsey and D. E. B. Fleming, Physiol. Meas., 2015, 36(12), 2443–2459.
91. K. Hagiwara, Y. Koike, M. Aizawa and T. Nakamura, Talanta, 2015, 144, 788–792.
92. A. A. Dias, M. Carvalho, M. L. Carvalho and S. Pessanha, J. Anal. At. Spectrom., 2015, 30(12), 2488–2495.
93. S. Pessanha, M. Carvalho, M. L. Carvalho and A. Dias, J. Trace Elem. Med. Biol., 2016, 33, 26–30.
94. V. Romero, M. Gryglicka, I. De La Calle, I. Lavilla and C. Bendicho, Microchim. Acta, 2016, 183(1), 141–148.
95. L. Borgese, F. Bilo, R. Dalipi, E. Bontempi and L. E. Depero, Spectrochim. Acta, Part B, 2015, 113, 1–15.
96. A. Figueiredo, T. Fernandes, I. M. Costa, L. Goncalves and J. Brito, J. Pharm. Biomed. Anal., 2016, 122, 52–58.
97. E. M. Sussman, P. Jayanti, B. J. Dair and B. J. Casey, Food Chem. Toxicol., 2015, 85, 10–19.
98. A. Al Zaki, J. Z. Hui, E. Higbee and A. Tsourkas, J. Biomed. Nanotechnol., 2015, 11(10), 1836–1846.
99. O. D. Hendrickson, S. G. Klochkov, O. V. Novikova, I. M. Bravova, E. F. Shevtsova, I. V. Safenkova, A. V. Zherdev, S. O. Bachurin and B. B. Dzantiev, Toxicol. Lett., 2016, 241, 184–192.
100. C. Aude-Garcia, F. Villiers, V. Collin-Faure, K. Pernet-Gallay, P. H. Jouneau, S. Sorieul, G. Mure, A. Gerdil, N. Herlin-Boime, M. Carriere and T. Rabilloud, Nanotoxicology, 2016, 10(5), 586–596.
101. T. Jager, H. Drexler and T. Goen, Arch. Toxicol., 2016, 90(1), 149–158.
102. Y. Anan, M. Kimura, M. Hayashi, R. Koike and Y. Ogra, Chem. Res. Toxicol., 2015, 28(9), 1803–1814.
103. E. Leese, J. Morton, E. Tan, P. Gardiner and V. Carolan, μLC-ICP-MS determination of unexposed UK urinary arsenic speciation reference values, J. Anal. Toxicol., 2014, 38, 24–30.
104. J. Moreda-Piñeiro, A. Cantarero-Roldan, P. Hermelo-Herbello, A. Moreda-Pineiro and P. Bermejo-Barrera, At. Spectrosc., 2015, 36(6), 237–246.
105. S. Stice, G. L. Liu, S. Matulis, L. H. Boise and Y. Cai, J. Chromatogr. B: Anal. Technol. Biomed. Life Sci., 2016, 1009, 55–65.
106. M. Marcinkowska, I. Komorowicz and D. Baralkiewicz, Anal. Chim. Acta, 2016, 920, 102–111.
107. S. J. M. Van Malderen, A. J. Managh, B. L. Sharp and F. Vanhaecke, J. Anal. At. Spectrom., 2016, 31(2), 423–439.
108. G. D. Pessôa, J. L. Capelo-Martinez, F. Fdez-Riverola, H. López-Fernandez, D. Glez- Peña, M. Reboiro-Jato and M. A. Z. Arruda, J. Anal. At. Spectrom., 2016, 31(4), 832–840.
109. J. P. Thieleke and C. Vogt, J. Anal. At. Spectrom., 2016, 31(6), 1198–1205.
110. D. N. Douglas, A. J. Managh, H. J. Reid and B. L. Sharp, Anal. Chem., 2015, 87(22), 11285–11294.
111. D. P. Bishop, D. Clases, F. Fryer, E. Williams, S. Wilkins, D. J. Hare, N. Cole, U. Karst and P. A. Doble, J. Anal. At. Spectrom., 2015, 31(1), 197–202.
112. O. Hachmöller, M. Aichler, K. Schwamborn, L. Lutz, M. Werner, M. Sperling, A. Walch and U. Karst, J. Trace Elem. Med. Biol., 2016, 35, 97–102.
