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

Contents

• 1 Reviews

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

• 3 Sample collection and preparation
  • 3.1 Collection, storage and preliminary preparation
  • 3.2 Digestion, extraction and preconcentration

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

• 5 Applications: clinical and biological materials
  • 5.1 Metallomics
  • 5.2 Imaging: LA-ICP-MS and XRF spectrometry
  • 5.3 Multielement applications
  • 5.4. Progress for individual elements

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

• 7 Applications: foods and beverages
  • 7.1 Progress for individual elements
  • 7.2 Single and multielement applications in food and beverages

• 8 Abbreviations

• Conflicts of interest

• References

---

Abstract

This update covers publications from the second half of 2017 to the middle of 2019. Techniques and applications relevant to clinical and biological materials, foods and beverages are discussed in the text, presenting key aspects of the work referenced, while the tables provide a summary of the publications considered. A tutorial review gives an encyclopaedic account of how ICP-MS has developed for analysis of metal-containing nanoparticles and colloids. While it is a relatively new area, interest in how LIBS has many clinical applications is such that this topic is expertly reviewed. As previously noted in recent Updates, the growing importance of XRF techniques is accompanied by useful review papers. Among the technical developments that featured during the year were yet more materials for preconcentration or speciation, a high-temperature torch as part of a novel integrated ICP-MS sample introduction system, and a tubular dielectric barrier discharge trap as an alternative to vapour traps such as quartz tube or graphite furnace, all of which lead to lower claimed detection limits. A new approach to imaging applications, offered by ICP-TOF-MS, also provides for lower LODs. Measurements of As and Hg continue to feature heavily in work with foods and beverages. Of note, is a collaborative project to validate a straightforward HG-ICP-MS procedure in view of anticipated EU legislation on the maximum allowable limits of iHg in marine samples. Several articles concerning trace elements and fertility and reproductive health were noted during this review period with publications focusing on single, or groups of elements, in maternal blood, follicular fluid, placental tissue, and progress of pregnancy. Two accounts describe unusual incidents of poisoning. Cigarettes, adulterated with Hg, were provided to the victim. Flu-like symptoms developed after smoking. A dog developed symptoms of Ba poisoning after eating a firework, it made a complete recovery, with prompt treatment. Results from population studies can sometimes produce ambiguous results but the levels of 13 elements in urine from healthy Italian adults look realistic.

---

1 Reviews

This latest Update adds to that from last year¹ and complements other reviews of analytical techniques in the series of Atomic Spectrometry Updates from the previous year (Tables 1 and 2).^2–6

Table 1 Clinical and biological materials

Element	Matrix	Technique	Sample treatment/comments	Reference
Al	Brain	ETAAS	Al found to be increased in the temporal lobe neocortex in Alzheimer’s disease (n = 186), Down’s syndrome (n = 24) and dialysis dementia syndrome (n = 27) compared to age- and gender-matched brains for the same anatomical region	84
As	Hair	HG-AFS	A slurry sample dilution with 5% HCl (v/v) was followed by HG for arsine generation which was then coupled with an in situ dielectric barrier discharge system to pre-concentrate As and reduce interferences prior to AFS. Reported detection limit was in the ultra-trace range (14 pg)	52
Ba	Dog serum	ICP-MS	Case of acute barium poisoning in a dog subsequent to ingestion of a common handheld pyrotechnic (sparkler)	89
Br	Saliva	ICP-MS	Microwave-assisted dissolution sample preparation method followed by ICP-MS. Limits of quantification achieved for Br and I were 0.052 μg mL^−1 for Br and 0.022 μg mL^−1 respectively	22
Cd	Breast milk	ICP-MS/MS	Breast milk from mothers in Norway (n = 300) was analysed. Number of amalgam fillings and fish consumption were significantly positively associated with Hg concentration (p < 0.001)	109
Cr	Skin	XRF	Portable hand-held XRF system	60
Cr	Blood	ICP-MS	Dual mobility cups are an implant type used for hip arthroplasty. In this study they were found to increase blood Cr but not Ni. It is noted that Co was not measured	79
Cu	Hair	FAAS	Microextraction process based on the chelation of Cu^II from aqueous solutions with 4-(2-thiazolylazo)resorcinol, which is then moved to the organic supramolecular phase by ultrasonic waves	25
Cu	Breast cancer cells	TOF-SIMS	TOF-SIMS combined with delayed extraction applied to probe Cu localisation (at subcellular resolution) in fixed MDA-MB-231 breast cancer cells	72
Cu	Blood, serum	TXRF, AAS, ICP-MS	AAS, ICP-MS and TXRF methods were compared. There was an approximately 80% negative bias for serum Se results obtained by TXRF compared to ICP-MS possibly due to the loss of volatile Se species during the drying process required for TXRF sample preparation	55
Fe	Rat skin	XRF	Portable hand-held XRF system	59
Gd	Cerebrospinal fluid	ICP-MS	Investigation into whether Gd penetrates human cerebrospinal fluid after MRI with a Gd-based contrast agent	92
Gd	Cerebrospinal fluid	ICP-MS	Accumulation of Gd in human cerebrospinal fluid after gadobutrol-enhanced MRI	91
Gd	Rat brain and plasma	ICP-MS	Gd-containing contrast agents injected into rats and deposition in brain measured	93
Gd	Skin and brain	LA-ICP-MS/MS	Both Gd and P were monitored with a mass shift of +16, corresponding to mono-oxygenated species, as well as Zn, Ca, and Fe on-mass. This method resulted in a significantly reduced background and improved limits of detection for P and Gd. These method adaptations allowed elemental bioimaging to be performed with resolution of 5 μm × 5 μm, allowing the detection of small Gd deposits in fibrotic skin and brain tumour tissue with diameters of approximately 50 μm	75
Gd	Dried blood spots	LA-ICP-MS	See main text (ICP-MS section)	43
Hg	Blood	ICP-MS	There was no association between maternal blood Hg levels and birth weight among 15 444 pregnant women in Japan	96
Hg	Hair, blood and urine	ID-GC-ICP-MS	Hg concentrations (including speciation) in individuals occupationally exposed to gaseous elemental Hg	68
Hg	Breast milk	ICP-MS/MS	See Cd, ref. 109	109
Ho	Mouse 3T3 fibroblast cells	LA-ICP-MS	Adherent 3T3 fibroblast cells were stained with two metal dyes (mDOTA-Ho, Ir-DNA-intercalator). Detection limits of 12 fg for Ir and 30 fg for Ho achieved and 57 ± 35 fg Ir and 1192 ± 707 fg Ho quantified per single cell	70
I	Saliva	ICP-MS	See Br, ref. 22	22
I	Urine and serum	ICP-MS	Median urine iodine in 72 pregnant women in Beijing was 125.5 g L^−1 (lower than that WHO limit for iodine deficiency (<150 g L^−1)) and median serum iodine level was 69.0 g L^−1	94
I	Breast cancer cells	ICP-MS	See main text (ICP-MS section)	42
Ir	Mouse 3T3 fibroblast cells	LA-ICP-MS	See Ho, ref. 70	70
Mg	Bone	μXRF	Study into the spatial distribution of trace elements in bone samples	76
Mn	Blood, fingernail	ICP-MS	Fingernail Mn was significantly correlated (Spearman’s rho = 0.44; p = 0.02) with self-reported occupational Mn exposure over the past 16 years in 60 Chinese male workers	115
Ni	Blood	ICP-MS	See Cr, ref. 79	79
Os	Mouse serum	SEC-ICP-MS	Serum-binding preferences of the organoruthenium compound plecstatin-1 and its isosteric Os analogue (two metallodrugs with anticancer activities)	97
Pb	Hair	LA-ICP-MS	Investigation of a Tl poisoning case	99
Pb	Bone	HR-CS-ETAAS	Direct determination of lead in bones using slurry sampling	28
Pb	Urine	GF-AAS	Ultra-trace analysis using silica-based amino-tagged nanosorbent which was ultrasonically dispersed in the sample of interest for the adsorption of Pb ions. Following the adsorption, the nanosorbent was removed from the sample and underwent the desorption process using a small volume of HCl before final Pb determination by GF-AAS	27
Pb	Bone	XRF	Portable device for bone Pb measurement tested in cortical and trabecular bones from 31 cadavers	58
Pb	Breast milk	ICP-MS/MS	See Cd, ref. 109	109
Ru	Mouse serum	SEC-ICP-MS	See Os, ref. 97	97
Sb	Urine	ICP-MS	Creatinine corrected Sb concentrations were significantly higher in pregnant women in China (n = 2093) with gestational diabetes mellitus compared with those without the condition (median value: 0.49 μg g^−1vs. 0.38 μg g^−1; p = 0.001)	85
Se	Plasma	UPLC-IDA-ICP-MS/MS	Se peptide quantification achieved by gradient elution UPLC followed by removal of organic solvent from eluent by addition of 20% O2 in Ar prior to introduction to the ICP plasma	62
Se	Blood, serum	TXRF, ICP-MS	See Cu, ref. 55	55
Se	Urine and breast milk	ICP-MS	Median daily Se intake in pregnancy in a cohort of women in New Zealand based on urinary excretion was 49 μg per day, with 59% of women below estimated average requirement. Median Se concentration in breast milk was 11 μg L^−1	116
Se	Blood	ICP-MS	See Hg, ref. 96	96
Sn	Blood and urine	HPLC-ICP-MS	Urinary trimethyl Sn shown to reflect blood trimethyl Sn in workers exposed to organic Sn species	117
Sr	Hair	MC-ICP-MS and ICP-MS	^87Sr/^86Sr ratio of hair was related to the ratio in tap water, which varied geographically across the USA	98
Th	Serum	ICP-MS	Serum Th significantly associated with risk of gestational diabetes mellitus (p < 0.05) in cohort of 3013 Chinese women	87
Tl	Hair	LA-ICP-MS	See Pb, ref. 99	99
Y	Bone	μXRF	See Mg, ref. 76	76
Zn	Blood, serum	TXRF, AAS, ICP-MS	See Cu, ref. 55	55
Zn	Bone	μXRF	See Mg, ref. 76	76
Various (10)	Blood and placenta	ICP-MS, CV-AAS (total Hg), GC-ECD (methylHg)	Total Hg and Sb levels in the cord blood were twofold higher than those in the maternal blood in a Japanese cohort (n = 594–650)	18
Various (11)	Urine	XRF	Multielemental analysis using a low-power benchtop system	118
Various (11)	Follicular fluid	ICP-MS/MS	See main text (ICP-MS section)	36
Various (13)	Urine	ICP-MS	Reference values for trace elements in the urine in the Italian general population	17
Various (31)	Placenta	ICP-MS	Trace element concentrations and relationships with birth outcomes	83
Various (4)	Breast milk	SEC-ICP-MS	Human whey protein fractions in breast milk determined by ultra-centrifugation. Identification of metals present in different protein fractions	61
Various (4)	Various rat tissues	ICP-MS	Distribution of toxic trace elements in various rat tissues following exposure to either single elements or mixtures of elements	119
Various (5)	Blood	ICP-MS	Stability study	21
Various (5)	Rat blood and heart	SR-TXRF	Quantification of element concentrations after total body irradiation with low (0.1 Gy) and high (2.5 Gy) doses	56
Various (5)	Liver	μSR-XRF	Quantitative trace element mapping in liver tissue from patients with Wilson’s disease	77
Various (6)	Joint fluid, blood and periprosthetic tissue	ICP-MS	In patients with total joint arthroplasty concentrations of metals decreased dependent on the distance of the tissue from the implant	80
Various (7)	Bone	AAS	Chemical separation of the components of bone tissue based on their selective solubility followed by determination of trace element concentrations by AAS	120
Various (8)	Lung	XRF microscopy	Characterisation of trace elements in environmental particulates in lung tissues	78
Various (9)	Hair	ICP-MS	Study showing decreased hair Mo/Co, Co/Fe^2+ ratios in patients with long-term dental Ti implants and amalgams	82

