2017 atomic spectrometry update – a review of advances in X-ray fluorescence spectrometry and its special applications

Margaret West *a, Andrew T. Ellis b, Christina Streli c, Christine Vanhoof d and Peter Wobrauschek c
a405 Whirlowdale Road, Sheffield S11 9NF, UK. E-mail: margaretwest@blueyonder.co.uk
b8 Burgess Close, Abingdon, OX14 3JT, UK
cVienna University of Technology, Atominstitut, Stadionallee 2, 1020, Vienna, Austria
dFlemish Institute for Technological Research (VITO), Boeretang 200, 2400 Mol, Belgium

Received 30th June 2017

First published on 21st July 2017


Abstract

This review describes advances in the XRF group of techniques published approximately between April 2016 and March 2017. With technique maturity, it is the sample rather than the instrumentation that limits its wider development, so regular readers are advised to consult our companion application updates for advances in XRF related to specific sample types. In this reconfigured update the Instrumentation section offers advances in hardware and software developments for off-site and laboratory investigations including synchrotron, TXRF and related techniques. However, publications that reflect increasing interest in the non-destructive advantages of X-ray systems for cultural heritage applications remain within this update.


1 Introduction

Over the years, this update has benefited from the insight of the authors in many different areas of XRF activity to offer our readers a timely, comprehensive and critical review of developments and applications. During the last 30 years laboratory systems, particularly WDXRF and EDXRF instrumentation have fully matured with sufficient capability for their intended use. In the past 10 years hand-held and μ-XRF spectrometry in particular have also reached the point where there are many very good commercially available systems. The sample rather than the instrument or technique is what limits its wider deployment. With technique maturity, the applications section in this review has overlaps with our companion application ASU reviews. It is therefore time for change, hence the applications section has been radically reduced. This review describes advances in the XRF group of techniques published approximately between April 2016 and March 2017 and continues the series of Atomic Spectrometry Updates in X-ray Fluorescence Spectrometry1 that should be read in conjunction with other related updates in the series.2–6 Readers interested in environmental applications and XRF work concerning geological and climate change should consult Butler et al.,6 the update on advances in environmental analysis. Applications related to XRF techniques dealing with clinical, biological and drugs are to be found in Taylor et al.2 whereas updates on chemical state and speciation are considered by Clough et al.4 Likewise, XRF related work on industrial materials and minerals, thin films, coatings and nano-materials are the province of Carter et al.5

Philip Potts has dedicated many years to offer his talents and insight into developments in hand-held, mobile, on-line XRF techniques and planetary exploration together with applications in geological, climate change, mining activities, archaeological and cultural heritage. Philip has decided to retire from the writing team; his expertise will be sorely missed and our thanks go to him for his commitment to X-ray fluorescence spectrometry.

Detection limits for portable X-ray spectrometers continue to improve giving rise to novel applications covering a broad range of matrices. Hand-held XRF technology is now related to the development of integrated sample preparation tools for the benefit of planetary science.

Researchers working at SR facilities continue to discover and increasingly recognise the benefit of X-ray analysis as an analytical tool, since comprehensive characterisation by XRF spectrometry in combination with XAS and XRD or tomography can be performed. Owing to its extreme sensitivity, quantitative mapping of elemental distributions via XRF microscopy (XFM) has become a key micro- and increasingly, a nano-analytical technique. Within the review period a large number of SRXRF applications were in the fields of medicine and biology, mainly due to the fact that the micro beamlines combine XRF element mapping (imaging) with XANES spectroscopy for speciation, so that not only 2D or 3D element images can be obtained, but also information on the chemical state of the respective element can also be established within one measurement. Over the past decade, advances in SRXRF instrumentation presented new opportunities for the 2D or 3D mapping of trace element distributions within intact specimens as the spatial resolution has improved from μm to nm.

The contributions reported in this review period consider not only fundamental TXRF spectrometric development and research, but also instrumentation and special applications in various fields of science, which evidently could not be studied without using TXRF and related techniques. The two fundamental techniques considered in this review are TXRF and GI-XRF spectrometries. Analysing a sample, intentionally placed on the reflector substrate, at a fixed incident X-ray angle below the critical angle of total reflection leads to the classical TXRF effect. Whereas the GI-XRF technique enables the user to characterise surface layers or implants with a variable grazing X-ray incidence angle around the critical angle and leads to a wide new field of application. The quantification procedure for chemical analysis by TXRF spectrometry is relatively simple, based on the addition of an internal standard to the sample and using pre-established relative elemental sensitivities for the spectrometer. In the case of GI-XRF a rather complicated mathematical formalism is involved for the quantitative surface analysis to determine important parameters: thickness, element or composition and density of layers as well as implantation dose and implantation profile.

In the case of detectors, the only real advances were in large area arrays for larger scale systems (high energy physics and synchrotrons). The energy resolution of existing silicon drift detectors (SDD) is now close to theoretical best so interest during this review period concerned array configurations.

Publications this year continue to reflect increasing interest in the non-destructive advantages of X-ray systems for cultural heritage applications.

2 Instrumentation

2.1 Hand-held, mobile, on-line XRF techniques and planetary exploration

Although detection limits are much higher for the portable XRF technique compared to other laboratory – based methods, its portability, ease of use and high throughput rate make it a valuable tool especially for field-based studies. With the purpose of improving detection limits for portable XRF measurements a new portable XRF spectrometer based on the use of double curved crystal optics was developed and evaluated by Guimaraes et al.,7 for measuring toxic elements in consumer goods and cultural products. Two models of the developed system (a pre-production and a final production unit) were investigated. The performance parameters including accuracy, precision and detection limits were characterised in a laboratory setting using CRMs and standard solutions. The LODs, based on solid matrices, were 1 μg g−1 for As and Pb, 1–2 μg g−1 for Hg, and 9–11 μg g−1 dry matter for Cd. Unfortunately, the measurement time was not included. The bias for these elements ranged from −10% to 11% for the pre-production, and −14% to 16% for the final production model. Five archived public health samples including herbal medicine products, ethnic spices and cosmetic products showed a good agreement between both instruments for the four key elements, and data were confirmed by ICP-OES after digestion. In a Chinese language contribution,8 the authors recognised the need to develop a portable analysis technique to monitor water quality. Based on adsorption with Purolite S930 chelating resin at pH 4 and thin-film sample preparation technique, a rapid and simple on-site analysis method was implemented for the determination of Cu, Fe, Mn, Ni, Pb and Zn in aqueous solutions utilising a hand-held EDXRF spectrometer. The LODs were between 6 and 19 μg L−1, precision tests carried out on multi-element mixed solutions showed that the RSDs (n = 10) were better than 15%. Comparative analysis with ICP-MS of real water samples confirmed the validity of the XRF method. Both the pre-concentration device and the XRF instrument were small, light, portable and could operate without external power supply.

This years’ review reflects a number of novel applications covering a broad range of different matrices. The feasibility of measuring As and Se contents in a single nail clipping was investigated by Fleming et al.,9 using a small-focus portable XRF instrument. The unit incorporated doubly curved crystal optics, providing a high intensity beam delivery to a spot size of only 1 mm diameter. The other key distinguishing feature was the ability to provide multiple monochromatic X-ray beams. In this particular application, the nearly mono-energetic beam of 17.4 keV (molybdenum Kα) was a very good match for the excitation of both As (K-edge of 11.9 keV) and Se (K-edge of 12.7 keV). Nail clipping phantom samples supplemented with As and Se to produce materials with 0, 5, 10, 15 and 20 μg g−1 were used for calibration purposes. The calculated LODs, when considering the Kα peak only, ranged from 0.210 μg g−1 Se under one measurement position to 0.777 μg g−1 Se under another. Compared with previous portable XRF nail clipping studies, the LODs were substantially improved for both elements. The new measurement technique had the additional benefits of being short in duration (∼3 min) and requiring only a single nail clipping. A field-portable XRF spectrometer configured in a low density plastics mode with thickness correction was used by Bull et al.,10 for direct measurement of trace elements in samples of dried marine macroalgae (Fucus serratus, Palmaria palmata and Ulva Iactuca). Detection limits for a 200 s counting time ranged from less than 5 μg g−1 for As and Pb in F. serratus and As in P. palmata to several tens of μg g−1 for Cd, Sb and Sn in all species tested. Independent measurements by ICP-MS following nitric acid digestion, revealed a direct and significant proportionality with XRF data, with slopes for the XRF-ICP relationships of respectively 1.0 for As, 2.3 for Cu, 2.4 for Pb and 1.7 for Zn that might be used to calibrate the instrument for direct measurements. Turner and Solman11 also used a hand-held XRF spectrometer in plastic mode configuration and together with a thickness correction algorithm, to analyse the elemental composition of marine litter. Accuracy, evaluated by analysing two reference polyethylene discs, was better than 15% for all elements that had been artificially impregnated into the polymer. Regarding the litter samples, LODs for a 120 s counting time, although element and material dependent, were generally lowest for plastics and painted items with concentrations of less than 10 μg g−1 for As, Bi, Br, Cr, Hg, Ni, Pb, Se and Zn. The quality of the XRF results significantly improved when applying thickness correction for certain elements (Ba, Cl, Cr, Cu, Fe, Sb, Ti, Zn) in all matrices tested. Comparative analysis with ICP spectrometry resulted in an overall slope XRF/ICP of 0.85. Guzzonato et al.12 improved the accuracy of a hand-held XRF spectrometer used to monitor brominated flame retardant (BFR) in waste polymers. Customised standard samples of specific BFRs in a styrenic polymer were used to perform an external calibration ranging from 0.08 to 12% m m−1 of Br, and cross-checking with LA-ICP-MS having similar LODs (0.0004% m m−1 for LA-ICP-MS and 0.0011% m m−1 for XRF). The authors developed a thickness-corrected calibration for the hand-held XRF instrument and the resulting correction was applied to 28 real samples and showed excellent accordance with measurements obtained via LA-ICP-MS (R2 = 0.9926). Interestingly, the authors proposed that expressing limit values for BFRs in waste materials in terms of Br rather than BFR concentration, would present a practical solution to the need for an accurate, yet rapid and inexpensive technique capable of monitoring compliance with limit values in situ. Non-destructive measurements of Ca and K in apple and pear samples using a hand-held XRF instrument were performed by Kalcsits,13 who showed that in situ hand-held XRF spectrometry might be a viable alternative to compliment traditional lab elemental analysis. The authors observed significant correlations between hand-held XRF measurements and concentrations determined using microwave plasma-AES lab analysis. Pearson correlation coefficients ranged from 0.73 to 0.97. Furthermore, the XRF measurements identified spatial variability in Ca and P concentrations on the surface of individual fruit, which might contribute to the development of localised nutritional imbalances. The benefits of portable XRF spectrometry to assist with skeletal analyses were demonstrated by Byrnes and Bush,14 who recognised that knowledge related to the analysis parameters such as X-ray penetration and exit depth were crucial. Analysis depth was determined by examining pure elements through known thicknesses of equine bone slices. The analysis depth for Sr in bone was determined to be 1.9 mm. The device was validated by the analysis of bone surfaces from a small unknown historic cemetery. In an excellent and well-documented publication, Rouillon and Taylor15 re-confirmed the widely accepted analytical capabilities of a field portable XRF spectrometer for the measurement of contaminated soil samples using a matrix-matched calibration. The calibrated instrument generated exceptional data quality from the measurement of 10 soil RMs. A matrix-matched calibration was evaluated against factory settings, showing an improvement in the element recoveries for all 11 elements when applying a matrix-matched calibration, with reduced measurement variation and detection limits in most cases. Measurement repeatability of reference values ranged between 0.2 and 10% RSD, while the majority (82%) of reference recoveries were between 90 and 110%. Measurement comparability with ICP-AES values was excellent for most elements. Parallel measurement of RMs revealed that both ICP-AES and ICP-MS measured Ti and Cr poorly when compared with the portable XRF technique. In the context of the recovery of Dy from end-of-life neodymium–iron boron magnets, Imashuku et al.16 demonstrated the use of a portable TXRF spectrometer in order to perform rapid on-site elemental analysis of Dy in these magnets. Especially the small sample volume required and its low detection limits (ppb level) made the TXRF technique very attractive. At first, the magnet was dissolved in hydrochloride acid, followed by an extraction of the Fe using 4-methyl-2-pentanone in order to avoid the Fe Kα spectral line overlap with the Dy Lα line. Yttrium oxide and a diluted standard solution of rubidium were added to the solution as internal standards. The Dy composition in the NdFeB magnet was determined from the measured intensities of Dy Lα, Y Kα and Rb Kα lines, the relative sensitivities of Dy and Y to Rb for the portable TXRF spectrometer, and the weights of the dissolved NdFeB magnet and yttrium oxide. The calculated Dy composition was in good agreement with that obtained by conventional ICP-AES and the authors were able to detect less than 1% m m−1 of Dy in a NdFeB magnet.

