2020 atomic spectrometry update – a review of advances in X-ray fluorescence spectrometry and its special applications

Contents

• 1. Introduction

• 2. Chemical imaging using X-ray techniques
  • 2.1 Computed tomography and 3D XRF techniques
  • 2.2 Laboratory-based 2D XRF techniques

• 3. Synchrotron and large scale facilities

• 4. Grazing X-ray techniques including TXRF spectrometry

• 5. Hand-held, mobile and on-line XRF techniques
  • 5.1 Hand-held and mobile XRF techniques
  • 5.2 On-line XRF techniques

• 6. Cultural heritage applications

• 7. Abbreviations

• 8. Conflicts of interest

• Acknowledgements

• References

---

Abstract

The XRF-CT technique continues to gain attention with respect to both its development and applications. The emergence of fast scanning approaches in conjunction with ED-detectors of high counting rate capability has allowed XRF-CT to evolve towards a truly 3D elemental imaging technique. At SR sources, the technique is regularly applied at sub-μm resolution together with complementary imaging modalities such as absorption and phase-contrast CT, XRD-CT, 2D/3D XAS and ptychography. The TXRF spectrometric analysis of suspensions and undigested samples (e.g. clay, cosmetics and nanoparticles) continues to receive much attention. In the analysis of fly ash samples, elemental sensitivities were affected by the angle of the measurement. In the use of TXRF spectrometry for the analysis of particle-like large specimens, the influence of the standing wave field was considered to be negligible. However, more work needs to be done on this source of uncertainty to improve our understanding of the influence of the standing wave field on these kind of samples. Novel equipment capable of very good angle resolution will be useful for such studies. Pre-concentration procedures are becoming more sophisticated with the introduction of new substrates for ligands capable of binding heavy metals and, in some cases, of being applied directly on the TXRF spectrometry reflector for immediate analysis. A sophisticated explanation of the excitation spectrum reported for μ-XRF spectrometry included a new description of the energy distribution of electrons in the target. Another highlight was the application of high-resolution monochromatic μ-XRF spectrometry using DCC optics for excitation and focusing as well as for increasing the spectral resolution of the Pu and U L-lines. Although the use of portable XRF spectrometry systems is now well established, new dedicated systems continue to be developed including those with a triaxial configuration. The focus is now primarily on the development and use of portable TXRF spectrometry systems for an expanding range of applications. In the area of cultural heritage applications, macroXRF spectrometry is a well-established imaging technique for the analysis of paintings that is now also being used to image stained-glass panels. The extensive study using a number of micro- and macro-techniques of Vermeer’s painting of The Girl with a Pearl Earring illustrated the power of this approach. Not only could the different shades of blue in the painting be characterised but also the materials and techniques used to achieve various effects in the painting.

---

1. Introduction

This review describes advances in the XRF spectrometry group of techniques published approximately between April 2019 and March 2020. The review is selective with the aim of providing a critical insight into developments in instrumentation, methodologies and data handling that represent a significant advance in XRF spectrometry. It is not the intention of the review to cover comprehensively the applications of XRF spectrometry techniques except in those cases where the non-destructive and remote sensing nature of XRF spectrometry analysis makes it particularly valuable and the method of choice. These applications concern samples which are irreplaceable and of great cultural value such as works of art and archaeological artefacts. For a wider appreciation of the applications of XRF spectrometry, this review should be read in conjunction with other related ASUs in the series, namely: environmental analysis;¹ clinical and biological materials, foods and beverages;² advances in atomic spectrometry and related techniques;³ elemental speciation;⁴ and metals, chemicals and functional materials.⁵

Regular readers of this annual review will note that regrettably Andy Ellis is no longer able to contribute after more than thirty years “service”. Andy has been involved with this review from its inception so his contribution has been immense. He had a deep knowledge of his subject and so acted as an invaluable mentor to new writers, both in terms of the science and in the specific requirements for writing an ASU review. He will be sorely missed.

All the ASU reviews adhere to a number of conventions. An italicised phrase close to the beginning of each paragraph is intended to highlight the subject area of that individual paragraph. A list of abbreviations used in this review appears at the end. It is a convention of ASUs that information given in the paper being reported on is presented in the past tense whereas the views of the ASU reviewers are presented in the present tense.

2. Chemical imaging using X-ray techniques

2.1 Computed tomography and 3D XRF techniques

2.1.1 X-ray fluorescence computed tomography (XRF-CT). The XRF-CT technique continues to gain attention with respect to both its development and applications. The emergence of fast scanning approaches in conjunction with ED-detectors of high counting rate capability has allowed XRF-CT to evolve towards a truly 3D elemental imaging technique. At SR sources, the technique is regularly applied at sub-μm resolution together with complementary imaging modalities such as absorption and phase-contrast CT, XRD-CT, 2D/3D XAS and ptychography.

Besides conventional scanning XRF-CT at synchrotron sources, full-field X-ray fluorescence (FF-XRF) imaging techniques that do not require microbeam scanning of the sample have become available as a result of the development of novel energy-dispersive CCD-based 2D detectors, a notable example of which is referred to as ‘color X-ray camera (CXC)’ or ‘SLcam’. The different schemes used by De Samber et al.⁶ for the FF-XRF imaging of biological samples included: 2D magnified imaging with pinhole optics using the ‘camera obscura’ principle; fixed-magnification 2D imaging using magnifying polycapillary detector optics; and 3D FF-XRF imaging using an X-ray sheet beam or CT. The different FF-XRF imaging modes, demonstrated using the crustacean Daphnia magna, achieved a spatial resolution level of 8–10 μm.

The use of XRF-CT with complementary techniques is becoming attractive in the field of material science for visualising changes in the structure and chemistry of microvolumes of interest. Gill et al.⁷ used a multimodal approach at the NSLS II Hard X-ray Nanoprobe (HXN) beamline to investigate intergranular (IG) corrosion of sensitised stainless-steel microsamples using synchrotron-based nanoscale X-ray imaging and to determine the microstructure and chemical composition at the sub-μm scale. NanoXRF-CT could be used to map quantitatively the distribution of Cr and Fe at grain boundaries and at cracked corrosion regions. A pixel resolution of 80 nm was used for 3D tomography measurement employing a fly-scan with a dwell time of 0.05 s per pixel. The combination of XRF-CT with differential phase-contrast-imaging allowed variations in composition and associated structural changes to be identified thereby shedding light on the role of grain boundaries in the failure of stainless steel in corrosive environments and on the mechanism of IG corrosion.

In the field of environmental science, Porfido et al.⁸ applied a novel combination of laboratory-based X-ray spectrometry techniques to investigate As bioaccumulation in earthworms (Eisenia andrei) exposed to six polluted soils with As concentrations of 58–13 330 mg kg⁻¹. Thin sections were mapped using μ-XRF spectrometry and intact dehydrated earthworms that had been exposed to various polluted soils were analysed using an integrated μ-CT, XRF-CT and confocal μ-XRF spectrometry instrument (Herakles scanner at Ghent University, Belgium). Comparison of the μ-XRF spectrometry and the XRF-CT results showed that no redistribution of elements or distortion of the microstructure had occurred within the earthworm body at the μm scale as a result of the fixation, embedding, and cutting procedures required as part of the sample preparation for the 2D μ-XRF spectrometry analysis.

Martin et al.⁹ carried out an important study by investigating uranium particulates contained within the material ejected by the explosion at Fukushima Daiichi Nuclear Power Plant Unit 1. The authors focused on the >100 μm size fraction of Si-based particles sampled within 5 km of the Fukushima site. The combined use of μ-XRF spectrometry, absorption contrast tomography and μ-XANES revealed U^IV-containing particulates trapped on the exterior surface of highly porous Si-based particles. For the XRF-CT acquisition, the particles were scanned through a beam (5 μm, 19 keV) rastered with a step size of 2.5 μm over the sample under each angular view.

The μ-XANES tomography technique as used by Mijovilovich et al.¹⁰ in plant physiology studies adds an extra energy scan to XRF-CT. In a study on infection with Turnip Yellow Mosaic Virus (TYMV), the Zn distribution and Zn speciation were measured in the non-accumulator plant Noccaea ochroleucum. The plant had been enriched in Zn to a concentration (5000 ppm) sufficiently high for 3D XRF spectrometry and XANES measurements. A spatial resolution of 5 μm was sufficient at PETRA III P06 for visualising the upper and lower epidermis thereby suggesting that biomineralisation occurred in the TYMV-infected samples to form crystallites strongly resembling Zn silicate.

A milestone in quantitative nanoscale XRF-CT imaging was¹¹ the complementary use of X-ray holographic nanotomography for 3D electron-density-mapping and XRF-CT for correlative 3D elemental-imaging. The ID16A Nano-Imaging beamline of the ESRF was used to produce nanoscale 3D mass concentration, mass density and mass fraction volumes from malaria-infected red blood cells measured under cryogenic vacuum conditions. The beamline provided an exceptionally bright X-ray nanofocus of 27 × 37 nm with a flux of 4 × 10¹¹ photons per s at 17 keV. The nanoscale resolution of the key endogenous elements Fe, P and S made it possible to reveal unambiguously the various cellular compartments, including the parasitic vacuole and the hemozoin crystals. Of particular interest was the comparison of the resolution, S/N and reconstruction fidelity for five Fe, P and S distributions in red blood cells.

In the field of cancer research, Conesa et al.¹² studied a new and highly potent Ir-based compound active at the sub-μm scale against a number of cancer cells. Cryogenic XRF-CT (resolution 70 nm) was used to complement soft X-ray tomography (resolution 50 nm) for the first time in order to explore the therapeutic potential of the compound. The distribution of the drug within the cell ultrastructure was mapped accurately with minimal perturbation of the cells. The elemental and structural imaging of cryopreserved samples has numerous advantages as free-cation intracellular flux, including potential metallodrug leakage from target organelles, is avoided and no chemical fixation is required. The method could be extremely valuable for the visualisation of the movement between organelles of key elements involved in biochemical processes in human cells.

A new concept of fully integrated thin-film solid-state interfacial de-alloying (thin-film-SSID) used¹³ electron microscopy, XRF-CT, XAS and atom probe tomography methods. The impressive 3D XRF nanotomography results for a Fe–Ni thin film (dealloyed with Mg by SSID at 460 °C for 30 minutes), obtained at the HXN beamline at NSLS-II, provided detailed distributions of Fe and Ni at the 10 nm scale within 250 nm to 1 μm thin-films. A Multilayer Laue Lens used for beam focusing at an energy of 12 keV provided the resolution of 10 nm required. The XRF-CT images, collected with a 2° step size in the angular range of 0–126° and a dwell time of 0.04 s, represented some of the highest resolution XRF-CT elemental images ever recorded.

The investigation of catalytic particles is becoming an important topic for X-ray structural and spectroscopic imaging. Bossers et al.¹⁴ used both X-ray ptychography and XRF-CT at the PETRA III P06 beamline to investigate with sub-μm 3D spatial resolution the fragmentation behaviour of individual polyolefin catalyst particles in the early stages of α-olefin polymerisation. The strength of using the two techniques to obtain both electron density and elemental distribution information was demonstrated by the imaging of a Ziegler–Natta catalyst particle (ca. 41 μm diameter) with a voxel size of 43.2 × 43.2 × 43.2 nm (ptychography) and 150 × 150 × 150 nm (XRF-CT). Another remarkable application in the field of catalysis was¹⁵ the use of single-particle XRF spectrometry and XRD analyses at a spatial resolution of 5 μm together with absorption tomography for studying nickel poisoning of a cracking catalyst. One of the primary sources of catalyst deactivation in the so-called fluid catalytic cracking process was Ni contamination from crude oil which promotes dehydrogenation–hydrogenation and speeds up coke growth. It was possible to observe the behaviour of Ni in these particles in situ for the first time and so detect that Ni hotspots (3500 ppm) were highly correlated with the γ-Al₂O₃ matrix throughout the entire particle. These experiments were performed at the Swiss Light Source (SLS) X05LA μ-XAS beamline.