---

Table 2 Foods and beverages

Element	Matrix	Technique(s)	Sample preparation/comments	Ref.
—	Rice	LIBS	LIBS coupled with linear discriminant analysis was proposed as the best tool for rice geographic origin classification	47
Al	Duplicated diet samples	ICP-MS, LC-ICP-MS	Daily intake of Al, total As, iAs and Pb in Japan were estimated in a duplicate diet study involving 319 subjects	107
As	Diet study	ICP-MS	A 3 day duplicate diet study assessed the intake of total As and iAs from food and drinking water for children and pregnant women	104
As	Duplicated diet samples	ICP-MS, LC-ICP-MS	See Al, ref. 107	107
As	Fish	HPLC-ICP-MS	Fish samples were extracted with 40 mg protease and 0.75 mL of 0.5% (v/v) HCl. Separation of 11 As species was achieved in less than 17 min on an anion-exchange column by gradient elution at 20 °C with (NH4)2CO3 and using kinetic energy discrimination to eliminate the spectral interference of ^40Ar^35Cl^+ on ^75Ar. The LODs ranged from 0.11 μg kg^−1 to 0.59 μg kg^−1. On spiked fish samples, repeatability, as RSD%, was between 1.1% and 7.6% and recoveries ranged from 91% to 106%	32
As	Fish tissue CRM	HPLC-ICP-MS	Graphene or graphene oxide, as the stationary phase, allowed the separation of inorganic and organic species of As, Hg and Se, with retention times from 12 min to 32 min. Recoveries were between 92% and 96%	63
As	Seafood, seaweed, CRMs	HG-ICP-MS	The selective conversion of iAs to volatile AsH3 was obtained using HCl (8 mol L^−1) and H2O2. Contributions from other As species were 20% (methylarsonate) and less than 1% (DMA, TMAO). The results were compared with those obtained with the reference method. A reproducibility study was conducted with three European laboratories	65
As	Seafood, seaweed	LC-ICP-MS	Samples were heated at 100 °C in 0.3 mol L^−1 HNO3. As species were determined by LC-ICP-MS, on an ODS column with an ion-pair reagent in the mobile phase. Performance parameters ranges were LOD: 0.0023–0.012 mg kg^−1; LOQ: 0.0077–0.042 mg kg^−1; repeatability: 3.0–7.4%: Intermediate precision: 4.4–7.4% and recoveries: 94–107%	103
Bi	Tap, bottled and mineral water	FAAS	An H2-assisted T-shaped slotted quartz tube-atom trap was applied to pre-concentrate Bi prior to FAAS determination. LOD and LOQ were 0.95 μg L^−1 and 3.2 μg L^−1, respectively	121
Ca	Food, RMs	ICP-MS/MS	ICP-MS/MS, aided by a collision/reaction cell, was applied to resolve interferences on the major Ca and Cl isotopes	37
Cd	Coconut water and milk	HR-CS-ETAAS	Higher concentrations of Cd and Pb were observed in 14 natural vs. 16 industrial coconut water samples (Cd: 0.42–18.72 μg L^−1vs. <0.06–1.49 μg L^−1; Pb: <0.70–36.32 μg L^−1vs. 6.57–29.02 μg L^−1). In 16 coconut milk samples, the levels of Cd and Pb ranged from <0.10 to 5.93 ng g^−1 and from <0.85 to 22.41 ng g^−1, respectively	122
Cd	Dark, milk and white chocolate, cocoa powder products	ICP-MS	Cd concentrations ranged from 0.00021 to 2.3 mg kg^−1, in chocolate products, and from 0.015 to 1.8 mg kg^−1 in cocoa powder products, present on the Japanese market	123
Cd	Food (corn flour, bean, cabbage), mineral water, CRMs	FAAS	Cd pre-concentration was achieved by ultrasound-assisted emulsification micro-extraction with solidification of floating organic droplet using 1-(2-thiazolylazo)-p-cresol as complexing reagent. Analysis of CRMs yielded recoveries between 90% and 100%	24
Cd	Human milk	ICP-MS	Cd, Hg and Pb content was determined in 300 human milk samples by means of triple quadrupole ICP-MS, after microwave-aided acid digestion. The findings were investigated in relation with dietary habits, smoking and amalgam fillings	109
Cd	Rice	LIBS	Rice sample preparation for LIBS was achieved by ultrasound assisted extraction in a HCl solution, dropping the solution on a glass slide, centrifugation and drying on a heater. This enhanced the LODs and LOQs for Cd (2.8 μg kg^−1 and 9.3 μg kg^−1) and Pb (43.7 μg kg^−1 and 145.7 μg kg^−1), compared with the conventional pellet method	48
Cd	Soil, vegetables, CRM	ETAAS	A new deep eutectic solvent consisting of a 1 + 2 (molar ratio) mixture of 1-decyl-3-methylimidazolium chloride–1-undecanol was applied with DLLME to extract and concentrate Cd, Hg and Pb from soil and vegetables samples prior to determination by ETAAS. Enrichment factors ranged from 114 to 172. LODs were between 0.01 and 0.03 μg kg^−1. Precision experiments (N = 7) yielded RSD% <4.1% (Hg, 0.80 μg kg^−1) and <6.6% (Cd and Pb, 0.20 μg kg^−1)	33
Cl	Food, RMs	ICP-MS/MS	See Ca, ref. 37	37
Cr	Drinking water, waste water, industrial waters, recipient waters	ICP-MS, HPLC-ICP-MS	The performance of a fully automated prepFAST IC system coupled to ICP-MS was investigated. For the speciation of Cr^III and Cr^VI in water, LODs were 7 ng L^−1 (Cr^VI) and 12 ng L^−1 (Cr^III) and the results compared with those of conventional HPLC-ICP-MS. The total content of 63 elements measured in a variety of samples (sludges, soils, organic waste, ashes, biological samples, paint) generally agreed with those obtained after digestion with aqua regia	34
Cu	Alcoholic and non-alcoholic drinks	HR-CS-FAAS	Simultaneous determination of Cu, Pb and Zn, using internal standardization with Ag to overcome matrix effects. LOQs were 0.016 mg L^−1 (Cu), 0.099 mg L^−1 (Pb) and 0.040 mg L^−1 (Zn), respectively. The RSD% was <4% and recoveries ranged from 92% to 111%	44
Cu	Food, herbs and spices, hair, tobacco and hashish, water, CRMs	FAAS	A supramolecular solvent consisting of THF and 1-decanol was investigated for the micro-extraction of Cu^II complexes with either 4-(2-thiazolylazo)resorcinol or 5-methyl-4-(2-thiazolylazo) resorcinol from a variety of food and environmental samples	25 and 124
Cu	Wine	Stripping potentiometry, AAS, AES	Stripping potentiometry, HPLC and solid-phase or liquid–liquid extraction followed by AAS or AES were evaluated for the speciation of Cu and Fe in wine	66
F	Fish and seafood	HR-CS-ETAAS	Fish fillets and seafood were dried at 110 °C and minced finely. Less than 1.2 mg were introduced into the graphite furnace with 20 μg of calcium. F Levels were determined from the molecular absorption of CaF, with an absolute LOD of 0.28 ng and a characteristic mass of 0.14 ng	125
Fe	Wine	Stripping potentiometry, AAS, AES	See Cu, ref. 66	66
Hg	Biological samples, CRMs	HPLC-ICP-MS	Samples were extracted with 10% (w/w) TMAH at 80 °C for 2 h prior to the chromatographic separation of MeHg and iHg on either adamantyl or octadecylsilyl columns, using isocratic elution. For MeHg, retention times were 6 and 4 min, respectively, and LODs were 0.08 ng g^−1 and 0.13 ng g^−1 (as Hg), respectively. Analysis of biological CRMs yielded results within the expanded uncertainties (k = 2) of the certified values. The RSD% for MeHg was <2% and recoveries were 101 ± 1% (MeHg) and 103 ± 3% (iHg)	64
Hg	Bottled natural mineral water	CV-AFS	Hg concentrations in the range from sub-ng to a few ng L^−1 were determined by CV-AFS in a total of 255 water samples	112
Hg	Fish (canned sardines)	CVG-AFS	After acid digestion of the sample, Hg and Se were determined simultaneously, with LODs of 0.33 ng g^−1 (Hg) and 9.18 ng g^−1 (Se), respectively	126
Hg	Fish tissue CRM	HPLC-ICP-MS	See As, ref. 63	63
Hg	Fish tissue RMs	CVG-MC-ICP-MS	To determine Hg isotope ratios for MeHg in fish tissues, MeHg and iHg were first separated off-line using RP HPLC in isocratic mode, with a mobile phase containing L-cysteine as a complexing agent. Subsequently, Hg species were completely oxidated, and the isotope ratio were determined by MC-ICP-MS using CVG with SnCl2 as the reducing agent. The reported combined standard uncertainties for the isotope ratios of ^199Hg : ^198Hg, ^200Hg : 1^98Hg, ^201Hg : ^198Hg and ^202Hg : ^198Hg in fish tissue RMs (BCR 463 and NIST SRM 1947) ranged from 0.10 parts per thousand to 0.22 parts per thousand	35
Hg	Fish tissue, water, CRM	FI CV-AAS	Speciation of Hg^2+ and MeHg was achieved by retention on an aminated Amberlite XAD-resin at pH 4 and subsequent separate elution with 10 mL 0.1% (m/v) thiourea in 3% (v/v) HCl (Hg^2+) and 10 mL of 6 mol L^−1 HCl. Hg^2+ was determined directly, whereas, for MeHg, an oxidation step with KMnO4 was necessary. LODs were 0.148 μg L^−1 (Hg) and 0.157 μg L^−1 (MeHg), respectively. The observed differences from certified values for Hg^2+, MeHg and total Hg ranged from −1.8% to −3.2%	30
Hg	Fish, CRMs	CV-AAS, HR-CS-AAS, HPLC-ICP-MS	After extraction from fish samples, with TMAH and HCl at 75 °C, and separation by RP-HPLC, Hg species (Hg^2+, MeHg, EtHg, phenHg) were converted to Hg^0 using post-column ultraviolet photochemical VG. The cold vapour was analysed either by conventional AAS or HR-CS-AAS, both using a quartz tube cell. Results on fish samples and fish-based CRMs were comparable to those obtained by L-cysteine extraction and HPLC-ICP-MS	45
Hg	Fish, water, CRM	ETAAS	Using ultrasound assisted liquid phase microextraction into 1 + 3 molar ratio choline–phenol as the deep eutectic solvent, MeHg and the Hg^2+ dithizone complex were separated. The enrichment factors were 34 and 18.3 for Hg^2+ and MeHg, respectively	26
Hg	Rice	AAS	The microwave-assisted extraction of MeHg from 1.5 g of rice powder, with 12 mL toluene–6 mL 30% HCl at 110 °C for 5 min, was followed by a back-extraction of MeHg into 2.0 mL of 1.5% L-cysteine, prior to analysis by thermal decomposition amalgamation-AAS. LOD and LOQ were 0.13 and 0.43 ng g^−1, respectively, and the RSD% was <5%	53
Hg	Soil, vegetables, CRM	ETAAS	See Cd, ref. 33	33
I	Milk (raw and processed)	ICP-MS	Iodine was extracted from cow, buffalo, goat, sheep and donkey milk in 0.6% (v/v) NH4OH prior to detection by ICP-MS. Intra-laboratory RSD% was 4.01%	110
I	Total diet	ICP-MS	Alkaline digestion and ICP-MS provided lower LOQs, allowing measurement of the I content in over 90% of 266 food samples, compared to 40% when analysed by perchloric acid digestion and spectrophotometry	106
Mn	Bioaccessible fractions of enteral nutrition formulas	AAS	Pre-concentration of Mn was achieved using CPE, under the following conditions: pH 10; chelating reagent concentrations of 0.36 mol L^−1 8-hydroxyquinoline, 0.09 M 1-phenyl-3-methyl-4-benzoyl-5-pyrazolone and 0.09 mol L^−1 APCD; surfactant 25% (w/v) Triton X-100; equilibration temperature 85 °C for 30 min. An LOD of 0.015 mg L^−1 and an LOQ of 0.050 mg L^−1 were obtained. Free Mn could be distinguished from previously bound Mn	127
Mo	CRMs (fresh water, sea water), food supplements	ICP-MS	The efficiency of sample introduction via photochemical vapour generation was in the range between 46% and 66%, compared to liquid nebulisation. An LOD of 1.2 ng L^−1 and RSD% of 3% (RSD) at 250 ng L^−1 were achieved	39
Pb	Alcoholic and non-alcoholic drinks	HR-CS-FAAS	See Cu, ref. 44	44
Pb	Coconut water and milk	HR-CS-ETAAS	See Cd, ref. 122	122
Pb	Duplicated diet samples	ICP-MS, LC-ICP-MS	See Al, ref. 107	107
Pb	Rice	LIBS	See Cd, ref. 48	48
Pb	Soil, vegetables, CRM	ETAAS	See Cd, ref. 33	33
Sb	Bottled mineral water	FI-HG-AAS	The speciation of Sb^III and Sb^V was achieved by means of the selective adsorption of Sb^III on a SiO2/Al2O3/SnO2 minicolumn. Sb^V was determined by difference between Sb^3+ and total Sb	67
Sb	Tap, bottled and mineral water	FAAS	An H2-assisted T-shaped slotted quartz tube-atom trap was applied to pre-concentrate Sb prior to FAAS determination. LOD and LOQ were 0.75 μg L^−1 and 2.49 μg L^−1, respectively	128
Se	Coconut water	PVG-FI-ETAAS	Se was determined in coconut water using photochemical vapour generation followed by FI of Se volatile species into the ETAAS integrated graphite platform, coated with iridium. LOD and LOQ were 0.65 μg L^−1 and 2.2 μg L^−1, respectively. Recovery from spiked coconut water samples ranged between 80% and 103% for Se^IV, Se^VI and SeMet	54
Se	Eggs	HG-AFS	Egg samples were digested overnight in a 1 + 1 mixture of concentrated HNO3–HClO4, then heated until clear at 200 °C. After cooling at room temperature, 5 mL of 6 mol L^−1 HCl was added and the solution heated again until clear, cooled and transferred to a 100 mL flask with 1 mL 10% potassium ferricyanide and 10% HCl. Se was determined by HG-AFS, using 1.5% KBH4 and 2% HCl as the carrier fluid. The LOD was 0.01 μg L^−1, and RSD% ranged from 0.07% to 0.72%. Recovery from spiked samples was between 96.12% and 99.1%	129
Se	Fish (canned sardines)	CVG-AFS	See Hg, ref. 126	126
Se	Fish tissue CRM	HPLC-ICP-MS	See As, ref. 63	63
Se	Food samples, CRMs	HR-CS-ETAAS	Ultrasound-assisted dispersive micro-SPE of Se on halloysite nanotubes used as solid sorbent, achieved a pre-concentration factor of 18. The RSD% was between 6 and 9%. CRMs and food samples (corn flour, rice flour, oat flour, buckwheat flour, garlic, cranberries, goji berries, and raisins) were analysed	29
Sr	Wine	LIBS	2 mL wine samples were dried on aluminium or silicon wafers with a 25 cm^2 surface area, to overcome problems with analysis of liquid samples	50
Zn	Alcoholic and non-alcoholic drinks	HR-CS-FAAS	See Cu, ref. 44	44
Zn	Lettuce (Lactuca sativa L.)	spICP-MS	The uptake of Zn, supplied as ZnO NPs or ZnCl2 by a model edible plant was investigated using spICP-MS. Two-dimensional chromatography (SEC and HILIC) followed by parallel ICP-MS and ESI-qTOF-MS/ESI-FT-Orbitrap-MS detection, allowed the investigation of Zn species and their distribution in the plant tissues	111
Various	Lyophilised seafood, food samples	LIBS	A two-step measurement procedure, consisting in recording two emission spectra with different delays between the laser pulse and the detector gate, was reported to improve the sensitivity of calibration-free LIBS to measure major and minor element concentrations	51
Various	Whisky	ICP-MS, CV-AAS	The content of chemical elements was investigated as a tool to distinguish between whiskies of various provenance	114
Various	Wine	ICP-MS	The determination of several elements, including Ba, Ca, Cs, Cu, Mg, Mn, Na, Ni, Pb and Zn, allowed classification of wines according to viticultural and oenological factors	113
Various (4)	Beer	XRF	The content of Ca, Cl, K and P in 52 samples of bottled beer from 11 countries were determined by XRF, to assess the contribution of beer to the dietary intake of these elements	108
Various (4)	Human milk	SEC-ICP-MS	Metalloproteins (Co, Cu, Mo, Se) in the whey protein fractions obtained by ultra-centrifugation of human milk were identified and quantified by SEC-ICP-MS. The protein profile was determined by N-PAGE and computer assisted image analysis	61
Various (5)	Meat	XRF, ICP-MS	The simultaneous determination by XRF of Ca, Cu, Fe, K and Zn in ground meat samples, suspended in an aqueous solution of Triton X-100 and polyvinyl alcohol, was compared with ICP-MS with acid digestion	57
Various (6)	Tap water	spICP-MS	spICP-MS was applied to detect the presence of particles containing Ag, Cu, Fe, Pb, Sn or Ti in tap water samples	41
Various (7)	Infant milk formulas	EDXRF	The concentrations of Br, Ca, Fe, K, Rb, Sr and Zn, determined by EDXRF in 28 infant milk formulas, were compliant with stated limits	130
Various (8)	Plant and food materials	LA-ICP-MS	After sample homogenisation and particle size reduction to 10 μm, As, Cd, Cu, Mn, Ni, Pb, Se and Zn were determined in pressed pellets of plant and food materials	23
Various (11)	Milk powder, CRM	FAAS, ETAAS, HG-AAS	An investigation based on central composite design revealed that optimal conditions for the digestion of milk powder samples, prior to determination of As, Cd, Cr, Cu, Fe, Hg, Mn, Ni, Pb, Se and Zn, were: final digestion temperature of 190 °C, HNO3 concentration at 56.8% (w/v) and digestion time of 8.47 min. Recoveries ranged between 92% and 108%, while repeatability (as RSD%) was <6.59%	131
Various (14)	Cattle blood serum, feed	ICP-MS	No differences were detected for the concentrations of 14 elements (As, Cd, Co, Cr, Cu, Fe, Hg, I, Mn, Mo, Ni, Pb, Se and Zn) in blood serum of cattle from organic and conventional farms, except for Cd. In concentrated feed, the levels of Cu, I, Se and Zn were significantly higher in samples from conventional farms	132
Various (15)	Total diet	—	Occurrence data for 15 elements (Ag, Al, As, Ba, Cd, Co, Cr, Ga, Ge, Ni, Sb, Sn, Sr, Te and V) in food for infants and toddlers were reported	105
Various (16)	Wine	ICP-MS	A high-temperature torch integrated sample introduction system, operating at 125 °C, with a liquid flow rate of 30 mL min^−1, allowed the direct aspiration of untreated wine samples into the ICP-MS nebuliser. 16 elements (As, Cd, Cu, Cr, Fe, Gd, Mn, Mo, Nd, Ni, Pb, Sm, Tb, Ti, V, Zn) were determined, with LODs ranging from 0.002 to 6 mg kg^−1, in 10 wine samples, and the results compared with those obtained by a conventional method. Sample throughput was 10 h^−1	38
Various (10)	Olives	ICP-OES	The chemometric technique k-NN, applied to data obtained for ten elements (Ca, Cu, Fe, K, Mg, Na, P, S, Se and Zn) in olive samples, allowed distinction between their organic or conventional origin	133

---

Other analytical reviews include a Journal of Analytical Atomic Spectrometry tutorial review that focuses on ICP-MS for the analysis of metal-containing nanoparticles and colloids.⁷ This excellent review provides a brief history of technical developments since 1983 leading to applications of the technique at the nanoscale. The extensive article (231 references) describes the ICP-MS techniques used to analyse natural and engineered NPs, either directly or with hyphenated systems, considers the benefits and pitfalls, including some practical advice, and gives some examples from recent literature. Within this rapidly developing field it provides an encyclopaedic account of a very important topic.

Three one-off analytical reviews were prepared by Chinese authors,^8–10 with two of them written in Chinese. The flourishing work from China was highlighted in our last Update,¹ and these three reviews address similar material as in our Updates, with much of the core material from Chinese authors.

Our recent Updates have also noted the rapid increase in applications of laser-induced breakdown spectroscopy (LIBS). Gaudiuso et al.¹¹ draw attention to applications of LIBS for diagnosis of disease and treatment, for on-line feedback during surgery and dentistry and for detection of bacteria and viruses. Some discussion of future applications is included in the review, referring to cancer diagnosis, chemometrics for discrimination of different tissues and elemental imaging. Such work is further discussed in Section 4.3 of this Update.

With reference to assessing the quality of food, Bianchi et al.¹² discuss how spectral imaging, isotope ratio-MS and other techniques are applied to food control and testing for food authenticity. Attention is drawn to the importance of quality control and quality assurance systems and to metrological traceability, concepts with which all analysts are well aware as is seen from two other reviews. Pashkova et al.¹³ reviewed techniques used to determine inorganic elements in milk, observing that EDXRF techniques predominate. However, they go on to recommend XRF for routine nutritional and quality assessment and for environmental monitoring in dairy and milk products. From an extensive body of publications, Braeuer and Goessler¹⁴ observed that there are more than 18 species of As in mushrooms and speculate on whether there might be a different pathway of metabolism to that found in seafoods. These reviews are discussed in greater detail in Sections 7.2.3 and 7.2.5, respectively.

2 Metrology, interlaboratory studies, reference materials and reference ranges

In 2019 the ‘Consultative Committee for Amount of Substance: Metrology in Chemistry and Biology’ (CCQM – http://www.bipm.org) celebrated its 25^th meeting. Established in 1995, its aim is to support the development of metrology in chemistry and biology. Although the same principles apply to measurements in physics and in other disciplines, the variety and complexity of chemical measurements, most of which are affected by the variability of sample composition and the analyte status, pose a greater challenge to the achievement of global comparability. Investigations related to the traceability and comparability of measurements of chemical elements and their species are an important part of the activities of the CCQM. Up to now, 184 international comparisons among national metrology institutes and other designated laboratories were organised in this area. These exercises addressed the measurement of elements and/or their species in samples ranging from pure solutions and items of known composition to natural items, such as, environmental, food and clinical specimens (list and individual reports available from http://www.bipm.org). Since its introduction, the capabilities of ICP-MS for lower detection limits and discrimination of species by their m/z have been an asset for the development of a metrological approach in this area. As part of a series of papers highlighting advances of metrology in chemistry and biology, Sargent et al.¹⁵ reviewed the role of ICP-MS over the years. Using examples from the international comparisons, organised by the CCQM, that involved measurements by ICP-MS, the authors described the impact of instrumental developments, new calibration strategies and implementation of primary reference methods, based on ID, on the accuracy of measurements of chemical elements and their species. In addition, they provided an insight into the on-going work, within the metrology network, and in new areas of measurements involving chemical elements. Accurate measurements of heteroatoms bound to proteins, by means of HPLC-ICP-MS, may allow the SI traceable quantitation of metalloproteins. Nanoparticles are increasingly exploited for energy, electronic, environmental, food and medical applications, although their use also raises concern over their effects on the environment and on human health. Efforts are in place to assess the comparability of measurements for the characterization and quantitation of nanoparticles, using spICP-MS. Finally, progress in sampling and calibration procedures suggest that LA-ICP-MS can now be considered sufficiently mature to allow an inter-laboratory comparison relating to solid sample analysis to be planned in the near future.

Certified reference materials are essential tools to demonstrate the traceability (preferably to the SI) of chemical measurements. Critical issues are a suitable match between the matrices and concentration levels of analytes in the CRMs and the test samples and acceptable uncertainties, associated with the CRMs certified values (ideally less than one third of the target uncertainty for the measurements). Recently, four new candidate reference materials, based on river bottom sediment and animal tissues (cormorant, Baltic herring and codfish), were analysed for their Cd and Pb concentrations in a collaborative exercise. Although aimed primarily to support environmental analyses, two of them could find application in the area of food testing as well. Krata and co-workers¹⁶ described in detail the procedures developed and applied to provide reference measurements of Cd and Pb in the candidate CRMs, using ID-ICP-MS, and the calculation of the associated expanded uncertainties. Samples, 0.4 g, were prepared by acid digestion in an “ultrawave” microwave digestion unit, with a ramp step to 240 °C in 20 min and a holding step at 240 °C for 10 min at a pressure of 140 bars. The optimal composition of the digestion mixture was determined by evaluating the recovery of Cd and Pb from CRMs based on sediments (IAEA-457 and HR-1), lobster tissue (LUTS-1) and bovine liver (SRM-1577c). The mixture of 3 mL HNO₃–1 mL HCl–1 mL HF provided the best results for the inorganic matrix, whereas, for tissue CRMs, digestion with 5 mL HNO₃ was optimal. After dilution of the digests and addition of Rh as the internal standard, the approximate concentrations of Cd and Pb were calculated by comparison with calibration curves. This simpler approach allowed, first, to assess the recovery with different digestion mixtures and, second, to determine the amount of enriched isotope spike needed to achieve optimal isotopic ratio values for ID measurements. The addition of known amounts of the isotopically enriched solutions was performed gravimetrically, directly after weighing of samples. Owing to the high saline content of the digests, removal of potentially interfering elements was achieved using an ion exchange resin AG1-X8 filled in a 3 × 0.8 cm column. Digests were diluted with 32% HCl and loaded onto the resin with sequential elutions of different concentrations of HCl and HNO₃. After evaluation of each fraction for matrix removal and element recovery, the 4^th fraction, for Cd, and the 3^rd fraction, for Pb, were analysed. The uncertainty of the measured values was evaluated starting from a set of equations representing the ID-ICP-MS mathematical model. Values for the input quantities (and associated standard uncertainties) in each equation were obtained either by measurement, mathematical calculation or from certificates and combined according to international guidelines, using dedicated software. The efforts spent in devising this rather complex procedure yielded the expected results in terms of documented traceability to the SI of the assigned values and expanded uncertainties between 1 and 5%. For the Baltic herring and codfish tissue CRMs, the Cd measured values and expanded uncertainties (k = 2) were, respectively, 337 ± 2 μg kg⁻¹ and 2.7 ± 0.4 μg kg⁻¹. The corresponding values for Pb were 108 ± 5 μg kg⁻¹ and 43 ± 2 μg kg⁻¹, respectively.

Exposure to environmental pollutants varies over time according to the diversification of anthropogenic activities, such as the increasing use of REE in the electronic industry. Potential noxious effects on the health of individuals and particularly on developing organisms raise concern and call for further investigation. The assessment of background levels of chemical elements in biological fluids and tissues is necessary to allow for comparison between different population groups and detection of trends. However, the concentrations of elements are influenced by numerous variables, including gender, age, life styles, ethnic origin and the environment. It is therefore important to continue to collect information on population reference ranges and their variations across time and geographical areas. In this year’s review, two papers reported investigations of the levels of selected elements in population groups. Among biological fluids, urine is easily accessible and most widely used for occupational exposure surveys. This was selected by Aprea et al.¹⁷ to investigate the background levels of 13 elements (Be, Cd, Co, Cr, Cu, In, Mn, Ni, Pb, Pt, Sb, Tl and V) in a sample of the Italian general adult population. The study involved 13 centres, distributed across Italy. Each centre collected two urine samples (in June and November of the same year) from 20 subjects, for an overall total of 120 men and 140 women, aged between 18 and 60 years. All subjects were non-smokers or had abandoned the habit for at least 5 years. In addition, information was gathered, by means of a questionnaire, on personal characteristics and other factors (occupational or extra-occupational activities, dietary habits, dental fillings containing amalgam, exposure to passive smoking and/or traffic) representing potential sources of exposure. Urine spot samples, collected in plastic containers without the addition of preservatives or stabilisers, were frozen at −18 °C. Prior to analysis, the samples were thawed, brought to room temperature, homogenised and diluted 1 : 5 with 1% HNO₃. Analyses were performed by ICP-MS in six different laboratories, who had successfully participated in 5 different inter-laboratory comparisons on urine samples.