The research group of Weindorf focused their research efforts on the development of applications for new technologies in field soil using proximal sensors including portable XRF spectrometry. In a first publication17 two proximal sensors; portable XRF spectrometry and visible near infrared diffuse reflectance spectroscopy (VIS-NIR DRS) were used to predict calcium carbonate equivalent, gypsum content, salinity as electrical conductivity 1[thin space (1/6-em)]:[thin space (1/6-em)]5 (soil[thin space (1/6-em)]:[thin space (1/6-em)]water) and Ca content of a diverse set (n = 116) of samples from arid soil in Spain. In a second publication18 both sensors were used for predicting the total carbon and total nitrogen content in organic soil matrices. Using advanced statistical techniques for data evaluation (penalised spline regression, support vector regression and random forest analysis), the outcome of both studies proved that a combined portable XRF spectrometry and VIS-NIR DRS approach was superior to those that used data from a single proximal sensor, enabling the prediction of several properties from two simple, rapid, non-destructive scans. In a third contribution,19 a portable XRF spectrometer was used to determine element abundances and related those to soil pH through predictive multiple linear regressions. Two operational XRF modes, soil mode and geochem mode, were utilised to scan frozen soils in situ and under laboratory conditions, respectively, after soil samples were dried and ground. Results showed that lab scanning produced optimal results with adjusted R2 of 0.88 and 0.87 and root mean squared errors of 0.33 and 0.34 between elemental data and lab-determined pH for soil mode and geochem mode, respectively. Summarily, portable XRF spectrometry could be used for in situ determination of soil pH in arctic environments without the need for sample modification and thawing.

The element composition of diverse planetary bodies is of fundamental interest to planetary science. Until now the mass and volume of focusing X-ray optics for X-ray imaging were too large for resource-limited in situ missions, and therefore near-target X-ray observations of planetary bodies were limited to simple collimator-type X-ray instruments. Hong et al.20 introduced a new Miniature lightweight X-ray Optics (MiXO) using metal–ceramic hybrid X-ray mirrors based on electroformed nickel replication and plasma thermal spray processes. This enabled compact, powerful imaging X-ray telescopes suitable for future planetary missions. With a few example configurations, the superior performance was demonstrated: efficient light-collection power (3× improvement relative to an alternative approach using micro-pore optics), high angular resolution (>10× improvement), high detection sensitivity (>10× improvement) and wide energy band coverage (up to 10–15 keV with multilayer coating). Young et al.21 provided a comprehensive look at the hand-held XRF spectrometer as a valuable tool for both terrestrial and planetary field geology, investigating strengths, weaknesses, and best practices for deploying the instrument. Having demonstrated that sample preparation and presentation have an important effect on the quality of the XRF data, the authors suggested that hand-held XRF technological advances should co-evolve with the development of integrated sample preparation tools – perhaps similar to the Rock Abrasion Tool on the Mars Exploration Rovers. The authors claimed that with these instrument and operational improvements, the use of hand-held, field-friendly geochemical tools could revolutionise terrestrial field geology and could become critical tools for future planetary field explorers.

To conclude, in this years’ review just one contribution related to online XRF analysis was brought to our attention.22 In order to accurately estimate health risk and control the source of pollution caused by cement raw meal, an XRF analyser provided with data acquisition software and a laser rangefinder, was developed to carry out measurements of heavy metals in cement raw meal directly above a conveyor belt. The analyser was mounted on a sled, which could effectively smooth the surface of cement raw meal and reduce the impact of surface roughness during online measurement. The laser rangefinder was mounted over the sled for measuring the distance between cement raw meal sample and the analyser. The LODs for the elements Ca, Cr, Fe, Pb, S and Ti were 246 mg kg−1, 33 mg kg−1, 37 mg kg−1, 47 mg kg−1, 118 mg kg−1, and 44 mg kg−1, respectively. The RSD for the elements mentioned was less than 10.7%. By comparison with results obtained by ICP-MS and a CHNS/O elemental analyser, relative deviation of the online XRF analyser was less than 7.4% for Ca, Fe, S and Ti, and between 1.71% and 12.1% for Cr and Pb.

2.2 Laboratory instruments and excitation sources

Perhaps because of the nowadays wide availability of competent commercially available systems, there was little publication activity during the review period in the field of μ-XRF spectrometry instrumentation although Guerra and colleagues23 presented an interesting study of mercury migration from filling amalgam in teeth using an impressive commercially available μ-XRF system. Their system was equipped with a 50 kV 300 μA rhodium anode microfocus X-ray tube coupled to a polycapillary lens delivering a 25 μm diameter spot onto the sample mounted on an XY stage, all located within a vacuum chamber. The system was able to be fitted with two X-ray sources with the second one typically delivering a 200 μm diameter spot with either an optic or a collimator and having the option of primary beam filters to reduce background and provide lower detection limits. The system was fitted with a state of the art 30 mm2 SDD with an energy resolution <145 eV at 5.9 keV, although that could be replaced with a 60 mm2 SDD or indeed have two separate SDDs operating simultaneously to increase measurement speed or improve sensitivity. Elemental maps and scans were performed using scans of 25 μm steps and 1 ms dwell time and the authors presented interesting maps showing the presence of high levels of Hg in the tooth pulp that had been in contact with the amalgam filling. Some conclusions were inconclusive regarding the migration and levels of amalgam elements but the benefits of having such a versatile μ-EDXRF spectrometer were clear from this study. An update to a previously described full field μ-XRF setup was described by Romano and co-workers24 for use in both benchtop XRF and PIXE configurations. The CCD detector was a commercially available 1024 × 1024 pixel device that was cooled to −85 °C and coupled with a 1[thin space (1/6-em)]:[thin space (1/6-em)]1 straight polycapillary, rather than a pinhole as was used in the original setup. The energy resolution of the detection system was reported to be 155 eV at 5.9 keV and the lateral resolution of the system was 55 μm. The detector was operated in a single photon counting mode and the authors developed algorithms to correct for event sharing and to combine frame data to provide energy dispersive maps by post-processing of the data. The 3 kW X-ray tube was fitted with a shutter that was synchronised to close during the CCD readout cycle, thereby reducing the possibility for multiple events and signal smearing in a frame. The system was able to operate at a frame rate of 15 to 20 frames per second but required a measurement time of 6700 s to produce a map of a honey bee for the elements Ca, Cr, Fe, K, Mn and Zn. The authors also reported the use of the camera system on a PIXE beamline to make element maps of cultural heritage samples.

There was also only a small amount of reported activity available to review on the topic of XRF Computed Tomography (XRF-CT). Unlike the typically medical instruments for XRF-CT, Laforce and co-workers25 built a laboratory scale XRF-CT instrument for use in earth and environmental science applications. The instrument comprised a microfocus X-ray tube directly coupled to a glass monocapillary X-ray optic that enabled a claimed spatial resolution of 20 μm to be achieved. A state of the art SDD was used, with a measured energy resolution of <130 eV at 5.9 keV. The authors fully characterised the system using reference samples then demonstrated the capability to image tomographically the accumulation of Cu in organs of a ecotoxicological model organism. Such high performance laboratory equipment is a welcome addition and is likely to be more readily accessible than the more powerful SRXRF imaging facilities. In a similar vein, Ricketts et al.26 developed a benchtop EDXRF system for quantitative analysis of gold NPs in biological samples. The system used a tungsten target X-ray tube operated at high enough kV to generate Au K series lines that enabled thicker samples to be measured than is the case with Au L series measurements. The spectrum background was reduced by means of a primary beam filter and the X-rays from the sample were collected by an HPGe detector that had high stopping power for Au K lines. Detection limits for gold NP between 0.2 and 0.6 g L−1 were achieved, which the authors claimed were an order of magnitude better than reported from other K-XRF systems. While not performing XRF-CT, this setup was shown to be capable of quantitative gold NP determination and it was much better suited than L-XRF systems, despite a 5 to 9 times lower sensitivity, to measurements of thicker samples due to the greater penetration of K lines. The novel and impressive combination of K-XRF and Compton imaging by means of a Compton camera for the imaging of gold NP in lung tissue was presented by Vernekohl and colleagues27 from Stanford University (Stanford, California). The authors claimed that the use of a cone beam with the Compton camera that had a low Z and a high Z detector for both energy and spatial detection of XRF and X-ray scatter provided better ability to image and a greater Au K line sensitivity than when using pencil beam excitation for XRF-CT alone. The authors described in detail the MC simulations used to check the validity of their approach and provided a highly comprehensive description of their simulations, data processing and experimental setup. The study showed that the Compton camera approach to XRF-CT imaging delivered a spatial resolution for discrete inclusions of around 5 mm in the centre of a human lung and an astounding acquisition time reduction of 45×, compared with conventional pencil beam XRF-CT. This would result in a substantial and very valuable potential reduction in dose for in vivo human measurements. Not surprisingly, based on their excellent results, the authors highly recommended the establishment of an experimental bench-top Compton camera setup to validate their theoretical findings and to demonstrate pre-clinical X-ray molecular imaging in large animals.