A laboratory scale prototype XRF-CT imaging system was based¹⁶ on the use of a polycapillary focusing optic optimised for the localisation of gold NPs in model biological systems. The optic, custom-designed for the prototype and with a focal spot size of 2.6 mm at a source-to-sample distance of 3 cm, achieved an order-of-magnitude improvement in flux-density over that achievable with standard collimation. The test objects were water phantoms with a diameter of 30 mm and with mm-sized regions with Au concentrations of 60, 80, and 100 μg mL⁻¹. The acquisition time of ca. 30 minutes was >30-fold shorter than those reported previously for similar systems.

2.1.2 Confocal 3D XRF spectrometry. In a review article by Streli et al.,¹⁷ advances in the use of 2D and 3D micro- and nano-XRF spectrometries were discussed for the investigation of trace-elements in both healthy and cancerous bone tissues. The three studies reviewed were: the investigation of resorbable Mg implants in bone using laboratory μ-XRF spectrometry; a study of Zn accumulation in mineralised human osteosarcoma tissue by SR confocal μ-XRF spectrometry; and nanoXRF spectrometry imaging of human bone samples with a spatial resolution of ca. 500 nm.

The new PUMA beamline at SOLEIL, characterised for confocal-XRF spectrometry and μ-XANES applications by Tack et al.,¹⁸ is a hard-X-ray imaging beamline primarily aimed at the investigation of cultural heritage materials. The beamline was optimised for 2D and 3D imaging with a microscopic resolution, using analytical techniques such as XRF spectrometry, XAS, XRD and phase-contrast imaging. The X-ray source was a 1.8 T wiggler insertion device with 8.98 keV critical energy. A beam size of 2 × 3 μm and a photon flux of 1.5 × 10¹⁰ photons per s at 18 keV could be attained using a KB-mirror optic. In the analysis of geological, biological and glass matrices, the LODs were 0.1–100 ppm for elements with Z = 17–83 when the confocal mode was used with an acquisition time of 1 s.

An inventive application of confocal 3D-XRF spectrometry was¹⁹ the retrieval of written information from covered printed papers based on the depth-selective detection of Fe and layer-by-layer reconstruction of the hidden text. A 16.5 μm excitation beam was used with a confocal depth resolution ranging from 47 to 37 μm in the energy range of 5 to 18 keV. Apart from presenting a method which could be of use in archaeology or in forensic science, an especially interesting aspect of this paper was the detailed characterisation of the instrument which used a Mo-anode X-ray tube at 30 kV and 1 mA. The flux density in the focal spot was estimated to be 1.41 × 10⁴ photon per s per μm².

Li et al.²⁰ demonstrated a simple approach to quantitative confocal XRF spectrometry which was based on the measured peak-to-background ratios to determine the elemental composition of CuSO₄ solutions quantitatively. The method was benchmarked by determining the concentration of Cu²⁺ ions in a series of standard solutions with Cu²⁺ concentrations of 0.1 to 0.5% (w/w). The confocal XRF spectrometry spectra were collected using a Mo rotating-anode X-ray generator (25 kV, 30 mA). The linear relationship between the peak-to-background intensity ratio and the Cu²⁺ concentration in the solution was independent of, for example, the thickness of the sample container wall and so could be used for quantitative determination of Cu²⁺ with relative errors of 1.2–3.3%.

One of the key parameters in confocal 3D XRF spectrometry is the actual volume probed as defined by the microbeam focal size and the acceptance of the detector polycapillary lens. A novel experimental assessment of this volume involved²¹ the use of particles that were significantly smaller than the interaction volume and sufficiently isolated to prevent the influence of nearby particles. Spherically shaped uranium particles with a diameter of 1–3 μm, originally produced for calibration in SEM, were used. The acceptance diameters along the three axes were ca. 25, 15 and 15 μm, corresponding to an effective probe volume of ca. 3000 μm³.

A refined procedure for the collection and processing of data from the depth-dependent mapping of oxidation states within solid particles was based²² on confocal μ-XRF/XANES imaging. The procedure was applied to a biochar particle reacted with Cr^VI-spiked water. From an experimental point of view, an interesting aspect of this work was the significantly improved depth resolution (2 μm) obtained when a Ge microchannel array optic was used instead of the usual polycapillary detector lens. This confocal arrangement at the APS beamline 20ID together with the 2 × 2 μm incident X-ray beam yielded a confocal volume of 8 μm³, which was substantially better than previously reported values (ca. 1000 μm³).

Szalóki et al.²³ optimised a versatile table-top confocal-XRF/Raman spectrometer for the elemental and chemical analysis of solid and liquid samples, primarily for daily analysis of toxic and/or radioactive materials in the nuclear industry. For biological samples, a miniature X-ray tube (2–3 W) provided a spot size of 3 mm and LODs of 3–100 ppm for elements with Z = 12–30. The confocal-XRF spectrometry module of this combined instrument was rigorously tested for quantitative elemental analysis based on the use of FPs.

2.2 Laboratory-based 2D XRF techniques

Laboratory-based 2D XRF spectrometry has become a versatile tool in many applications for which a spatial resolution of several μm to mm is needed. In the period covered by this review, however, only a handful of papers presented a substantial component of analytical development. Although a comparison of element contents can usually be achieved quite easily when the matrix composition does not change dramatically, it is more challenging to quantify element amounts in each pixel correctly. Smit and Prokes²⁴ parameterised their table-top μ-XRF spectrometry set-up (Mo tube, 50 KV; polycapillary optic with 15 μm focal spot at 5.9 keV and 76 mm length) for the analysis of hand written pigments. An initial FP monochromatic description of the X-ray-sample interaction was used and modelling was limited to ideal thin and thick samples only. In a second step, the emission spectrum of the tube was described using the Bethe–Heitler formalism in order to characterise the Bremsstrahlung spectrum. A particular emphasis was put on estimating the energy distribution of the electrons in the target. The transmission of the polycapillary optic was modelled taking into account the critical angle of total reflection and the curvatures. The effective thickness of the anode was determined by an iterative experimental approach to produce an exponential factor to account for the absorption of the anode. This factor was used to correct the spectrum obtained from the emission formalisms and the transmission function of the polycapillary optic. In the analysis of pigments, the ratio of element concentrations to that of Si had a bias of 5% when compared to results obtained by PIXE.

Luo et al.²⁵ presented the use of μ-XRF spectrometry for studying the distribution of micro nutrients and Pb in plant seeds. The semi-quantitative approach compared count rates and RGB maps corrected by Compton and Rayleigh scattering and was capable of distinguishing significant differences in the elemental distributions of seeds from different plant species. The 15 μm focal spot size of the custom-made 50 W Rh anode μ-XRF spectrometry set-up was sufficient for resolving different organs of the seeds. In general, Ca and K were located in the seed embryo whereas Cu, Fe, Mn and Zn were concentrated near the radicle and newly developed roots. Most of the Pb was located in the newly developed roots. Use of Pb L₃ edge XANES performed at the 15 U1 and 14 W1 beamlines of the Shanghai Synchrotron Radiation Facility identified the Pb species present as Pb₅(PO₄)₃Cl, Pb₃(PO₄)₂, Pb(Ac)₂, a fulvic acid complex and hexadecyl mercaptide Pb.

The μ-XRF spectrometry technique can be used not only for imaging but also for the screening of microscopic sample deposits. Kim et al.²⁶ evaluated a prototype monochromatic μ-XRF (MMXRF) spectrometry instrument equipped with a doubly curved crystal (DCC) and SDD for the fast estimation of U content in swipes. Results obtained by combining the MMXRF spectrometry data with those obtained by γ-spectrometry overestimated the ²³⁵U enrichment slightly (2–12%). A comprehensive evaluation was presented by McIntosh et al.²⁷ for a powerful MMXRF spectrometry prototype built at the Los Alamos National Laboratory and named the high resolution X-ray (hiRX) instrument. The set-up incorporated a Rh anode tube (50 W), an SDD, a source DCC to provide monochromatic excitation at 20.2 keV and four DCCs on the detection side (one for the detection of U and three for the detection of Pu at their characteristic lines of 13.61 and 14.27 keV, respectively). The Pu concentration could be determined in a volume as little as 7 μL of a U-rich sample (Pu : U mass fraction 1 : 40). The analysis of RMs showed that the deviation from the reference value was <10%.

3. Synchrotron and large scale facilities

The development of SR-induced XRF spectrometry continues with the aim of increasing spatial resolution, sensitivity and throughput when used with other techniques. Typically, scanning μ-XRF spectrometry is combined with scanning micro- and nano-XRD and XAS and various absorption- and phase-contrast imaging methods at hard-X-ray micro- and nano-probe facilities. This section focuses on the most recent technical and methodological developments and highlights a few outstanding applications of SR-induced XRF spectrometry.

During the period covered by this Update, a number of review papers on the current trends in instrumental developments provided comprehensive overviews on applications in diverse scientific fields. In their review on applications of SR-based X-ray techniques for material characterisation, Rahimabadi et al.²⁸ covered X-ray powder diffraction, grazing-incidence XRD, small- and wide-angle X-ray scattering, XAS and various XRF spectrometry imaging techniques at X-ray microprobe and nanoprobe facilities. The combined use of XRF spectrometry with XAS was illustrated through four applications: studies on bone structure influenced by osteoarthritis; investigation of distribution and chemical state of Co and Cr in human tissues originating from implants made of CoCr alloys; analysis of fractionated atmospheric aerosols with respect to their Cr and Mn content and oxidation states; and Co distribution in Co-implanted ZnO nanowires deposited on a Si substrate.

In an overview of recent advances in trace element analysis of environmental samples by X-ray based techniques, Terzano et al.²⁹ summarised the principles of the main synchrotron-based analytical techniques employed to study trace elements in environmental samples. These techniques included XRF microscopy, spatially resolved XAS, XRF tomography and confocal detection. The most recent developments in laboratory- and synchrotron-based instrumentation were considered with particular attention given to X-ray sources, detectors and optics. Applications discussed were the analysis of soils, sediments, waters, wastes, living organisms, geological samples and atmospheric particulates. Consideration of sample preparation was also included.

von der Heyden³⁰ gave a highly detailed review on the principles and applications of various synchrotron-based spectroscopic and imaging techniques for the analysis of ores. Techniques discussed were those commonly applied in the fields of fundamental and applied ore-geology and included XRF spectrometry and XRF mapping, XANES, EXAFS, XPS and XAS-CT. A valuable overview was given of the historical importance of SR-based X-ray techniques in the geosciences. Specific criteria to be considered for selecting beamlines at various SR facilities for ore related research were the techniques available, beam size, beam energy and elemental detection range. Examples were given of applications at various beamlines so this review provided an outstanding source of information for geoscientists planning to perform experiments at SR facilities.

A review of spatially resolved synchrotron XRF spectrometry and XAS techniques for probing trace elements in human tissues included³¹ the principles of these techniques. Special emphasis was placed on research into the distribution and biochemistry of trace elements in human tissues, in particular the important topic of radiation-induced damage in biological samples. Applications included the analysis of tissues from the brain and nervous system, bone, teeth, hair, skin and internal organs. A clear trend in recent publications was the complementary use of XRF spectrometry and XAS with other synchrotron-based methods such as FTIR, Raman spectroscopy, STXM, XMCD, XRD and ptychography.

Hidalgo et al.³² reviewed imaging and mapping techniques widely used to explore the fundamental properties of lead halide perovskites used in perovskite solar cells (PSCs). The fundamental limits of techniques such as electron microscopy, atomic force microscopy, optical spectroscopy and synchrotron-based micro- and nano-XRF spectrometry mapping were presented. Synchrotron-based XRF spectrometry was used to explore the distributions of Cl, Fe and I impurities as well as the defect tolerance of iron contamination in lead halide PSCs. Typical spatial resolution was ca. 50 nm. A very interesting development was the complementary use of nanoXRF spectrometry mapping and XBIC measurements to reveal elemental heterogeneities and their role in the performance of solar cells.