113. T. S. Lum, C. L. Ho, Y. K. Tsoi, C. H. Siu, P. Y. K. Yue, W. Y. Wong and K. S. Y. Leung, Int. J. Mass Spectrom., 2016, 404, 40–47.
114. S. G. Elci, Y. Jiang, B. Yan, S. T. Kim, K. Saha, D. F. Moyano, G. Y. Tonga, L. C. Jackson, V. M. Rotello and R. W. Vachet, ACS Nano, 2015, 10(5), 5536–5542.
115. D. R. Roberts, S. M. Lindhorst, C. T. Welsh, K. R. Maravilla, M. N. Herring, K. A. Braun, B. H. Thiers and W. C. Davis, Invest. Radiol., 2016, 51(5), 280–289.
116. T. J. Shepherd, W. Dirks, N. M. W. Roberts, J. G. Patel, S. Hodgson, T. Pless-Mulloli, P. Walton and R. R. Parrish, Environ. Res., 2016, 146, 145–153.
117. M. Bonta, J. J. Gonzalez, C. D. Quarles, R. E. Russo, B. Hegedus and A. Limbeck, J. Anal. At. Spectrom., 2016, 31(1), 252–258.
118. T. Tian, J. C. Zhang, H. Z. Lei, Z. Ying, J. Y. Shi, J. Hu, Q. Huang, C. H. Fan and Y. H. Sun, Nucl. Sci. Tech., 2016, 26(3), 6.
119. O. Hachmöller, A. G. Buzanich, M. Aichler, M. Radtke, D. Dietrich, K. Schwamborn, L. Lutz, M. Werner, M. Sperling, A. Walch and U. Karst, Metallomics, 2016, 8(7), 648–653.
120. M. J. Ceko, K. Hummitzsch, W. M. Bonner, J. B. Aitken, K. M. Spiers, R. J. Rodgers and H. H. Harris, Microsc. Microanal., 2015, 21(3), 695–705.
121. D. Bourgeois, B. Burt-Pichat, X. Le Goff, J. Garrevoet, P. Tack, G. Falkenberg, L. Van Hoorebeke, L. Vincze, M. A. Denecke, D. Meyer, C. Vidaud and G. Boivin, Anal. Bioanal. Chem., 2015, 407(22), 6619–6625.
122. Y. Kashiv, J. R. Austin, B. Lai, V. Rose, S. Vogt and M. El-Muayed, Sci. Rep., 2016, 6, 9.
123. N. Espallargas, C. Torres and A. I. Munoz, Wear, 2015, 332, 669–678.
124. I. Catelas, E. A. Lehoux, I. Hurda, S. J. Baskey, L. Gala, R. Foster, P. R. Kim and P. E. Beaule, Clin. Orthop. Relat. Res., 2015, 473(12), 3903–3914.
125. M. Cadossi, A. Mazzotti, N. Baldini, S. Giannini and L. Savarino, J. Biomed. Mater. Res., Part B, 2016, 104(1), 204–209.
126. M. Sakamoto, H. Watanabe, H. Higashi and H. Kubosawa, Orthopedics, 2016, 39(1), E162–E165.
127. C. Harrington, C. McKibbin, M. Rahanu, D. Langton and A. Taylor, Measurement of titanium in hip-replacement patients by inductively coupled plasma optical emission spectroscopy, Ann. Clin. Biochem., 2016 DOI:10.1177/0004563216662292.
128. E. Lukina, A. Laka, M. Kollerov, M. Sampiev, P. Mason, P. Wagstaff, H. Noordeen, W. W. Yoon and G. Blunn, Spine J., 2016, 16(3), 380–388.
129. L. Golz, A. C. Knickenberg, L. Keilig, S. Reimann, S. N. Papageorgiou, A. Jager and C. Bourauel, Journal of Orofacial Orthopedics/Fortschritte der Kieferorthopädie, 2016, 77(2), 85–93.
130. F. Amini, M. Shariati, F. Sobouti and V. Rakhshan, Am. J. Orthod. Dentofacial Orthop., 2016, 149(5), 666–672.
131. N. Simsek, E. T. Bulut, F. Ahmetoglu and H. Alan, J. Mater. Sci.: Mater. Med., 2016, 27(3), 6.
132. Y. X. Wang, W. Feng, Q. Zeng, Y. Sun, P. Wang, L. You, P. Yang, Z. Huang, S. L. Yu and W. Q. Lu, Environ. Health Perspect., 2016, 124(4), 468–476.