An interesting geometry for a laboratory EDXRF spectrometer was described28 in which the incidence angle was set to 90° and the detector angle to almost 0°. This grazing exit setup was used to improve detection limits for the determination of trace elements in soils. The authors reported the new geometry significantly improved accuracy, precision and detection limits over the conventional 45°/45° geometry used in their alternate setup, although absence of a detector collimator in the latter system would have been the cause of a significant and unnecessary increase in spectrum background in that system. Samples were prepared from dried, ground and sieved soils that were pressed into boric acid backing. Although, sadly, no details of spectrometer components, operating parameters, measurement times or data processing were provided, the authors showed detection limits to be improved by a factor of around 5 and accuracy to be transformed to something analytically useful compared with their original EDXRF spectrometer setup. The use of Cartesian geometry for secondary target and polarisation excitation is not new, but was used in conjunction with a sample deposition and drying method to obtain impressive LOD values (200 s) for trace elements in water.29 Unfortunately, no spectrometer operating parameters were included for the system equipped with a 50 kV, 50 W palladium target X-ray tube and an SDD but detection limits (200 s) of around 0.1 μg g−1 were obtained for elements in the Z range 23 to 38 and 0.5 and 2 μg g−1 for K and Mg respectively. An in-house secondary target/polarisation EDXRF system was described30 in which a powerful 100 kV, 600 W X-ray tube was used in conjunction with polarisation targets. The authors optimised the operating parameters for the determination of Se in biological samples and achieved a detection limit of 0.1 μg g−1, which they claimed was an order of magnitude better than previously reported for EDXRF measurements of biological samples.

A MC simulation of excitation source voltage and primary beam filtration for in vivo quantification of a gadolinium marker in tumour tissue was described by Santibanez and co-workers.31 The authors modelled a typical medical X-ray source with a tungsten target and capable of a maximum of 90 kV and 100 mA, used in conjunction with various thickness copper foil primary beam filters. The operating voltage of the X-ray tube was varied between 54.5 and 80.5 kV with the copper foil thickness being varied from 0.78 to 6 mm in order to characterise the output spectrum for its central energy and FWHM value. Details of the simulation were presented and the authors established that an output spectrum centred on 58 keV and with a FWHM of 9.2 keV achieved by use of a 2.9 mm thick copper primary beam filter was optimum for the LOD of Gd in tumours at tissue depths typically expected. Overall, the modelling revealed the optimum settings of kV and copper filter thickness for human studies and improved the detection limits for Gd by a factor of 5.7 or 2.3 times respectively compared to two Am241 radioisotope-excited systems. In addition, the skin doses delivered in the 16 s measurement time achieved with the optimised X-ray tube system were much lower than with radioisotope sources and more than an order of magnitude below the maximum recommended levels for in vivo XRF exposure of humans.

Finally, a pyroelectric X-ray generator was used32 in a tube-excited EDXRF configuration equipped with an SDD having an uninspiring energy resolution of 201 eV at 8.04 keV. The authors reported that the thicknesses of the pyroelectric crystal and of the X-ray tube target were optimised for the energy region 1 to 27 keV, which contained the characteristic X-ray lines of Cu and Ta. Although few real details were available, the potential of highly compact and low power, portable EDXRF systems using such sources remains of interest.

2.3 Synchrotron and large scale facilities

The ongoing trend, during the review period was the combination of various techniques with XRF spectrometry. At the Canadian Light Source, a medium-energy microprobe end-station at the soft X-ray micro-characterisation beamline (SXRMB) was established.33 This new microprobe end-station, optimised for the medium energy range (1.7–5 keV), was recently commissioned and made available to general users. The performance (energy range, beam size and flux) as well as examples in chemical imaging and μ-XAFS for elements such as Ca, P and S in soil and biological samples were highlighted. At the Australian Synchrotron XFM beamline, simultaneous XRF and scanning XRD microscopy was realised allowing quantitative high-resolved visualisation.34 Both step- and fly-scanning modes in simultaneous XRF-XRD could be performed. Combined SR-CT and XRF-CT was performed at the Imaging and Medical beam line also at the Australian Synchrotron.35 A lower resolution, but wider field-of-view was achieved by means of 3D XRF imaging combined with SR-CT and a 2D reconstruction of an iodine map from a phantom was presented. The XRF data were obtained during a CT scan using a single point detector, while transmission data were simultaneously collected using an area detector. A “maximum likelihood expectation-maximisation iterative algorithm” was used to reconstruct the fluorescence map.

Several diseases are correlated with trace element deficiencies or accumulation. Hachmoller et al.36 reported element bioimaging and speciation analysis for the investigation of Wilson’s disease using μ-XRF and XANES spectroscopy at the BAMline at BESSY2, Berlin. A liver biopsy specimen from a Wilson’s disease patient was analysed by means of μ-XRF spectrometry to determine the elemental distribution. Firstly, a bench-top μ-XRF system was used for a coarse scan of the sample. The resulting distribution maps of Cu and Fe enabled the identification of a specific region of interest, which was then analysed by SRXRF spectrometry with a beam size of 4 mm. Due to the higher sensitivity of SRXRF, distribution maps of additional elements such as Zn and Mn were also obtained. The XANES technique was performed to identify the oxidation states of Cu in Wilson’s disease as a mixture of CuI and CuII.

The inherent structural heterogeneity of biological specimens causes a number of problems for analytical techniques to assess the element composition of a sample, especially in the case of quantitative XRF spectrometry. Surowka et al.37 reported novel approaches for correction of the matrix effects in quantitative SRXRF element imaging of human substantia nigra (SN) tissue. The authors showed that differences in density with possible variation in thickness after freeze drying of thin samples could significantly influence the quantitative XRF results. The SN tissue samples of various thicknesses were mounted onto silicon nitride membranes with the main goal being to derive several semi- and fully-quantitative methods to correct for the mass thickness effects. These topographic studies on the dried specimens demonstrated that the drying procedure was accompanied by an around 80% reduction in the thickness of the sample. A scheme to correct the mass thickness effects on XRF intensities from these structures was presented. The increased levels of Fe found in the SN of Parkinson’s disease patients with a potential interlink between molecular changes is not fully understood. Ortega et al.38 hypothesised that Fe content and distribution could be modified in cellulo, in cells over-expressing alpha-synuclein. Using PIXE and nano-SRXRF imaging, the authors were able to quantify and describe the Fe distribution at the subcellular level. They showed that in neurons exposed to excess iron, the mere over-expression of human alpha-synuclein resulted in increased levels of intracellular Fe and in Fe redistribution from the cytoplasm to the perinuclear region within alpha-synuclein-rich inclusions. The authors concluded that their results linked two characteristic molecular features found in Parkinson’s disease, the accumulation of alpha-synuclein and the increased levels of Fe in the SN. Surowka et al.39 combined in situ imaging of structural organisation and element composition of SN neurons in the elderly. The human dopaminergic system in general, and SN neurons in particular, were implicated in the pathologies underlying the human brain aging. The interplay between aberrations in the structural organisation and element composition of SN neuron bodies has recently gained in importance. Selected elements (Ca, Cu, Fe, Zn) were found to trigger oxidative-stress-mediated aberration in their molecular assembly due to concomitant protein (alpha-synuclein, tau-protein) aggregation, gliosis and finally oxidative stress. The authors were able to demonstrate an integrated approach to the analysis of the structural organisation, assembly and metals’ accumulation in two distinct areas of SN: in the neuromelanin neurons and neuropil. By using the highly brilliant SR source at PETRA III and the Kirkpatrick–Baez nano-focus spectrometer, large area histological brain slices were scanned with sub-neuronal spatial resolution, taking advantage of continuous motor movement and reduced acquisition time available with that setup. Element analysis using SRXRF spectrometry was combined with X-ray Phase Contrast Imaging (XPCI) to correct for inherent aberrations in the samples’ density and thickness. Based on the raw SRXRF spectra, the authors observed the accumulation of Ca, Cl, Cu, Fe, K, P, S and Zn predominantly in the SN neurons. However, after mass thickness correction, the distributions of Cl became significantly more uniform. Simultaneously with the XRF signal, the SAXS signal was recorded by a pixel detector positioned in the far-field, enabling fast online computation of the dark field and differential phase contrast. The data demonstrated that the SN neurons and neuropil produced excellent contrast which was due to their differing mass density and scattering strength, indicative of differences in local structure and assembly therein. The results showed that combined SRXRF-XPCI-SAXS experiments could robustly serve as a unique tool for understanding the interplay between the chemical composition and structural organisation that might drive the biochemical age-related processes occurring in the human dopaminergic system. Bille et al.40 investigated XFM artefacts in element maps of topologically complex samples. Despite the known angular dependence of XRF, topological artefacts remain an unresolved issue when using X-ray micro- or nano-probes. The authors investigated the origin of the artefacts in XRF imaging caused by the limited travel distances of low energy XRF emission from the light elements. Human Embryonic Kidney (HEK293T) cells were mapped and exemplary results with biological samples obtained, using a soft X-ray scanning microscope installed at ELETTRA (Trieste, Italy) when testing a mathematical model based on detector response simulations, and for proposing an artefact correction method based on directional derivatives. The authors claimed that despite the peculiar and specific application, the methodology could be easily extended to hard X-rays and to set-ups with multi detector array systems when the dimensions of surface reliefs were within the size of the probing beam.

Bergamaschi et al.41 developed a multi-platform freeware data-processing software (MMX-I) for multimodal X-ray imaging and tomography. The software platform aimed to treat different scanning imaging techniques: XRF, phase, absorption and dark field in any combination, thereby providing an easy-to-use data processing tool for the X-ray imaging user community. A dedicated data input stream that could run even on a standard PC dealt with the input and management of large datasets (several hundred GB) collected during a typical multi-technique fast scan at the Nanoscopium beamline at SOLEIL, France. The authors claimed this was the first software tool aimed at treating all of the modalities of scanning multi-technique imaging and tomography experiments, which seems to be a significant improvement for users.

Bone is a dynamic tissue which exhibits complex patterns of growth as well as continuous internal turnover (i.e. remodeling). Tracking such changes can be challenging and a high resolution imaging-based tracer could provide a powerful new perspective on bone tissue dynamics. Panahifar et al.42 developed a 3D labelling of newly formed bone using SR Ba K-edge subtraction (KES) imaging. The authors reported a “proof-of-principle” for the use of Ba as an alternative tracer with a higher K-edge energy (37.441 keV), albeit for ex vivo imaging at the moment, which enabled application in larger specimens and has the potential to be developed for in vivo imaging of preclinical animal models. New bone formation within growing rats in 2D and 3D was demonstrated at the Biomedical Imaging and Therapy bending magnet beamline of the Canadian Light Source synchrotron. Comparative XRF imaging confirmed those patterns of uptake detected by KES imaging. This initial work provided a platform for the further development of this tracer and its future exploitation for in vivo application. Rauwolf et al.43 reported the comparison of a confocal and a colour X-ray camera setup for μ-SRXRF system for the investigation of thin structures in bone samples. In their quest to find the ideal SR-induced imaging method for the investigation of trace element distributions in human bone samples, experiments were performed using both a scanning confocal μ-SRXRF setup on the FLUO beamline at ANKA Karlsruhe, Germany and a full-field colour X-ray camera setup on the BAMline at BESSY-II (Berlin, Germany). As Zn is a trace element of interest in bone, the setups were optimised for its detection. The setups were compared with respect to count rate, required measurement time and spatial resolution. The authors concluded that the ideal method depended on the element of interest. For Ca, with a low energy of 3.69 keV for its K alpha XRF line, the colour X-ray camera provided a higher resolution, whereas for Zn, only the confocal μ-SRXRF setup was able to sufficiently image the distribution.