The observed ranges (5^th to 95^th percentile), based on the averages of the results on the two urine samples provided by each subject, were as follows (concentrations all are μg L⁻¹)

Be	Cd	Co	Cr	Cu
<0.010–0.034	0.055–0.900	0.077–2.24	0.050–0.600	5.01–24.0
In	Mn	Ni	Pb	Pt
<0.010–0.013	0.040–1.53	0.372–4.44	0.170–2.64	0.010–0.022
Sb	Tl	V
0.010–0.095	0.060–0.759	0.010–0.855

The authors acknowledged that the size of the group studied was too small to allow for the evaluation of the effects of variables such as gender, age and potential sources of exposure. In addition, owing to the various factors affecting urinary excretion of creatinine, such as muscle mass, physical exercise and meat consumption, the attempt to take into account the individual dilution of spot urine samples by standardising the observed values by gram of urinary creatinine, met with little success. Despite the effort required in collecting 24 h urine samples, these are still the best option for the purpose of assessing background and reference values of chemical elements in urine for the general population taking into account the potential effects of physiological and lifestyle factors.

The developing child is particularly vulnerable to the effects of toxins and consequently prenatal exposure even to low levels of chemicals is a cause for concern and research. Iwai-Shimada and colleagues¹⁸ carried out a birth cohort study in an urban area of Japan, involving 687 pregnant women, to investigate maternal uptake and placental transfer to the foetus of several elements (As, Bi, Cd, Cu, Hg, Pb, Sb, Se, Sn and Zn) and Hg species (MeHg, iHg). Samples of maternal blood were collected at 28 weeks of pregnancy and cord blood at delivery. For placentas, a sample from the lower parts of the root of the cord tissue, collected at delivery, was considered representative and homogenised before storing at −80 °C until analysis. All elements except Hg and Hg species, were determined by ICP-MS after microwave-aided acid digestion. After measurements of total Hg levels by CV-AAS and of MeHg by GC with ECD, iHg concentrations were determined by subtraction. For all elements and species, the distributions of the observed values were skewed. For Bi and Sn, most data were below the LOD, so these elements were excluded from further analysis. The authors reported the median and the 25^th and 75^th percentile concentrations for all other elements in maternal and cord blood and in placenta tissue. Both total Hg and MeHg showed a strong correlation between their respective levels in maternal and cord blood (Spearman’s rank correlation coefficients 0.78 and 0.77), whereas iHg did not. On the other hand, only moderate to weak correlations were observed for As, Pb and Se concentrations in maternal and cord blood and no significant association for any of the other elements. Concentrations of As, Cd, Pb, Se, Zn and total Hg were higher in placenta than in maternal or cord blood, suggesting active mechanisms preventing the transfer of these elements to the developing foetus. Cord to maternal blood median ratios ranged from 0.40 (Cu) to 1.09 (Se), except for Hg (1.81) and Sb (1.95), raising concern for the exposure of the foetus to these elements. The range of Hg levels in maternal blood (25^th to 75^th percentiles: 3.89–7.59 ng g⁻¹) was high, compared to 0.678 ng mL⁻¹ (geometric mean) observed in the United States, 0.64 ng mL⁻¹ (median) and 2.24 ng L⁻¹ in Taiwan. Most of the Hg was present as MeHg (range: 3.68–7.15 ng g⁻¹). Sb levels in maternal and cord blood showed median values (0.20 and 0.39 ng g⁻¹, respectively) lower than those observed in studies in Germany and Western Australia, but a similar skewed distribution reaching much higher maximum values (7.99 and 6.40 ng g⁻¹, respectively). Notwithstanding the large amount of data produced in this study, several questions remain open and it can be concluded that further research is necessary to clarify the mechanisms of placenta transfer and identify the best biomarker (maternal or cord blood) to assess the exposure of the child to toxic elements.

3 Sample collection and preparation

3.1 Collection, storage and preliminary preparation

Most readers will be familiar with the NHANES projects and reports from the CDC. Recognising the very low concentrations of many of the elements that are measured and the authoritative nature of the reports, the laboratory responsible for the trace element determinations have described the procedures used to clean the materials required for sample collection and analysis.¹⁹ This involves screening consumables such as sample collection devices, cleaning all items and testing the washings, and working in an ultra-clean environment. The procedures adopted may not be possible to replicate in all laboratories but the precautions noted in this report should always be borne in mind when planning and executing studies and reporting results. A report by Roos²⁰ considered the precautions to be taken when collecting blood and CSF specimens for studies of neurodegenerative disorders. The discussion included the ambient air, clothing worn by staff, possible contamination from patient’s skin and collection devices. The author suggested that polypropylene collection tubes may not be entirely suitable and recommended PFA vials as a superior alternative.

A second publication from the CDC laboratory examined the stability of blood samples collected for the measurement of Cd, Hg, Mn, Pb and Se.²¹ Human blood samples with K₂EDTA as anticoagulant were spiked, where necessary, to produce pools with low, high and elevated concentrations of these elements. Portions were dispensed into polypropylene cryovials and stored at −70 °C, −20 °C, 4 °C, 23 °C and 37 °C for up to 36 months. Analysis of samples, together with controls and SRMs, were performed by ICP-MS. Those stored at 37 °C and 23 °C were clotted after 4 and 12 months, respectively. Data analysis showed that results from samples stored for 4 months at 37 °C were equivalent to those from the samples kept at −70 °C for 36 months. Similarly, equivalence for the samples at 23 °C was shown for up to 10 months, while storing at 4 °C and −20 °C gave results that were equivalent at 36 months to those from the samples kept at −70 °C.

Examples of simple preliminary preparation for analysis included nothing more than centrifugation prior to measurement of Br and I in saliva by ICP-MS²² and formation of pressed pellets of plant or food materials, for LA-ICP-MS.²³

3.2 Digestion, extraction and preconcentration

The keywords associated with recent work relating to extraction of analytes from biological samples are ultrasound and microextraction.

Following a simple digestion of beans and cabbage leaves, Cd in the aqueous sample was extracted and concentrated by a process described as ultrasound-assisted emulsification microextraction with solidification of floating organic droplet.²⁴ To 5 mL of the digestate, 100 μL of the complexing reagent, 1-(2-thiazolylazo)-p-cresol, dodecanol and 1 mL of a pH 9.0 buffer were added and the mixture sonicated for 3 minutes. After centrifugation, it was placed in an ice bath to solidify the organic phase which was then removed, thawed and diluted with EtOH/HNO₃ to reduce the viscosity for analysis by FAAS. The LOD was 0.60 μg L⁻¹ and the precision was 3.0% RSD at a concentration of 10 μg L⁻¹. Recoveries from three CRMs were 90 to 100%.

Alothman et al.²⁵ measured concentrations of Cu in samples of hair, a range of spices and other sample types, by microsampling FAAS, following a process they developed for extraction. The Cu in aqueous solutions prepared from the samples was chelated at pH 6 with 0.1% 4-(2-thiazolylazo) resorcinol, and then transferred to the organic supramolecular phase with ultrasonic energy applied for up to 8 minutes. The LOD was reported as 1.13 μg L⁻¹.

A fast, and sensitive ultrasound-assisted choline-phenol-based (1 : 3, molar ratio) deep eutectic solvent (DES) was prepared and used in a liquid phase microextraction procedure by Thongsaw et al.²⁶ to determine Hg concentrations in water and fresh-water fish samples. Following pH adjustment, 10 mL of sample solution was mixed with 100 μL of 0.1% dithizone chelating reagent followed by 500 μL of DES solution. After addition of 500 μL tetrahydrofuran, the sample was transferred to an ultrasonic bath for 2 minutes to accumulate the DES phase as a turbid solution. A supernatant was isolated by centrifugation at 4032 × g for 10 minutes and the DES phase was collected for analysis by ETAAS. The LODs were less than 0.1 ng mL⁻¹ with RSDs less than 4.1%. The accuracy of the developed method was verified by analysis of NIST SRM-1566b oyster tissue.

To determine concentrations of Pb in urine and water Sobhi et al.²⁷ first prepared and characterised (using FTIR, XRD, TEM, TGA and BET) a silica-based amino-tagged nanosorbent (MCM-41@NH₂). The nanosorbent was ultrasonically dispersed in the sample to enable adsorption of Pb, which was desorbed using a small volume of HCl. The Pb was then measured by ETAAS. The pH, sorbent content, ionic strength, sonication time and desorption solvent parameters were investigated and optimised. A linear range of 0.1–1.0 μg L⁻¹ for Pb was achieved with an assay precision of 4.8–9.2% and accuracy of 92–110%.

Husakova et al.²⁸ were interested in measuring the concentration of Pb in bone by HR-CS-ETAAS. Rib bone of otters were cleaned, freeze dried and ground to a particle size of less than 315 μm diameter. Slurries were prepared from 100 mg powdered bone (moistened with drops of 2% Triton X-100 in EtOH) in either glycerol or 5% HNO₃. After 2–10 minutes sonication, 25 μL of the slurry and 10 μL of a chemical modifier (0.5 g L⁻¹ of Pd and 25 g L⁻¹ of citric acid) were injected into the graphite furnace. Results were verified by analysis of NIST SRM 1486 Bone Meal. An LOD of 9.1 μg kg⁻¹ and characteristic mass of 7.6 pg were obtained. The concentrations of Pb in otter bones ranged from 0.30 to 2.05 mg kg⁻¹, n = 10.

High-resolution continuum source graphite furnace atomic absorption spectrometry was also used by Krawczyk-Coda²⁹ in a procedure developed to determine concentrations of Se in food samples. After microwave acid digestion, the Se was ultrasonically-assisted adsorbed onto halloysite nanotubes and the samples analysed. A preconcentration factor of 18 was obtained and the reported LOD was 0.01 μg L⁻¹ with precision of 6–9% RSD. Results from the analysis of NIST SRM 1567a (wheat flour) and NIST SRM 1549 (non-fat milk powder) were in good agreement with the certified values. Concentrations of Se were reported for samples of corn flour, rice flour, oat flour, buckwheat flour, garlic and goji berries (0.025–0.045 μg g⁻¹). Values for cranberries and raisins were below the LOD.

New adsorbants continue to be investigated. Amberlite XAD-series resins are useful materials for SPE and recent work with Amberlite XAD-4 modified by introducing amino groups into the aromatic ring (AAXD-4), showed that this sorbent has a high affinity for metal ions and ion exchange capacity. Also, a polyaniline sorbant with amino groups (PANI) showed a high affinity for Hg^II and MeHg^I species. On the basis of these properties and the structural similarities present between PANI and the AAXAD-4, Caylak et al.³⁰ therefore hypothesised, that AAXAD-4 would be a useful sorbent for Hg ions. It was demonstrated that Hg ions were retained on the AAXAD-4 resin column at pH 4, and that sequential quantitative elution of Hg^II and MeHg^I was achieved using 10 mL of 0.1% (m/v) thiourea in 3% (v/v) HCl and 10 mL of 6 mol L⁻¹ HCl, respectively. Using FI-CV-AAS detection limits for Hg^II and MeHg^I ions were 0.148 and 0.157 μg L⁻¹, respectively. The method was validated by analysis of CRM NRCC-DORM-4 and was applied to a batch of fish and mussel samples. A new sensitive method to measure concentrations of Sb in water by ETAAS was developed using N-benzoyl-N-phenylhydroxylamine and 1-butyl-3-methylimidazolium hexafluorophosphate ([C(4)mim][PF₆]) in a single drop microextraction procedure.³¹ Under the optimized conditions, the LOD and the EF of Sb^III were 0.01 μg L⁻¹ and 112, respectively with precision of 4.2% RSD at 0.5 μg L⁻¹.

In a procedure for the speciation of As in fish developed by Zhao et al.³² compared water, aqueous acidic solutions, aqueous organic solvents and enzyme solutions as extraction media prior to analysis by HPLC-ICP-MS. Optimal results were found using 40 mg protease with 0.75 mL 0.5% (v/v) HCl as the extraction agent. The major part of the publication refers to the chromatography conditions which enabled the detection of 11 As species.

The use of a DES extractant was referred to above. Habibollahi et al.³³ elected to use a similar approach, with a DES consisting of 1-decyl-3-methylimidazolium chloride and 1-undecanol in a molar ratio of 1 : 2. Samples were digested using HNO₃ and HClO₄ and an aliquot was mixed with 50 μL of the DES, containing 15 μL diethyldithiophosphoric acid (chelating agent), which was maintained at 55 °C in a water bath. Following addition of 400 mg NaCl, the mixture was shaken and centrifuged to separate into phases. The tube was transferred to an ice bath to solidify the fine droplets of DES. The procedure was applied to the determination of Cd, Hg and Pb in soil and vegetables irrigated with treated municipal wastewater with detection by ETAAS. The enrichment factors for the three metals were 114–172 and the LODs were 0.01–0.03 μg kg⁻¹.

4 Progress with analytical techniques

4.1 Mass spectrometry

Two new speciation analysis methods were reported during the review period. The first was an automated all-in-one system for the measurement of total metals in environmental and biological samples and chromium species in water.³⁴ Separation of Cr^III and Cr^VI was achieved by LC using an AEC column preceded by a syringe driven speciation module and prepFAST M5 unit for online dilutions. The addition of NH₃ to the ICP-MS carrier solution was found to improve detection limits, which were impressive at 12 ng L⁻¹ and 7 ng L⁻¹ for Cr^III and Cr^VI, respectively. In a second study, a novel approach for Hg speciation was described using HPLC followed by CV generation and MC-ICP-MS for the analysis of fish tissue samples.³⁵ Mercury species, including MeHg and iHg, were separated using RP-HPLC in isocratic mode with a mobile phase containing the complexing agent L-cysteine. Isolated Hg species were then oxidised to remove potential interferences caused by the presence of L-cysteine, before CV generation with SnCl₂ as reducing agent which improved the S/N of MC-ICP-MS. Fish tissue CRMs containing MeHg were used to confirm accuracy and precision at environmentally relevant concentrations with combined standard uncertainties ranging from 0.10 to 0.22 parts per thousand.

This year saw further advances in the use of ICP-MS/MS for the analysis of both clinical and food samples. Galusha et al.³⁶ described a multi-element ICP-MS/MS (with octopole reaction system) method for the measurement of 11 elements in follicular fluid at trace and ultra-trace concentrations. The LODs achieved for ultra-trace elements ranged from 5.6 ng L⁻¹ for Cd to 0.11 μg L⁻¹ for Mo. A total of 197 samples were analysed and Co, Cu, Mn, Mo, Sr, and Zn were detected in all. In another study a novel method for the determination of Ca and Cl in food samples was developed using ICP-MS/MS with H₂ as reaction gas to overcome the ⁴⁰Ar⁺ and ¹⁶O¹⁸O¹H⁺ interferences on ⁴⁰Ca⁺ and ³⁵Cl⁺, respectively.³⁷

Novel sample introduction techniques were reported employed with advantages that included increased analytical sensitivity, improved sample throughput, reduced interferences and the ability to perform nanoparticle analysis. A high-temperature torch formed part of a novel integrated sample introduction system coupled to ICP-MS for the measurement of 16 trace elements in undiluted wine.³⁸ The sample introduction system consisted of a MicroMist nebuliser joined to a 9 cm³ single-pass spray chamber heated by a copper coil with the temperature of the chamber wall controlled achieved using a thermocouple. The system worked on the principle of low sample consumption through complete aerosol evaporation before introduction into the ICP, achieving close to 100% analyte transport efficiency. The advantages of this method, when compared to conventional approaches, were reduced interferences associated with complex organic matrices (e.g. matrix effects, plasma degradation and soot deposition), low detection limits (0.002 to 6 μg kg⁻¹) and good sample throughput as no sample preparation was required (approximately 10 per hour). A photochemical VG sample introduction method was developed for the measurement of Mo in fresh water, seawater and dietary supplements by ICP-MS.³⁹ The approach achieved an LOD of 1.2 ng L⁻¹ and good precision (3% RSD at 250 ng L⁻¹). Accuracy was verified by analysis of suitable CRMs. The recent trend for the analysis of single cells using ICP-MS continued with a study by Wang et al.⁴⁰ to investigate the uptake and distribution of Au nanoparticle in HeLa cells. A novel on-line droplet chip system was utilised for sample introduction to the ICP-MS. Further information on Updates to NP analysis can be found in a review article published by Meermann et al.⁷ summarising advances in the analysis of metal-containing NPs and colloids by ICP-MS.

Tap water samples collected from three buildings were analysed using an optimised sp-ICP-MS technique to determine Ag, Cu, Fe, Pb, Sn, and Ti-containing particles that may have originated from source water or generated in situ (e.g. by corrosion) in the distribution systems.⁴¹ Concentrations of Fe, Pb and Sn-containing particles were 26–890, 0.06–4.8 and 0.11–14 ng L⁻¹, respectively. No Ag or Ti particulates were detected in the samples. Concentrations of Pb in the first 125 mL fraction collected were three times higher than those in the subsequent samples. Detection of the Cu particles required modification of the sample introduction system (direct self-aspiration into the nebuliser) to reduce matrix interaction with the auto-sampler tubing. The Cu particles were detected in 50% of the analysed samples at 15–136 ng L⁻¹. While all the metal concentrations were below the health advisory levels, this study showcases the feasibility and first application of spICP-MS to monitor metal-containing particles in tap waters.

Unusual clinical applications of ICP-MS this year included development of an immunoassay and dried blood spot analysis with the introduction of internal standard during sampling. A method for measuring transferrin in breast cancer cell lines by an ICP-MS linked sandwich immunoassay was reported by Alonso-Garcia and co-workers.⁴² In this novel approach, cell lysates were mixed with both iodinated and biotinylated anti-transferrin antibodies. Streptavidin-coated magnetic microparticles were added to capture the immuno-complex and after washing, measurement of I by ICP-MS was carried out. The advantages of this technique, when compared with conventional immunoassays, included a wider dynamic range, enhanced sensitivity, reduced possibility of matrix effects and interferences and higher sample throughput. Sensitivity was particularly important for the measurement of transferrin in breast cancer cell lines as analyte concentrations in this sample type are orders of magnitude lower than in serum. Gadolinium-based contrast agents were spiked into blood prior to LA-ICP-MS quantitative analysis of dried blood spots.⁴³ The workflow was simplified by placing the internal standard (necessary to correct for inhomogeneous blood drying) into the capillary prior to blood sampling, enabling mixing to occur with internal standard prior to deposition onto the filter paper. Optimisation of laser fluence ensured particles larger than μm size were removed from the dried blood spot ablation aerosol leading to increased transport efficiency of particles to the plasma and optimal signal stability as larger particles would not be fully atomised and ionised in the plasma.