Both nanofocused X-ray and electrical measurements, including XBIC, were used³³ at the NanoMAX Beamline (MAX IV, Sweden) which employs a KB mirror pair to focus the X-rays down to ca. 50 nm. At a beam energy of 15 keV, the scanning system’s mechanical stability was sufficient to collect scanning XRD and XRF spectrometry maps in conjunction with XBIC fly-scan measurements on single InP nanowire devices (200 nm diameter) with acquisition times of only 10–100 ms per pixel.

A key aspect of XRF spectrometry instrumentation for mapping applications at SR sources is the implementation of large-solid-angle, high-count-rate detection systems coupled with high-speed scanning modes. Hafizh et al.³⁴ described a fast SDD X-ray spectrometer called ARDESIA which was developed specifically for synchrotron applications, in particular XRF spectrometry and XAS applications with good energy resolution at count rates of a few Mcps. The system featured a 2 × 2 monolithic array of 5 mm-pitch SDDs cooled with a double Peltier scheme and coupled to a four-channel CUBE charge preamplifier. An excellent count rate capability of 3.17 Mcps at a deadtime of approximately 30% was achieved while maintaining satisfactory energy resolution of 259.8 eV at the energy of the Cu Kα line. The detector was capable of detecting the fluorescence emission of low Z elements down to C (Z = 6, Kα = 277 eV). This system was demonstrated³⁵ for high-throughput fluorescence mode EXAFS applications at the BM-08 “LISA” CRG beamline at ESRF.

The incorporation of a new Rococo 2 XRF spectrometry detector into the cryogenic sample environment at the PETRA III P06 beamline was presented by Rumancev et al.³⁶ The detector had a four sensor-field cloverleaf design optimised for the investigation of planar samples and operating in a backscattering geometry. The large solid angle of 1.1 sr enabled XRF spectrometry measurements at high count rates of up to 10⁶ cps per sensor. The energy resolution of 129 eV (Mn Kα) at 10 000 cps was not affected significantly at the highest count rates.

Other developments in XRF spectrometry detection involved the improvement of energy resolution using efficient crystal-based dispersion. Scordo et al.³⁷ described a high-resolution multielement XRF spectrometry instrument using a Von Hamos spectrometer with a highly annealed pyrolytic graphite crystal. Peak resolutions as low as 3.60 ± 0.05 eV and 4.02 ± 0.08 eV for the individual Cu(Kα_1,2) and Fe(Kα_1,2) line measurements could be achieved. At higher energies, the best resolutions for Mo(Kα_1,2) and Nb(Kβ) were 19.4 and 39.9 eV, respectively.

De Pauw et al.³⁸ demonstrated a high-resolution method for the non-destructive detection of REEs. The characteristic X-rays emitted by the sample were dispersed by a fixed Ge(111) analyser crystal over the active area of a pnCCD detector thereby enabling the high-energy-resolution detection of X-rays differentiated by their corresponding Bragg angles. The energy resolution for the Ti Kα fluorescence line was 12 eV and the LODs were as low as 0.50 ppm for REE L-lines at the PETRA III P06 beamline.

The truly atomic-resolution XRF holography data obtained by Shirakata et al.³⁹ at the Photon Factory BL-6 beam line were possible primarily as a result of the rapid developments in detection methods. Using X-ray beam energies of 9.2–13.2 keV, Cu Kα holograms were collected for CuInSe₂ and CuGaSe₂ compounds with chalcopyrite structures. The local atomic structure could be reconstructed with Å spatial resolution. Some two decades after its first demonstration by Tegze and Faigel,⁴⁰ this powerful 3D element-specific imaging technique is now gaining momentum for exploring atomic structure with the highest spatial resolution possible.

4. Grazing X-ray techniques including TXRF spectrometry

The aim of this section is to highlight contributions that have furthered our knowledge of analytical aspects of the TXRF and G-XRF spectrometry techniques. The majority of the publications included are dedicated to novel instrumental developments, knowledge-based improvements in standardisation and analytical procedures, assessments of figures of merits and innovation in sample preparation.

Publication of ISO standard 20289 for the TXRF spectrometry analysis of water, in particular drinking and ground waters, represented⁴¹ a major advance in the use of the technique. The paper summarised the standard and proposed that the validity of the TXRF spectrometry measurement be related to the threshold mass of the residue. The mass limit for ensuring negligible matrix effects was defined as the upper end of the linear calibration curve.

The pre-concentration of analytes reduces LODs quite significantly not only by increasing the analyte concentration but also by removing matrix components. However, these procedures add to the sample handling and so increase the potential for errors. A procedure⁴² in use for many years at the laboratories of the German Hydrographic agency is capable of achieving sub-ppb LOQs. This procedure involves complexation with sodium dibenzyldithiocarbamate to pre-concentrate metal ions in sea water and removal of the salt matrix. The final elution and drying steps were eliminated by de la Calle et al.⁴³ by absorbing the metal complexes with a pyrrolidine dithiocarbamate ligand directly onto ultra-thin graphene membranes prepared on a quartz reflector (120–160 nm). The LODs of 0.3–0.6 μg L⁻¹ were significantly lower than those (3–12 μg L⁻¹) achievable without pre-concentration. Whereas good recoveries (70–100%) from mineral-, ground- and waste-waters could be achieved, those from sea water were poor (30–40%) due to the incomplete removal of the salt matrix. A three-fold increase in the preconcentration of Hg²⁺ ions from a 25 μg L⁻¹ aqueous solution was achieved⁴⁴ when dithizone-functionalised polymer membranes were used in place of plain PVC membranes. This increase was shown by Hg L₃-edge XANES analysis to be a result of the complexation of the Hg²⁺ ions. It was possible to distinguish between free Hg²⁺ ions adsorbed on a non-functionalised membrane and the Hg²⁺ complexed with the dithizone. The ability to evaluate the capacity of novel materials to adsorb or complex Hg provides the potential for developing this preconcentration approach further.

The interference of the U Lα (13.61 keV) and Rb Kα lines (13.40 keV) can be a major source of error in the determination of U in environmental samples such as drainage waters from the Fukushima Daiichi Nuclear Power Plant site. A significant improvement in LODs (0.042 μg L⁻¹) was achieved⁴⁵ for the determination of U in the presence of Rb by using suspension-assisted SPE. Following adsorption of ca. 100% of U^VI from aqueous solution onto suspended graphene oxide, the particles were filtered and re-suspended in HNO₃ with Ga added as IS. The enrichment factor of U over Rb in the suspension was 150.

A major advantage of TXRF spectrometry for the analysis of suspensions is that there is no need for sample digestion. This is especially advantageous for refractory materials. However, the quality of results is strongly dependent on individual characteristics of the material such as particle size, size distribution and drying procedures. Haberl et al.⁴⁶ studied the influence of different digestion and suspension sample preparation procedures on the precision and accuracy of the TXRF spectrometry analysis of fly ash and so were able to provide sound recommendations on how to minimise the sample preparation procedure. Recoveries of elements present predominantly in the large particle fraction (ca. 60 μm diameter) were poor (40–60%) if the sample was not ground. Minimal sample preparation was, however, needed for elements present in the small particle fractions (4 μm and 500 nm diameters) and so accurate and precise TXRF spectrometry analysis was possible. An accurate and precise determination of Cu, Pb, Rb and Zn concentrations could be obtained by preparing a suspension of the fly ash in water (recoveries of 90–110%). Samples suspended in 1% (w/w) HNO₃ instead of water gave acceptable results for Ca, K, Mn and Sr (recoveries of 70–90%). Grinding was only necessary for the determination of Fe and Ti concentrations. The use of angle scans following digestion of samples showed that the poorer precision and accuracy were due to inhomogeneities in the sample. Angle scans are commonly used in the surface analysis of Si wafers but this approach is less common in environmental studies involving TXRF spectrometry. Such scans have the potential to contribute considerably to a better understanding of the interaction of the standing-wave field of TXRF spectrometry with suspension-derived specimens. Suspension-based sample preparation with subsequent TXRF spectrometry analysis was successfully applied to clay samples by Allegretta et al.⁴⁷ The recoveries were 80–120% for most elements and the within-laboratory precision was <20%. The low sample mass requirement of only 50 mg was significantly less than the 5 g required in, for example, WDXRF spectrometry analysis. As to be expected for these kinds of materials, matrix effects were significant so RSFs needed to be established using aluminosilicate RMs. A different approach to account for matrix effects in geological-like samples, as applied by Bilo et al.⁴⁸ to samples of mining waste, involved the determination of a correction factor using data from the ICP-AES analysis of digested samples as reference. The recovery of elements such as Cd, Pb and Zn were improved from 60–70% to 95–98%. Although the indirect determination of F as CaF₂ in fluxes is not uncommon, the use of TXRF spectrometry analysis as described by Kowalkiewicz and Urbaniak⁴⁹ was very interesting and applicable to the analysis of other suspended samples. The particle size of the CaF₂ obtained after lixiviation of other Ca species was <10 μm. Although there was no significant influence of sonication on the uncertainty of the measurement, the particulate nature of the analyte did increase uncertainty. The poorer analytical precision for elements present as a solid than that for elements in solution was attributed to the pipetting of particles and inhomogeneous drying of the specimen.

The ability of TXRF spectrometry to probe minute amounts of sample material (<1 mg) is of particular advantage in the analysis of nuclear material as the risk of exposure to radiation is reduced. Dhara and Misra⁵⁰ presented a comprehensive review paper on the performance and analytical procedures of TXRF spectrometry in this field. In the analysis of cosmetics, Margui et al.⁵¹ presented a simpler, more rapid and more cost-effective alternative to the commonly used ICP-AES analysis of digested samples. As there was no requirement for sample pretreatment, the analysis time was reduced and the sample size of 0.01–0.02 g was considerably lower than that needed for digestion (1 g). The LODs were 1–10 mg kg⁻¹ and recoveries 90–110% for all elements except Cr (119%) and Pb (130%). The use of TXRF spectrometry is particularly valuable for the analysis of NPs which are often obtained only in low quantities.^52,53 The analysis by XRD, XPS, HR-TEM and TXRF spectrometry of <0.025 mmol of Pt in Pt–Sn NPs dispersed in 4 mL toluene showed⁵⁴ that the NPs had ordered intermetallic phases which changed in composition from the surface to the core. This TXRF spectrometry analysis enabled the bulk composition of the particles to be determined in only 1 μL of the NP solution. This was considerably less than the volume required by the other techniques used for the NP characterisation such as HR-TEM (5 μL), XPS (60–120 μL) and XRD (>250 μL). Allegretta et al.⁵⁵ demonstrated the microanalytical capabilities of TXRF spectrometry by developing a procedure to analyse synthetic caesium lead-halide NPs (<0.35 mmol Pb in 2 mL of hexane) in which 10 μL was pipetted onto preheated quartz carriers.

A poor deconvolution of the characteristic lines can be a major source of error in TXRF spectrometry for which energy dispersive detectors with an energy resolution of ca. 130 eV are usually used. Gonzales et al.⁵⁶ overcame such problems by using the Cd Lα and K Kα lines in a multivariate PLS approach to determine Cd and K in clays. The biases in Cd and K concentrations of 3–12% when using an energy range of 2.98–3.48 keV were lower than those (5–22%) obtained with conventional deconvolution.

In a fundamental parameter approach, Garmay et al.⁵⁷ used the dependence of the Rayleigh and Compton scattering intensity ratios on the low-mass matrix content to determine light-element concentrations in mineral water. Rayleigh scattering intensities were used for the sample mass determination. A detailed description of this procedure was presented and this will enable other researchers to apply this approach in their studies. An FP approach was also used⁵⁸ to quantify contamination on the surface of high purity silicon wafers. Although this procedure did not require the application of an external calibration or the addition of an IS, additional angle-scan measurements were necessary to solve the theoretical equations.

A quite interesting observation on species-dependent intensity changes of the M- and L-lines from Au⁰ and AuCl₃ was reported by Fernandez-Ruiz.⁵⁹ Whereas, the AuCl₃ M-lines were more intense than those of the Au⁰ species, the AuCl₃ L-lines were somewhat less intense. Absorption effects and the instrumentation were excluded as the source of these species-specific intensity differences.