133. F. Breier, D. Loader, G. Wasilewicz, W. Brenner, P. Sator, R. Feldmann, A. Stella, W. Osterode, G. Falkenberg, A. Steiner and W. Jurecka, Hautarzt, 2016, 67(1), 64–68.
134. S. Hulo, A. Radauceanu, N. Cherot-Kornobis, M. Howsam, V. Vacchina, V. De Broucker, D. Rousset, M. Grzebyk, M. Dziurla, A. Sobaszek and J. L. Edme, Int. J. Hyg. Environ. Health, 2016, 219(1), 40–47.
135. A. Atila, Z. Halici, E. Cadirci, E. Karakus, S. S. Palabiyik, N. Ay, F. Bakan and S. Yilmaz, Mater. Sci. Eng., C, 2016, 58, 1082–1089.
136. C. Burton, Y. B. Dan, A. Donovan, K. Liu, H. L. Shi, Y. F. Ma and C. P. Bosnak, Clin. Chim. Acta, 2016, 452, 142–148.
137. D. H. Tsai, Y. F. Lu, F. W. DelRio, T. J. Cho, S. Guha, M. R. Zachariah, F. Zhang, A. Allen and V. A. Hackley, Anal. Bioanal. Chem., 2015, 407(28), 8411–8422.
138. X. L. Shi, Y. T. Zhu, W. D. Hua, Y. L. Ji, Q. Ha, X. X. Han, Y. Liu, J. W. Gao, Q. Zhang, S. D. Liu, K. L. Ren, X. C. Wu, H. Y. Li and D. Han, Nano Res., 2016, 9(7), 2097–2109.
139. R. Tagne-Fotso, A. Leroyer, M. Howsam, B. Dehon, C. Richeval, C. Nisse and N.-P.-d. Hlth Examination Ctr, J. Toxicol. Environ. Health, Part A, 2016, 79(6), 245–265.
140. R. C. Fachehoun, B. Levesque, P. Dumas, A. St-Louis, M. Dube and P. Ayotte, Food Addit. Contam., Part A: Chem., Anal., Control, Exposure Risk Assess., 2015, 32(9), 1501–1511.
141. T. Lech, Am. J. Forensic Med. Pathol., 2015, 36(3), 227–231.
142. Y. Gao, M. Jiang, Y. Q. Ma, S. Wu, W. L. Li, X. G. Yang, Y. X. Li, X. B. Jing and H. Y. Jiang, Anti-Cancer Drugs, 2016, 27(2), 77–83.
143. E. Humann-Ziehank, M. Ganter and B. Michalke, J. Trace Elem. Med. Biol., 2016, 33, 14–20.
144. B. Campanella, M. Onor, A. D'Ulivo, R. Giannecchini, M. D'Orazio, R. Petrini and E. Bramanti, Sci. Total Environ., 2016, 548, 33–42.
145. A. S. Saljooghi, M. Babaie, F. D. Mendi, M. Zahmati and Z. S. Saljooghi, Toxicol. Ind. Health, 2016, 32(1), 83–88.
146. M. Golasik, M. Herman, M. Olbert, T. Librowski, J. Szklarzewicz and W. Piekoszewski, Toxicol. Lett., 2016, 247, 56–61.
147. P. F. Jin, X. L. Liang, L. F. Xia, F. Jahouh, R. Wang, Y. M. Kuang and X. Hu, J. Trace Elem. Med. Biol., 2016, 33, 73–80.
148. P. Li, L. M. Li, J. Xia, S. Cao, X. Hu, H. Z. Lian and S. Ji, J. Sep. Sci., 2015, 38(23), 4043–4047.
149. A. Kowalski and M. Frankowski, Regul. Toxicol. Pharmacol., 2016, 73(1), 396–400.
150. G. Li, D. Schoneker, K. L. Ulman, J. J. Sturm, L. M. Thackery and J. F. Kauffman, J. Pharm. Sci., 2015, 104(12), 4197–4206.