4.2 Atomic absorption and atomic emission spectrometry

Although AAS is not extensively employed for applications within the scope of this Update it has been noted that there is a continuing interest where it is possible to take advantage of the speed of analysis and simplicity of the technique. This is best seen in relation to publications that particularly refer to CS-AAS and to atom traps. An example of these positive features is given by the work of Fernandez-Lopez et al.⁴⁴ who report the simultaneous determination of three elements in non-alcoholic beverages, beers, wines and spirits using a single calibration line. The approach, to measure concentrations of Cu, Pb and Zn by HR-CS-FAAS involved the use of Ag as the internal standard for calibration. These workers achieved LOQs of 0.016 mg L⁻¹, 0.099 mg L⁻¹ and 0.04 mg L⁻¹ for Cu, Pb and Zn respectively, precision less than 4% RSD and recoveries were in the range of 92–111%. It was found that Cu concentrations were below 1.28 and 0.15 mg L⁻¹ in alcoholic and non-alcoholic drinks respectively, Zn levels were similarly below 0.37 and 12.3 mg L⁻¹ while Pb was, reassuringly, not detected in any samples.

In Section 3.2, two sample preparation procedures were described where the final measurements were made using HR-CS-ETAAS. These were for Pb in bone samples²⁸ and Se in foods.²⁹ The LOD for Pb was 9.1 μg kg⁻¹ with a characteristic mass of 7.6 pg. For Se, the LOD was 0.01 μg L⁻¹ with precision of 6–9% RSD. These figures of merit compare well against other techniques.

In a novel investigation, conventional CV-AAS and HR-CS-CV-AAS were compared as detector systems following HPLC and photochemical Hg VG.⁴⁵ The LODs with the HR-CS-AAS analysis were superior to that of the more usual CV-AAS system and equivalent to those obtained using an AFS detector. The detection limits were 0.47, 0.84, 0.80 and 2.0 μg L⁻¹ for iHg, MeHg, EtHg and PhHg, respectively. Results were reported for these Hg species in samples of fish and in CRMs.

4.3 Laser induced breakdown spectroscopy

Continuing the trend from previous Updates,^1,46 a number of publications highlighting LIBS as an important tool for food analysis has continued. In two works by Yang and co-workers,^47,48 LIBS was applied to rice samples for different purposes. In the first study,⁴⁷ geographical origin classification was tested with a number of statistical tools and LIBS data from 20 rice samples. One of the benefits of LIBS is the minimal sample preparation required, enabling quick and direct analysis with sufficient discrimination power to differentiate between samples. The researchers found that LDA provided the most robust model. In the second paper,⁴⁸ Cd and Pb were determined in rice using LIBS, achieving good limits of detection and quantification for the technique: 2.8 and 9.3 μg kg⁻¹ for Cd and 43.7 and 145.7 μg kg⁻¹ for Pb. In this case, a simple sample pre-treatment process was applied utilising ultrasound-assisted extraction in dilute hydrochloric acid with drops placed on a glass slide and evaporated. The work highlighted that LIBS has the ability to determine toxic elements in food products at relevant levels. This is contrast to Nam et al.⁴⁹ who developed a method for the separation and quantification of As^III and As^V using LIBS. Anion exchange polymer filters were used to selectively separate As^III and As^V and collected for analysis. The As^III was then converted to As^V for measurement. Excellent spike recoveries were obtained from standards and spiked water CRMs (97–99%), however the LOD was found to be 10 mg kg⁻¹. Whilst the method was simple, quick and cost effective, the lack of suitable sensitivity for the analysis of real-world samples restricts its usefulness. Following on from work reported in the 2019 Update,¹ Bockova et al.⁵⁰ applied the surface-assisted LIBS method to the analysis of Sr in wine, including a comparison of different substrates (aluminium and silicon). The wine was simply prepared by allowing 2 mL to dry using an infrared lamp for 30 min. A number of Sr lines were tested (Sr I 460.733 nm and 481.188 nm; Sr II 407.771 nm and 421.552 nm) whilst calibration was achieved by spiking Sr standards into the wine matrix and internal correction with Ca (Ca I 646.257 nm; Ca II 373.690 nm). It was found that the Sr I line at 460.733 nm with the silicon wafer provided the lowest LOD at 0.053 mg L⁻¹ with an R² of 0.994. Recovery tests showed that both aluminium and silicon substrates performed well when Ca was used for normalisation and the RSDs were on average <5%. The data demonstrated the approach was suitable for use but did not extend to the analysis of multiple real samples.

In a publication by Chen and co-workers,⁵¹ the feasibility of calibration-free LIBS analysis for food products was assessed. Based on the assumption of local thermodynamic equilibrium, calibration-free measurement is possible but typically this has only been applied to major elements at sufficiently high concentration levels due to the S/N. In this work, a two-step measurement process was established to capture major, minor and trace elements. It was tested with three freeze-dried seafood samples providing data for bulk elements (%) such as C, H, N and O, major elements (high ppm) such as Ca, Mg, Na and P, through to minor and trace elements (low ppm) such as Al, As, B, Cd, Cu, Fe, Si, Sr and Zn. The study showed promising results but it lacked in the use of reference materials or corroborative data from alternative established techniques.

A review by Gaudiuso et al.¹¹ highlighted the use of LIBS as an emerging and potentially important technique for the clinical field. The article covered a variety of applications such as monitoring and diagnosis of disease, detection of pathogens, forensic cases, and for surgical and dental operations. It also considered developing areas such as bioimaging and cancer diagnosis. The review provided a good overview of the benefits of LIBS for human and animal health studies.

4.4 Vapour generation procedures and atomic fluorescence spectrometry

Liu and colleagues previously described a tubular dielectric barrier discharge (DBD) trap as an alternative to vapour traps such as quartz tube or graphite furnace. With an AFS detector they reported measurement of 0.1 ng L⁻¹ As from 20 mL water. The same workers subsequently developed a DBD trap with a three-concentric-tube design that afforded an even lower LOD of 2.8 pg from a sample volume of 2 mL. Liu et al.⁵² have now reported the application of this technology to the determination of As in hair and microbiological samples, prepared as slurries in 5% HCl. The instrumentation mixed the slurry sample with 8 g L⁻¹ KBH₄ in 1.5 g L⁻¹ KOH. An Ag-air gas flow transfers AsH₄ vapour to the DBD system where it was retained by an 11 kV discharge. With a switch to a 13 kV discharge under H₂–Ar the hydride was released and passed to the flame in an AF spectrometer. The LODs for hair and microbiological sample preparations were 14 and 8 pg, respectively from 2 mL sample. Recovery of As from spiked samples was 93–102%.

Concerns around dietary exposure to MeHg usually focus upon intakes from fish and other seafoods. Rice, by contrast, is seen as a potential source of inorganic As. Luo et al.⁵³ considered that as rice is a staple component of the diet in many countries, a simple, rapid, sensitive procedure to determine the MeHg content of rice and other cereals should be readily available. A sample of 1.5 g rice powder was extracted with 12 mL toluene and 6 mL 30% HCl at 110 °C for 5 min under microwave irradiation. From 4 mL of this solution, the MeHg was back extracted into 2.0 mL 1.5% L-cysteine solution and the concentration determined by thermal decomposition amalgamation-AAS on a direct mercury analyser. The method was validated by analysis of the CRMs TORT-2 and NIST SRM 2976, achieving 97.6 and 93.8% recovery, respectively. Sensitivity I was an order of magnitude lower than other techniques previously used for this application, achieving an LOD and LOQ of 0.13 and 0.43 ng g⁻¹ respectively. Results from analysis of real samples by the developed procedure and by HPLC-ICP-MS were comparable.

A photochemical device to produce Hg vapour, with detection by HR-CS-AAS, was described in Section 4.2.⁴⁵ This system was found to have increased sensitivity compared to the more usual line source instrumentation with performance equivalent to that seen with AFS systems. The analyser was used together with HPLC to determine four Hg species from fish samples. A second report which featured photochemical vapour generation referred to the determination of Se. This work, by Potes et al.⁵⁴ concentrated on optimisation of the conditions for vapour generation and detection. Volatile Se species, generated in the UV reactor, were separated from the condensed phase in a gas–liquid separator and were swept into a graphite furnace with an integrated graphite platform surface coated with Ir as a permanent modifier. Formic acid 0.44 mol L⁻¹ was selected as the low molecular weight organic compound and addition of 14 mmol L⁻¹ HNO₃ and 145 mmol L⁻¹ H₂O₂ increased the sensitivity. Pre-reduction of Se^VI to Se^IV was performed on-line in a quartz spiral tube coated with a thin film of TiO₂. The insertion of Ar at a flow rate of 115 mL min⁻¹, before the reactor inlet, improved the sensitivity, compared with the introduction of Ar after the reactor.

4.5 X-ray spectrometry

A comprehensive review of recent advances in X-ray spectrometry³ complements the applications with clinical and biological materials, foods and beverages covered in this Update. Imaging applications of X-ray spectrometry are described separately in Section 5.2 below. In addition to this, a review of the analysis of milk and dairy products using XRF spectrometry was published by Pashkova et al.¹³ covering the instrumentation, sample preparation and calibration approaches. A number of examples were also given, providing a well-rounded overview of this technique for a specific food group.

In a publication by Jeffery et al.,⁵⁵ a validation study for the determination of Cu, Se and Zn in serum samples using a benchtop TXRF spectrometer was described, with comparisons to both AAS and ICP-MS. The sample preparation was relatively straight forward, requiring 100 μL of serum mixed with 10 μL of 0.03% PVA and Ga as the internal standard, vortexed then 10 μL pipetted onto a glass disk and dried prior to measurement. The acquisition time was 1000 s and calibration was achieved using Ga and pre-set response factors. Fit for purpose results were obtained for LOD/LOQ, linearity and precision. Recovery was assessed through the comparison with AAS and ICP-MS for a group of patient samples, spike recoveries (n = 3) and analysis of EQA samples. In general, Cu and Zn performed adequately with slight positive biases (4.6–9.6%) in comparison to AAS and ICP-MS observed for the patient samples, spike recoveries of 115% for Cu and 111% for Zn and acceptable results for the EQA samples. However, Se demonstrated a significant negative bias of 28% for the patient samples which increased to 80% for the EQA samples – no spike recovery data was presented. It was postulated that the drying process caused losses due to the volatility of Se compounds, especially considering that the majority of EQA samples were prepared by spiking serum with SeO₂, revealing that further investigation was required for Se. Although the authors concluded that TXRF spectrometry was suitable for the analysis Cu and Zn in serum with a number of advantages such as simple sample preparation, calibration-free, ease of installation in a clinical laboratory and minimal running costs, there were still a number of disadvantages such as the significant analysis time per sample (over 16 min) and a small autosampler (currently only 25 positions). Mota et al.⁵⁶ applied SR-TXRF spectrometry in the investigation of changes in trace element levels after exposure to radiation at low and high doses, representing environmental exposure and medical intervention e.g. radiotherapy. Rat models were used and split into three groups: control, low dose at 0.1 Gy and high dose at 2.5 Gy (n = 6 in all groups). Whole blood and left ventricle heart tissue were collected 24 hours after exposure. The samples were digested with nitric acid, mixed with internal standard (Ga), and 5 μL pipetted on to a Perspex® slide and evaporated. The elements determined were Ca, Fe, K, S and Zn. A number of significant trends (either P < 0.05 or P < 0.01) were observed due to irradiation, such as increased Ca and Zn levels and decreased Fe and K. The authors concluded that the results indicated an oxidative stress reaction to irradiation and postulated that potential side effects of cancer treatments such as radiation-induced cardiac disease could be linked to the trace element behaviour, suggesting that further research was warranted.

In an unusual application, Dalipi et al.⁵⁷ applied TXRF spectrometry to differentiate fresh meat from mechanically separated meat. After a simple sample preparation of mixing ground chicken meat with Triton X-100 and PVA in water, the levels of Ca, Cu, Fe, K and Zn were determined, comparing fresh meat with mixtures prepared with various proportions of mechanically separated meat up to 100% as well as other meat products. With PCA, it was found that Ca, in addition to Fe and K, were key to characterising the presence of the mechanically separated products. The elemental concentrations were confirmed by ICP-MS after acid digestion, indicating that the method offered a quick and simple approach to support the assessment of food quality.

During this review period, a number of papers have highlighted the use of portable XRF instruments applied to clinical samples. In fact, this review would not be complete without a publication concerning the in vivo determination of Pb in bone. Specht and co-workers⁵⁸ investigated the effect of analysing different bone types (trabecular and cortical bone) using a portable system, as Pb has different residence times in each type. However, due to the low energy of the system, only the outer surface of the bone is measured which is primarily cortical even when targeting trabecular bone. The researchers also considered tissue thickness as an additional parameter. The work was conducted on cadavers and the portable XRF spectrometer was calibrated using bone phantoms doped with Pb and tissue phantoms. A robust statistical approach was used in the calculations, taking into consideration the uncertainty of the measurements. The results showed no significant differences between bone types (tibia, patella, skull, ankle and index finger) or across different positions of the bone (tibia and skull). This indicated that comparison to data obtained from other studies at different bone sites could be compared. Two publications^59,60 have focussed on the use of portable systems for the analysis of skin. In the first, Dao et al.⁵⁹ investigated the feasibility of Fe measurements in skin as a proxy for levels in internal organs. Using rat models dosed with iron dextran, the skin, liver and kidney were analysed utilising a conventional XRF spectrometer which showed a strong correlation between the skin levels and liver (R² = 0.92), with a moderate relationship between the skin and kidney (R² = 0.65). The researchers then measured the skin samples with the handheld device and compared the results with the liver and kidney data from the conventional system, finding similar correlations; R² = 0.91 between liver and skin and R² = 0.83 between kidney and skin. Thus Fe levels in skin assessed by portable XRF spectrometers could be extrapolated to liver as a fast, non-invasive test. However, it must be considered that the experiments used Fe dopants which may not reflect endogenous levels. The second article by Afzal et al.⁶⁰ described the feasibility of portable XRF spectrometry for the determination of Cr in skin tissues. The system was calibrated using resin bone phantoms doped with Cr up to 250 ppm for K_α and K_β X-ray peaks, achieving an LOD of 5 μg g⁻¹ Cr with a 300 s acquisition time. The approach was applied to five different body parts on four cadavers, with detectable levels found in 90% of the sites analysed. The study reported levels between 3 and 20 μg g⁻¹ Cr, some of which were below the LOD. The authors also noted that the phantoms may not be appropriate for calibration for Cr due to the low energy of the X-rays and relatively low penetration (mean free path of photons was estimated as 0.4 mm) as human skin thickness can vary between 1.35 to 1.8 mm depending on location and gender. It was also stated that the effective dose delivered to the skin was significant (201 mSv) due to the small beam size and proximity to the skin which would be considered too high for living subjects. In conclusion, there was a lack of validation data and a number of concerns raised by the authors themselves that questions the value of this research alongside the clinical relevance of Cr skin levels.

5 Applications: clinical and biological materials

5.1 Metallomics

A comprehensive review of recent advances in inorganic speciation⁶ complements the applications with clinical and biological materials, foods and beverages covered in this Update. In biological systems metals and other elements are often bound to, or associated with, proteins. In many situations this is a reflection of their function, transport or storage. To understand the biology of these interactions it is necessary to characterise the nature of the binding. To this end Acosta et al.⁶¹ used SEC to separate protein fractions in the whey compartment of human breast milk, collected 12 months post partum, and used ICP-MS to identify metals in those fractions. In addition, Native-PAGE was used to give further information relating to the size of the proteins. The results showed that eight protein fractions, with molecular weights of between 68.878–1228.277 kDa, could be identified. The first fraction, possibly serum albumin, had detectable levels of Mn while the second fraction also contained Co, Cu and Se.

The number of therapeutic peptides is increasing every year. While attractive as potential drugs and drug delivery systems they have short circulating plasma half-lives and limited bioavailability owing to fast enzymatic and chemical degradation. With this growing interest, sensitive and selective quantitation methods are becoming increasingly important. Analysis of peptides in plasma is complicated by the similarities of the analytes with endogenous compounds and the complex sample matrix. Usually LC-MS/MS is the method of choice for identification of peptides but needs costly species-specific and isotopically labelled peptide standards for quantitation of each analyte. ICP-MS is a sensitive and robust technique and quantitation by an inorganic elemental standard is possible. Sulfur is naturally present in the amino acids, methionine and cysteine and Se in their Se-containing analogues. Gronbaek-Thorsen et al.⁶² reported on the extensive development for a UPLC-ICP-MS/MS procedure, with post column IDA for quantification. Using the peptide penetratin and its Se-containing analogue, the chromatography and the ICP-MS operating conditions were optimised to give results for linearity, precision, LOD, LOQ, recovery and accuracy. The LODs were 1.5 μg Se per L (0.019 μM) and 16 μg S per L (0.49 μM). The method was applied for a stability study, showing rapid degradation of the peptides in plasma.

Elemental speciation is often accomplished by physical or temporal separation from a support or other medium. Useful developments in techniques for speciation follow on from new, or modifications to, existing materials. Graphene and graphene oxide were used by Cheng et al.⁶³ as the stationary phase in HPLC-ICP-MS speciation assays for a range of elements. The authors showed that iHg and organoHg coordinate with 2-thiosalicylic acid (TSA) to form Hg-TSA complexes and were retained by graphene oxide owing to its strong pi electron stacking capability for TSA. Organoarsenicals were also separated, using long alkyl quaternary ammonium compounds, which have a similar strong pi interaction. Successful separation of inorganic As and Se species was due to an electrostatic attraction by aromatic quaternary ammonium cations in the mobile phases. Four Hg species were separated within 12 minutes As species within 32 minutes and Se species within 20 minutes. Efficient resolution of iodate, iodide, bromate, bromide, chromic acid and chromate were also demonstrated. Properties of adamantyl-type and octadecylsilyl-type columns were exploited by Narukawa et al.⁶⁴ to separate iHg and MeHg in a number of biological CRMs with detection by HPLC-ICP-MS. The Hg species were simply extracted with 10% w/w TMAH solution at 80 °C for 2 h and injected onto the column. Complete separation was achieved within 6 minutes (adamantyl column) or 4 minutes (octadecylsilyl column) and the LODs, as Hg, were 0.08 and 0.13 ng g⁻¹, for the respective columns. Results were in good agreement with the certified values. The analytical precision was less than 2% RSD and recoveries of MeHg and iHg were 101 ± 1 and 103 ± 3%, respectively. In a similar study,³⁰ previously described in Section 3.2, used a column filled with aminated Amberlite XAD-4 resin to speciate iHg and MeHg but with detection and measurement by CV-AAS. The reported LODs were 0.148 and 0.157 μg L⁻¹ respectively. In a quite separate approach, an anion-exchange polymer membrane was used by Nam et al.⁴⁹ to separate the iAs species As^III and As^V in water. When the sample was applied to the membrane, As^V was retained while As^III passed through. Following drying of the membrane at 40 °C for 2 h, the trapped As^V was determined using the LIBS technique. Oxidation of As^III in the filtrate to As^V, followed by application to the membrane, enabled this fraction to be measured. The selectivity of the procedure was verified by analysing each of the fractions by HPLC-ICP-MS, which confirmed complete retention of As^V by the membrane with no As^III contamination.