In a comparison of TXRF spectrometry and ICP-AES for the analysis of gold NPs in human cancer cells, the ability to analyse unfiltered samples gave⁶⁰ TXRF spectrometry the advantage of simple sample preparation and improved accuracy. Whereas the recovery for the TXRF spectrometry analysis of unfiltered samples was ca. 99%, that for the ICP-AES analysis of filtered samples was only 27%. The study demonstrated the superior robustness of TXRF spectrometry for the analysis of samples difficult to mineralise completely.

The performance of a prototype GI-XRF spectrometer with an angle resolution of 0.014° ± 0.003° was described by Szwedowski-Rammert et al.⁶¹ The profiles for a well-defined nickel–carbon multilayer component obtained using GI-XRF and GE-XRF spectrometries matched simulated profiles very well. The range of elements detectable by the G-XRF spectrometry approach was greater than that of the commonly used XRR measurements. A Kr contaminant was present predominantly on the surface of the carbon layers. A portable low-power (4 W) GE-XRF spectrometry instrument provided⁶² impressive LODs of 4–300 ng. The instrument was comprised of a low-power X-ray tube (gold anode), a SiPIN detector, a quartz reflector and an X–Y table with a rotary stage for sample positioning. Analysis of NIST SRM1577b (bovine liver) revealed a bias of 2–12% for Cu, Fe and Zn and 22–36% for the lighter elements P, K and S. Kokkoris et al.⁶³ considered SR GI-XRF spectrometry to be faster in general than RBS/EBS for the study of Ar deeply implanted in silicon crystals. The challenge of the analysis lay in the low concentrations of implanted Ar ions which were expected to implant in a relatively broad structure at the limit of their penetration (<120–150 nm from the Si wafer surface). The projected ranges of Ar ions measured by both techniques (maxima at 2388 Å and 2080 Å) agreed well considering the broad in-depth distribution (quasi Gaussian with an SD of 1000 Å) of the implants.

5. Hand-held, mobile and on-line XRF techniques

5.1 Hand-held and mobile XRF techniques

The portable XRF spectrometry technique undoubtedly occupies a prominent place for conducting in situ analysis over a wide range of applications. A comprehensive review considered⁶⁴ the in situ XRF spectrometry analysis technology as used in China and covered the development of XRF spectrometry analysers as well as the improvement in data processing and progress in applications. Another contribution, on the evolution of hand-held XRF spectrometry instruments, showed⁶⁵ that current modern instrumentation is sufficiently sophisticated to allow rapid, on-site decision-making without the need for specialist analytical expertise.

Portable XRF spectrometry systems developed specifically for cultural heritage applications included⁶⁶ a low-cost, lightweight system which could be used to perform 2D elemental mapping of large areas (maximum 35 × 35 cm). A measuring head, composed of an X-ray source (max. 50 kV, max. 200 μA) and a detector (Si-PIN, 192 eV FWHM at the Fe Kα line), was mounted on a 3-axis stage with movement reproducibility better that 0.03 mm. The final elemental map resolution of 1.4 mm was determined experimentally. The system performance was demonstrated using a reference canvas painted with different materials and pigments of known composition and with many geometrical forms. A transportable device for simultaneous XRD/XRF analysis was developed⁶⁷ to perform in situ non-invasive analyses. The curved position-sensitive detector for XRD measurements had a 180 mm radius and a 120° 2θ range with a reflection-mode asymmetric geometry. In addition, a SDD was used for simultaneous XRF spectrometry analyses. This device was configured for powder samples and objects smaller than 200 mm as well as for larger flat objects (e.g. paintings). The design, optical features and characterisation of the system, as well as some examples of application, were presented.

Pessanha et al.⁶⁸ proposed a simple, sustainable and reliable triaxial portable XRF spectrometry method for multi-elemental analysis (Co, Cr, Cu, Fe, Mn, Ni, V and Zn) of mining water samples. The XRF spectrometry system was composed of a molybdenum target X-ray tube (50 kV, 1 mA), a secondary yttrium target disk and a SDD (active area 50 mm², 160 eV FWHM at Fe Kα). The spot size was 17 × 12 mm (ellipsoid). To preconcentrate the sample, 200 μL was deposited onto a commercial filter paper retainer and, after drying, the loaded filter was analysed directly in the field using a 600 s measurement time. Although the LODs (2–18 mg L⁻¹) of the portable triaxial XRF spectrometry system were significantly poorer than those (0.1–0.3 mg L⁻¹) obtained by analysis using TXRF spectrometry, they were acceptable for mining water samples with high metal concentrations. The method was applied to the analysis of various types of mine water samples characterised by a high salt content (14–89 mg L⁻¹). A major advantage of this direct analysis was that the potential precipitation of salts and metals during transportation and storage was avoided.

The development and use of portable TXRF spectrometry systems is a growing area of research. A portable, automated and low-cost TXRF spectrometry system featured⁶⁹ a compact arrangement consisting of a low power (50 W) X-ray tube with a molybdenum target, a compact Si-PIN detector, quartz reflectors, lucite monochromator and micrometric tables driven by step motors controlled through the Arduino programming software. The method was suitable for the quantitative determination of Ca, Cr, Cu, Fe, K, Mn, Ni, P, Rb, S, Se, Ti and Zn concentrations in milk samples. The NIST SRM 1577b (bovine liver) was used for validation prior to the analysis of 15 commercially available bovine milk samples collected in the industrialised region of Sao Paulo (Brazil). The relative error between the measured value and the reference value was below 7% for Ca, Cu, Fe, K, Mn and Zn whereas those for P, Rb and S were below 22%. The LODs for the major elements (K, P and S) were 34–2.8 μg g⁻¹ and those for the minor elements (Ca, Cu, Fe, Mn, Rb and Zn) were 1.08–0.08 μg g⁻¹. A portable TXRF spectrometry system equipped with a molybdenum target X-ray tube and a Si-PIN detector was used⁷⁰ to gain further insight into the relationship between altered metal homeostasis and the development of Alzheimer’s disease. Six structures from the brains of mice induced with Alzheimer’s disease were dissected and digested and, following addition of Ga as IS, an aliquot (10 μL) was pipetted onto a precleaned quartz carrier and dried. The methodology was validated through the analysis of NIST SRM 1577B (bovine liver). The concentrations of Fe, K, P, Rb, S and Zn were measured at very low μg g⁻¹ levels. There were statistical differences in the concentrations in brain regions studied between the Alzheimer’s disease and control groups. A new TXRF spectrometry methodology approach proposed for the determination of U concentrations in drainage water could be used⁷¹ in uranium handling such as at the decommissioning of the Fukushima Daiichi Nuclear Power Plant. The LOD of 1.4 μg L⁻¹ obtained using a benchtop TXRF spectrometer in a total analysis time of 10 minutes corresponded to less than a hundredth of the effluent standard value stipulated under Japanese law. Even when portable TXRF spectrometry equipment was used, a sensitivity of one tenth of the effluent standard value could be obtained if the sample was condensed using a simple evaporator thereby making the detection of U in drainage waters feasible. The total analysis time of 25 minutes included condensation time, drying time and measurement time. It was proposed that the portable TXRF spectrometry system be used to define the correct dilution for safe transport of samples and that the benchtop TXRF spectrometry device be used for more precise analysis.

A fast methodology for the determination of uranium in wounds involved⁷² the XRF spectrometry analysis of contaminated blood collected on filter paper. A small amount (7 μL) of U-containing blood was applied to a 220 μm thick filter paper which was placed on a 3 μm Mylar film and covered with a 50 μm polypropylene adhesive tape to prevent contamination. A desktop XRF spectrometry device (tungsten target X-ray tube, 50 kV, 1 mA) with an incident X-ray beam of 10 mm diameter was used so the X-rays covered the whole filter paper (5.5 mm diameter). Distilled water containing 10 μg g⁻¹ U was also analysed in order to evaluate the influence of interfering elements. The results demonstrated that the U Lα peak could be successfully separated from the Rb Kα and Br Kβ peaks, both elements being present in blood samples. The LOD for the determination of U in blood samples was 0.45 μg g⁻¹ when a measurement time of 600 s was used. In this study, a commercially available hand-held XRF spectrometry device was used⁷³ for the determination of Pt and U. A 100 μm primary nickel filter was the most suitable for performing simultaneous high-sensitivity analyses when using a model of a needle-puncture wound contaminated with Pt and U. For a 60 s measurement time, a minimum detectable radioactivity of 30 and 829 Bq for Pu and U, respectively, was obtained by using the LOD equation and converting the concentration into radioactivity. Optimising the material and thickness of the filter therefore improved the LOD by ca. 10%.

In general, the determination of heavy metals in organic amendments, such as compost, manure, biofertiliser and sludge, is based on the aqua regia-soluble fraction of the sample. In order to be able to use a commercially available portable XRF spectrometry system for such analyses, it was necessary⁷⁴ to use a set of 32 dry and finely ground samples to establish calibrations based on aqua-regia-extractable concentrations. The data processing involved multiple linear equations that corrected the readings obtained for each element using the readings obtained for elements such as Al, Ca, Fe, K, S and Si. The aqua-regia-soluble contents of the elements Cr, Cu, Fe, K, Mn, P, Pb, S, Sr and Zn could be determined quantitatively, as confirmed by a paired t-test of actual and predicted aqua-regia-soluble values. Moreover, the organic matter content could be derived from the XRF spectrometry data using an adjusted equation starting from the reported ‘balance value’ which is defined as the difference between 100% and the sum of all the measured concentrations. Portable XRF spectrometry was employed⁷⁵ in a study to evaluate 74 compost samples from the USA and Canada using random forest regression to predict the compost cation exchange capacity (CEC). The elements with the highest influence on the prediction were Cu and Zn for which the calibration and validation statistics (relative to laboratory-measured CEC) were R² = 0.90, RMSE = 5.41 meq. 100 g⁻¹ (n = 60) and R² = 0.60, RMSE = 8.07 meq. 100 g⁻¹ (n = 12), respectively. Although additional validation is necessary, these first results proved that a single 60–90 s portable XRF spectrometry scan could produce data useful in predicting numerous compost chemical parameters such as heavy metal concentrations, salinity, pH and CEC.

Fleming et al.⁷⁶ continued to investigate the use of human nail clippings for assessing the element status in individuals. The determination of Zn in single nail clipping phantoms (300 s measurement time) was performed using a small diameter (ca. 1 mm) mono-energetic X-ray beam from a portable XRF spectrometry device. The 22 nail clipping phantoms analysed had the same Zn concentration of ca. 40 μg g⁻¹ but different physical characteristics. Three different normalisation approaches, used to account for variation in the amounts of sample interrogated by the X-ray beam, produced similar results and successfully reduced the RSD from 41% to 12–13%. An RSD of 6.2% could be achieved if results were corrected for thickness. Repeatability of measurement from individual clippings was excellent with RSDs of ca. 1%. Further work would be focused on the analysis of real nail clippings.

5.2 On-line XRF techniques

In mineral flotation processing, information on the slurry elemental grade (SEG) is of great importance. Although the simplicity of EDXRF spectrometry makes it suitable for on-line measurements, the significant influence of water in slurries decreases accuracy. A novel on-line detection method was based⁷⁷ on a model for correcting XRF spectrometry intensities. Initially, the minimum slurry velocity was derived using a mineral-particle-settling model in order to verify uniformity of pulp content. Then, under the condition of limited velocity, the effect of water content on the XRF spectrometry intensity was established so that a general model could be set up to convert the elemental slurry EDXRF spectrometry intensity to the equivalent dry specimen intensity. Using this procedure, the RSD of the target’s SEG dropped by more than 60 and 40% for Pb and Zn, respectively. Meanwhile, the average relative error of the SEG was reduced by more than 80% for Pb and 90% for Zn. The accuracy and stability of on-line slurry grade detection were therefore significantly improved.