Some papers within the review period dealt with improvements in X-ray optics at large scale facilities as demonstrated by Brown et al.44 who experimentally determined the relative efficiency of spherically bent germanium (111) and quartz (1011) crystals. The authors used the EBIT-I electron beam ion trap at the Lawrence Livermore National Laboratory and a duplicate Orion High Resolution X-ray Spectrometer. The L-shell X-ray photons from highly charged molybdenum ions generated in EBIT-I were simultaneously focused and Bragg reflected by each crystal, both housed in a single spectrometer, onto a single CCD X-ray detector. The flux from each crystal was then directly compared. The authors were able to show that the germanium crystal had a reflection efficiency significantly better than the quartz crystal. However, its energy resolution was significantly worse. Moreover, they found that the spatial focusing properties of the germanium crystal were worse than those of the quartz crystal. Gotta et al.45 published their remarks on the use of a Johann spectrometer for exotic-atom research. The general properties of a Johann-type spectrometer equipped with spherically bent crystals were described leading to simple “rules of thumb” for its practical use. The results were verified by comparison with Monte-Carlo studies.

Uncertainties in atomic fundamental parameters such as mass attenuation coefficients cross sections and fluorescence yields, are of decisive importance in element quantification using XRF techniques, particularly high precision SRXRF spectrometry. The Physikalisch-Technische Bundesanstalt (PTB) group at BESSY in Berlin46 investigated the influence of the photoelectric cross section (PCS) on accurate quantification. Two contrary models for the photon energy dependence of the L-subshell PCS present in the literature were discussed. The authors concluded that depending on the excitation conditions, these PCSs and, thus, the derived quantitative results could differ significantly if the wrong PCS model was employed. The same group47 published their experimental determination of the oxygen K-shell fluorescence yield using thin silica and alumina foils by employing the radiometrically calibrated XRF instrumentation of the PTB. Multiple excitation photon energies were applied to record XRF spectra of all four samples. The resulting value, omega (O, K-shell) = 0.00688 ± 0.00036 was found to be almost 20% higher than the commonly used value in the Krause tables. In addition, the achieved total uncertainty budget for the new experimental value was reduced significantly in comparison with available literature data. Also at Indus 2, India, the L-XRF cross-sections for W at excitation energies 12, 14, 15 and 16.5 keV were measured.48 A large area SDD with an energy resolution of 138 eV at 5.96 keV X-rays was employed for the analysis. The experimental results were compared with the theoretical estimates of Krause (1979), Campbell (2003) and Puri (1993) and existing experimental results (Barrea and Bonzi, 2001) with the present results found to be closest to the Puri data.

Recent developments in biological X-ray microscopy have allowed structural information and element distribution to be simultaneously obtained by combining X-ray ptychography and XRF microscopy. Jones et al.49 combined two distinct measurements of ultrastructure and element distribution, with each measurement performed under optimised conditions. The authors were able to determine molar concentrations from 2D images, allowing an investigation into the interactions between the environment-sensing filopodia in fibroblasts and extracellular calcium.

Crawford et al.50 reported their development of an XRF flow cytometer that could determine the total metal content of single cells. Capillary action or pressure was used to load cells into hydrophilic or hydrophobic capillaries, respectively. Once loaded, the cells were transported at a fixed vertical velocity past a focused X-ray beam and XRF spectrometry was then used to determine the content of metal in each cell. By making single-cell measurements, the population heterogeneity for metals in the μM to mM concentration range on femto-litre sample volumes could be directly measured, which is normally difficult using most analytical methods. This approach was used to determine the metal composition of various individual bovine red blood cells (bRBC), 31 individual 3T3 mouse fibroblasts (NIH3T3) and 18 Saccharomyces cerevisiae (yeast) cells with an average measurement frequency of about 4 cells per min. Surprisingly, broad metal distributions were found.

One of the very few papers dealing with quantification using nano-XRF data was published by the group of Laszlo Vincze. The authors reported51 their experiments on using SRXRF spectrometry to probe intracellular element concentration changes during neutrophil extracellular trap formation. High pressure frozen, cryo-substituted microtome sections of 2 μm thickness containing human neutrophils (white blood cells) were analysed using nano-SRXRF spectrometer with a spatial resolution of 50 nm. In order to gain insight into metal transport during this process, precise local determination of elemental content was performed reaching limits of detection of 1 ng g−1. Mean weight fractions within entire neutrophils, their nuclei and cytoplasms were determined for the three main elements Cl, P and S, but also for the 12 following trace elements: Br, Ca, Co, Cu, Fe, K Mn, Ni, Pb, Se, Sr and Zn. The authors emphasised nano SRXRF spectroscopy as an enabling analytical technique to study changing (trace) element concentrations throughout cellular processes.

The distributions of elements within cells are of prime importance in a wide range of basic and applied biochemical research. Kashiv et al.52 performed imaging of trace element distributions in single organelles and subcellular features. An example was the role of the subcellular Zn distribution arising from Zn homeostasis in insulin-producing pancreatic beta cells and the development of type 2 diabetes mellitus. The authors combined TEM with μ-and nano-SRXRF spectrometry to image for the first time the natural element distributions, including those of trace elements, in single organelles and other subcellular features. Detected elements included Ca, Cd, Cl, Co, K, Ni and Zn. Cell samples were prepared by a technique that minimally affected the natural element concentrations and distributions, and did not require optical fluorescent indicators. The authors suggested their method could be applied to all cell types and provide new biochemical insights at the single organelle level not available from organelle population level studies.

Trace metals play important roles in biological function and XFM provides a way to quantitatively image their distribution within cells. The veracity of these measurements is dependent on appropriate sample preparation. Jin et al.53 presented a method to preserve element content in adherent mammalian cells for analysis using SRXRF microscopy. Using mouse embryonic fibroblast NIH/3T3 cells as an example, the authors compared various approaches for the preparation of adherent mammalian cells for XRF imaging under ambient temperature. Direct side-by-side comparison showed that plunge-freezing-based cryoimmobilisation provided more reliable preservation than conventional chemical fixation for most biologically important elements including Cl, Cu, K, P, S, Zn and possibly Ca in adherent mammalian cells. Although cells rinsed with fresh media had a great deal of extracellular background signal for Cl and Ca, the described approach maintained cells at the optimal physiological status before rapid freezing and it did not interfere with the XRF analyses of other elements. If chemical fixation had to be chosen, the combination of 3% paraformaldehyde and 1.5% glutaraldehyde preserved Cu, Fe, S and Zn better than either fixative alone. When chemically fixed cells were subjected to a variety of dehydration processes, air drying was proved to be more suitable than other drying methods such as graded ethanol dehydration and freeze drying. This first detailed comparison for XFM showed how detailed quantitative results could be affected by the choice of cell preparation method.

Hesse et al.54 reported full-field Ca K-edge XANES spectroscopy on cortical bone at the micrometre-scale and observed that polarisation effects revealed mineral orientation. The authors performed XANES spectroscopy in both transmission and XRF full-field modes (FF-XANES) on human bone tissue in both healthy and diseased conditions and for different tissue maturation stages. The authors observed that the dominating spectrum differences originating from different tissue regions, which were well pronounced in the white line and post-edge structures, were associated with polarisation effects. These polarisation effects dominated the spectrum variance and must be well understood and modelled before measuring the very subtle spectrum variations related to the bone tissue variations itself. However, the authors noted that these modulations in the fine structure of the spectra could potentially be of high interest for quantifying orientations of the apatite crystals in highly structured tissue matrices such as bone. Due to the extremely short wavelengths of X-rays, FF-XANES overcomes the limited spatial resolution of other optical and spectroscopic techniques that use visible light. Since the field of view in FF-XANES was rather large, the acquisition times for analysing the same region were short compared with XRD techniques. The results on the angular absorption dependence were verified by both site-matched polarised Raman spectroscopy, which had been shown to be sensitive to the orientation of bone building blocks and by mathematical simulations of the angular absorbance dependence. The data suggested that neither the anatomical site (tibia vs. jaw) nor pathology (healthy vs. necrotic jaw bone tissue) affected the averaged XANES spectrum shape.

2.4 TXRF and related techniques

The TXRF intensity dependence on position of dried residue on sample carriers in the determination of halogens in liquid samples, such as drinking water and environmental water samples, was described by Tabuchi and Tsuji.55 The authors found it was difficult to determine Br and Cl because they were lost as volatile hydrogen halide compounds when using an acidic internal standard solution. A new method was proposed in using a conventional calibration curve method for the determination of halogens but without internal standard, despite the fact that the absence of an internal standard would mean that there was no longer any compensation for the XRF intensity dependency on the relative position to the detector of the dried residue. The position of the droplet of the sample solution was carefully controlled surprisingly simply by using an air blower in order to place the dried residue at the most effective position on the sample carrier, which resulted in a linear calibration curve for Cl. Using a table-top TXRF instrument, the authors achieved a LOD for Cl of 63 ppb (ng mL−1). A sample preparation method for non-aqueous suspensions was proposed by Sharanov et al.56 who used viscous liquids e.g. ethylene glycol and glycerol, as a dispersion phase. The authors reported the sedimentation stability of suspensions in these liquids was higher than in water. The authors claimed that ethylene glycol and glycerol could be efficiently used to prepare samples for TXRF analysis. The usual method to establish the relative sensitivity curve for TXRF methods uses multi-element solutions, which may be purchased or prepared in the laboratory.