Two publications referring to the determination of As in fish are of interest. In the context of anticipated EU legislation on the maximum allowable limits for iAs in marine samples, an established method suitable for high-throughput screening of iAs content in seafood samples was evaluated by a group of four expert laboratories.⁶⁵ The procedure uses microwave-assisted extraction with diluted HNO₃ and H₂O₂ and selective HG with ICP-MS. The work included the introduction and validation of modifications to the original method in order to improve selectivity for iAs. Essentially, the high concentration of HCl (8 M) for HG together with H₂O₂ at the same concentration used for extraction, leads to a selective conversion of iAs to volatile AsH₃ that is released and transported to the detector. The analytical parameters were optimised to minimise any contribution from DMA^V. The method was applied to analysis of eight seafood CRMs, namely TORT-3, DORM-3, DORM-4, DOLT-4, DOLT-5, PRON-1, SQID-1 and ERM-CE278k, and to various seaweed samples. Good agreement was obtained with the certified concentrations, whilst the LOD of 2 μg kg⁻¹ for iAs is sufficient for food testing and precision was 2–8%. Interlaboratory results were in good agreement (5–7%) for higher iAs concentration samples but at levels close to the LOD, this increased to 61–77%. The authors noted that the procedure can be used with alternative detection systems such as AAS, AFS and ICP-OES. The work reported by Zhao et al.³² compared extraction procedures using aqueous acid solvents, aqueous organic solvents and enzymes. It was concluded that 40 mg protease with 0.75 mL 0.5% HCl as the extraction agent provided the optimal results. Extracted samples were analysed by AEC and ICP-MS. The method was reported to provide simultaneous determination of eleven arsenic species with LODs in the range of 0.11–0.59 μg kg⁻¹ and repeatability values for spiked fish samples of 1.1–7.6%. The recoveries from spiked fish samples were 91–106%.

Continuing with analysis of fish samples, two further reports described interesting developments for the speciation of Hg. Reference has already been made to the novel work of Linhart et al.⁴⁵ who applied photochemical HG and HR-CS-AAS for the detection and measurement of iHg, MeHg, EtHg and PhHg (Sections 4.2 and 4.4). Nine sample extraction procedures were investigated but only the use of TMAH and HCl at 75 °C was compatible with the proposed separation and detection steps. This approach provided high extraction efficiency and no changes in Hg speciation. The method was validated by analysis of the CRMs DOLT-4 (dogfish liver) and ERM-CE464 (tuna fish) and the results were comparable to those given by a reference method with L-cysteine extraction and HPLC-ICP-MS determination. The LODs; iHg (0.47 μg L⁻¹), MeHg (0.84 μg L⁻¹), EtHg (0.80 μg L⁻¹) and PhHg (2.0 μg L⁻¹), and precision were comparable with those given by AFS.

Natural variations in the isotopic composition of Hg can be a very useful tracer for studying its biogeochemistry. Species-specific Hg isotopic data may be used to trace pathways in the environment and to assess potential impact on ecosystems. This data in fish can help in revealing sources by relating isotopic signatures to those formed during biotransformation processes and can also be useful for understanding mechanisms of in vivo MeHg metabolism. Isotopic data is best obtained with MC instrumentation and, of the techniques available for speciation of Hg, only GC has previously been hyphenated to MC-ICP-MS. However, the necessary derivatisation is prone to causing species transformation. For this reason, Entwisle et al.³⁵ developed a HPLC-CV-MC-ICP-MS procedure to obtain precise and accurate Hg isotope ratio measurements of MeHg in fish tissues, at environmentally relevant concentrations. The chromatography was in isocratic mode using RP-HPLC with a mobile phase containing L-cysteine as a complexing agent. The separate species were oxidised to enable the use of CV with SnCl₂ as the reducing agent for maximum reactivity. Measurements of ¹⁹⁹Hg : ¹⁹⁸Hg; ²⁰⁰Hg : ¹⁹⁸Hg; ²⁰¹Hg : ¹⁹⁸Hg and ²⁰²Hg : ¹⁹⁸Hg; of the fish CRMs BCR 463 and NIST SRM 1947 had combined standard uncertainties of 0.10 to 0.22 parts per thousand. The measured results for these isotope ratios agreed with the CRM indicative values. This validated method for accurate species-specific Hg isotope ratio measurements in fish tissues will allow determination of natural variations in the isotopic composition of MeHg and iHg.

As has been done for iAs in marine samples, there are moves by agencies such as the EU and the US EPA to establish maximum allowable limits for the carcinogen Cr^VI. In response to these concerns, reliable, accurate methods are necessary for samples types such as drinking water, waste water, industrial waters and recipient waters. In what is described as a fully automated total metals and chromium speciation single platform introduction system for ICP-MS, Quarles et al.³⁴ investigated the use of a commercial device, the prepFAST IC. This is a sample delivery system which, in this work, was used in conjunction with ICP-MS. In one mode, a sample is diluted and pumped onto an anion exchange column which enabled pre-concentration and then separation of Cr^III and Cr^VI, before leading to the ICP-MS. In the second mode, the column is bypassed and diluted samples are transferred directly to the ICP-MS for ‘total metals’ determination. The device can prepare dilutions from a single stock solution of Cr^III and Cr^VI to calibrate the assay. The LODs for Cr^III and Cr^VI were determined as 0.012 μg L⁻¹ and 0.007 μg L⁻¹ respectively, approximately 40-fold lower than was found with a conventional HPLC-ICP-MS arrangement. Results were obtained for the concentration of Cr^VI in 50 water samples using the prepFAST IC and HPLC-ICP-MS. Concentrations were from 0.70 to 32.4 μg L⁻¹, with an average bias of 1.9% between the two data sets.

For reports involving beverages other than drinking water refer to Cu and Fe in wine⁶⁶ and Sb in bottled mineral water.⁶⁷ The details are discussed in Section 7.2.

An account is given in Section 5.4.8 of how ID-GC-ICP-MS was used by Queipo-Abad and colleagues⁶⁸ to investigate the body burden of iHg, MeHg and EtHg in a group of workers who were exposed to gaseous Hg three years previously.

5.2 Imaging: LA-ICP-MS and XRF spectrometry

Within this review period, a number of articles investigating the analysis of single cells with LA-ICP-MS were published, highlighting the cutting edge research with the latest technological developments in this field. Van Malderen et al.⁶⁹ successfully demonstrated a spatial resolution of 1 μm for 3D cell imaging using the latest LA system with an enhanced ablation cell and improved washout capability which was combined with ICP-MS and advanced algorithms for data handling. Human cervical carcinoma cells were used as models to develop the approach and two staining agents (uranyl acetate and an Ir-based DNA intercalator) were applied then embedded. A series of 132 individual 2 μm slices were produced and each one analysed by 2D LA-ICP-MS. In order to realise the cell locations in 3D, multiple registration approaches were applied, with the outline of the section itself acting as the waypoint, providing good agreement with μ-CT estimations. The final result was visualisation of the cells in 3D space with high resolution and derivation of the nucleus dimensions. Lohr et al.⁷⁰ performed a systematic comparison between single spot analysis and imaging mode for the detection of metal tags in single cells. Both measurement options have advantages and disadvantages, with single spot analysis being more time efficient but lacking in spatial resolution, whilst the imaging approach provides greater detail at the expense of instrument time. The workers used mouse 3T3 fibroblast cells stained with two dyes tagged with metals ions to target specific sites, namely mDOTA-Ho linking to free thiol groups in proteins which are ubiquitously present and an Ir-DNA-intercalator to bind to DNA in cell nuclei. Despite the different approaches, the results obtained were similar in terms of S/N and the average intensity per pixel for both Ho and Ir. Histograms of the integrated intensity versus the number of cell events were not significantly different. Whilst high RSDs were found, up to 53% for Ho and 68% for Ir in the single spot mode, the authors attributed this to the limited number of cells analysed (30–50) considering that several hundred cells would be required to overcome the effects of natural variation. However, as the process is currently manual, significant advances in pattern recognition and software automation are required to make this viable. There was also an attempt at quantification using spiked nitrocellulose membranes to calculate the absolute mass of Ho and Ir which was compared to the total concentrations obtained via acid digestion of approximately 60 000 cells and ICP-MS analysis. Recoveries of 54% of Ir and 358% Ho were found leading the authors to conclude further work was required to understand how the sample preparation process affects this recovery.

Another technological development that is receiving published attention is the use of time-of-flight MS for imaging applications. Bauer et al.⁷¹ utilised ICP-TOF-MS with the latest LA system equipped with enhanced ablation cell and transfer line for rapid quantitative bioimaging. The combination of fast sample introduction and reduced washout times with the quasi-simultaneous multi-elemental analysis provided a powerful tool for quantitative determination. In this work, kidney samples from rats treated with cisplatin were used to evaluate two calibration approaches (external calibration and IDA) as well as isotope ratio precision. The system was able to analyse a 70 mm² area in just 5 h, up to 5-fold quicker than conventional systems. The instrumental setup included a desolvation system for the introduction of aqueous standards enabling online IDA as well as external calibration, with both approaches leading to average concentrations which agreed within 2%. The IDA method provided some further benefits as a definitive method with SI traceability and good precision from the measurement of a single calibrant, saving time over external calibration. The Pt bioimages generated enabled detailed examination of kidney tissue sections from cisplatin treated rats. The authors also reported some further benefits of the quasi-simultaneous TOF system – the mass range from 0 to 255 m/z was captured in 30.3 μs permitting the identification other elemental trends from the same dataset. Patterns of interest were found for Cu, Ni and W. An additional finding was the observation of ‘black spots’ in the images which were eventually correlated to droplets reaching the plasma causing signal fluctuations – the reduction in Pt intensity associated with a decrease in ⁴⁰Ar₂ and increase in ¹⁸O. In conclusion, the system was able to produce high resolution, ultra-fast, multi-elemental and quantitative bioimages. Blockhuys and co-workers⁷² utilised TOF-SIMS for the analysis of Cu in breast cancer cells for bioimaging purposes. They applied the delayed extraction methodology to achieve sub-cellular spatial resolution whilst maintaining sensitivity and mass resolution for the determination of both Cu isotopes. A Bi₃ ion gun was used as the primary ion source for simultaneous detection of Cu isotopes and organic compounds, achieving a spatial resolution of approximately 400 nm at a mass resolution of 3500, whilst C₆₀ sputtering was applied for depth profiling to 600 nm enabling the researchers to build 3D images of cancer cells. The results identified Cu accumulation in the nuclei of the cells which was contrary to their hypothesis of localisation within cell membranes. The work demonstrated that TOF-SIMS has the capability to provide valuable insights into the role of essential elements and organic compounds with cancer progression at the sub-cellular level.

McAllum and Hare⁷³ published a review highlighting the importance of bioimaging techniques for neurological research and its development over time as a critical tool for understanding the role of trace elements in neuroscience. It covered relevant analytical techniques including LA-ICP-MS, μXRF and PIXE with key examples of applications. There was also consideration given to the current limitations and remaining challenges within these techniques and a view to future developments. Ali et al.⁷⁴ described the complementary use of FTIR and LA-ICP-MS for imaging analysis in rat brain tissue slices. A significant advantage of FTIR is its non-destructive nature allowing further analysis on the same sample to build a better picture. Quantitative data for Cu, Fe and Zn by LA-ICP-MS were given whilst relative intensity maps were provided for C, Ca, Mg, P and S. Although the data presented was relatively limited (only 6 male rats, partial quantitative data), it does highlight the benefit of combining multiple analytical techniques. Clases et al.⁷⁵ described the use of ICP-MS/MS combined with LA to investigate the deposition of Gd in skin and brain tissues from patients exposed to Gd-based MRI contrast agents. The use of oxygen reaction gas was applied and detection of Gd and P occurred at a mass shift of +16, whilst Ca, Fe and Zn were on mass. For Gd and P, significantly improved limits of detection were obtained through this measurement approach which enabled the production of bioimages with a spatial resolution of 5 μm × 5 μm. Deposits in fibrotic skin and brain tumour tissue, approximately 50 μm in diameter, showed Gd accumulation which was correlated with Ca and P, indicating co-precipitation. The work demonstrated the benefits of the ICP-MS/MS technology with improved limits of detection and multi-element capability, applied to the investigation of Gd deposition mechanisms in the body.

The use of X-rays is a regular feature in this review section. In this period, two works have highlighted the role of XRF spectrometry for bioimaging applications. Streli et al.⁷⁶ provided an overview of μXRF and nanoXRF spectrometry for the analysis of bone samples. They summarised their research findings from three distinct projects: μXRF system specially designed for light elements to investigate the degradation and distribution of Mg and Y in bone from Mg-based implants, spatial distribution of Zn in immature bone cancer (osteosarcoma) using induced confocal SR-μXRF spectrometry and quantitative backscattered electron imaging to distinguish between healthy and diseased bone, and thirdly, SR-nanoXRF spectrometry with a resolution of approximately 500 nm. The report included discussion on the advantages of μXRF and nanoXRF spectrometry as well as the remaining challenges still to overcome. The use of SR-μXRF spectrometry was also applied by Osterode et al.⁷⁷ for the investigation of trace element distribution in liver tissues from patients with Wilson’s disease. Samples from twelve patients and seven controls were analysed and the results compared to traditional histological staining methods (separate tissue sections but within 20–30 μm) in addition to AAS analysis for total Cu. Quantitation for Cu, Fe, P, S and Zn was achieved via external standardisation using a Ge foil and normalised net peak intensities using the fundamental parameter method. The resultant bioimages showed significant differences between the patient group and the controls, particularly for the hepatocytes and fibrotic areas where the average Cu level was enhanced by 20-fold and 3-fold respectively for the patient samples. Additionally, Cu was inhomogeneously distributed in the regeneration nodules. Correlations were observed between Zn and Cu or Fe and S, and differences in the Cu/S were found between hepatocytes and fibrotic areas. Although the quantitation by SR-μXRF spectrometry was described, the correlation of the Cu levels against AAS gave a coefficient of 0.79 but the large RSDs from the imaging method should be taken into consideration against the single bulk total Cu measurement. Overall, the work demonstrated that archived samples could be used to obtain valuable trace element data and more detail than histological staining alone, providing further insight to the disease progression.

Gianoncelli et al.⁷⁸ combined SR-nanoXRF spectrometry and soft X-ray microscopy to characterise the composition of airborne particles/dust in lung tissues. Samples were selected from individuals exposed to environmental particles from urban areas. The imaging results enabled the researchers to differentiate single particles from aggregates ranging from the nano to the micro scale. The elemental composition showed Ti associated with K, O and Si, links to Al and Si, whilst Fe and Zn were correlated with O. Furthermore, the particulate morphology could be observed from the simultaneous absorption and phase contrast images which enabled the detection of a high density of nanoparticles or high concentration of heavy elements. The work provided a useful insight into the composition inhaled particles in lung samples from urban environments using two complementary techniques.

5.3 Multielement applications

5.3.1 Specimens analysed to investigate metallic implants and biomaterials. The number of publications referring to joint implants has decreased considerably as the use of metal on metal joints is no longer popular. A design of hip prosthesis favoured for patients with a high risk of dislocation, dual mobility cups, which have a cup formed of 19% Cr and 13–15% Ni, a ceramic head, and a titanium alloy stem, has not previously been studied for possible release of metal ions. Marie-Hardy et al.⁷⁹ investigated 16 patients two years after hip arthroplasty. Clinically, there were no complications or any signs of loosening. Preoperative concentrations of Cr and Ni were less than 1 μg L⁻¹ (1.5 in one subject) and 0.2 μg L⁻¹, respectively. Two years later, the concentrations of Cr were 0.4–0.7 μg L⁻¹ and were 0.2–5.6 μg L⁻¹ for Ni (four were >2 μg L⁻¹). These levels are comparable to those observed following other orthopaedic implants such as those used in the spine. Kuba et al.⁸⁰ used ICP-MS to determine concentrations of Co, Cr, Mo, Nb, Ti and V in blood, joint fluid and periprosthetic tissues from 117 patients. The cohort comprised patients who had received total hip or knee arthroplasty with metal on polyethylene or ceramic on polyethylene prostheses. The results led to six conclusions;

• metals released from implants are mainly deposited in periprosthetic tissues,

• concentrations of metals in blood are poor markers of effects occurring in local tissues,

• accumulation of metals in tissues increases with the time since implantation,

• concentrations of metals decrease dependent on the distance of the tissue from the implant,

• concentrations of metals in joint fluid and tissues are greater in patients with loosened implants compared to those with stable joints, and

• concentrations of metals are also greater in patient groups with aseptic loosening compared to those with stable implants.

All of which might have been predicted from previous work with patients having metal on metal implants.

The biological safety of dental biomaterials has long been of concern, with investigations relating to mercury amalgam carried out more than 40 years ago. Of more recent interest are the possible effects of titanium dental implants. Peri-implantitis is a variety of periodontitis or gum infection associated with dental implants. The most common implants are prepared from titanium and it has been suggested that the peri-implantitis infection may be exacerbated by the titanium. Pettersson et al.⁸¹ therefore used ICP-MS to determine the concentrations of Ti in biopsies from 13 patients with peri-implantitis and 11 with periodontitis. Two biopsies from each group were analysed by light microscopy, TEM and SEM to investigate the presence of particulate Ti. The Ti concentrations in the peri-implantitis tissue (mean ± SD) were determined to be 98.7 ± 85.6 μg g⁻¹ while the control tissues contained 1.2 ± 0.9 μg g⁻¹. Particulate metal was identified in peri-implantitis and control biopsies, but element analysis showed Ti in only the peri-implantitis tissue. The Ti in peri-implant mucosa may aggravate inflammation and possibly reduce the prognosis for treatment. In a second study, Cabana-Munoz et al.⁸² measured concentrations of Co, Cu, Fe, Hg, Mo, Sn and Zn in hair samples by ICP-MS and also reported the plasma malonaldehyde status (a marker of free radical activity) from 130 subjects. No information is provided for the sample preparation of the hair specimens or of whether the authors consider hair analysis to be a reliable indicator of status. The samples were from those with both implants and amalgam (54), those with just amalgam (51) and controls (25). All except the controls had the implants and/or amalgam in place for at least 10 years. Unfortunately, although there is a large amount of data, the conclusions are somewhat nebulous and a more focused approach may be more informative.

5.3.2 Biological fluids and tissues. It has long been recognised that some elements are harmful to the process of conception and a successful pregnancy. Essential trace element status is also considered to be important and assessments of trace element status feature in some pre-conception programmes. Two reports from the last year are relevant to human reproduction. Analysis of the fluid that surrounds the developing oocyte within the ovary may be useful in situations of women seeking in vitro fertilisation to establish if there are factors that may have adverse effects on oocyte viability and subsequent pregnancy outcomes. Therefore, Galusha et al.³⁶ evaluated the suitability of a method using ICP-MS/MS to measure concentrations of As, Cd, Co, Cu, Hg, Mn, Mo, Pb, Se, Sr, and Zn in human follicular fluid. The detection limits were from 5.6 ng L⁻¹ for Cd to 0.11 μg L⁻¹ for Mo. For 197 specimens of follicular fluid, Pb was present in 84% of samples, while Co, Cu, Mn, Mo, Sr, and Zn were detected in all samples. As a consequence of a consistently high incidence of low birth weight babies in Chattanooga, Tennessee, a study was established to investigate relationships between metals in placental tissue and birth weight, gestational age, birth weight centile, placental weight, birth length and head circumference.⁸³ This involved measurement of 37 elements in 374 placental samples using ICP-MS. Of the elements determined, only 14 were at quantifiable levels in at least 86% of the samples. Significant associations with birth weight were shown for Co, Fe, Mn, P and Rh. Placental Rh concentrations were negatively correlated with almost all infant parameters. The authors do not speculate on how or why the results should be relevant to Chattanooga.

5.4. Progress for individual elements

5.4.1 Aluminium. This section of the review begins with a familiar topic: Al and neurological disease. The accumulation of Al in the brain tissues of Alzheimer’s disease sufferers was first noted many decades ago and has remained an active area of study. McLachlan et al.⁸⁴ have added to the evidence base by measuring Al in brain tissue from sufferers of Alzheimer’s disease and 13 other neurological conditions. Samples from the temporal lobe neocortex were taken postmortem and Al was measured in the tissues using ETAAS. Concentrations of Al were raised in samples from Alzheimer’s disease, dialysis dementia syndrome and Down’s syndrome sufferers, but were normal in sufferers of other diseases and in healthy control subjects.