6. Cultural heritage applications

The XRF spectrometry technique is widely implemented in cultural heritage applications such as the analysis of pottery for provenance studies, of pigments to provide insights into mural painting techniques and of artefacts such as coins, porcelain, icons and jewellery. This review is not intended to provide a comprehensive coverage of all such applications but highlights those studies involving new instrumentation or methodology. Our companion ASU review on advances in the analysis of metals, chemicals and functional materials⁵ provides additional information on cultural heritage applications. The role of analytical chemists in cultural heritage research was highlighted⁷⁸ in an interesting review. Analytical techniques valuable for research on pigments, dyes, binders and coatings and the dating and cleaning of artworks included XRF spectrometry.

MacroXRF spectrometry is nowadays a well-established imaging technique that allows the visualisation of the distribution of chemical elements at and below the surface of paintings. Several contributions were presented during the macroXRF spectrometry workshop in Trieste (Italy) in September 2017.⁷⁹ In a study of 15–17^th century icons from the collection of the National Museum in Krakow, Poland, measurements were performed⁸⁰ with a scanner able to scan large surface areas of up to 80 × 60 cm. The X-ray spots of 600–700 μm were applied. The distances between the object and the measurement head (X-ray tube, 50 kV and 0.6 mA, and detector) ranged between 0.8 and 1.5 cm. The dwell times were 15–85 ms per pixel. In this study, macroXRF spectrometry measurements facilitated the determination of inorganic pigments and revealed the presence of a scarcely legible inscription painted with orpiment. A detailed inspection of spatial elemental distribution maps provided evidence for the presence of three different types of metal foil (silver, gold and gold–silver double foil). In a second contribution,⁸¹ two large-dimension paintings by Bartolomé Esteban Murillo were characterised by the combination of macroXRF spectrometry scanning, XRF spectrometry point analyses and stratigraphic analyses. A portable XRF spectrometer (33.4 kV, 80 μA and 200 s) was used for the point analysis to elucidate the pigments composition. The macroXRF spectrometry equipment was composed of a microfocus X-ray tube (rhodium target, 30 W, 50 kV, 0.6 mA), focused with a polycapillary with a 50 μm spot diameter at 1.5 cm distance, and two SDDs operating simultaneously in a time-list event mode. The system allowed a 110 × 70 cm area to be scanned continuously at the maximum speed of 100 mm s⁻¹. Measurements of the Murillo paintings were performed with a resolution of 1 mm per pixel and with a dwell time of 10 ms. The results showed that a similar palette was used in both paintings and so Murillo’s painting technique could be better understood. That previous conservation treatments had taken place was shown by the presence of pigments (zinc white, Prussian blue, etc.) not contemporary with the artist’s own palette. Rembrandt’s painting Old Woman Praying (1629–1630) is the most valuable and exceptional work of art of the Residenzgalerie Salzburg, Austria. It was painted on a gilded copper plate with dimensions of only 15.5 × 12.2 cm and probably belonged to a series of three small-scale tronies. The painting was characterised⁸² using a macroXRF spectrometry scanner operated at 40 kV, 100 mA and 1 s measurement time for each spot with 1 mm step size. As the size of the painting (155 × 122 mm) exceeded the area (100 × 100 mm) that could be analysed, four scans were required to capture the majority of the work’s surface. The mappings revealed several unexpected observations and indicated a different structure for this painting than that used for The Laughing Man and the Self Portrait. The gilding was applied directly onto the copper plate but there were three areas with no gilding. It seems likely that in these sections, the gold was purposely removed to provide a different, darker, effect.

A portable full-field imaging XRF spectrometer was developed⁸³ for cultural heritage research. Unlike most macroXRF spectrometry instruments, it was not based on a scanned polycapillary optic but instead collected the spectral data over a broad area with both spatial and spectral discrimination provided by a direct-illumination CCD camera coupled to a square-pore micro-channel Plate optic. The latter (overall size 20 × 20 mm, thickness of 1.2 mm) was placed 50 mm from the CCD. Spatially resolved XRF spectrometry data from an area of 13 × 13 mm could be collected. Two X-ray tubes (palladium target, operated at 30 kV, 3 W) illuminated the sample from opposite sides of the camera to limit topographic contrast and illumination gradients. The artwork was placed at a distance of 25 mm from the instrument. This prototype was used to study a painting by Gustave Caillebotte, on-site in the exhibition room of the museum. The data showed the artist’s choice of pigments for the colours and the background and revealed the presence of chemical impurities. Evidence for alteration compounds was found in some areas of the painting. Although the area covered by the instrument was limited in comparison with most scanning macroXRF spectrometry devices currently available, the portability and ease of deployment made this system a practical alternative for targeted studies of artworks in the field.

The macroXRF spectrometry technique is a macroscopic imaging technique originally developed for easel paintings and recently made available to conservators of glass. The first real-life application of macroXRF spectrometry imaging to stained-glass panels was⁸⁴ the study of six panels installed in the Saint Bavo cathedral (Ghent, Belgium) since their creation in 1555–1559 AD. The analysis involved an XRF spectrometry measurement head mounted on a software-controlled X–Y motor stage with a travel range of about 60 × 57 cm. The measurement head was equipped with a 50 W rhodium target X-ray tube and a SDD that was positioned at a 20° angle with respect to the incident X-ray beam. During the measurement, a stable distance of ca. 1 cm was maintained between the edge of the measuring head and the panel. The divergent X-ray beam had an effective spot size of 700 μm. The macroXRF spectrometry results provided readily interpretable information on glass type, colouring and alteration processes. In particular, the chemical imaging technique allowed the surviving original glass panes to be distinguished unambiguously from later additions, thereby ensuring a correct historical understanding. However, the qualitative nature of the technique allowed only panes within the same panel to be differentiated. Quantitative information on the glass composition will require further analysis.

In research studies of historical paintings, a number of micro- and macro-techniques is often applied to collect the relevant information. Vermeer’s painting of The Girl with a Pearl Earring was the object of investigation in two contributions. The ultramarine in the headscarf was examined⁸⁵ by microanalyses using SEM-EDX, FTIR-ATR and SR sulfur K-edge XANES. These microanalyses showed that Vermeer used high quality ultramarine to paint the Girl’s blue headscarf. The entire painting was imaged using multispectral-IR reflectography, macroXRF spectrometry, RIS and digital microscopy to reveal the distribution of materials of the headscarf. The analyses and reconstructions led to the hypothesis that the blue headscarf originally contained a wider range of different blue colour shades: an opaque light blue for the left (lit) zone, a slightly brighter opaque blue for the middle zone and a deep dark blue-green glaze with alternating blue-green glazing brushstrokes for the shadow zone now largely compromised by paint degradation. The second contribution⁸⁶ described the materials and techniques Vermeer used to accomplish the smooth flesh tones and facial features of the Girl. The macroscopic study demonstrated that Vermeer built up the face, beginning with distinct areas of light and dark. He then smoothly blended the final layers to create almost seamless transitions. The soft contour around her face was created by leaving a ‘gap’ between the background and the skin. A further study of paint cross-sections revealed that Vermeer intentionally used different qualities or grades of lead white in the flesh paints. Synchrotron macroXRF spectrometry at two energies was successfully employed⁸⁷ to reveal and identify the landscape underlying Exit From the Theater, a late work by Honore Daumier completed around 1863. Scans were performed at two different energies: 12.9 keV which was below the Pb L₃-edge (13.035 keV) and then at 38.5 keV which was above the Ba K-edge. New details in the As, Cd, Cu, Sb and Sn XRF spectrometry maps revealed a previously concealed mountainscape thereby rendering the landscape recognisable as a view sketched by explorer John Harming Speke during his expedition to the source of the river Nile (1857–59). A comparison of SR- and laboratory-scale macroXRF spectrometry scans illustrated the essential role that tunable SR played in recovering the underlying landscape scene. Defeyt et al.⁸⁸ presented the main results of a study conducted on René Magritte’s La toile de Penelope (1958) which included the feet lying underneath the current picture. In order to obtain further details on the hidden composition and to get a better understanding of the distribution of the pigments, a macroXRF spectrometry scanning of the total canvas (24.6 × 14.8 cm) was conducted. The XRF spectrometry scan was completed in ca. 8 h by using a system equipped with a silver target X-ray tube (40 kV, 120 μA) and a SDD (25 mm², resolution of 130 eV at 5.9 keV). The scanning step and speed were 1 and 3 mm s⁻¹, respectively. The woman’s feet revealed through the IRR and XRR images matched perfectly the feet in the diagram Magritte made of the 1954 variant of L’evidence eternelle.

The size, shape and application techniques of gold leaf in 14–15^th century panel paintings with gold backgrounds attributed to artists working in Florence, Siena and Fabriano were directly visualised⁸⁹ using in situ scanning macroXRF spectrometry with a spot size of 450–1000 μm. The high spatial resolution of the Au Lα (9.712 keV) element distribution maps enabled an accurate and reproducible measurement of the shape and size of individual gold leaves to be made. The results suggested that the dimensions of the gold leaves related to the place of production on the Italian peninsula and the period in which they were created, while the degree of overlap related to the individual hand of a gilder or artist. Pessanha et al.⁹⁰ demonstrated how Monte Carlo simulation of a simple stratigraphy system (gilding over lead white mordant) could be used as a tool for the completely non-invasive determination of gold leaf thickness in artworks. Initially, this methodology was calibrated using model samples of simple stratigraphy, namely pure gold leaves of 1, 2 and 2.5 μm thickness covering a thick lead sheet. Subsequently, the gilding thickness in three panel paintings belonging to the Museum of Christian Art in Old Goa (India) was measured and found to be compatible with known leaf-beating procedures. Nevertheless, the authors recognised that further simulations would be necessary as significant contributions of Ag and/or Cu would change the mass attenuation coefficients and density of the gilding layer.

Three miniature portraits, all dated to the beginning of the 19^th century and all painted in watercolour on ivory, were investigated⁹¹ using macroXRF spectrometry, μ-XRD and SEM-EDS. The macroXRF spectrometry scanner allowed scans of large surface areas (up to 80 × 60 cm) to be made with an adjustable lateral resolution (650 to 50 μm) set by appropriate focusing of the excitation beam with polycapillary optics. Variable conditions were applied to collect the relevant data. A rhodium target X-ray tube (50 kV, 600 μA) and a SDD were utilised. The elemental maps visualised all secondary interventions and obtained analytical information from all layers, including the ivory support. A new approach to the SEM-EDS analysis of miniatures involving the innovative use of water vapour (corresponding to an environment with 10% relative humidity) instead of nitrogen in the specimen chamber was also tested so allowing measurements to be made close to normal atmospheric conditions. In this way it was possible to identify directly Prussian blue from its nitrogen content. By using combinations of these methods, partial overpaintings and retouchings or signatures added later could be localised.

Although the XRF spectrometry scanners now available produce incredibly good results, they are expensive and relatively bulkier than hand-held non-invasive portable instrumentation. Therefore, the development of less expensive and less complex XRF spectrometry scanners and data analysis tools remains a focus of study. A multipurpose XRF spectrometry scanner for in situ analyses used⁹² control/acquisition/analysis software developed from only open-source software. Strong features of the system were the ease of use, high portability, good performance and ultra-low radiation dispersion. The system was conceived and designed as an open system suitable for development of individual components such as multi-detectors or for integration into other analytical techniques. The paper included a detailed description of the configuration and its main features were characterised based on recent cultural heritage applications. Other researchers found⁹³ that much of the information collected by a scanner might be redundant. Based on the concept that uniform portions of the painting could be treated as a unity, a procedure was proposed to combine a high-resolution visible image with compositional measurements carried out on a grid with a much smaller number of points. A computer program, named SmART_scan, was adapted to work with a limited number of representative points selected by the user. The ease of application, the short processing time, the thorough control of the data treatment and the quality of the compositional maps produced made this a valuable alternative approach. Examples of maps relative to mockups, real paintings, cross sections and 3D objects were shown.