Technical advances continue in the area or SR-TXRF spectrometry and Wrobel et al.57 described novel instrumentation including control software based upon a LabVIEW interface with Tango control system for the multi-technique X-ray spectrometry IAEA beamline end-station at ELETTRA Sincrotrone, Trieste, Italy. The key components at the core were an ultra-high vacuum chamber that included a seven-axis motorised manipulator for sample and detector positioning with different types of X-ray detectors available plus optical cameras. The beamline end-station allowed measurements using several related X-ray techniques such as μ-XRF, TXRF, GI-XRF, GE-XRF, XRR and XAS. This format of custom modules provided a powerful, user-friendly tool for control of the entire end-station hardware components. A new setup58 including an automatic sample changer chamber for SR-induced TXRF spectrometry and XANES was commissioned at BESSY II (BAMline, Berlin, Germany). The authors emphasised that these techniques were valuable tools for element determination and speciation, especially where sample amounts were limited (<1 mg) and concentrations were as low as ng mL−1. The newly installed chamber allowed reliable sample positioning, remote sample changing and evacuation of the fluorescence beam path. Results showed accurate determination of element concentrations in the NIST water CRM 1640 with achievable LOD <100 fg (10 pg mL−1 for Ni). Using a powerful combination of SR-TXRF-XANES on different Re species resulted in the discovery of a previously unknown Re7+ species. It is well known that GI-XRF can be successfully used to characterise layered and implanted structures. Due to the downscaling of the process size for semiconductor devices, junction depths as well as layer thicknesses are reduced to a few nanometers, i.e. the length scale where GI-XRF is highly sensitive. Ingerle et al.59 from the Atominstitut Vienna, Austria group together with international cooperation partners developed and published a software package graphical user interface taking all phenomena and making the necessary correction such as beam divergence, detection geometry, roughness of the interface and the fundamental ambiguity problem into account. The obtained curves were correlated to the layer thickness, depth distribution and mass density of the elements in the sample. However, the evaluation of these measurements was ambiguous with regard to the exact distribution function for the implants as well as for the thickness and density of nm-thin layers. In order to overcome this ambiguity, GI-XRF was combined with XRR. This was straightforward, as both techniques used similar measurement procedures and the same fundamental physical principles. The combined analysis removed ambiguities in the determined physical properties of the studied sample and, being a correlative spectroscopic method, also significantly reduced experimental uncertainties of the individual techniques. The paper reported the approach for a correlative data analysis, based on a concurrent calculation and fitting of simultaneously recorded GI-XRF and XRR data. Based on this approach the developed JGIXA (Java Grazing Incidence X-ray Analysis), represented a multi-platform software package equipped with a GUI and offered various optimisation algorithms. The data evaluation approach was benchmarked by characterising metal and metal oxide layers on silicon as well as arsenic implants in silicon. The results of the different optimisation algorithms were compared to test the convergence of the algorithms. Simulations for iron NPs on bulk silicon and on a W/C multilayer were presented, using the assumption of an unaltered X-ray standing wave above the surface. A similar subject was discussed by Tiwari and Das60 presenting a continuation and extension of his earlier work on the development of a software platform CATGIXRF, as a solution to provide non-destructive evaluation of nanostructured materials. Here an interactive GUI for the CATGIXRF program was described. The newly developed GUI interface facilitates determination of microstructural parameters on an angstrom length scale for the nanostructured thin layered materials using SR as well as laboratory X-ray sources. It allowed combined analysis capabilities for both the XRR and GI-XRF data simultaneously, thus enabling a greater sensitivity for the determination of microstructural parameters such as thickness, interface mixing and roughness of a thin film medium with improved accuracies. The utility and various newly added salient features of the GUI-CATGIXRF program were described by providing example calculations as well as analysing experimentally a few thin film structures with different surface–interface properties. The influence of the interface morphology of a deeply buried layer in periodic multilayer structures was studied by Das et al.61 Long-term durability of a thin film device was strongly correlated with the nature of the interface structure associated between different constituent layers. Synthetic periodic multilayer structures were primarily employed as artificial X-ray Bragg reflectors in many applications, and their reflection efficiency was predominantly dictated by the nature of the buried interfaces between the different layers. It was demonstrated that the applicability of the combined analysis approach of the XRR and GI-XRF measurements for the reliable and precise determination of a buried interface structure inside periodic X-ray multilayer structures was perfectly suited. An X-ray standing wave (XSW) field generated under Bragg reflection condition was used to probe the different constituent layers of the W–B4C multilayer structure at 10 keV and 12 keV incident X-ray energies. The results showed that the XSW-assisted XRF measurements were markedly sensitive to the location and interface morphology of a buried layer structure inside a periodic multilayer structure. The cross sectional TEM results obtained on the W–B4C multilayer structure demonstrated the overall reliability and accuracy of the XSW method. The authors claimed that the developed method would also be applicable for non-destructive characterisation of a wide range of thin film-based semiconductor and optical devices. Biomolecules are often organised as functional thin layers in interfacial architectures, the most prominent examples being biological membranes. Biomolecular layers also play important roles in context with biotechnological surfaces, for instance, when they are the result of adsorption processes. For the understanding of many biological or biotechnologically-relevant phenomena, detailed structural insight into the involved biomolecular layers is required. Schneck et al.62 used XSW-XRF spectrometry, (here better named GI-XRF) to localise chemical elements in solid-supported lipid and protein layers with near-Angstrom precision. The technique complemented XRR experiments that merely yielded the layers bulk density profiles. While earlier work mostly focused on relatively heavy elements, typically metal ions, the authors showed that it was also possible to determine the position of the comparatively light elements P and S, which were found in the most abundant classes of biomolecules and were therefore particularly important. The authors claimed that the need for artificial heavy atom labels, the main obstacle to a broader application of high-resolution XSW-XRF spectroscopy in the fields of biology and soft matter, was avoided.

In addition to determining the quantity of impurities on a surface, TXRF spectrometry can reveal information about the vertical distribution of contaminants by using GI-XRF as discussed by Singh et al.63 In this study, two samples were intentionally contaminated with copper in non-oxygenated and oxygenated ultrapure water resulting in impurity profiles that were either automatically dispersed in a thin film or particle-like surface, respectively. The surface concentration of the copper contaminants was analysed using the SR-TXRF facility on the 6-2 beamline at the Stanford Synchrotron Radiation Laboratory (Stanford, California) at an incident beam energy of 11.2 keV. In order to determine if the Cu was deposited as a thin surface layer or as particles on the Si surfaces, the Cu fluorescence signal was measured as a function of incident beam angle from 0.01 to 0.3°. In the sample, where Cu was present as nanoparticles, AFM measurements were also taken to confirm the particle size distribution. The concentration profile of the samples immersed into deoxygenated ultrapure water was calculated using a theoretical concentration profile representative of particles, yielding a mean particle height of 16.1 nm. However, the resulting theoretical profile suggested that a distribution of particle heights existed on the surface. The fit of the angular distribution data was further refined by minimizing the residual error of least-squares fit employing a model with a Gaussian distribution of particle heights about the mean height. The presence of a height distribution was also confirmed with the AFM measurements.

2.5 X-ray detectors

Design details and initial results were reported64,65 for an interesting array of silicon drift detectors (SDDs) destined for use in low energy SRXRF measurements on the TwinMic spectro-microscopy beamline at Elettra SR facility in Trieste, Italy. The 234 mm2 array comprised eight sensor cells, four of which were square and four were half-size triangular cells, fabricated on a 450 μm thickness wafer that was fully depleted and had an entrance window suitable for energies <1 keV. The sensor was combined with a custom ASIC readout system and the measured energy resolution was 158 eV at 5.9 keV; although the authors indicated that the readout system had limited this figure and further optimisation was expected. Initial results were shown of clearly resolved K spectra for F, Mg and Si. The tested detector provided a four-fold increase in count rate over the previous detector system and the intended geometry for the next system was described as four trapezoidal detector elements, each of six SDD sensors, which the authors anticipated would provide an order of magnitude increase in count rate over the previous system. The use of a 3D device simulation tool was reported66 for the design and electric field simulation of a SDD structure in which concentric sensor rings were combined with a so-called spiral biasing adapter. The authors claimed their new design generated an optimum electric field potential that was better than a conventional SDD and that heat dissipation was improved but there were no devices built or tested yet in order to confirm these claims. A SDD design was reported67 in which so-called PureB guard ring structures were incorporated to permit devices to be operated at a higher bias potential of up to 1100 V yet with very low leakage current. Such a high bias voltage would be particularly valuable for thicker devices for increased X-ray stopping power but we must await the fabrication and testing of real devices before effectiveness of the interesting new design and simulation can be established.

During the review year there was continuing interest in pixellated X-ray detectors for high speed imaging applications, especially in High Energy Physics and SR setups. A group based at CHESS (Ithaca, NY, USA) reported68 the experimental verification of their high speed pixel array comprising 256 × 384 pixels bump-bonded to a custom ASIC, showing it to be capable of a frame rate up to 6.5 MHz. This very impressive frame rate could only be maintained for 8 to 12 frames before the full readout was made at 1 kHz. Although this operating mode was well-suited to the high rate, bunched beams often encountered at the latest-generation SR beamlines. An additional feature of this sensor array was the 500 μm thick pixels that gave a useful increase of X-ray stopping power compared with the usual 300 μm thickness. The integration and characterisation of the so-called MAIA pixellated array detector installed on beamline P06 at DESY (Hamburg, Germany) was described by Boesenberg et al.69 The detector was capable of sub-μm spatial resolution and a very impressive count rate of 10 MHz, providing excellent detection limits, leading to impressive results in a range of material science, chemistry and environmental μ-XRF applications. The details were provided70 of a custom ASIC used in the high dynamic range hybrid pixel detector system used at the European Free Electron Laser facility (Hamburg, Germany). In addition to a high count rate of up to 104 cps, the authors reported that the front end ASIC was also able to meet the exacting radiation tolerance requirements of this high power experiment. An interesting paper by Egan and co-workers71 described a novel post-processing algorithm for gain correction and energy calibration of pixellated ED X-ray detectors. The authors’ robust and flexible algorithm made use of correlation optimised warping, which enabled alignment of shifted data sets by linear stretching and compression to match a reference spectrum. The authors were able to correct spectra of Ce K series lines from a raw energy resolution of 2.45 keV (at 34.72 keV) to a value of 1.11 keV, which compared well with the mean value of 1.00 keV for all pixels measured individually. The authors claimed that their algorithm was robust in correcting low count rate spectra and that it had a major benefit of not requiring prior knowledge of peak shapes or quality of data, making it suitable for in-line processing and for many types of system simultaneously collecting X-ray spectrum data. Doering and co-workers72 also used post-processing of signals from pixellated sensor arrays coupled to CMOS readout devices. In this instance the processing was offline but the authors reported they could correct pixel charge sharing and recover the pixel energy resolution but at the expense of reduced quantum efficiency. Although dealing with XRD systems, an interesting review73 compared CCD, CMOS and pixellated semiconductor X-ray detectors in terms of delivered data quality. The authors concluded that the XPAD pixel detector was the best, despite its relatively low efficiency (37%) for Mo Kα X-rays and noted that even this disadvantage would be removed if and when higher stopping power sensor materials such as gallium arsenide were used.

Other candidates for sensor materials with increased X-ray stopping power are cadmium telluride or cadmium zinc telluride that are more frequently used in gamma ray detectors. In that vein, Choi et al.74 described the design and optimisation of an analog shaping pre-filter for a digital pulse processor coupled with a CdTe X-ray detector used in XRF spectrometry. Initial results showed an energy resolution of 4.97 keV at 53 keV, which is unimpressive and concentrates on higher energies than those measured in most XRF systems. Trapped charge polarisation effects are a common problem in CdZnTe X-ray detectors when subjected to higher count rate, which prompted Pekarek and co-workers75 to shine a IR-LED emitting at 1200 nm onto the sensor in order to neutralise the accumulated deep charge in the sensor. The authors reported their method to be effective in removing peak artefacts and deformation that would typically arise from incomplete charge collection when the sensor was polarised due to high input count rate. In a comparison76 of Si PIN and CdTe X-ray detectors the authors highlighted the very well-known differences between these two types of detector used in XRF spectrometry, highlighting the incomplete charge collection problems due to hole trapping/polarisation. The authors were able to improve the energy resolution for the distorted peaks in the spectrum from the CdTe detector by means of a rise time discriminator incorporated into their digital pulse processor – a solution commonly used in higher energy X and gamma ray detection systems.