151. J. D. Strickland, W. R. LeFew, J. Crooks, D. Hall, J. N. R. Ortenzio, K. Dreher and T. J. Shafer, Toxicology, 2016, 355, 1–8.
152. B. E. S. Costa, L. M. Coelho, C. S. T. Araujo, H. C. Rezende and N. M. M. Coelho, J. Chem., 2016, 2016, 1427154.
153. N. Yogarajah and S. S. H. Tsai, Environ. Sci.: Water Res. Technol., 2015, 1(4), 426–447.
154. W. C. Zu and Z. H. Wang, J. Hazard. Mater., 2016, 304, 467–473.
155. J. Abolhasani, R. H. Khanmiri, M. Babazadeh, E. Ghorbani-Kalhor, L. Edjlali and A. Hassanpour, Environ. Monit. Assess., 2015, 187(9), 11.
156. H. Y. Liu, W. Z. Zhu, B. Y. Lu, Z. H. Wei and D. X. Ren, J. Dairy Sci., 2015, 98(12), 8359–8367.
157. A. Ohki, T. Nakajima, S. Hirakawa, K. Hayashi and H. Takanashi, Microchem. J., 2016, 124, 693–698.
158. T. I. Todorov and P. J. Gray, Food Addit. Contam., Part A: Chem., Anal., Control, Exposure Risk Assess., 2016, 33(2), 282–290.
159. R. Marin-Martinez, X. Barber, C. Cabrera-Vique, A. A. Carbonell-Barrachina, E. Vilanova, V. M. Garcia-Hernandez, E. Roche and E. Garcia-Garcia, Food Addit. Contam., Part B, 2016, 9(1), 66–71.
160. A. O. Mehder, E. Yilmaz, A. Sungur, M. Soylak and Z. A. Alothman, At. Spectrosc., 2015, 36(6), 254–260.
161. N. Limchoowong, P. Sricharoen, S. Techawongstien and S. Chanthai, Food Chem., 2016, 200, 223–229.
162. E. Blicharska, J. Flieger, K. Oszust, M. Frac, R. Swieboda and R. Kocjan, Anal. Lett., 2015, 48(14), 2272–2287.
163. X. Y. Liu, H. O. Qiu, Y. J. Huang, Z. Y. Tang, S. Z. Tang and Z. L. Zhu, Food Anal. Methods, 2016, 9(1), 1–6.
164. A. Ariano, A. Lo Voi, M. D'Ambola, R. Marrone, D. Cacace and L. Severino, J. Food Prot., 2015, 78(12), 2253–2256.
165. X. X. Xie, C. C. Feng, M. D. Ye and C. J. Wang, Food Anal. Methods, 2015, 8(7), 1739–1745.
166. T. L. Gerke, B. J. Little and J. B. Maynard, Sci. Total Environ., 2016, 541, 184–193.
167. D. R. S. Middleton, M. J. Watts, E. M. Hamilton, E. L. Ander, R. M. Close, K. S. Exley, H. Crabbe, G. S. Leonardi, T. Fletcher and D. A. Polya, Sci. Rep., 2016, 6, 11.
168. M. Birka, C. A. Wehe, O. Hachmöller, M. Sperling and U. Karst, J. Chromatogr. A, 2016, 1440, 105–111.
169. G. H. O. Rocha, R. S. Lini, F. Barbosa, B. L. Batista, V. C. D. Souza, S. B. Nerilo, E. Bando, S. A. G. Mossini and P. Nishiyama, Int. Arch. Occup. Environ. Health, 2015, 88(7), 875–880.
170. C. M. S. Almeida, A. C. Almeida, M. Godoy, T. D. Saint'Pierre and J. M. Godoy, J. Braz. Chem. Soc., 2015, 27(6), 1026–1031.
171. R. M. Barbosa, E. S. de Paula, A. C. Paulelli, A. F. Moore, J. M. O. Souza, B. L. Batista, A. D. Campiglia and F. Barbosa, J. Food Compos. Anal., 2016, 45, 95–100.