5.4.2 Antimony. Here we review the first of a series of papers studying the association of blood and urine trace element concentrations with outcomes in pregnant Chinese women and their babies. Zhang et al.⁸⁵ investigate the association of Sb exposure during pregnancy with gestational diabetes mellitus (GDM). Urine samples were collected from 2093 women enrolled in the Tongji Maternal and Child Health Cohort study in Wuhan China. Arsenic, Cd and Sb were measured in urine samples with ICP-MS (LOD 0.0090 μg L⁻¹, 0.0039 μg L⁻¹ and 0.0015 μg L⁻¹ respectively). All participants were screened for GDM with an oral glucose tolerance test between 24 and 28 weeks of pregnancy. Increased urine Sb concentrations were significantly associated with increased risk of GDM; women in the highest tertile of Sb concentration had a 1.92-fold (95% CI 1.42 to 2.60, p-value for trend < 0.001) higher risk of GDM than the lowest tertile. This effect was independent of pre-pregnancy BMI, age, gravidity, and foetal sex. Interestingly, high urine Sb concentrations were associated with type two diabetes mellitus in the NHANES study, previously carried out in the USA. Taken together, these studies suggest a link between low-level Sb exposure and impaired glucose homeostasis that warrants further investigation.

5.4.3 Arsenic. The second instalment of the series of Chinese papers was an investigation into the effect of maternal As exposure on pregnancy outcomes. Wang et al.⁸⁶ published a study of 3194 Chinese women-child pairs from the Ma’anshan-Anhui birth cohort, the first of two papers in this review using the same patient population (see Zhu et al.⁸⁷ in the Tl section). Maternal serum was collected between the 4th and 27th week of pregnancy and serum As was measured with HG-AFS (LOD 0.02 μg dL⁻¹). The participants were divided into two groups: low serum As (75th centile and below ≤6.68 μg L⁻¹) and high serum As (above the 75th centile >6.68 μg L⁻¹). Female babies born to high serum As mothers had increased odds of being small-for-gestational-age (adjusted odds ratio 1.46 (1.03 to 2.07), p-value 0.034) but male babies were not affected (adjusted odds ratio 1.18 (0.74 to 1.88), p-value 0.49). There was also increased incidence of moderate-to-late preterm delivery in the high serum As group (adjusted odds ratio 1.47 (1.03 to 2.09), p-value 0.034).

Microbiology does not feature often in this section of the review but here Hoen et al.⁸⁸ investigated the effect of As exposure on the gut microbiome of babies. Stool and urine samples were collected from 204 6 week old babies, and urine total As was measured with ICP-MS (LOD 0.05 μg L⁻¹). Mean urinary As was 0.5 μg L⁻¹ in exclusively breast-fed babies, 0.9 μg L⁻¹ in combination fed babies and 1.2 μg L⁻¹ in exclusively bottle-fed babies. Bacteria in the stool samples were identified by ribosomal DNA sequencing. Differences in the gut microbiome were significantly associated with urine As in formula fed males but not in formula fed females or exclusively breastfed babies of either sex.

5.4.4 Barium. Stanley et al.⁸⁹ presented an unusual case of Ba toxicity in a dog that ate a sparkler. Although many Ba salts are insoluble and are not absorbed when ingested, soluble Ba salts such as Ba(NO₃)₂ and its combustion products BaO, Ba(OH)₂, and BaCl₂ are readily absorbed. Sparklers and other fireworks often contain Ba(NO₃)₂. The dog’s symptoms (muscle paralysis, heart arrhythmia, hypokalemia and respiratory acidosis) were similar to those previously noted in human Ba poisonings. The dog’s serum Ba concentration (measured by ICP-MS) was 2000 μg L⁻¹, more than three orders of magnitude above the concentration expected in unexposed animals. Thankfully, prompt treatment of the dog by its veterinarian owner allowed it to make a full recovery.

5.4.5 Boron. It has been two years since B last featured in this section of the review and we welcome back this somewhat under-investigated and under-appreciated element. Hjelm et al.⁹⁰ studied the effect of boron exposure on infant growth, continuing from similar work on foetal growth that appeared in ASU two years ago.⁴⁶ The study participants were 177 mother–child pairs living in a region of the Argentinean Andes with high and variable drinking water B concentrations (377–16 076 μg L⁻¹). Maternal blood and urine were collected during and after pregnancy, along with breast milk, infant urine and infant blood; B was measured in the samples with ICP-MS. There was a strong correlation between maternal serum and breast milk (r_s = 0.94). In multivariate models adjusted for confounding, higher concentrations of B in infants’ urine were significantly associated with lower body weight and lower head circumference in babies at both 0–3 months and 3–6 months of age. These findings suggested that high B intake in early infancy affects growth, and complements these researchers’ previous finding that B exposure affects foetal development. The authors noted that these findings should be confirmed in different populations and at older ages and we look forward to reviewing these papers in future ASU editions.

5.4.6 Gadolinium. Gadolinium has continued to be an element of interest in this review period with ongoing concerns around the safety of Gd-based contrast agents (GBCAs) in medical imaging. The review period saw two similar studies assessing the accumulation of macrocyclic Gd-based contrast agents in the CSF. Both Nehra et al.⁹¹ and Berger et al.⁹² used ICP-MS to measure Gd in the CSF collected from patients after contrast MRI with gadobuterol or gadoterate respectively. In both studies Gd accumulated in the CSF of patients rapidly after contrast agent infusion, irrespective of the patient’s blood–brain barrier integrity. Gd concentrations dropped quickly within the first 48 to 100 hours post infusion, with low but measurable concentrations seen at later time points (Nehra et al.⁹¹ at 24 days, and Berger et al.⁹² at 50 days).

Robert et al.⁹³ published an impressive study comparing the brain retention of Gd after linear and macrocyclic contrast agents were administered to rats. The animals were given 5 doses of either gadodiamide (linear) or gadoterate (macrocyclic) over 5 days and then followed up for 12 months with MRI and Gd measurement in the blood and brain tissues. Total Gd concentrations were measured by ICP-MS, macrocyclic Gd species with SEC-ICP-MS and the parent contrast agent with HILIC-ICP-MS. No Gd from either contrast agent was detected in blood after 5 months and brain concentrations after gadoterate treatment returned to background levels after 5 months. However, after gadodiamide treatment 75% of Gd detected in cerebellum at the first time point remained after one year with almost all of that Gd bound to macromolecules.

Clases et al.⁷⁵ described the imaging of Gd in tissues with LA-ICP-MS/MS. The samples analysed were from patients who had previously received GBCAs: skin from a patient with nephrogenic systemic fibrosis and brain from a patient with glioblastoma. In both tissues Gd co-localised with Ca, P and Zn in agreement with previous imaging studies. This supports the hypothesis that Gd is liberated from GBCAs and deposits as insoluble mixed phosphates in tissues.

5.4.7 Iodine. Yu et al.⁹⁴ described the development and validation of a simple ICP-MS method for measuring I in serum and urine with an LOQ of 0.87 μg L⁻¹. Iodine was measured in serum and urine from 72 healthy pregnant women from Beijing. Median urine I was 125.5 μg L⁻¹, with 55.6% of samples below the WHO sufficiency guideline of 150 g L⁻¹ suggesting this population is mildly I deficient. Median serum I was 69.0 g L⁻¹ and showed relatively poor correlation with urine I (r 0.39).

5.4.8 Mercury. Hitosugi et al.⁹⁵ reported an unusual case of attempted murder with Hg. The victim’s cigarettes were doctored with elemental Hg and then smoked using an electrical heated tobacco product. Within a few hours the victim noticed flu-like symptoms and contacted police after noticing a metallic object in an unused cigarette. On the day of presentation, the victim’s serum Hg was 99 μg L⁻¹ (measured by AAS) and the presence of Hg in the remaining cigarettes was confirmed by EDXRF.

Queipo-Abad et al.⁶⁸measured Hg species in blood, hair and urine from workers exposed to Hg 3 years prior. Samples were obtained from 17 workers at a zinc manufacturer who were exposed to Hg vapour in an industrial accident plus control samples from their closest relatives. Inorganic Hg, EtHg and MeHg were measured in the samples by ID-GC-ICP-MS. There were no significant differences between the two groups except for MeHg, which was higher in the blood of exposed individuals. Interestingly, the relationship of hair MeHg to blood MeHg differed between the exposed and unexposed individuals although the cause of this difference is unclear.

Kobayashi et al.⁹⁶ studied the effect of maternal Hg and Se on foetal growth. Total Hg and Se concentrations were measured in the blood of 15 444 pregnant women enrolled in The Japan Environment and Children’s Study. The blood samples were taken in the second and third trimester and were analysed by ICP-MS (LOQ for Hg and Se: 0.13 ng g⁻¹ and 2.2 ng g⁻¹ respectively). The blood Hg and Se concentrations were not significantly associated with birth weight, the odds of small-for-gestational-age, birth length, birth chest circumference or birth Ponderal index. There was a significant association between high blood Hg and low birth head circumference, although the difference was small.

5.4.9 Osmium and ruthenium. Papers describing Ru and Os-containing anticancer drugs have featured regularly in recent ASU editions and this work is continued in this review period. Klose et al.⁹⁷ studied the in vivo distribution of the Ru-containing drug Plecstatin-1 and its Os analogue. Ru and Os were measured in blood and serum from drug-treated mice using ICP-MS, finding the drugs in the serum rather than blood cell fractions. Ru and Os co-eluted with transferrin and albumin when serum was analysed with SEC-ICP-MS/MS (in oxygen mode). These results were confirmed with in vitro incubation of the drugs with human serum, suggesting that the drugs bind to albumin and transferrin in vivo.

5.4.10 Strontium. Tipple et al.⁹⁸ presented an interesting study investigating whether the ratio of Sr isotopes found in human hair samples are linked to provenance. The ⁸⁷Sr : ⁸⁶Sr ratio found in biological materials is governed by the geology of local bedrock, and has been used to track the origins of archeological, food and forensic specimens. In this study 98 human hair samples and corresponding drinking water samples were collected from 48 municipalities in 16 states of the USA. The ⁸⁷Sr : ⁸⁶Sr ratio was measured in washed hair samples using MC-ICP-MS and was found to correlate strongly with the drinking water ⁸⁷Sr : ⁸⁶Sr ratio (R = 0.93, p < 0.0001) when specific exclusion criteria were applied. These results suggest that measuring the ⁸⁷Sr : ⁸⁶Sr ratio of samples from an individual may help indicate their recent whereabouts.

5.4.11 Thallium. Ash et al.⁹⁹ described how LA-ICP-MS analysis of hair was used to investigate a historic case of Tl poisoning. The victim was a female college student in Beijing who developed symptoms over approximately 1 year between 1994 and 1995. Symptoms included gastric pain, vomiting, numbness, paresthesia, delirium, seizures and coma; with two distinct bouts of hair-loss and hospitalisation. The victim survived after treatment with Prussian-blue but was left with life-changing sequelae. Thallium was measured in the victim’s hair with LA-ICP-MS, showing both chronic exposure and acute events that matched the chronology of symptoms. The highest Tl concentration found was ∼530 ng g⁻¹, approximately 200-fold higher than the basal concentration. There was also evidence of Pb exposure in addition to Tl. Although the case remains unsolved the study demonstrates how hair analysis with LA can add useful temporal information in poisoning cases.

We bring this section of the review to a close with the final instalment of the Chinese pregnancy series. Zhu et al.⁸⁷ studied the association of Tl with gestational diabetes mellitus (GDM). The study included 3013 pregnant Chinese women from the Ma’anshan-Anhui birth cohort. Thallium was measured in serum samples taken in the first trimester by ICP-MS. Thallium concentrations ranged from 0.011 μg L⁻¹ to 0.232 μg L⁻¹ with a median value of 0.062 μg L⁻¹ and was significantly associated with both age and pre-pregnancy BMI. Higher Tl concentrations were associated with GDM (P-value 0.007) but this effect was not significant when models were adjusted for pre-pregnancy BMI. When patients were stratified by age, serum Tl was associated with GDM even after adjustment for pre-pregnancy BMI in the >30 year subset (highest quintile serum Tl had odds-ratio of 1.05–5.05 of GDM).

6 Applications: drugs and pharmaceuticals, traditional medicines and supplements

Rebiere et al.¹⁰⁰ addressed a situation that is not widely recognised although they suggest that it is increasing in prevalence. This refers to falsified medical products and the risk to the health of patients associated with such items. The workers suggested that multivariate analysis of EDXRF spectra, using a properly validated classification model, allowed for the authentication of suspect samples. By way of example, five suspect samples of Plavix (R) were investigated. The XRF spectra were studied using chemometrics (PCA and Soft Independent Modelling of Class Analogies). A classification model was validated with positive and negative samples and four of the samples were shown to be fake.

Two publications refer to guidelines for limits of elemental impurities in drugs and over-the-counter medicines. Mattiazzi et al.¹⁰¹ suggested that absolute quantification is not necessary when screening these materials and that what is needed is a procedure to establish whether or not a threshold concentration has been exceeded. It was proposed that HR-CS-ETAAS offers a suitable technique to evaluate elemental impurities in drug active ingredients or drug final products. An alternative approach was proposed by Pinheiro et al.¹⁰² who described a simple dilute-and-shoot procedure with analysis by ICP-AES. Four liquid pharmaceutical samples were diluted 10-fold in 0.14 mol L⁻¹ HNO₃. Calibration was achieved using Bi, Ge and Y internal standards or one-point standard addition. Experiments produced recoveries of 86 to 116% when the best internal standard was employed for each analyte, and 78 to 119% with the one-point standard addition. Precision was less than 9% RSD. Taken together with similar studies as reported in our recent Updates,¹ there are now many suitable procedures available to screen medicines for elemental impurities.

7 Applications: foods and beverages

7.1 Progress for individual elements

7.1.1 Arsenic. Much of the developmental work reported in recent years has been directed towards methods for speciation of As. In the last year, only the work of Zhao et al.³² falls within this category. The reported procedure used the enzyme protease to extract the As species from fish tissue, separation by AEC and measurement by ICP-MS. The method was able to separate 11 species of As within 17 minutes. Unsurprisingly, the main species found in real samples was AsB with only trace amounts of iAs.

Other publications of interest focus on toxic inorganic As species. In a collaborative project among a small group of reference laboratories a procedure for fast screening of iAs in seafood was validated.⁶⁵ The As was measured by CV-ICP-MS and the extraction and VG conditions were optimised so that almost all of the AsH₄ formed was derived from the iAs in the samples. The LOD of 2 μg kg⁻¹ for iAs is sufficient for food testing. A second report considered determination of iAs in seaweed and seafoods. Matsumoto-Tanibuchi et al.¹⁰³ extracted the As by heating samples at 100 °C in 0.3 mol L⁻¹ HNO₃. Species were separated by LC-ICP-MS using an octadecylsilane column (see Section 7.2.6). In Sections 2 and 7.2.1 of this Update, the contribution of Hijiki seaweed to the iAs in the diet of Japanese children and pregnant women is discussed.¹⁰⁴

7.1.2 Mercury. Four different approaches to the measurement of MeHg in foods were published in this review period, three refer to Hg in seafood and the other to HG in rice. All present results of validation work rather than reporting concentrations in real samples. Accurate species-specific Hg isotope ratio measurements (¹⁹⁹Hg : ¹⁹⁸Hg; ²⁰⁰Hg : ¹⁹⁸Hg; ²⁰¹Hg : ¹⁹⁸Hg and ²⁰²Hg : ¹⁹⁸Hg) involved development of an HPLC-CV-MC-ICP-MS procedure³⁵ An LC-CV-AAS method exploited the properties of aminated Amberlite-XAD-resin to capture iHg and MeHg, followed by elution with 0.1% thiourea in 3% HCl (iHg) and then 6 M HCl (MeHg).³⁰ Measurement of Hg by photochemical CV-HR-CS-AAS following extraction of Hg with TMAH and HCl at 75 °C and RP-HPLC.⁴⁵ A rapid, sensitive procedure to determine the MeHg content of rice involved extraction of Hg into an L-cysteine solution and measurement by thermal decomposition amalgamation-AAS on a direct mercury analyser.⁵³ Details are described in Sections 4.4 and 5.1.

7.2 Single and multielement applications in food and beverages

7.2.1 Dietary intake. Infant and toddler foods were investigated for Ag, Al, As, Ba, Cd, Cr, Ga, Ge, Ni, Sb, Sn, Sr, Te and V as part of the 1^st French total diet study (TDS).¹⁰⁵ Ready to eat meals, whether meat-, fish-, or vegetable-based were the most ‘contaminated’ for most elements studied except Ga, Sb and V. The highest Al concentrations were found in soups/purées and cereal-based foods (653 and 630 μg kg⁻¹, respectively), while fruit purées the highest concentration was Sn (424 μg kg⁻¹). Sadly, chocolate-based foods contributed considerably to higher concentrations of Al, Cd, Co, Cr and Ni in the category ‘sweet and savoury biscuits and bars, dairy-based desserts and croissant-like pastries’, although infant/follow-on formulae and growing-up milk had relatively low mean Al, As and Cd concentrations compared with the other infant food categories (220, 1.80, 0.51 μg kg⁻¹, respectively).

Iodine in USA TDS samples from 2013 were re-analysed by ICP-MS using alkaline extraction, as well as HClO₄ digestion followed by spectrophotometric detection for historic comparison.¹⁰⁶ Out of 266 samples separated into 11 groups, bread, eggs and dairy showed the highest I concentrations with fruit and vegetables having the lowest. A duplicate diet study in Japan over three days assessed Al, As/iAs and Pb in 319 subjects using ICP-MS.¹⁰⁷ The median intake was calculated as 41.1, 2.31, 0.260 and 0.0928 μg kg⁻¹ d⁻¹, respectively. Although Al, total As and Pb intake varied with gender, this was explained by the amounts consumed, while differences in total As varied with age, location and community type (urban, farming, fishing) due mainly to fish and seaweed consumption. A similar study¹⁰⁴ using the same protocol reached the same conclusion for total and iAs. Dietary intakes among Japanese children (n = 104) and pregnant women (n = 101) in two different cities were assessed through analysis of food and water using ICP-MS. Estimated intakes of total and iAs were significantly higher in children at 8.46 ± 3.02 and 1.74 ± 1.07 μg per kg bw per week compared to pregnant women (20.07 ± 3.53 and 8.46 ± 3.02 μg per kg bw per week). Although drinking water was not considered the source of high As, both total and iAs were higher in frequent consumers of hijiki seaweed.

Curiously Styburski et al.¹⁰⁸ have assessed beer as a potential source of macroelements (Ca, Cl, K, P) given social license frowns on alcoholic beverages. Bottled beer (52 samples from 10 countries, covering five styles but predominantly lager or pilsner) were analysed on a sample volume of 6 mL using EDXRF. Beer was shown to furnish a reasonable proportion of Ca (0.05–0.31 g L⁻¹) only, up to 15.5% based on the USA daily requirements, from one 500 mL bottle; the highest concentrations were from German and Armenian beers. The remainder of the studied elements were shown to provide up to 3% of the daily need from a 500 mL bottle. Halloysite nanotubes (HNT) were developed²⁹ to provide a preconcentration technique for low level determination of Se using HR-CS ETAAS (described above). Samples (500 mg) were digested (20 min, 300 W) with HNO₃ (65%, 5 mL) after moistening with H₂O₂ (30%, 1 mL) before preconcentration with the HNTs. Four types of flour, and several types of dried berries were analysed using this new procedure. Corn, rice and buckwheat flour samples purchased in Poznan, Poland, had Se concentrations between 0.025 and 0.031 μg g⁻¹ whereas oat flour had considerably higher values at 0.228 μg g⁻¹. Cranberries and raisin had concentrations <LOD (0.01 μg L⁻¹). The RSDs (replicate n = 5) of each sample were between 4 and 8%.