Tablecloth drawings by Lagoa Henriques (1923–2009) were investigated⁹⁴ using μ-Raman and μ-XRF spectrometries to inform future preservation. The μ-Raman measurements enabled identification of calcite and anatase as fillers in the paper support and carbon black in the ink and contributed to an assessment of the state of conservation of the drawings. Furthermore, the μ-XRF spectrometry measurements (spot size 25 μm) allowed the clay fillers and sizing materials in the paper support to be identified. In addition, the 2D elemental maps obtained by μ-XRF spectrometry indicated an accumulation of Ca, K, P and S in coffee stains on the tablecloth. Daly et al.⁹⁵ used macroXRF spectrometry scanning, RIS, μ-Raman spectrometry and FORS to examine two drawings by the artist Redon. An XRF spectrometry scanner equipped with a 30 W rhodium polycapillary X-ray tube (50 kV, 600 μA) was used to collect macroXRF spectrometry maps. Each scan was performed with a spot diameter of 120–450 μm, a step size of 150–420 μm and a dwell time of 8–12 ms per step. Six distinct black drawing materials and a yellow pastel were identified in his work Apparition, underscoring the complexity of Redon’s noir drawings. In Head within an Aureole, the artist used only two black drawing materials and three colour pastels (two pink and one blue). These observations provided a glimpse into the transitional period of Redon’s career.

The strength of macroXRF spectrometry scanning to probe materials in subsurface layers not only allows visualisations of hidden paintings to be made but also small-scale painted illustrations and decorative elements found in illuminated manuscripts to be studied at high spatial resolution. Turner et al.⁹⁶ demonstrated that macroXRF spectrometry scanning can visualise underdrawings in materials, such as iron-gall ink, lead point and pigmented inks/paints and so can be used as a useful complement to traditional IR imaging. The scanning XRF spectrometer (rhodium target X-ray tube, 40–50 kV and 500–600 μA; 40 keV spectral range; SDD, 30 mm²) used polycapillary optics to focus the X-rays onto illuminated manuscripts. Preset working distances of 7–20 mm produced nominal spot sizes of 120–530 μm. However, macro-XRF spectrometry could not effectively visualise some underdrawing materials such as chalks and silver point due to the poor excitation by the polycapillary focusing optics of elements with absorption edges at either very low or higher energies. Therefore, the use of complementary techniques will remain essential. The macroXRF spectrometry technique, in combination with FORS, was used by Tiburcio et al.⁹⁷ to study two illuminated manuscripts from the 15^th century Portuguese royal court. The macroXRF spectrometry operating conditions were: rhodium target X-ray tube (40 kV, 20 μA); no filter; 1 mm collimation; a pixel time of 2000 s; a map step of 500 × 500 μm; and a SDD (17 mm²). Interesting findings for two representative illustrations were the use of a dark-brown ochre in a putto carnation (one of the first dark-skin angels’ representation in Portuguese illuminations) and the use of two painting techniques and different blue colours in a single folium. Although the two illustrations were created 20–40 years apart, hand-held EDXRF spectrometry analysis of the writing inks revealed a common, improvised, way of working. Serhrouchni et al.⁹⁸ used a multi-analytical approach to study four illuminated manuscripts (14^th to 19^th centuries) from the Royal Library of Rabat (Morocco). Elemental distribution was obtained using a μ-XRF spectrometer equipped with a rhodium target X-ray tube (50 kV and 200 μA), an SDD (30 mm² sensitive area) and a polycapillary lens with a spot size as small as 25 μm. The maps showed the use of Cu, Fe and vermillion in blue, black and red inks, respectively. Arsenic and Pb were identified in orange inks in the 17^th and 19^th century manuscripts, respectively. Quantitative characterisation of the paper supports obtained by triaxial geometry EDXRF spectrometry revealed high levels of Cl (2900–6400 μg g⁻¹), K (1200–3400 μg g⁻¹) and S (1100–1700 μg g⁻¹).

The analysis of the different glasses used in émail champlevé provided⁹⁹ new insights into the mixing of glass for the preparation of enamel powder by high-medieval goldsmiths. Colour measurements demonstrated that all production centres used glass of very similar hues but with large differences in colour saturation. The elemental μ-XRF spectrometry distribution (50 kV, 600 μA, 25 μm spot size) was measured in fly-scan mode in which the object moved continuously with respect to the measuring head. The complete spectrum was acquired from each data point and a hyperspectral map of the areas recorded. The spot distance of 20 μm resulted in a measuring time of 0.37 ms per pixel. The significant variations found in contents of Co, Mn, Pb and Sb oxides suggested that more than one glass material was used to prepare the powder for enamelling. A portable EDXRF spectrometry system was used¹⁰⁰ to identify the chemical composition of the colouring agents present in the overglaze-blue enamel decoration on Arita porcelains (Japan) in order to determine their geographical area of origin. The measurement parameters were: 35 kV and 80 μA; acquisition time of 100 s; no filter; and a distance of 1 cm between sample and detector. The analysed area was about 1 mm². The analysis depth was typically 40 μm for a lead-rich glaze and 500 μm for a lime-alkali glaze. The study revealed that Western technology exerted a continuous influence on Japanese art. In particular, newly developed synthetic pigments, believed to have been used solely on paintings, were also employed for enamel decoration in Arita kilns. Moreover, a chromium-based chromophore was detected that was employed for yellow enamelling on a porcelain lidded-cup decorated with a Western-inspired design and fired in Arita kilns in the 19^th century.

Spectral multiband imaging, paired with in situ FORS and in situ portable XRF spectrometry analysis, gave¹⁰¹ important insights into the painting techniques at the monastery church of Müstair (Switzerland) and at the church of St. Benedict in Malles (Italy). Clear similarities in the palette of pigments appeared which included mainly materials typically used in medieval mural paintings. Of particular interest was the use of Egyptian blue and ultramarine blue that made these paintings among the first in which the precious lapis lazuli pigment had been used in Europe. The surprising presence of As indicates the possible use of orpiment for the creation of the wall paintings. A multi-analytical approach was adopted¹⁰² for the characterisation of a sacred statuette replica of the Nossa Senhora da Conceicao Aparecida, a Brazilian saint. The portable XRF spectrometry measurements (molybdenum target X-ray tube, 45 kV and 600 μA; SDD; 200 s measurement time) were made at 66 points distributed throughout the piece in the wood structure, in the ground layer and on the pigment layer. In addition, elemental mapping was performed using the μ-XRF spectrometry technique on a small fragment containing both the pigment and the ground layer. The μ-XRF spectrometry system was composed of a rhodium target X-ray tube (45 kV, 600 μA), polycapillary optics (25 μm spot size) and an SDD. The sample mapping was performed using a 12.5 μm Al filter, 50 μm pixel size and a measurement time of 30 ms per pixel under vacuum (20 mbar). The major elements in the pigment (Fe and Mn) indicated the probable use of Burnt Sienna pigment. The XRD and Raman analyses showed that the substrate layer was made of calcium carbonate (calcite) and a binder material (probably rabbit glue). Molari and Appoloni¹⁰³ characterised the elemental composition of pigments used in The Portrait of a Young Gentleman, a wood panel (64 × 42.5 cm) painted by Lucas Cranach the Elder in 1539. A total of 40 spectra was collected using a portable XRF spectrometry system (silver target X-ray tube, 3 mm silver collimator) working at 28 kV and 1 μA and with an acquisition time of 300 s. In addition, some nanometric pigment samples were collected during the restoration process for analysis in the laboratory by TXRF spectrometry (50 kV, 600 μA, acquisition time 300 s). Some of the original materials identified were lead white, chalk, bone black, white vitriol, azurite, vermillion and earth pigments such as red, yellow and brown ochres. Both the materials present in the original painting and those used in subsequent interventions could be characterised.

The use of XRF spectrometry is common practice in the analysis of coins, alloys and steel samples. A large-scale study of Archaic (pre-479 BC) Athenian silver coins from six museum collections around the world was conducted¹⁰⁴ by analysing 788 coins with a benchtop EDXRF spectrometer (rhodium target X-ray tube, 50 kV). A 6 × 4 mm ellipse on both the obverse and reverse was analysed in air for 100 s with a spinner operated at 1 Hz. Element quantification was performed with a FP matrix model supplemented with data from the analysis of matrix-matched CRMs. Principal Component Analysis revealed compositional patterns including at least one Au, Cu, Pb triplet with strong clustering of data. This study provided an important new understanding about Athenian history in the late sixth century BC through information on the exploitation of Lavrion silver and revealed when Lavrion was the metal source for Wappenmunzen and Owl coinage. Five institutions using nine tube-based instruments of seven types participated¹⁰⁵ in an interlaboratory study to analyse copper alloys in heritage materials using the CHARMed PyMca protocol. A set of 12 RMs not used for calibration was analysed following the protocol. The reproducibility achieved in this study was 65–83% better than that obtained in a similar study carried out in 2010. The protocol allowed the consistent reporting of concentrations for 15 elements and was considered to generate sufficiently accurate quantitative results with a well-characterised precision to aid researchers in the interpretation of data. Bottaini et al.¹⁰⁶ presented the results of the elemental analysis of a group of six Islamic oil lamps (10^th to the 13^th centuries AD) found at different sites in Southern Portugal. The handling of EDXRF spectrometry data using a Monte Carlo simulation code named “X-ray Monte Carlo” revealed that the six oil lamps were covered by a structure composed of three different layers (protective layer + corrosion patina + alloy). The bulk metal was a copper-based alloy with a great variability in the concentrations of elements such as As, Fe, Pb, Sn and Zn. This approach was especially suitable for the non-destructive analysis of artefacts with a multi-layered structure because the composition of the original alloys could be estimated and the structure and composition of each layer could be characterised without the need for removing samples or cleaning the artefacts’ surfaces. Pollnitz et al.¹⁰⁷ analysed two signposts from the Auschwitz-Birkenau State Museum (Poland) collection to gain information about the materials used, the object’s genesis and damage effects. The fragile condition of these cultural heritage objects made their preservation difficult and comprehensive studies for paint conservation were lacking. Therefore, inorganic constituents were identified using a commercially available hand-held XRF spectrometer and SEM-EDS and organic materials were analysed by FTIR spectrometry. The inorganic pigments detected were titanium white, zinc white and several mineral extenders. The synthetic organic pigments and binders detected were paint materials newly adopted for use at the time of manufacture.

7. Abbreviations

2D	two dimensional
3D	three dimensional
AD	anno domini
AES	atomic emission spectrometry
APS	Advanced Photon Source
ASU	Atomic Spectrometry Update
ATR	attenuated total reflectance
BC	before Christ
CCD	charge coupled device
CEC	cation exchange capacity
CRM	certified reference material
CT	computed tomography
DCC	doubly curved crystal
EBS	elastic backscattering spectrometry
ED	energy dispersive
EDS	energy dispersive X-ray spectrometry
EDX	energy dispersive X-ray
EDXRF	energy dispersive X-ray fluorescence
ESRF	European Synchrotron Radiation Facility
EXAFS	extended X-ray absorption fine structure
FF	full-field
FORS	fibre optics reflectance spectrometry
FP	fundamental parameter
FTIR	Fourier transform infrared
FWHM	full width at half maximum
GE-XRF	grazing exit X-ray fluorescence
GI-XRF	grazing incidence X-ray fluorescence
G-XRF	grazing X-ray fluorescence
HR	high resolution
HXN	Hard X-ray Nanoprobe
ICP	inductively coupled plasma
IG	intergranular
IR	infrared
IRR	infrared reflectography
IS	internal standard
KB	Kirkpatrick–Baez
LOD	limit of detection
LOQ	limit of quantitation
MMXRF	monochromatic micro-X-ray fluorescence
NIST	National Institute of Standards and Technology
NP	nanoparticle
NSLS	National Synchrotron Light Source
PETRA	Positron-Electron Tandem Ring Accelerator
PIXE	particle-induced X-ray emission
PLS	partial least squares
PSC	perovskite solar cell
PUMA	Photons Utilisés pour les Matériaux Anciens
PVC	poly(vinyl chloride)
RBS	Rutherford backscattering spectrometry
REE	rare earth element
RGB	red green blue
RIS	reflectance imaging spectroscopy
RM	reference material
RSD	relative standard deviation
RSF	relative sensitivity factor
RMSE	root mean square error
SD	standard deviation
SDD	silicon drift detector
SEG	slurry elemental grade
SEM	scanning electron microscopy
SLS	Swiss Light Source
S/N	signal-to-noise ratio
SOLEIL	Source optimisée de lumière d’énergie intermédiaire du LURE
SPE	solid phase extraction
SR	synchrotron radiation
SRM	standard reference material
SSID	solid-state interfacial de-alloying
STXM	scanning transmission X-ray microspectroscopy
TEM	transmission electron microscopy
TXRF	total reflection X-ray fluorescence
TYMV	Turnip Yellow Mosaic Virus
WDXRF	Wavelength dispersive X-ray fluorescence
XANES	X-ray absorption near edge structure
XAS	X-ray absorption spectroscopy
XBIC	X-ray beam induced current
XMCD	X-ray magnetic circular dichroism
XPS	X-ray photoelectron spectroscopy
XRD	X-ray diffraction
XRF	X-ray fluorescence
XRR	X-ray reflectometry
Z	atomic number

8. Conflicts of interest

There are no conflicts to declare.