The five papers on cryogenic X-ray detectors available to this year’s review all arose from the 16th International Workshop on Low Temperature Particle Detectors held in July 2015 in Grenoble, France. Perhaps the most interesting and relevant to XRF spectrometry was a paper77 describing the combination of superconducting tunnel junction (STJ) sensors with a silicon pixel absorber array. The purpose of this new design was to increase X-ray stopping power to 95% in the energy range up to 10 keV. The authors reported a measured energy resolution of 82 eV FWHM at 1.74 keV, which is not dissimilar to an SDD, allowing them to acquire a useful S X-ray absorption edge spectrum from a solution of just 0.1% m m−1 sulfur in a soda lime glass sample. Of the other papers, one used mathematical modelling to establish a proposed expression for correction of self-recombination in STJ detectors,78 while the other three covered transition edge sensor (TES) microcalorimeters; in dark matter searches;79 coupled with TES anti-coincidence detectors for use in future space X-ray telescopes80 and Lamb-shift measurements at storage rings.81 All in all, it is fair to say that cryogenic X-ray detectors have now fully migrated from laboratory X-ray possibilities to the high energy physics margins.

Finally, returning to the more prosaic area of corrosion barriers for laboratory X-ray detector windows, Rowley and co-workers82 fabricated amorphous carbon coatings by low temperature sputtering, which was compatible with the window fabrication process itself, and found them to be mechanically robust enough and to have excellent low energy transmission. We await the evaluation of these coatings in the often challenging X-ray spectrometer environments they are designed for.

2.6 Quantification and data processing

In EDXRF spectrometry, as for most analytical spectrometric techniques, the correction of background is an important part of spectrum processing, which itself is an essential precursor to performing matrix corrections in quantitative XRF analysis. Recognising this, Zhao et al.83 proposed a new background estimation method using an iterative wavelet transform algorithm. Based on results for a simulated spectrum and a measured spectrum, the authors concluded that their method was of great significance for improving accuracy in spectrum processing. A comparison was made84 of SHAPE, Fourier transform and measured background removal to establish the optimum method for background correction of spectra from a tube excited EDXRF system. The measurement of a scattered background spectrum, in which there were no characteristic X-ray peaks, was found to be the best method. However, it should be recognised that such an approach is hardly new and cannot correct for background shelf and tail contributions from the detector itself and may also require a matrix-matched background spectrum in order to be effective. An interesting and detailed study85 of spectra from the EDXRF spectrometer on the Curiosity Mars rover revealed non-linearity issues and satellite peaks that distorted the spectrum and led to increased peak-fit residuals from spectrum processing using the GUPIX code. The effects were most notable for X-ray emission peaks of light elements such as the K lines of Al, Mg, Na and Si. The apparent shift in these low energy lines was regarded as more likely to arise from imperfections in the detector and, in particular, the pulse processor than from the presence of weak satellite peaks. Empirical corrections for the local variation of detector energy calibration were established, by allowing the light element Kα and Kβ peak positions to shift independently of each other, and then incorporated into a modified GUPIX code, which significantly improved the quality of peak fitting and accuracy for K lines of Al, Mg, Na and Si by a few percent. Improvements for higher energy lines were much less and the overall quantitative analysis analytical error was still strongly dominated by the sample heterogeneity of the Martian soil. The authors did, however, recommend the inclusion of corrections for the presence and influence of satellite peaks in accelerator-based PIXE spectra. The shifts in L series line positions and relative intensities were studied86 for a large number of compounds of the REE Gd, Nd, Sm and Tb. Unlike the study above of light element X-ray peaks in the spectra from the Mars lander instrument, the authors ascribed all the peak shifts from literature values to be due to chemical (atomic environment) effects, which is an important reminder of the ability to measure such effects and of compensating for them in spectrum processing algorithms. As if to underline this point, Jabua et al.87 used an extremely high resolution WDXRF spectrometer with Johann geometry to measure peak positions and line widths of the Kα1 and Kα2 emission lines from manganese compounds. Measurements from pure metallic manganese established that the Kα1 and Kα2 peak energies could be measured with an impressive precision of ±8.4 meV, while peak positions in compounds could typically be measured with a precision of 10 to 20 meV.

A novel method for calculating the excitation spectrum needed for fundamental parameter (FP) matrix corrections was described.88 In order to get around the problem of making direct measurements of the very high intensity beam produced by capillary optics in μ-XRF systems the authors used an iterative reverse-FP calculation method using spectra from homogeneous pure element samples. The method had the benefit of not needing any complex model or knowledge of the capillary optic and its effect on the actual X-ray tube output and could be applied generally provided high quality homogeneous pure element samples can be obtained. The quantitative results from an FP correction algorithm using the calculated excitation spectrum were shown to be comparable with alternative approaches and were 5% relative for major components and 10% relative for minor elements.

An interesting study was reported89 in which a chemometrics approach based on Partial Least Squares (PLS) regression was used with EDXRF spectra in an attempt to estimate the maximum soil temperature reached during slash and burn clearance of forest in Brazil. Reference samples were prepared from virgin forest samples, adjacent to the burn sites, that were heated in a muffle at 5 selected temperatures for various times. A PLS model was established using three EDXRF spectra per sample – the whole spectrum 1–40 keV, the “element peak” section 1–10 keV and the scatter section 18–40 keV. The latter, scatter portion, provided the best model, probably as it best reflected the soil–ash ratio, and the resulting model gave an usable correlation between calculated and measured burn temperature. When applied to unknown measurements of two slash and burn field samples for which the maximum burn temperature had been measured using embedded thermocouples, a useful indication of maximum burn temperature could be obtained. This was a limited initial study and shows some promise although each model would require local (matrix specific) reference samples to be generated and measured. A method using non-negative matrix factorisation was described by Alfeld and co-workers90 for the near real-time interpretation of XRF maps obtained from SRXRF imaging of a two-phase titanium alloy. The method was able to correct for absorption effects where one of the phases overlaid another and the method was verified using MC simulations of the layered system. The authors claimed their approach was suitable for several stacks of differing composition and enabled much improved interpretation of maps for challenging samples in which X-ray absorption effects were strong. Similarly, in the case of absorption effects in XRF-CT images, Jiang et al.91 used a Seddon algorithm and three types of reconstruction algorithm to successfully correct for image distortions in a PMMA phantom containing a 0.6 cm diameter gold NP region. The results were verified using MC simulations of the model system.

3 Cultural heritage applications

In a wide range of publications related to non-destructive analysis and characterisation of pigments, minute paint samples, and/or entire paintings the XRF technology is very well implemented and remains an active area. With reference to the art work from the seventeenth to the early twentieth century painters, Janssens et al.92 presented a comprehensive overview encompassing the use of laboratory and SR-based instrumentation in combination with several variants of the XRF technique as a method of elemental analysis and imaging, as well as with the combined use of XRD and XAS. Beside the investigation at microscopic level of a relatively limited number of minute paint samples, several methods for macroscopic, non-invasive imaging were highlighted. Those based on XRF scanning and full-field hyperspectral imaging appear to be very promising. An excellent overview of in situ non-invasive investigations of paintings by portable instrumentation was presented by Brunetti et al.93 The “MOLAB” setup, a unique collection of integrated mobile instruments, greatly contributed to demonstrate that it was possible to obtain satisfactory results in the study of a variety of heritage objects without sampling or moving them to a laboratory. Several non-invasive methods by portable equipment, including XRF spectrometry, mid-and near-FTIR, UV-VIS and Raman spectroscopy, as well as XRD, were discussed and examples of successful applications were given.

The strength of the SR-macro (MA)-XRF technique as a powerful and fast technique to identify pigments in fragments of Roman mural paintings, with no need of sample preparation, was demonstrated by Debastiani et al. In a first contribution94 analyses were performed at the FLUO beamline (ANKA, Karlsruhe Institute of Technology) to provide qualitative and quantitative element information about the sample, and in the 2D mode, lateral distribution of the elements. The measurements revealed the presence of Ca, Cu, Fe, K, Pb, among other elements, which could be correlated with red, blue, yellow and green pigments. In a second contribution95 the SR-MA-XRF technique was combined with μ-SRXRF, SRXRD and Raman spectroscopy. Correlation between SR-MA-XRF, μ-SRXRF elemental map distributions and optical images of scanned areas was mainly found for the elements Ca, Fe and K. With XRF, Fe and K were identified as correlated with green pigment, but in samples from two sites, Mendig Lungenkarchen and Weissenthurm “Am guten Mann”, also Cu was detected in minor concentration. As reviewer, it has to be said that although the performed analyses produced valuable results, nowadays these measurements can be conducted using an in situ MA-XRF scanning device, which is already commercially available. The advantages and limitations of the combined use of portable UV-VIS Fibre Optics Reflectance Spectrometry (FORS) and an XRF-XRD portable instrument for the non-invasive characterisation of pigments from Roman wall paintings from Seville (Spain) was described by Garofano et al.96 The portable XRF-XRD apparatus was based on a 4 mm diameter beam from a copper anode X-ray source (40 kV, 700 mA), which impinged on the coloured surface of the mural painting fragments at an angle of 10°. The source was equipped with a 15 μm Ni filter to attenuate the CuKβ line, providing a pseudo-monochromatic X-ray beam. The irradiated area of the sample was approximately 4 × 3 mm2, providing information on the different hues of the fragments. The complete identification of all pigments was only achieved by combining data from the different techniques. The characterisation of ancient pigments from a 19th century Korean Buddhist painting was conducted by Kim et al.,97 by combining portable EDXRF measurements with TOF-SIMS. The XRF results revealed the characteristic elements in the pigments such as Cu, Hg, Pb and S, which were commonly used in that period. Combining the XRF and TOF-SIMS results, it was shown that the white, green, and red areas in the Buddhist painting were painted with lead white, malachite, atacamite, cinnabar, and mars red as single or mixed pigments.

In the framework of restoration and conservation treatment decisions, it is important to gain information about the paint stratigraphy in paintings without taking additional samples. Non-invasive confocal-XRF analysis was therefore carried out by Eveno et al.,98 to examine several areas of the Virgin’s blue robe in a Renaissance painting of Caroto. Using this technique, it was possible to identify the pigments and order of application of paint by the artist. All the pigments identified were considered to be artist-applied, and on this basis no further cleaning was carried out and conservation proceeded with retouching only of minor areas of loss or damage. Confocal 3D-XRF spectrometry was also used by Nakano et al.99 to identify pigments for painting restoration or art appraisal. The in-house developed confocal 3D μ-XRF spectrometer, incorporating two polycapillary X-ray lenses, showed a depth resolution of 54 μm at Au Lβ (11.4 keV), as estimated by scanning of a thin 10 μm gold foil. To confirm the analytical performance, several pigment mixtures of Emerald green and Prussian blue were analysed. The XRF intensities of Cu and Fe in the pigment mixtures showed good correlation with the percentage content of Prussian blue. In addition, the replica paint in Daubigny’s garden by Vincent van Gogh was measured and the authors succeeded in visualising a black cat drawing which was hidden inside the painting.