172. A. L. Marriott, I. D. McCarthy, A. L. Ramsay and S. R. N. Chenery, Mar. Ecol.: Prog. Ser., 2016, 546, 183–195.
173. H. Shirkhanloo, M. Ghazaghi and H. Z. Mousavi, J. Pharm. Biomed. Anal., 2016, 118, 1–8.
174. N. R. Kendall, H. R. Holmes-Pavord, P. A. Bone, E. L. Ander and S. D. Young, Vet. Rec., 2015, 177(19), 493.
175. S. Prochazkova and R. Halko, Anal. Lett., 2016, 49(11), 1656–1668.
176. G. K. Das, D. S. Anderson, C. D. Wallis, S. A. Carratt, I. M. Kennedy and L. S. Van Winkle, Nanoscale, 2016, 8(22), 11518–11530.
177. Y. Yan, C. B. Zhang, J. T. Zhang, T. J. Zhang, W. Y. Zhou, J. Zeng and W. X. Chen, Clin. Lab., 2016, 62(4), 719–725.
178. K. Flouda, J. M. Dersch, C. Gabel-Jensen, S. Sturup, S. Misra, M. Bjornstedt and B. Gammelgaard, Anal. Bioanal. Chem., 2016, 408(9), 2293–2301.
179. N. Zuverza-Mena, R. Armendariz, J. R. Peralta-Videa and J. L. Gardea-Torresdey, Front. Plant. Sci., 2016, 7, 11.
180. B. A. Sadee, M. E. Foulkes and S. J. Hill, Food Addit. Contam., Part A: Chem., Anal., Control, Exposure Risk Assess., 2016, 33(3), 433–441.
181. Y. T. Huang, J. H. Shan, B. Fan, Y. He, S. M. Xia, Y. F. Sun, J. Lu, M. Wang and F. Z. Wang, Anal. Methods, 2015, 7(20), 8896–8900.
182. E. Bralatei, S. Lacan, E. M. Krupp and J. Feldmann, Anal. Chem., 2015, 87(22), 11271–11276.
183. S. V. da Silva, R. S. Picoloto, E. M. M. Flores, R. Wagner, N. Richards and J. S. Barin, Food Chem., 2016, 190, 364–367.
184. Y. Zhang, X. F. Mao, J. X. Liu, M. Wang, Y. Z. Qian, C. L. Gao and Y. H. Qi, Spectrochim. Acta, Part B, 2016, 118, 119–126.
185. B. Markiewicz, I. Komorowicz and D. Baralkiewicz, Talanta, 2016, 152, 489–497.
186. N. Altunay, R. Gurkan and U. Orhan, Food Addit. Contam., Part A: Chem., Anal., Control, Exposure Risk Assess., 2015, 32(9), 1475–1487.
187. D. Trak, E. Kenduzler and Y. Arslan, Anal. Lett., 2016, 49(10), 1589–1599.
188. M. A. Morgano, R. F. Milani and A. A. M. Perrone, Food Anal. Methods, 2015, 8(9), 2301–2307.
189. R. Gurkan, S. Korkmaz and N. Altunay, Talanta, 2016, 155, 38–46.
190. G. Ploegaerts, C. Desmet and M. Van Krieken, J. Food Compos. Anal., 2016, 45, 66–72.
191. M. A. Taher, F. Pourmohammad and H. Fazelirad, J. Sep. Sci., 2015, 38(23), 4153–4158.
192. F. Barahona, I. Ojea-Jimenez, O. Geiss, D. Gilliland and J. Barrero-Moreno, J. Chromatogr. A, 2016, 1432, 92–100.
193. E. Q. Oreste, A. O. de Souza, C. C. Pereira, M. T. Lisboa, M. J. A. Cidade, M. A. Vieira, S. Cadore and A. S. Ribeiro, Food Anal. Methods, 2016, 9(3), 777–784.
194. T. Dasbasi, S. Sacmaci, A. Ulgen and S. Kartal, Food Chem., 2016, 197, 107–113.
195. K. Miranda, A. L. Vieira, M. A. Bechlin, F. M. Fortunato, A. Virgilio, E. C. Ferreira and J. A. G. Neto, Food Anal. Methods, 2016, 9(7), 1887–1894.
196. F. D. Muller, C. Hackethal, R. Schmidt, O. Kappenstein, K. Pfaff and A. Luch, Food Addit. Contam., Part A: Chem., Anal., Control, Exposure Risk Assess., 2015, 32(11), 1959–1964.

---