9. Acknowledgements

Writing of the ASU is funded by JAAS.

10. References

1. J. R. Bacon, O. T. Butler, W. R. L. Cairns, J. M. Cook, C. M. Davidson, O. Cavoura and R. Mertz-Kraus, J. Anal. At. Spectrom., 2020, 35(1), 9–53.
2. A. Taylor, A. Catchpole, M. P. Day, S. Hill, N. Martin and M. Patriarca, J. Anal. At. Spectrom., 2020, 35(3), 426–454.
3. E. H. Evans, J. Pisonero, C. M. M. Smith and R. N. Taylor, J. Anal. At. Spectrom., 2020, 35(5), 830–851.
4. R. Clough, C. F. Harrington, S. J. Hill, Y. Madrid and J. F. Tyson, J. Anal. At. Spectrom., 2019, 34(7), 1306–1350.
5. S. Carter, R. Clough, A. Fisher, B. Gibson, B. Russell and J. Waack, J. Anal. At. Spectrom., 2019, 34(11), 2159–2216.
6. B. De Samber, O. Scharf, G. Buzanich, J. Garrevoet, P. Tack, M. Radtke, H. Riesemeier, U. Reinholz, R. Evens, K. De Schamphelaere, G. Falkenberg, C. Janssen and L. Vincze, J. Anal. At. Spectrom., 2019, 34(10), 2083–2093.
7. S. K. Gill, M. Y. Ge, H. F. Yan, K. Sasaki, Z. X. Liang, H. Isaacs, K. Kisslinger, L. Ecker and Y. S. Chug, J. Electrochem. Soc., 2019, 166(11), C3320–C3325.
8. C. Porfido, I. Allegretta, O. Panzarino, B. Laforce, B. Vekemans, L. Vincze, E. de Lillo, R. Terzano and M. Spagnuolo, Environ. Sci. Technol., 2019, 53(18), 10961–10968.
9. P. G. Martin, M. Louvel, S. Cipiccia, C. P. Jones, D. J. Batey, K. R. Hallam, I. A. X. Yang, Y. Satou, C. Rau, J. F. W. Mosselmans, D. A. Richards and T. B. Scott, Nat. Commun., 2019, 10, 2801,  DOI:10.1038/s41467-019-10937-z.
10. A. Mijovilovich, A. Mishra, D. Bruckner, K. Spiers, E. Andresen, J. Garrevoet, G. Falkenberg and H. Kupper, Spectrochim. Acta, Part B, 2019, 157, 53–62.
11. Y. Yang, F. Fus, A. Pacureanu, J. C. da Silva, W. De Nolf, C. Biot, S. Bohic and P. Cloetens, Anal. Chem., 2019, 91(10), 6549–6554.
12. J. J. Conesa, A. C. Carrasco, V. Rodriguez-Fanjul, Y. Yang, J. L. Carrascosa, P. Cloetens, E. Pereiro and A. M. Pizarro, Angew. Chem., Int. Ed., 2020, 59(3), 1270–1278.
13. C. H. Zhao, K. Kisslinger, X. J. Huang, M. Lu, F. Camino, C. H. Lin, H. F. Yan, E. Nazaretski, Y. Chu, B. Ravel, M. Z. Liu and Y. C. K. Chen-Wiegart, Mater. Horiz., 2019, 6(10), 1991–2002.
14. K. W. Bossers, R. Valadian, S. Zanoni, R. Smeets, N. Friederichs, J. Garrevoet, F. Meirer and B. M. Weckhuysen, J. Am. Chem. Soc., 2020, 142(8), 3691–3695.
15. M. Gambino, M. Vesely, M. Filez, R. Oord, D. F. Sanchez, D. Grolimund, N. Nesterenko, D. Minoux, M. Maquet, F. Meirer and B. M. Weckhuysen, Angew. Chem., Int. Ed., 2020, 59(10), 3922–3927.
16. D. Vernekohl, M. Ahmad, X. J. Dai, W. Zhao, K. Cheng and L. Xing, Med. Phys., 2019, 46(12), 5696–5702.
17. C. Streli, M. Rauwolf, A. Turyanskaya, D. Ingerle and P. Wobrauschek, Appl. Radiat. Isot., 2019, 149, 200–205.
18. P. Tack, B. Bazi, B. Vekemans, T. Okbinoglu, F. Van Maldeghem, S. Goderis, S. Schoder and L. Vincze, J. Synchrotron Radiat., 2019, 26, 2033–2039.
19. X. P. Sun, X. Y. Zhang, X. Y. Wang, Y. B. Wang, Y. F. Li, S. Q. Peng, S. K. Shao, Z. G. Liu, H. Z. Shang and T. X. Sun, X-Ray Spectrom., 2020, 49(2), 267–273.
20. Y. F. Li, X. Y. Zhang, Y. B. Wang, X. P. Sun, Z. G. Liu and T. X. Sun, Radiat. Phys. Chem., 2019, 162, 168–171.
21. P. Poths, E. Chinea-Cano, N. Dzigal, I. G. Darby, J. Osan and R. Padilla-Alvarez, X-Ray Spectrom., 2019, 48(5), 553–560.
22. P. Liu, C. J. Ptacek, D. W. Blowes, Y. Z. Finfrock, M. Steinepreis and F. Budimir, Anal. Chem., 2019, 91(8), 5142–5149.
23. I. Szalóki, T. Pintér, I. Szalóki, G. Radócz and A. Gerényi, J. Anal. At. Spectrom., 2019, 34(8), 1652–1664.
24. Z. Smit and R. Prokes, X-Ray Spectrom., 2019, 48(6), 682–690.
25. L. Q. Luo, Y. T. Shen, Y. Ma, C. F. Fu, J. Liu and Y. Zeng, X-Ray Spectrom., 2019, 48(5), 401–412.
26. S. H. Kim, J. Y. Song, J. Y. Jeon, J. Y. Choi and H. Seo, Appl. Radiat. Isot., 2019, 148, 132–137.
27. K. G. McIntosh, G. J. Havrilla, R. F. Gilmore Jr and M. K. Holland, Spectrochim. Acta, Part B, 2018, 148, 1–7.
28. P. S. Rahimabadi, M. Khodaei and K. R. Koswattage, X-Ray Spectrom., 2020, 49(3), 348–373.
29. R. Terzano, M. A. Denecke, G. Falkenberg, B. Miller, D. Paterson and K. Janssens, Pure Appl. Chem., 2019, 91(6), 1029–1063.
30. B. P. von der Heyden, Ore Geol. Rev., 2020, 117, 103328,  DOI:10.1016/j.oregeorev.2020.103328.
31. M. R. Gherase and D. E. B. Fleming, Crystals, 2020, 10(1), 12,  DOI:10.3390/cryst10010012.
32. J. Hidalgo, A. F. Castro-Mendez and J. P. Correa-Baena, Adv. Energy Mater., 2019, 9(30), 1900444,  DOI:10.1002/aenm.201900444.
33. L. Chayanun, S. Hammarberg, H. Dierks, G. Otnes, A. Bjorling, M. T. Borgstrom and J. Wallentin, Crystals, 2019, 9(8), 432,  DOI:10.3390/cryst9080432.
34. I. Hafizh, G. Bellotti, M. Carminati, G. Utica, M. Gugiatti, A. Balerna, V. Tullio, G. Borghi, A. Picciotto, F. Ficorella, N. Zorzi, A. Capsoni, S. Coelli, L. Bombelli and C. Fiorini, X-Ray Spectrom., 2019, 48(5), 382–386.
35. I. Hafizh, G. Bellotti, M. Carminati, G. Utica, M. Gugiatti, A. Balerna, V. Tullio, G. O. Lepore, G. Borghi, F. Ficorella, A. Picciotto, N. Zorzi, A. Capsoni, S. Coelli, L. Bombelli and C. Fiorini, J. Instrum., 2019, 14, P06027,  DOI:10.1088/1748-0221/14/06/p06027.
36. C. Rumancev, A. Grafenstein, T. Vopel, S. Stuhr, A. R. von Gundlach, T. Senkbeil, A. J. Garrevoet, L. Jolmes, B. Konig, C. Falkenberg, S. Ebbinghaus, W. H. Schroeder and A. Rosenhahn, J. Synchrotron Radiat., 2020, 27, 60–66.
37. A. Scordo, L. Breschi, C. Curceanu, M. Miliucci, F. Sirghi and J. Zmeskal, J. Anal. At. Spectrom., 2020, 35(1), 155–168.
38. E. De Pauw, P. Tack, M. Lindner, A. Ashauer, J. Garrevoet, B. Vekemans, G. Falkenberg, F. E. Brenker and L. Vincze, Anal. Chem., 2020, 92(1), 1106–1113.
39. S. Shirakata, N. Happo and S. Hosokawa, Phys. Status Solidi A, 2019, 216(15), 1800971,  DOI:10.1002/pssa.201800971.
40. M. Tegze and G. Faigel, Nature, 1996, 380, 49–51.
41. L. Borgese, F. Bilo, S. Federici, E. Margui, T. Hase, Y. Y. Huang, B. Beckhoff and L. E. Depero, Surf. Interface Anal., 2020, 52(3), 119–123.
42. P. Freimann and D. Schmidt, Spectrochim. Acta, Part B, 1989, 44(5), 505–510.
43. I. de la Calle, T. Ruibal, I. Lavilla and C. Bendicho, Spectrochim. Acta, Part B, 2019, 152, 14–24.
44. N. Kallithrakas-Kontos and S. Foteinis, Global NEST J., 2019, 21(4), 471–476.
45. H. Takahashi, Y. Izumoto, T. Matsuyama and H. Yoshii, X-Ray Spectrom., 2019, 48(5), 366–374.
46. J. Haberl, S. Fromm and M. Schuster, Spectrochim. Acta, Part B, 2019, 154, 82–90.
47. I. Allegretta, B. Ciasca, M. D. R. Pizzigallo, V. M. T. Lattanzio and R. Terzano, Appl. Clay Sci., 2019, 180, 105201,  DOI:10.1016/j.clay.2019.105201.
48. F. Bilo, L. Borgese, G. Pardini, E. Margu, A. Zacco, R. Dalipi, S. Federici, M. Bettinelli, M. Volante, E. Bontempi and L. E. Depero, J. Anal. At. Spectrom., 2019, 34(5), 930–939.
49. Z. Kowalkiewicz and W. Urbaniak, Spectrochim. Acta, Part B, 2020, 164, 105736,  DOI:10.1016/j.sab.2019.105736.
50. S. Dhara and N. L. Misra, TrAC, Trends Anal. Chem., 2019, 116, 31–43.
51. E. Margui, R. Dalipi, L. Borgese, L. E. Depero and I. Queralt, Anal. Chim. Acta, 2019, 1075, 27–37.
52. M. Meyns, F. Iacono, C. Palencia, J. Geweke, M. D. Coderch, U. E. A. Fittschen, J. M. Gallego, R. Otero, B. H. Juarez and C. Klinke, Chem. Mater., 2014, 26(5), 1813–1821.
53. H. Kloust, C. Schmidtke, A. Feld, T. Schotten, R. Eggers, U. E. A. Fittschen, F. Schulz, E. Poselt, J. Ostermann, N. G. Bastus and H. Weller, Langmuir, 2013, 29(15), 4915–4921.
54. A. J. Erdt, C. Gutsche, U. E. A. Fittschen, H. Borchert, J. Parisi and J. Kolny-Olesiak, CrystEngComm, 2019, 21(21), 3363–3373.