Janssens et al.100 studied Rembrandt’s painting ‘Saul and David’ (125 × 158 cm), using non-destructive imaging by means of macroscopic XRF scanners. In 2013 the entire surface of the painting was scanned over a period of 3 days, at an average speed of 50 cm2 per hour (0.2 s per pixel). A series of seven large scans measuring 60 × 60 cm were performed as well as three smaller scans, to slightly extend the border of the investigated area. This resulted in elemental distribution images 1671 × 1314 pixels in size of 17 elements. The measurements revealed that three types of cobalt-containing materials were present in the painting. The presence of two varieties of smalts not only supported the re-attribution of the painting to Rembrandt and the idea that the picture was painted in two phases, but also that the second phase was painted by Rembrandt as well. While macroscopic XRF analysis visualises the element distribution in an area of about 80 × 60 cm2, the microscopic measurements are complementary and focused on smaller areas down to 100 × 100 μm2. To illustrate the interplay between X-ray and IR-based spectroscopic methods for macroscopic, mobile and microscopic pigment identification and imaging, Janssens et al.101 showed an example of their combined use for the full characterisation of the fifteenth century work of art named ‘Christ with singing and music-making Angels’ (Royal Museum of Fine Arts, Antwerp, Belgium), by Hans Memling. Analytical techniques like MA-XRF and VIS-NIR might be very useful for art historians and for art restorers alike, revealing information hitherto inaccessible. To complement the in plane information with in depth data, the use of highly-specific microscopic imaging methods such as combined μ-XRF/XRD and/or vibrational spectroscopies was shown to be very appropriate. The complementary features of the XRF and Raman techniques both in terms of spectroscopic information and propagation depth of the radiation, enabled Mosca et al.,102 to study the spatial distribution of pigments in paintings. The macro-XRF mapping of selected areas of the model painting was carried out with a portable XRF spectrometer, coupled to an XY translator stage covering an area of maximum 10 × 10 cm2. The spectrometer had an energy resolution below 135 eV, and allowed measurements within a spot size of about 1.0 mm on the painted surface at a working distance of about 1.4 cm. The XRF mapping showed the distribution of pigments in the whole stratigraphy of the painting, while the integration of the molecular information provided by Raman mapping yielded more exact pigment identification for some organic pigments and most inorganic artist materials, and it was useful for the correct attribution of a pigment to a superficial or to a hidden layer. To overcome unfavourable LODs and limited speciation possibilities of conventional spectroscopic techniques, Vermeulen et al.103 used SR-based techniques while working on painting cross-sections taken from a 17th-century painting by the Flemish artist Daniel Seghers and an 18th-century French Chinoiserie. Micro-SRXRF mapping analysis performed on a visually degraded orpiment-containing paint stratigraphy revealed that As was distributed throughout the entire cross-section, while μ-XANES measurements demonstrated that the As was present in both arsenite (AsIII) and arsenate (AsV) forms. Based on the XANES results and the high resolution maps of As presented, a qualitative difference in the stability of various types of arsenic sulfide pigments was observed. Reiche et al.104 studied the painting L’Homme blesse (80.8 × 97.0 cm2) by Gustave Courbet kept at the Musee d’Orsay in Paris, by X-ray radiography, SEM-EDS observation of paint cross sections and confocal μ-XRF analyses at locations where the cross section samples were taken. This study allowed the establishment of the paint palette used by Courbet for the three paint compositions. Eight or more paint layers could be evidenced. Elemental maps of As, Ba, Ca, Cu, Cr, Fe, Hg, Pb and Zn, consisting of 100 × 100 pixels, were acquired by MA-XRF spectrometry with a step size of 1.4 mm corresponding to a 140 × 140 mm2 area. Eleven scans were recorded. The obtained data were discussed in combination with depth profiles, obtained by confocal μ-XRF measurements, on strategic points where three painting compositions overlap. The order of the three successive compositions of this painting was determined. Thurrowgood et al.105 used the Maia 384 detector array at the XFM beamline (Australian Synchrotron) to enable the rapid collection of high-definition SRXRF data over an area of 42.6 × 26.7 cm2. The authors collected 31.6 megapixel scanning XRF element maps and reported a novel image processing methodology utilising these maps to produce a false colour representation of a “hidden” portrait by Edgar Degas. Ruberto et al.106 developed a portable XRF scanning spectrometer which could perform both element mapping, on areas up to 20 × 20 cm2, and single spot analyses. The subject of their analysis was the painting “La Muta” by Raffaello Sanzio, one of the “Old Masters” of the Italian Renaissance. Beyond identification of the painting palette, the new system allowed the researchers to deduce Raffaello’s use of bone black pigment and to identify various instances of “pentimenti” (underlying image in a painting, evidence of revision by the artist). Non-invasive in situ analyses were performed by Bonizzoni et al.,107 on seven choir books from the old library of the Certosa di Pavia. Pigments and binders were examined exploiting the synergy between four complementary techniques, namely, EDXRF spectrometry, FORS, FTIR and μ-Raman spectroscopy. The application of techniques with different penetration depths allowed the authors to establish the stratigraphic sequences without physical sampling. Moreover, some other painting techniques, such as the presence of mixtures or superimposed layers, could be related to different authors, allowing to go beyond the use of the same pigments by different authors. Additionally, gilding raw materials and applying technique were a clear hint to the different masters.

For the characterisation of pigments in a wide range of materials the XRF technique among other complementary techniques is commonly applied and well documented in this years’ review. A novel application of SR-XFM for the non-destructive micro-resolution element analysis of natural mineral pigments on Aboriginal Australian objects was presented by Popelka-Filcoff and co-workers.108 Although this technique required the transport and handling of an object to a synchrotron facility, the use of a high sensitivity XRF detector such as the Maia-384, allowed the radiation dose to the object to be as low as possible and yet returned amounts of data orders of magnitude greater than other traditional X-ray methods. Estimates of pigment thickness could be calculated. In addition, based on the element maps of the pigments (area of approximately 5 × 3 cm), further conclusions could be drawn on the composition, mixtures and uses of natural mineral pigments and whether the objects were made using traditional or modern methods and materials. Non-invasive XRF and complementary techniques were used by Garrote et al.,109 in the polychrome characterisation of the doors, shutters and ceiling in the Casa de Pilatos of Seville (Spain). The portable XRF measurements provided good information on the different elements present in the polychrome. Combining these data with the SEM-EDS study of cross-sections facilitated the characterisation of all layers and pigments from the support to the most external layer. A multi-analytical approach, including portable μ-XRF spectrometry, was applied by Vataj et al.,110 for the characterisation of glass mosaic tesserae recovered from the archaeological excavation of the mosaics at the early Christian basilicas in Albania. The results showed that compounds containing Co, Cu, Fe, Mn, Fe, Pb and Sn were used as colourants in the tesserae of different colours. The performance of a portable spectrometer for non-invasive XRF analysis of art objects, from the compositional mapping of the sample’s surface, rather than the usual point analysis, was proposed by Alberti et al.111 The translator stage could cover a maximum area of 10 × 10 cm2 with an accuracy of 10 μm. Compared with other XRF mapping devices, the authors claimed that the compact head and the absence of any X-ray optics made it easily usable for in situ studies and assured a good sensitivity even to lighter elements. The performance of this XRF system was evaluated by identifying the key elements in a variety of fragments of frescoes coming from the Villa dei Quintili in Rome (Italy), both in situ and in the laboratory. Results obtained by Raman spectroscopy supported the XRF results, and allowed the unambiguous identification of the pigments. Coccato et al.112 had a closer investigation of some specific instrumental parameters, i.e. sample positioning and cleaning of the sample carriers, in order to perform pigment particle analysis with a TXRF spectrometer. For the study of microscopic particles sampled on a dry cotton swab, they investigated the effect of positioning of the deposited sample. Element count ratios (Fe K α/Mn Kα) of a sample deposited in a controlled way, as regards both size and position on the sample carrier, showed a relative error of 4% for centred samples, increasing to 17% at 4 mm from the centre. A multi-analytical approach mainly based on the use of non-invasive spectroscopic techniques, but also on the additional use of a destructive technique (ICP-MS), was successfully applied by Marcaida et al.113 They characterised the mordant used to manufacture two Pompeian pink and one purple lake pigments recovered from the excavation of the archaeological site of Pompeii and preserved in the Naples National Archaeological Museum (MANN, Italy). A quick elemental in situ screening analysis of the two pink lake pigments in the MANN was performed using a commercial hand-held EDXRF spectrometer, showing the presence of an aluminosilicate enriched in Cu and Pb. In the laboratory, μ-XRF measurements of ten repetitive punctual point measurements of each lake pigment with a lateral resolution of 1 mm were performed, and confirmed the data obtained with the hand-held system. A hand-held XRF device might be of importance when the possibility of extracting micro-samples of archaeological remains will not be allowed. Guerra et al.114 provided new insights into the red and green pigments in the illuminated floral charter of Setubal, Portugal, by combined use of portable EDXRF and μ-Raman spectrometries. The EDXRF instrument collimated the outgoing radiation by a tantalum collimator, resulting in a 5 mm beam diameter at the sample surface. The sample was placed at a distance of 55 mm from the X-ray tube and at 10 mm from the detector’s beryllium window. The system was equipped with two laser pointers to identify the central point of the X-ray beam, which was positioned at a 90° angle from the detector axis. This geometry allowed a significant background reduction on Compton scattering. The results showed that the red and green pigments were particularly puzzling, as the widely used mercury- and copper-based pigments, vermillion and malachite, respectively, were not found in the illuminated frontispiece. Instead, the cheaper lead-based pigment minimum was used in the King’s flag, while a mixture of copper sulphates was found for the green colour, identified by means of μ-Raman spectroscopy. Manukyan et al.115 presented complementary XRF mapping, on macro-, millimetre- and micron-scale, and Raman spectroscopy to analyse pigments in a rare medieval Breton manuscript. The analysis was performed on 12 illustrated leaves (samples) out of the 92 which were recovered by Rare Books and Special Collections at the University of Notre Dame. The results from different leaves confirmed that pigments and inks of illustrated leaves belong to the same palette. The results also showed the pigments utilised in illustrations, text, and borders were identical indicating that the manuscript was prepared in a single setting, by a single artisan or a small number of artisans working closely. The micro-chemical characterisation of surface paintings of ceramic sherds from Michoacán (Mexico) were studied by Jadot et al.,116 who applied a combined analytical approach using μ-XRF spectrometry with SEM-EDS, μ-XRD and μ-Raman spectroscopy. The obtained results allowed a first insight in understanding the techniques used for pigment preparations and ceramic decorations of the Palacio and Milpillas phases, but also in identifying the origin of the raw materials used in the pottery production.