7.2.2 Human milk and infant formula. Heavy metal content of human milk was correlated to lifestyle of Norwegian mothers.¹⁰⁹ Samples (n = 300) from the Norwegian Human Milk Study were prepared using microwave-assisted digestion; Cd, Hg and Pb were determined using QQQ-ICP-MS. Median concentrations (range) were 0.057 (0.017–12), 0.20 (<0.058–0.89) and <0.67 (<0.2–7.5) μg kg⁻¹, respectively. Multiple linear regression, for Cd and Hg, and logistic regression, for Pb >LOQ were used to examine predictors for correlation with lifestyle factors such as diet, smoking and amalgam tooth fillings. No significant correlations were found for Cd; Pb in human milk was associated with consumption of liver and kidney from game animals and Hg with number of amalgam fillings (45% variance) and fish consumption (10% variance), specifically Atlantic halibut lean fish, mussels, scallops and crab.

7.2.3 Dairy products. Adequate I consumption is required in both infants and adults, and has been the focus of research in recent years. Niero et al.¹¹⁰ used extraction into NH₄OH (0.6% v/v) followed by I determination using ICP-MS of several types of raw and processed milk. Processed and partially skimmed cow milk had the highest concentration (359.42 μg kg⁻¹) while, curiously, raw milk had the lowest (166.92 μg kg⁻¹), although this was most probably due to variations in sample origin rather than technological differences. Milk from other species provided a wide range of I concentrations: raw goat 575.42 μg kg⁻¹; raw buffalo 229.82 μg kg⁻¹; sheep 192.64 μg kg⁻¹; donkey 7.06 μg kg⁻¹. Overall reproducibility, as RSD, was 4.01% (n = 45, 3 d, 3 operators). Although inorganic elements account for less than 1% of milk composition, assessment of its nutritional value by XRF was reviewed by Pashkova et al.¹³ They noted routine industry applications predominantly use EDXRF techniques over WDXRF despite potential improvements in resolution and sensitivity. Sample preparation invariably involved sample evaporation and pelletisation although some uses of TXRF allowed liquid analysis with dilution. The review authors recommended XRF for routine nutritional and quality assessment as well as for environmental monitoring in dairy and milk products.

7.2.4 Cereals, flour and rice. Dietary intake of MeHg, through consumption of staple cereals, is receiving growing attention, although fish is considered the primary source. A thermal desorption amalgamation (TDA)-AAS method with prior microwave digestion was developed for determination of MeHg in rice.⁵³ The accuracy and reliability of TDA-AAS was corroborated by HPLC-ICP-MS measurements of the L-cysteine back-extraction solutions. The MeHg and total Hg concentrations were determined in rice samples from two provinces of China (Jilin and Hubei). Average values were: for Jilin (n = 5) 0.75 ng g⁻¹ MeHg (range 0.49–1.19 ng g⁻¹), 1.76 ng g⁻¹ total Hg (range 1.43–2.42 ng g⁻¹), and Hubei (n = 10) 3.21 ng g⁻¹ MeHg (range 0.44–6.87 ng g⁻¹) and 6.02 ng g⁻¹ total Hg (range 3.02–16.08 ng g⁻¹). The higher concentrations of both MeHg and total Hg were ascribed to widespread non-ferrous mining and smelting operations. The calculated probable daily intakes derived from this and other surveys reveal that MeHg exposure through rice consumption does not cause significant health risk to adults but the children in Hubei province have unacceptable exposure risk. This study had limited sample numbers (n = 15) but indicated that further research is needed. The origin of rice samples was determined using LIBS and standard multi-dimensional statistics.⁴⁷ Twenty samples from 10 rice-producing regions, from north to south along the eastern seaboard of China, and also from Thailand, were studied. For each rice origin, four test samples were prepared, and 50 spectra were recorded at 25 surface locations with a 40 mJ laser pulse energy, 10 Hz frequency and gate width/delay of 3 μs/1.5 μs. Most of the valuable spectral lines with high intensities were around the regions of 240–320 nm, 350–450 nm, and 700–880 nm. Sample categorization was best achieved with Support Vector Machine data analysis achieving over 99% classification although this technique had the weakest modelling efficiency. The authors did not share which particular spectral features could be used to differentiate the rice samples.

7.2.5 Vegetables, fruits, mushrooms and nuts. The topic of As speciation in mushroom was reviewed (144 references) by Braeuer and Goessler¹⁴ citing over 18 species found. The preferred analytical technique remains HPLC-ICP-MS although HG-AFS, and also (μ-)XANES for spatial resolution information, have been used. It was questioned whether fungi can actually carry out their own bio-transformations and stated that although well-designed experiments have been performed, no firm conclusions have been made; potentially this is due to high total As level being employed. The authors suggested that it was always important to identify unknown species in order to elucidate mechanisms and suggested that the discovery of arsenoaldehydes might confirm the conversion of AC to AsB, and that further thio-arsenical research, would be of interest. They also question whether AsB formation follows the same pathway as in marine organisms, particularly with only DMA being found in some species.

The fate of Zn in plants was investigated through the proxy model of lettuce Lactuca sativa L. for the first time by Wojcieszek et al.¹¹¹ Total Zn was the same in plants supplemented with ZnCl₂ or ZnO NPs and using sp-ICP-MS the latter’s absence in leaves and stems indicate that NPs are dissolved rapidly in growth medium and therefore do not pose a health issue. Speciation of the Zn using HPLC (SEC and HILIC) with ICP-MS and ESI-MS/MS detection in the leaves indicated that 70% of Zn was complexed with nicotinamide.

7.2.6 Fish and seafood. Speciation of As in seaweed and seafood was measured by HPLC-ICP-MS using an octodisilane column with a mobile phase containing an ion-pair reagent.¹⁰³ Among the dried seaweed products, brown algae akamoku, hijiki, and mozuku contained significantly high levels of iAs. The intestinal organs of oyster, sardine and scallop contained higher As concentrations than muscle tissue.

7.2.7 Drinking water and non-alcoholic beverages. Although it is a widespread, highly toxic persistent pollutant with adverse health effects on humans, concentrations of Hg in bottled water have always been reported as below the method detection limits. Varde et al.¹¹² surveyed natural mineral water sources from 18 Italian regions and measured the total Hg concentrations by CV-AFS. A total of 255 waters were tested, representing 164 brands of bottled water. Concentrations were sub-ng L⁻¹ to a few ng L⁻¹. It was observed that concentrations were related to the environmental characteristics of the springs and to the impact of human activities. Higher concentrations were found in waters coming from regions with former mining and/or natural thermal and volcanic activity. Estimates of the Hg intake by population (adults, children and toddlers) from drinkable mineral waters were well below the European and Italian provisional tolerable value of 1 μg L⁻¹.

The need for Sb speciation is driven by the emerging understanding of the impact of Sb^IIIused in the manufacture of pet food and beverage packaging. A new method developed by de Oliveira et al.⁶⁷ used the ternary oxide SiO₂/Al₂O₃/SnO₂ in a minicolumn to selectively preconcentrate Sb^III ions prior to determination by HG-AAS. The preconcentration factor was 136 and the LOD and LOQ were 0.17 and 0.50 μg L⁻¹, respectively with only 20 mL of sample. Analysis of mineral water stored in blue PET bottles showed that the Sb^III concentration ranged from 0.5 to 1.0 μg L⁻¹.

7.2.8 Alcoholic beverages. A simple and accurate method for Cu and Fe speciation in wine is key for understanding shelf-life and risk of developing S-off odours and oxidation. Kontoudakis et al.⁶⁶ compared SPE-ICP-AES, LLE-ETAAS (as well as electrochemical techniques) to correctly ascertain the forms of both oxidation states of Cu and Fe. They concluded that the former technique was the best using a non-polar styrene-divinylbenzene cartridge (100 μm, 1 g/6 mL) and a strong cation exchange cartridge as SPE adsorbants. In white wine Fe was predominantly complexed with organic acids (tartaric, malic, citric) at low pH (ca. 3.2) while in red wines there was increased residual Fe which was correlated with phenolic compounds such as hydroxycinnamic acids and catechins. The LLE ETAAS using NH₄SCN/MIBK approach perturbed the delicate Cu–S complexes to allow for meaningful interpretation. The direct ICP-MS analysis of undiluted wine, which is problematic due to presence of organic and salts matrix, has been achieved successfully using a high temperature torch and low flow rate.³⁸ The torch was operated at a temperature of 125 °C with a low flow rate of 30 μL min⁻¹ supplied from a double-pass Scott-type spray chamber thermostatted to 2 °C. Problems of soot or salt deposition on the interfaces were mitigated with this arrangement. Using external calibration with aqueous standards containing 10% EtOH determination of As, Cd, Cr, Cu, Gd, Fe, Mn, Mo, Nd, Ni, Pb, Sm, Tb, Ti, V, Zn was comparable with digestion and normal ICP-MS analysis. Classification of 34 white and red wines from Rioja, Spain, using 19 elements, determined with internal calibration by ICP-MS after acid digestion, was an example of this year’s offering in this common type of study.¹¹³ The experimental design was convoluted, with different combinations of grape varieties and subregions being present, and incorporating other factors such as different vineyard foliar spray application rates, use of oak barrel ageing and SO₂ addition. Despite the severely reduced sample numbers which would allow strong conclusions to be drawn for these different factors, the authors suggest that differentiation of varieties can be made using Ca, Mg, Mn and Sr; geographical zones using Ba, Cu, Ni, Sr; soil types: Cs and Pb; N foliar sprays with Pb, Ni, Mn and Zn; that Cs, Cu, Mg, and Pb in wines characterized SO₂ addition; and that Cs and Na differentiate wines stored in barrel. With counterfeit spirits – particularly whisk(e)y – accounting for a large proportion of luxury goods fraud globally, and despite some strong techniques to detect this, Pawlaczyk et al.¹¹⁴ have sought to redress the low number of papers on their trace element content. Twenty samples (2 mL) of blended (n = 15) and single-malt whisk(e)y (n = 5) from Scotland (Lowland, Highland, Speyside; n = 18), Ireland (n = 1) and USA (n = 1) were first microwave-assisted digested with HNO₃ then determined with ICP-TOF-MS. Sadly, the authors admitted that no particular metal fingerprints were recognised for different geographical Scottish regions.

8 Abbreviations

3D	three dimensional
AAS	atomic absorption spectrometry
AB	arsenobetaine
AC
AEC	anion exchange chromatography
AES	atomic emission spectrometry
AF	atomic fluorescence
AFS	atomic fluorescence spectrometry
AsB
ASU	Atomic Spectrometry Update
BMI	body mass index
bw	body weight
CDC	Centers for Disease Control and Prevention
CRM	certified reference material
CS	continuum source
CSF	cerebrospinal fluid
CV	cold vapour
DES	deep eutectic solvent
DMA	dimethylarsonic acid
DMF	dimethylformamide
ECD	electron capture detector
EDTA	ethylenediamine tetraacetic acid
EDXRF	energy dispersive X-ray diffraction
EF	enrichment factor
EPA	Environmental Protection Agency
EQA	external quality assessment
ESI	electrospray ionisation
ETAAS	electrothermal atomic absorption spectrometry
EtHg	ethylmercury
EtOH	ethanol
EU	European Union
FAAS	flame atomic absorption spectrometry
FI	flow injection
FTIR	Fourier transform infrared
GAW	Global Atmosphere Watch
GC	gas chromatography
GDM	gestational diabetes mellitus
HG	hydride generation
HPLC	high performance liquid chromatography
HG	hydride generation
HR	high resolution
iAs	inorganic arsenic
IBMK	isobutyl methyl ketone
ICP-AES	inductively coupled plasma atomic emission spectrometry
ICP-MS	inductively coupled plasma mass spectrometry
ICP-OES	inductively coupled plasma optical emission spectrometry
ICP-QMS	ICP-quadrupole MS
ID	isotope dilution
IDA	isotope dilution analysis
iHg	inorganic mercury
LA	laser ablation
LC	liquid chromatography
LDA	linear discriminant analysis
LIBS	laser-induced breakdown spectroscopy
LLE	liquid–liquid extraction
LOD	limit of detection
LOQ	limit of quantification
MC	multicollector
MeHg	methyl mercury
MRI	magnetic resonance imaging
nanoXRF	nano X-ray fluorescence
NHANES	National Health and Nutrition Examination Survey
NIST	National Institute of Standards and Technology
NP	nanoparticle
PAGE	polyacrylamide gel electrophoresis
PCA	principal component analysis
PFA	perfluoroalkyl
PVA	polyvinyl alcohol
REE	rare earth elements
RF	reversed phase
RSD	relative standard deviation
SD	standard deviation of mean
SEM	scanning electron microscopy
SEC	size exclusion chromatography
SF	sector field
SIMS	secondary ion mass spectrometry
S/N	signal-to-noise ratio
SPE	solid phase extraction
spICP-MS	single particle ICP-MS
SRM	standard reference material
TEM	transmission electron microscopy
TGA	thermogravimetric analysis
THF	tetrahydrofuran
TMAH	tetramethylammonium hydroxide
TOF	time-of-flight
TXRF	total reflection XRF
UPLC	ultra high performance liquid chromatography
UV	ultraviolet
VG	vapour generation
WDXRF	wavelength-dispersive X-ray fluorescence
XANES	X-ray absorption near-edge structure
XRD	X-ray diffraction
XRF	X-ray fluorescence

Conflicts of interest

There are no conflicts to declare.