55. I. Allegretta, R. Giannelli, R. Grisorio, G. P. Suranna and R. Terzano, Spectrochim. Acta, Part B, 2020, 164, 105750,  DOI:10.1016/j.sab.2019.105750.
56. G. M. Gonzales, R. D. Castillo and J. Y. Neira, X-Ray Spectrom., 2019, 48(6), 700–707.
57. A. V. Garmay, K. V. Oskolok, O. V. Monogarova and N. V. Alov, Spectrochim. Acta, Part B, 2019, 152, 74–83.
58. I. Szaloki, G. Radocz and A. Gerenyi, Spectrochim. Acta, Part B, 2019, 156, 33–41.
59. R. Fernandez-Ruiz, Spectrochim. Acta, Part B, 2020, 165, 105773,  DOI:10.1016/j.sab.2020.105773.
60. G. Mankovskii and A. Pejovic-Milic, Spectrochim. Acta, Part B, 2020, 164, 105764,  DOI:10.1016/j.sab.2020.105764.
61. V. Szwedowski-Rammert, J. Baumann, C. Schlesiger, U. Waldschlager, A. Gross, B. Kanngiesser and I. Mantouvalou, J. Anal. At. Spectrom., 2019, 34(5), 922–929.
62. R. S. Santos, D. F. Oliveira, J. T. Assis and M. J. Anjos, Radiat. Phys. Chem., 2020, 167, 108245,  DOI:10.1016/j.radphyschem.2019.03.049.
63. M. Kokkoris, E. G. Androulakaki, M. Czyzycki, M. Erich, A. G. Karydas, J. J. Leani, A. Mighori, E. Ntemou, V. Paneta and S. Petrovic, Nucl. Instrum. Methods Phys. Res., Sect. B, 2019, 450, 144–148.
64. L. Q. Ge and F. Li, X-Ray Spectrom., 2020, 49(4), 458–470.
65. Analytical Methods Committee AMCTB No. 89, Anal. Methods, 2019, 11(18), 2498–2501.
66. P. Campos, C. R. Appoloni, M. A. Rizzutto, A. R. Leite, R. F. Assis, H. C. Santos, T. F. Silva, C. L. Rodrigues, M. H. Tabacniks and N. Added, Appl. Radiat. Isot., 2019, 152, 78–85.
67. V. Aguilar, J. L. Ruvalcaba-Sil, L. Bucio and E. M. Rivera-Munoz, Eur. Phys. J. Plus, 2019, 134(6), 286,  DOI:10.1140/epjp/i2019-12652-8.
68. S. Pessanha, E. Margui, M. L. Carvalho and I. Queralt, Spectrochim. Acta, Part B, 2020, 164, 105762,  DOI:10.1016/j.sab.2019.105762.
69. U. B. de Araujo, A. C. M. da Costa, D. F. de Oliveira, E. F. O. de Jesus, M. J. Anjos, E. T. Marsico, C. D. Carneiro, R. D. R. Ribeiro and R. T. Lopes, Radiat. Phys. Chem., 2019, 156, 216–221.
70. D. S. Almeida, M. M. Brigido, M. J. Anjos, S. T. Ferreira, A. S. Souza and R. T. Lopes, X-Ray Spectrom., 2019, 48(5), 452–464.
71. T. Matsuyama, Y. Izumoto, K. Ishii, Y. Sakai and H. Yoshii, Front. Chem., 2019, 7, 00152,  DOI:10.3389/fchem.2019.00152.
72. Y. Izumoto, T. Matsuyama, K. Ishii, Y. Sakai, Y. Oguri and H. Yoshii, X-Ray Spectrom., 2019, 48(5), 438–442.
73. K. Ishii, Y. Izumoto, T. Matsuyama, K. Fukutsu, Y. Sakai, Y. Oguri and H. Yoshii, X-Ray Spectrom., 2019, 48(5), 360–365.
74. R. Lopez-Nunez, F. Ajmal-Poley, J. A. Gonzalez-Perez, M. A. Bello-Lopez and P. Burgos-Domenech, Int. J. Environ. Res. Public Health, 2019, 16(22), 4317,  DOI:10.3390/ijerph16224317.
75. B. Li, S. Chakraborty, M. F. G. Sosa, N. Y. O. Kusi and D. C. Weindorf, Compost Sci. Util., 2018, 26(4), 271–278.
76. D. E. B. Fleming, S. L. Crook and C. T. Evans, Appl. Radiat. Isot., 2019, 145, 170–175.
77. H. Q. Li, Z. Q. Xu, W. D. Wang, Y. B. Liu and S. Zhang, Miner. Eng., 2019, 131, 14–22.
78. M. D. L. de Castro and A. Jurado-Lopez, Talanta, 2019, 205, 120106,  DOI:10.1016/j.talanta.2019.07.001.
79. F. P. Romano and K. Janssens, X-Ray Spectrom., 2019, 48(4), 249–250.
80. M. Walczak, D. Tarsinska-Petruk, M. Plotek, M. Goryl and M. P. Kruk, X-Ray Spectrom., 2019, 48(4), 303–310.
81. A. Kriznar, F. J. Ager, C. Caliri, F. P. Romano, M. A. Respaldiza, M. A. Gomez-Moron, L. Nunez and R. Magdaleno, X-Ray Spectrom., 2019, 48(5), 482–489.
82. K. Uhlir, M. Gironda, L. Bombelli, M. Eder, N. Aresi, G. Groschner and M. Griesser, X-Ray Spectrom., 2019, 48(4), 293–302.
83. P. Walter, P. Sarrazin, M. Gailhanou, D. Herouard, A. Verney and D. Blake, X-Ray Spectrom., 2019, 48(4), 274–281.
84. S. Legrand, G. Van der Snickt, S. Cagno, J. Caen and K. Janssens, J. Cult. Herit., 2019, 40, 163–168.
85. A. van Loon, A. A. Gambardella, V. Gonzalez, M. Cotte, W. De Nolf, K. Keune, E. Leonhardt, S. de Groot, A. N. P. Gaibor and A. Vandivere, Heritage Sci., 2020, 8(1), 25,  DOI:10.1186/s40494-020-00364-5.
86. A. van Loon, A. Vandivere, J. K. Delaney, K. A. Dooley, S. De Meyer, F. Vanmeert, V. Gonzalez, K. Janssens, E. Leonhardt, R. Haswell, S. de Groot, P. D’Imporzano and G. R. Davies, Heritage Sci., 2019, 7(1), 102,  DOI:10.1186/s40494-019-0344-0.
87. L. M. Smieska, J. Twilley, A. R. Woll, M. Schafer and A. M. DeGalan, Microchem. J., 2019, 146, 679–691.
88. C. Defeyt, F. Vandepitte, E. Herens and D. Strivay, Heritage Sci., 2019, 7(1), 96,  DOI:10.1186/s40494-019-0338-y.
89. D. MacLennan, L. Llewellyn, J. K. Delaney, K. A. Dooley, C. S. Patterson, Y. Szafran and K. Trentelman, Heritage Sci., 2019, 7, 25,  DOI:10.1186/s40494-019-0303-9.
90. S. Pessanha, M. Manso, V. Antunes, M. L. Carvalho and J. M. Sampaio, Spectrochim. Acta, Part B, 2019, 156, 1–6.
91. J. Hradilova, D. Hradil, M. Pech, P. Bezdicka, V. Nedela, E. Tihlarikova and P. Targowski, Microchem. J., 2020, 153, 104371,  DOI:10.1016/j.microc.2019.104371.
92. F. Taccetti, L. Castelli, C. Czelusniak, N. Gelli, A. Mazzinghi, L. Palla, C. Ruberto, C. Censori, A. Lo Giudice, A. Re, D. Zafiropulos, F. Arneodo, V. Conicella, A. Di Giovanni, R. Torres, F. Castella, N. Mastrangelo, D. Gallegos, M. Tascon, F. Marte and L. Giuntini, Rendiconti Lincei. Scienze Fisiche e Naturali, 2019, 30(2), 307–322.
93. J. D. Martin-Ramos and G. Chiari, J. Cult. Herit., 2019, 39, 260–269.
94. M. Manso, R. Simao, L. Arruda, M. L. Carvalho and F. A. B. Pereira, Radiat. Phys. Chem., 2020, 167, 108301,  DOI:10.1016/j.radphyschem.2019.04.051.
95. N. S. Daly, M. Sullivan, L. Lee, J. K. Delaney and K. Trentelman, Heritage Sci., 2019, 7, 43,  DOI:10.1186/s40494-019-0286-6.
96. N. K. Turner, C. S. Patterson, D. K. MacLennan and K. Trentelman, X-Ray Spectrom., 2019, 48(4), 251–261.
97. C. Tiburcio, S. Valadas, A. Cardoso, A. Candeias, C. Barreira and C. Miguel, Microchem. J., 2020, 153, 104455,  DOI:10.1016/j.microc.2019.104455.
98. G. I. Serhrouchni, L. Hajji, M. Talbi, A. Lhassani, M. El Kouali, S. Pessanha, M. L. Carvalho and M. Manso, Eur. Phys. J. Plus, 2019, 134(10), 542,  DOI:10.1140/epjp/i2019-12896-2.
99. S. Rohrs, L. Lambacher and R. Tagle, Archaeometry, 2019, 61(2), 374–391.
100. R. Montanari, N. Murakami, M. F. Alberghina, C. Pelosi and S. Schiavone, J. Cult. Herit., 2019, 37, 94–102.
101. G. Cavallo, M. Aceto, R. Emmenegger, A. T. Keller, R. Lenz, L. Villa, S. Worz and P. Cassitti, Archaeological and Anthropological Sciences, 2020, 12(3), 73,  DOI:10.1007/s12520-020-01024-2.
102. F. Sanches, R. C. Nardes, H. S. G. Filho, R. S. dos Santos, O. M. O. de Araujo, A. S. Machado, T. Calgam, R. Bueno, C. Canellas, E. A. S. Goncalves, J. T. Assis, R. P. de Freitas, D. F. de Oliveira, R. T. Lopes and M. J. dos Anjos, Radiat. Phys. Chem., 2020, 167, 108266,  DOI:10.1016/j.radphyschem.2019.04.016.
103. R. Molari and C. R. Appoloni, Radiat. Phys. Chem., 2019, 165, 108413,  DOI:10.1016/j.radphyschem.2019.108413.
104. G. Davis, D. B. Gore, K. A. Sheedy and F. Albarede, J. Archaeol. Sci., 2020, 114, 105068,  DOI:10.1016/j.jas.2019.105068.
105. A. Heginbotham, D. Bourgarit, J. Day, J. Dorscheid, J. Godla, L. Lee, A. Pappot and D. Robcis, Archaeometry, 2019, 61(6), 1333–1352.
106. C. Bottaini, J. Mirao, A. Candeias, H. Catarino, R. J. Silva and A. Brunetti, Eur. Phys. J. Plus, 2019, 134(7), 365,  DOI:10.1140/epjp/i2019-12894-4.
107. G. Pollnitz, M. Schreiner, W. Vetter, B. Pichler and A. Jastrzebiowski, Heritage Sci., 2019, 7(1), 98,  DOI:10.1186/s40494-019-0339-x.

---

Footnote

† Joint review coordinators.

---