References

1. A. Taylor, N. Barlow, M. P. Day, S. Hill, N. Martin and M. Patriarca, J. Anal. At. Spectrom., 2019, 34(3), 426–459.
2. S. Carter, R. Clough, A. Fisher, B. Gibson, B. Russell and J. Waack, J. Anal. At. Spectrom., 2018, 33(11), 1802–1848.
3. C. Vanhoof, J. R. Bacon, A. T. Ellis, L. Vincze and P. Wobrauschek, J. Anal. At. Spectrom., 2018, 33(9), 1413–1431.
4. J. R. Bacon, O. T. Butler, W. R. L. Cairns, J. M. Cook, R. Mertz-Kraus and J. F. Tyson, J. Anal. At. Spectrom., 2019, 34(1), 9–58.
5. E. H Evans, J. Pisonero, C. M. M. Smith and R. N. Taylor, J. Anal. At. Spectrom., 2019, 34, 803–822.
6. R. Clough, C. F. Harrington, S. J. Hill, Y. Madrid and J. F. Tyson, J. Anal. At. Spectrom., 2019, 34, 1306–1350.
7. B. Meermann and V. Nischwitz, J. Anal. At. Spectrom., 2018, 33(9), 1432–1468.
8. J. Hu, P. Yang and X. D. Hou, Appl. Spectrosc. Rev., 2018, 54, 180–203.
9. L. Hang, Z. Y. Xu, W. Hang and B. L. Huang, Spectrosc. Spectral Anal., 2019, 39(5), 1329–1339.
10. Q. X. Sun, X. Wei, X. Liu, T. Yang, M. L. Chen and J. H. Wang, Spectrosc. Spectral Anal., 2019, 39(5), 1340–1345.
11. R. Gaudiuso, N. Melikechi, Z. A. Abdel-Salam, M. A. Harith, V. Palleschi, V. Motto-Ros and B. Busser, Spectrochim. Acta, Part B, 2019, 152, 123–148.
12. F. Bianchi, M. Giannetto and M. Careri, TrAC, Trends Anal. Chem., 2018, 107, 142–150.
13. G. V. Pashkova, A. N. Smagunova and A. L. Finkelshtein, TrAC, Trends Anal. Chem., 2018, 106, 183–189.
14. S. Braeuer and W. Goessler, Anal. Chim. Acta, 2019, 1073, 1–21.
15. M. Sargent, H. Goenaga-Infante, K. Inagaki, L. Ma, J. Meija, A. Pramann, O. Rienitz, R. Sturgeon, J. Vogl, J. Wang and L. Yang, Metrologia, 2019, 56(3), 034005.
16. A. A. Krata, M. Wojciechowski, M. Kalabun and E. Bulska, Microchem. J., 2018, 142, 36–42.
17. M. C. Aprea, P. Apostoli, M. Bettinelli, P. Lovreglio, S. Negri, L. Perbellini, A. Perico, M. C. Ricossa, F. Salamon, M. L. Scapellato and I. Iavicoli, Toxicol. Lett., 2018, 298, 177–185.
18. M. Iwai-Shimada, S. Kameo, K. Nakai, K. Yaginuma-Sakurai, N. Tatsuta, N. Kurokawa, S. F. Nakayama and H. Satoh, Environ. Health Prev. Med., 2019, 24(1), 35.
19. C. D. Ward, R. J. Williams, K. Mullenix, I. Syhapanha, R. L. Jones and K. Caldwell, At. Spectrosc., 2018, 39(6), 219–228.
20. P. M. Roos, J. Trace Elem. Med. Biol., 2019, 52, 48–52.
21. D. S. Tevis, J. M. Jarrett, D. R. Jones, P. Y. Cheng, M. Franklin, N. Mullinex, K. L. Caldwell and R. L. Jones, Clin. Chim. Acta, 2018, 485, 1–6.
22. D. L. Novo, J. E. Mello, F. S. Rondan, A. S. Henn, P. A. Mello and M. F. Mesko, Talanta, 2019, 191, 415–421.
23. E. M. Papaslioti, A. Parviainen, M. J. R. Alpiste, C. Marchesi and C. J. Garrido, Food Chem., 2019, 274, 726–732.
24. F. D. Dias, L. B. Peixoto, L. A. Meira and J. A. Barreto, Anal. Methods, 2018, 10(35), 4257–4263.
25. Z. A. Alothman, M. A. Habila, E. Yilmaz and M. Soylak, At. Spectrosc., 2018, 39(3), 106–111.
26. A. Thongsaw, Y. Udnan, G. M. Ross and W. C. Chaiyasith, Talanta, 2019, 197, 310–318.
27. H. R. Sobhi, A. Mohammadzadeh, M. Behbahani and A. Esrafili, Microchem. J., 2019, 146, 782–788.
28. L. Husakova, T. Sidova, L. Ibrahimova, M. Svizelova and T. Mikysek, Anal. Methods, 2019, 11(9), 1254–1263.
29. M. Krawczyk-Coda, Food Anal. Methods, 2019, 12(1), 128–135.
30. O. Caylak, S. G. Elci, A. Hol, A. Akdogan, U. Divrikli and L. Elci, Food Chem., 2019, 274, 487–493.
31. X. S. Huang, M. X. Guan, Z. L. Z. Lu and Y. P. Hang, Int. J. Anal. Chem., 2018 DOI:10.1155/2018/8045324.
32. F. Zhao, Y. M. Liu, X. Q. Zhang, R. Dong, W. J. Yu, Y. F. Liu, Z. M. Guo, X. M. Liang and J. H. Zhu, J. Chromatogr. A, 2018, 1573, 48–58.
33. M. H. Habibollahi, K. Karimyan, H. Arfaeinia, N. Mirzaei, Y. Safari, R. Akramipour, H. Sharafi and N. Fattahi, J. Sci. Food Agric., 2019, 99(2), 656–665.
34. C. D. Quarles, M. Szoltysik, P. Sullivan and M. Reijnen, J. Anal. At. Spectrom., 2019, 34(2), 284–291.
35. J. Entwisle, D. Malinovsky, P. J. H. Dunn and H. Goenaga-Infante, J. Anal. At. Spectrom., 2018, 33(10), 1645–1654.
36. A. L. Galusha, A. C. Haig, M. S. Bloom, P. C. Kruger, A. McGough, N. Lenhart, R. Wong, V. Y. Fujimoto, E. Mok-Lin and P. J. Parsons, J. Anal. At. Spectrom., 2019, 34(4), 741–752.
37. X. Z. Zhou and H. W. Liu, Spectrosc. Spectral Anal., 2018, 38(11), 3567–3571.
38. C. Cerutti, C. Sanchez, R. Sanchez, F. Ardini, M. Grotti and J. L. Todoli, J. Anal. At. Spectrom., 2019, 34(4), 674–682.
39. J. Soukal, R. E. Sturgeon and S. Musil, Anal. Chem., 2018, 90(19), 11688–11695.
40. H. Wang, B. B. Chen, M. He, X. T. Li, P. Y. Chen and B. Hu, Talanta, 2019, 200, 398–407.
41. A. K. Venkatesan, B. T. Rodriguez, A. R. Marcotte, X. Y. Bi, J. Schoepf, J. F. Ranville, P. Herckes and P. Westerhoff, Environ. Sci.: Water Res. Technol., 2018, 4(12), 1923–1932.
42. F. J. Alonso-Garcia, E. Blanco-Gonzalez and M. Montes-Bayon, Talanta, 2019, 194, 336–342.
43. S. Kroger, M. Sperling and U. Karst, J. Trace Elem. Med. Biol., 2019, 51, 50–56.
44. L. Fernandez-Lopez, B. Gomez-Nieto, W. J. Gismera, M. T. Sevilla and J. R. Procopio, Spectrochim. Acta, Part B, 2018, 147, 21–27.
45. O. Linhart, A. Kolorosova-Mrazova, J. Kratzer, J. Hranicek and V. Cerveny, Anal. Lett., 2019, 52(4), 613–632.
46. J. H. Li, Z. H. Li, C. J. Wang, Y. L. Yang and X. Y. Zhang, Fuel Process. Technol., 2018, 178, 312–321.
47. P. Yang, R. Zhou, W. Zhang, S. S. Tang, Z. Q. Hao, X. Y. Li, Y. F. Lu and X. Y. Zeng, Appl. Opt., 2018, 57(28), 8297–8302.
48. P. Yang, R. Zhou, W. Zhang, R. X. Yi, S. S. Tang, L. B. Guo, Z. Q. Hao, X. Y. Li, Y. F. Lu and X. Y. Zeng, Food Chem., 2019, 272, 323–328.
49. S. H. Nam, S. W. Kwon and Y. Lee, Anal. Lett., 2018, 51(17), 2833–2846.
50. J. Bockova, A. M. Roldan, J. Yu and P. Veis, Appl. Opt., 2018, 57(28), 8272–8278.
51. C. T. Chen, D. Banaru, T. Sarnet and J. Hermann, Spectrochim. Acta, Part B, 2018, 150, 77–85.
52. M. T. Liu, T. P. Liu, J. X. Liu, X. F. Mao, X. Na, L. Ding, G. Y. Chen and Y. Z. Qian, J. Anal. At. Spectrom., 2019, 34(3), 526–534.
53. Q. Luo, X. Gao, Y. N. Hu, K. D. Lin, X. J. Wang and H. F. Cheng, Anal. Methods, 2019, 11(10), 1361–1370.
54. M. D. Potes, F. V. Nakadi, C. F. G. Frois, M. G. R. Vale and M. M. da Silv, Microchem. J., 2019, 147, 324–332.
55. J. Jeffery, A. R. Frank, S. Hockridge, H. Stosnach and S. J. Costelloe, Ann. Clin. Biochem., 2019, 56(1), 170–178.
56. C. L. Mota, A. Pickler, D. Braz, R. C. Barroso, A. Mantuano, C. Salata, S. C. Ferreira-Machado, C. C. Lau and C. E. de Almeida, J. Instrum., 2018, 13, C04021.
57. R. Dalipi, R. Berneri, M. Curatolo, L. Borgese, L. E. Depero and E. Sangiorgi, Spectrochim. Acta, Part B, 2018, 148, 16–22.
58. A. J. Specht, A. S. Dickerson and M. G. Weisskopf, Environ. Res., 2019, 172, 273–278.
59. E. Dao, M. P. Zeller, B. C. Wainman and M. J. Farquharson, J. Trace Elem. Med. Biol., 2018, 50, 305–311.
60. F. S. M. Afzal, A. Al-Ebraheem, D. R. Chettle, E. D. Desouza, M. J. Farquharson, J. M. O’Meara, A. Pidruczny, B. C. Wainman and F. E. McNeill, Nucl. Instrum. Methods Phys. Res., Sect. B, 2018, 433, 1–9.
61. M. Acosta, S. Torres, L. Marino-Repizo, L. D. Martinez and R. A. Gil, J. Pharm. Biomed. Anal., 2018, 158, 209–213.
62. F. Gronbaek-Thorsen, S. Sturup, B. Gammelgaard and L. H. Moller, J. Anal. At. Spectrom., 2019, 34(2), 375–383.
63. H. Y. Cheng, W. W. Zhang, Y. C. Wang and J. H. Liu, Microchim. Acta, 2018, 185(9), 425.
64. T. Narukawa, T. Iwai, K. Chiba and J. Feldmann, Anal. Sci., 2018, 34(11), 1329–1334.
65. K. Marschner, A. H. Petursdottir, P. Bucker, A. Raab, J. Feldmann, Z. Mester, T. Matousek and S. Musil, Anal. Chim. Acta, 2019, 1049, 20–28.
66. N. Kontoudakis, L. M. Schmidtke, M. Z. Bekker, M. Smith, P. A. Smith, G. R. Scollary, E. N. Wilkes and A. C. Clark, Food Chem., 2019, 274, 89–99.
67. L. L. G. de Oliveira, G. O. Ferreira, F. A. C. Suquila, F. G. de Almeida, L. A. Bertoldo, M. G. Segatelli, E. S. Ribeiro and C. R. T. Tarley, Food Chem., 2019, 294, 405–413.
68. S. Queipo-Abad, P. R. Gonzalez, E. Martinez-Morillo, W. C. Davis and J. I. G. Alonso, Sci. Total Environ., 2019, 672, 314–323.
69. S. J. M. Van Malderen, T. Van Acker, B. Laforce, M. De Bruyne, R. de Rycke, T. Asaoka, L. Vincze and F. Vanhaecke, Anal. Bioanal. Chem., 2019, 411(19), 4849–4850.
70. K. Lohr, H. Traub, A. J. Wanka, U. Panne and N. Jakubowski, J. Anal. At. Spectrom., 2018, 33(9), 1579–1587.
71. O. B. Bauer, O. Hachmoller, O. Borovinskaya, M. Sperling, H. J. Schurek, G. Ciarimboli and U. Karst, J. Anal. At. Spectrom., 2019, 34(4), 694–701.
72. S. Blockhuys, P. Malmberg and P. Wittung-Stafshede, Biointerphases, 2018, 13(6), 06E412.
73. E. J. McAllum and D. J. Hare, Spectrochim. Acta, Part B, 2019, 156, 20–32.
74. M. H. M. Ali, F. Rakib, V. Nischwitz, E. Ullah, R. Mall, A. M. Shraim, M. I. Ahmad, Z. K. Ghouri, D. McNaughton, S. Kuppers, T. Ahmed and K. Al-Saad, Appl. Sci., 2018, 8(12), 2436.
75. D. Clases, S. Fingerhut, A. Jeibmann, M. Sperling, P. Doble and U. Karst, J. Trace Elem. Med. Biol., 2019, 51, 212–218.
76. C. Streli, M. Rauwolf, A. Turyanskaya, D. Ingerle and P. Wobrauschek, Appl. Radiat. Isot., 2019, 149, 200–205.
77. W. Osterode, G. Falkenberg, P. Ferenci and F. Wrba, J. Trace Elem. Med. Biol., 2019, 51, 42–49.
78. A. Gianoncelli, C. Rizzardi, D. Salomon, V. Canzonieri and L. Pascolo, Spectrochim. Acta, Part B, 2018, 147, 71–78.
79. L. Marie-Hardy, P. O’Laughlin, M. Bonnin and T. A. S. Selmi, Orthop. Traumatol. Surg. Res., 2018, 104(8), 1179–1182.
80. M. Kuba, J. Gallo, T. Pluhacek, M. Hobza and D. Milde, J. Biomed. Mater. Res., Part B, 2019, 107(2), 454–462.
81. M. Pettersson, J. Pettersson, A. Johansson and M. M. Thoren, J. Oral Rehabil., 2019, 46(2), 179–188.
82. M. E. Cabana-Munoz, J. M. Parmigiani-Izquierdo, F. C. Alonso and J. J. Merino, J. Clin. Med., 2019, 8(1), 86.
83. C. K. Mikelson, J. Troisi, A. LaLonde, S. J. K. Symes, S. W. Thurston, L. M. DiRe, C. D. Adair, R. K. Miller and S. M. Richards, Environ. Res., 2019, 168, 118–129.
84. D. R. C. McLachlan, C. Bergeron, P. N. Alexandrov, W. J. Walsh, A. I. Pogue, M. E. Percy, T. P. A. Kruck, Z. D. Fang, N. M. Sharfman, V. Jaber, Y. H. Zhao, W. H. Li and W. J. Lukiw, Mol. Neurobiol., 2019, 56(2), 1531–1538.
85. G. F. Zhang, X. Y. Wang, X. Zhang, Q. Li, S. Z. Xu, L. Huang, Y. Zhang, L. X. Lin, D. Gao, M. Wu, G. Q. Sun, Y. Song, C. R. Zhong, X. F. Yang, L. P. Hao, H. Y. Yang, L. Yang and N. H. Yang, Environ. Int., 2019, 123, 164–170.
86. H. Wang, J. Li, X. Zhang, P. Zhu, J. H. Hao, F. B. Tao and D. X. Xu, Toxicol. Appl. Pharmacol., 2018, 356, 114–119.
87. B. B. Zhu, C. M. Liang, S. Q. Yan, Z. J. Li, K. Huang, X. Xia, J. H. Hao, P. Zhu and F. B. Tao, J. Trace Elem. Med. Biol., 2019, 52, 151–156.
88. A. G. Hoen, J. C. Madan, Z. G. Li, M. Coker, S. N. Lundgren, H. G. Morrison, T. Palys, B. P. Jackson, M. L. Sogin, K. L. Cottingham and M. R. Karagas, Sci. Rep., 2018, 8(1), 12627.
89. M. K. Stanley, K. Kelers, E. Boller and M. Boller, J. Vet. Emerg. Crit. Care, 2019, 29(2), 201–207.
90. C. Hjelm, F. Harari and M. Vahter, Environ. Res., 2019, 171, 60–68.
91. A. K. Nehra, R. J. McDonald, A. M. Bluhm, T. M. Gunderson, D. L. Murray, P. J. Jannetto, D. F. Kallmes, L. J. Eckel and J. S. McDonald, Radiology, 2018, 288(2), 416–423.
92. F. Berger, R. A. Kubik-Huch, T. Niemann, H. R. Schmid, M. Poetzsch, J. M. Froehlich, J. H. Beer, M. J. Thali and T. Kraemer, Radiology, 2018, 288(3), 703–709.
93. P. Robert, S. Fingerhut, C. Factor, V. Vives, J. Letien, M. Sperling, M. Rasschaert, R. Santus, S. Ballet, J. M. Idee, C. Corot and U. Karst, Radiology, 2018, 288(2), 424–433.
94. S. L. Yu, Y. C. Yin, Q. Cheng, J. H. Han, X. Q. Cheng, Y. Guo, D. D. Sun, S. W. Xie and L. Qiu, Scand. J. Clin. Lab. Investig., 2018, 78(6), 501–507.
95. M. Hitosugi, M. Tojo, M. Kane, N. Shiomi, T. Shimizu and T. Nomiyama, Int. J. Leg. Med., 2019, 133(2), 479–481.
96. S. Kobayashi, R. Kishi, Y. Saijo, Y. Ito, K. Oba, A. Araki, C. Miyashita, S. Itoh, M. Minatoya, K. Yamazaki, Y. A. Bamai, T. Sato, S. Yamazaki, S. F. Nakayama, T. Isobe, H. Nitta, T. Kawamoto, H. Saito, N. Yaegashi, K. Hashimoto, C. Mori, S. Itoh, Z. Yamagata, H. Inadera, M. Kamijima, T. Nakayama, H. Iso, M. Shima, Y. Hirooka, N. Suganuma, K. Kusuhara, T. Katoh and The Japan Environment and Children’s Study, Environ. Int., 2019, 125, 418–429.
97. M. H. M. Klose, A. Schoberl, P. Heffeter, W. Berger, C. G. Hartinger, G. Koellensperger, S. M. Meier-Menches and B. K. Keppler, Monatsh. Chem., 2018, 149(10), 1719–1726.
98. B. J. Tipple, L. O. Valenzuela, T. H. Chau, L. H. Hu, C. P. Bataille, L. A. Chesson and J. R. Ehleringer, Rapid Commun. Mass Spectrom., 2019, 33(5), 461–472.
99. R. D. Ash and M. He, Forensic Sci. Int., 2018, 292, 224–231.
100. H. Rebiere, A. Kermaidic, C. Ghyselinck and C. Brenier, Talanta, 2019, 195, 490–496.
101. P. Mattiazzi, D. Bohrer, E. Becker, C. Viana, P. C. Nascimento and L. M. Carvalho, Talanta, 2019, 197, 20–27.
102. F. C. Pinheiro, A. I. Barros and J. A. Nobrega, Microchem. J., 2019, 146, 948–956.
103. E. Matsumoto-Tanibuchi, T. Sugimoto, T. Kawaguchi, N. Sakakibara and M. Yamashita, J. AOAC Int., 2019, 102(2), 612–618.
104. N. Mise, M. Ohtsu, A. Ikegami, A. Mizuno, X. Y. Cui, Y. Kobayashi, Y. Nakagi, K. Nohara, T. Yoshida and F. Kayama, Food Addit. Contam., Part A, 2019, 36(1), 84–95.
105. R. Chekri, E. Le Calvez, J. Zinck, J. C. Leblanc, V. Sirot, M. Hulin, L. Noel and T. Guerin, J. Food Compos. Anal., 2019, 78, 108–120.
106. T. I. Todorov, T. Smith, A. Abdalla, S. Mapulanga, P. Holmes, M. Hamilton, T. Lewis and M. McDonald, Food Anal. Methods, 2018, 11(11), 3211–3223.
107. A. Hayashi, F. Sato, T. Imai and J. Yoshinaga, J. Food Compos. Anal., 2019, 77, 77–83.
108. D. Styburski, K. Janda, I. Baranowska-Bosiacka, A. Lukomska, K. Dec, M. Goschorska, B. Michalkiewicz, P. Zietek and I. Gutowska, Eur. Food Res. Technol., 2018, 244(10), 1853–1860.
109. M. Vollset, N. Iszatt, O. Enger, E. L. F. Gjengedal and M. Eggesbo, Sci. Total Environ., 2019, 677, 466–473.
110. G. Niero, M. Franzoi, V. Vigolo, M. Penasa, M. Cassandro, C. Boselli, G. Giangolini and M. De Marchi, J. Dairy Sci., 2019, 102(6), 4808–4815.
111. J. Wojcieszek, J. Jimenez-Lamana, K. Bierla, M. Asztemborska, L. Ruzik, M. Jarosz and J. Szpunar, J. Anal. At. Spectrom., 2019, 34(4), 683–693.
112. M. Varde, A. Servidio, G. Vespasiano, L. Pasti, A. Cavazzini, M. Di Traglia, A. Rosselli, F. Cofone, C. Apollaro, W. R. L. Cairns, E. Scalabrin, R. De Rosa and A. Procopio, Chemosphere, 2019, 219, 896–913.
113. E. P. Perez-Alvarez, R. Garcia, P. Barrulas, C. Dias, M. J. Cabrita and T. Garde-Cerdan, Food Chem., 2019, 270, 273–280.
114. A. Pawlaczyk, M. Gajek, K. Jozwik and M. I. Szynkowska, Molecules, 2019, 24(7), 1193.
115. D. Rolle-McFarland, Y. Z. Liu, J. Q. Zhou, F. Mostafaei, Y. Z. Zhou, Y. Li, Q. Y. Fan, W. Zheng, L. D. H. Nie and E. M. Wells, Int. J. Environ. Res. Public Health, 2018, 15(7), 1341.
116. Y. Jin, J. Coad, J. L. Weber, J. S. Thomson and L. Brough, Nutrients, 2019, 11(1), 69.
117. G. Ichihara, M. Iida, E. Watanabe, T. Fujie, T. Kaji, E. Lee and Y. Kim, J. Occup. Health, 2019, 61(3), 257–260.
118. U. Majewska, P. Lyzwa, A. Kubala-Kukus, D. Banas, J. Wudarczyk-Mocko, I. Stabrawa and S. Gozdz, Spectrochim. Acta, Part B, 2018, 147, 121–131.
119. S. E. Orr, M. C. Barnes, H. S. George, L. Joshee, B. Jeon, A. Scircle, O. Black, J. V. Cizdziel, B. E. Smith and C. C. Bridges, J. Toxicol. Environ. Health, Part A, 2018, 81(24), 1246–1256.
120. S. N. Danilchenko, Y. V. Rogulsky, A. N. Kulik and A. N. Kalinkevich, J. Appl. Spectrosc., 2019, 86(2), 264–269.
121. S. Bodur, S. Erarpat, D. S. Chormey, C. Buyukpinar and S. Bakirdere, Anal. Lett., 2019, 52(3), 539–548.
122. L. B. Paixao, G. C. Brandao, R. G. O. Araujo and M. G. A. Korn, Food Chem., 2019, 284, 259–263.
123. Y. Kataoka, T. Watanabe, K. Hayashi and H. Akiyama, Food Hyg. Saf. Sci., 2018, 59(6), 269–274.
124. N. Ozkantar, M. Soylak and M. Tuzen, At. Spectrosc., 2019, 40(1), 17–23.
125. N. Ozbek and S. Akman, Anal. Lett., 2018, 51(17), 2776–2789.
126. D. L. F. da Silva, M. A. P. da Costa, L. O. B. Silva and W. N. L. dos Santos, Food Chem., 2019, 273, 24–30.
127. M. A. Iturbide-Casas, G. Molina-Recio and F. Camara-Martos, Food Anal. Methods, 2018, 11(10), 2758–2766.
128. T. Unutkan, I. Koyuncu, C. Diker, M. Firat, C. Buyukpinar and S. Bakirdere, Bull. Environ. Contam. Toxicol., 2019, 102(1), 122–127.
129. L. Lu, Y. Li, M. J. Jing, M. D. Ma, Y. H. Peng, S. Y. Qin and L. A. Li, Spectrosc. Spectral Anal., 2019, 39(2), 607–611.
130. C. Papachristodoulou, M. C. Tsiamou, H. Sakkas and C. Papadopoulou, J. Food Compos. Anal., 2018, 72, 39–47.
131. E. Evgenakis, C. Christophoridis and K. Fytianos, Environ. Sci. Pollut. Res., 2018, 25(27), 26766–26779.
132. I. Orjales, C. Herrero-Latorre, M. Miranda, F. Rey-Crespo, R. Rodriguez-Bermudez and M. Lopez-Alonso, Animal, 2018, 12(6), 1296–1305.
133. M. J. Hidalgo, M. T. Pozzi, O. J. Furlong, E. J. Marchevsky and R. G. Pellerano, Microchem. J., 2018, 142, 30–35.

---