The analysis of artefacts is a recurrent topic in this application section. About 100 metal artefacts from the tomb of the Lady of Cao, were studied by Cesareo et al.,117 using various portable systems including EDXRF spectrometry. Gold objects and gold areas of nose decorations were characterised as having approximately the same composition, that is, Ag 16% ± 3%, Au 79.5% ± 2.5%, and Cu 4.5% ± 1.5%, while silver objects and silver areas of the same nose decorations showed widely varying results, and a systematic high Au concentration. Their findings suggested there might be reason to suspect at least some of them were subject to a depletion gilding process. A collection of 39 metallic artefacts recovered from archaeological sites in Southern Portugal was studied by μ-EDXRF by Vidigal et al.,118 to identify compositions and the use of metal among ancient communities. The results showed copper with variable amounts of As and very low content of other impurities, such as Fe, Pb or Sb. Moreover, nearly half of the collection was composed of arsenical copper alloys, and an association was found between arsenic content and typology because the weapons group (mostly daggers) presented higher values than tools (mostly awls). Lopes et al.119 introduced an alternative way for gold thickness determination of coatings in cultural heritage objects, combining portable XRF data and partial least square regression. Gold layers with thicknesses determined by Rutherford backscattering spectrometry were used as reference samples to produce a calibration model and to check the methodology before its application to unknown artefacts. This methodology was successfully evaluated by analysing in situ XRF measurements on a gilding frame in Brazil and on two pre-Columbian artefacts from Chavin culture in Peru. The thickness layer of the gilding frame, the six golden heads and the golden vessel was 1.15 μm ± 0.14 μm, 1.48 μm ± 0.21 μm and 1.31 μm ± 0.63 μm, respectively. The high deviation of the golden vessel was attributed to the heterogeneity of the golden layer over the Ag base. Comelli et al.120 investigated the origin of the iron dagger blade from the sarcophagus of the ancient Egyptian King Tutankhamun (14th century BCE). The composition of the blade (iron plus 0.58% m m−1 Co and 10.8% m m−1 Ni) was accurately determined through portable XRF spectrometry, and strongly supported its meteoritic origin. From different archaeological sites in Romania bronze artefacts were studied by Macovei and Popescu,121 in order to make a link with geological sources of raw material and to elucidate what kind of alloy their ancestors used. For that purpose, a commercially available μ-XRF instrument was used to determine the elemental composition of the different bronze samples and to perform element distribution mapping. Their research showed that the Romanian bronzes were considered complex tin bronzes with As, Cd, Fe, Ni, Pb, Sb and Si, presenting a notable heterogeneity. With the collected data the authors were able to conduct some source allocations. Figueiredo et al.122 investigated four metal axes from the Early/Middle bronze Age in Western Iberia, using a combination of μ-XRF and neutron imaging techniques, namely 2D radiography and 3D tomography. The μ-XRF analyses were performed with a portable spectrometer, with a low-power (30 W) molybdenum anode X-ray tube and a polycapillary lens delivering a spot size of about 70 μm in diameter, and a SDD with an energy resolution of 160 eV (Mn-Kα, FWHM). The μ-XRF data allowed the distinction among copper and bronze axes, and provided data about the composition of early bronzes for which data are scarce. The neutron imaging study allowed, for the first time, the visualisation of internal heterogeneities in early bronze axes, namely pores and large voids. The analyses of artefacts by XRF spectrometry combined with Monte Carlo (MC) simulations using the XRMC code was a topic of interest for Brunetti, Schiavon and co-workers. At first,123 XRF measurements using a polychromatic beam were performed to characterise the bulk chemical composition of restored (i.e., cleaned) and unrestored multilayered Peruvian metallic artefacts belonging to the twelfth-and thirteenth-century AD funerary complex of Chornancap-Chotuna in northern Peru. The multilayered structure was represented by a metal substrate covered by surface corrosion patinas and/or a layer from past protective treatments. The results led to a reformulation of previous hypotheses about the structure and composition of the metal used to create the Peruvian artefacts under investigation. Secondly,124 the chemical composition of a unique bronze artefact known as the “Cesta” (“Basket”) belonging to the ancient Nuragic civilisation of the Island of Sardinia, Italy, was analysed by combining XRF spectroscopy with MC simulations. The XRF measurements revealed that the handles of the object were composed of brass while the other parts were of bronze, suggesting the handles were a later addition to the original object. In order to determine the bronze bulk composition without the need for removing the outer patina, the artefact was modelled as a two layer object within the MC simulations. Finally,125 this MC simulation approach was also applied to characterise the element composition of a series of famous Iron Age small scale archaeological bronze replicas of ships (known as the “Navicelle”) from the Nuragic civilisation in Sardinia, Italy. As before, each sample was modelled as a multilayered object comprising two or three layers, depending on the sample, that enabled the authors to determine the bronze alloy composition together with the thickness of the surface layers without the need for previously removing the surface patinas.

For the study of the manufacturing processes of a group of gold jewels from the El Carambolo treasure, Scrivano et al.126 developed a new portable μ-XRF system. The μ-XRF set-up (Rh anode, 50 kV, 800 mA) with a polycapillary lens had a 50 μm FWHM spot size at 7.5 keV. The SDD had a detector sensor of 25 mm2 and a thickness of 500 μm, with an energy resolution of 127 eV at 5.9 keV. Eleven jewels were analysed in order to characterise the composition of the employed alloys, to identify the manufacturing processes and to discuss the hypothesis about the production workshop of the treasure. This μ-XRF set-up also allowed the analysis of small details like decoration elements and joining areas. Furthermore, the results demonstrated the capability of the new setup to substitute satisfactorily other micro-analytical techniques, such as μ-PIXE when the sample cannot be brought to a laboratory. Troalen and Guerra127 applied a multi-technique approach (portable XRF spectrometry, PIXE and SEM-EDS), to analyse a gold necklace and penannular earrings from tomb 296 at Riqqa (Egypt), together with eight penannular earrings from other find-spots of the same period. Analysis of jewellery items from tomb 296 at Riqqa revealed the use of high-purity gold alloys and electrum alloys, while the other earrings investigated were found to be made of electrum with high Ag content. Two earrings conserved in different museums were shown to be originally a pair. A timely and valuable remark was made by Blakelock,128 that care has to be taken when judging a gold object by its surface analysis. In his study, he considered whether non-destructive surface analysis could be useful in grouping the very large numbers of fragmentary pieces in the Anglo-Saxon Staffordshire Hoard. Surface and subsurface analyses of 16 objects were carried out by a portable μ-EDXRF spectrometer and SEM-EDS. The results indicated that there were significant but inconsistent levels of enrichment of the Au at the surface of many of these objects, due to the loss of both Cu and Ag. On the other hand, in three cases, an increase in Ag at the surface compared to the core metal was detected; which might tentatively be explained by redeposition from contact with silver objects during burial.

The XRF technology continues to be of interest for the analysis of pottery fragments and shards of ceramics and porcelain. A confocal 3D μ-XRF set-up combined with two individual polycapillary lenses was used by Yi et al.,129 to analyse the layers of painted pottery fragments from the Majiayao Culture (3300 BC-2900 BC). The element depth profiles of Ca, Fe and Mn were consistent with those obtained using an optical microscope. The images showed that the distribution of Ca was approximately homogeneous in both painted and unpainted regions. In contrast, Mn appeared only in the painted regions. Meanwhile, the distributions of Fe in the painted and unpainted regions were not the same. The surface topography showed that the pigment of dark-brown region was coated above the brown region. To get more insights into the chemical characteristics of glaze and the colour mechanism of the copper-red porcelain from Changsha Kiln (AD 7th–10th century), China, Li et al.130 analysed a shard of an opaque glaze porcelain with red pigments using a commercial μ-EDXRF spectrometer, XANES, μ-XRD, microscopy and SEM methods. The EDXRF results showed that Cu played an important role as red colourant. By using XANES and μ-XRD, the Cu valence was considered as 0+. The authors suggested that metallic copper in nanoparticles might be the major reason for the red colour.

Quantitative determinations and imaging in different structures of buried human bones from the XVIII-XIXth centuries in order to identify post mortem contamination was performed by Guimaraes et al.,131 using an in-house tri-axial geometry EDXRF setup and a commercial benchtop μ-XRF system. The more accurate tri-axial geometry setup was used to quantify Br, Ca, Cu, Fe, Pb, Sr and Zn in pressed powder bone pellets (n = 9 for each bone). Cluster analysis of these data showed a clear association between some bones regarding Br, Cu, Fe, Pb and Zn content but not a correlation between cortical and trabecular bone. The element distribution of Cu, Pb and Zn was assessed using the benchtop μ-XRF system, showing that contamination was mostly on the surface of the bone confirming that it was related to the burial shroud covering the individuals. Choudhury et al.132 demonstrated the advantages of SR-based confocal XRF imaging as a tool to generate high spatial-resolution element maps from fragile archaeological bone samples showing excellent details of element incorporation into bone microstructures. The implementation of confocal XRF imaging not only avoided physical thin sectioning of the sample, which was needed for standard XRF imaging, but also produced element maps showing accumulation of Pb in bone microstructures with excellent detail. The resultant high resolution was due to confocal detection, which facilitated not only optical sectioning of samples but also superior rejection of scattered background.

The combination of a portable XRF spectrometer with other complementary techniques was described in several contributions on the characterisation of cultural heritage objects. The conservation state of a sandstone entrance arch of 18th century La Galea Fortress, located in Getxo, northern Spain, was assessed by Morillas et al.,133 who conducted in situ spectroscopic analysis using two portable instruments; a commercially available hand-held EDXRF system and a portable Raman spectrometer. The EDXRF analyses indicated that the sandstone used for the entrance arch construction was the same as the one used for construction of the tower from La Galea Fortress. The analyses revealed that the sandstone contained Al, As/Pb, Ca, Cu, Fe, K, Mn, S, Si, Ti, Rb, Sr, Zn and Zr. Thanks to the use of the portable Raman spectrometer, the original composition of the sandstone was also confirmed. Through the analysis of 24 historic objects of garden statuary and ornamentation from three eighteenth and nineteenth century manufacturers, Karran and Colston134 successfully evaluated the use of portable XRF spectroscopy, and, more specifically, element profiles, in identifying, and differentiating between the products of manufacturers such as Coade, Blashfield and Doulton. Despite the inherent heterogeneity of these materials, it was shown that discrimination was nevertheless possible using portable XRF spectrometry, primarily due to the significant differences observed across a range of elements at both major and trace levels. Conventional metallographic and non-destructive methods, among which hand-held XRF spectrometry, were used by Carl and Young,135 to examine a 20th century sacrificial silver-plated “Century” fork from the Dallas Museum of Art collection in order to compare the effectiveness and validity of each method. The hand-held XRF experiments allowed a qualitative identification of the three alloying elements in the base metal (Cu, Ni, and Zn) and also a quick, semi-quantitative estimation of the plating thickness at multiple locations on the “Century” fork. Although all the non-destructive methods gave varying degrees of information, none of them alone would be sufficient in order to fully characterise a silver-plated cultural heritage object. Nevertheless, the authors aimed to create a database that could be used to effectively and efficiently characterise silver-plated objects using only hand-held XRF measurements.

To conclude, an interesting application was presented by Sessa et al.,136 who investigated the possible origins of sulfur in 19th century salted paper photographs using a portable μ-XRF spectrometer equipped with a rhodium target X-ray tube, SDD and a CCD camera. Dedicated thin-film sulfur standards were prepared for calibration. Optimal measurement conditions to detect sulfur in salted paper prints were derived, resulting in an acquisition time of 200 s, a 30 kV voltage, and a 1300 μA current in a helium atmosphere, without using primary beam filters. Based on the analysis of a salted paper print made in the laboratory strictly following a 19th century procedure and of two artistic salted paper photographs, the authors concluded that the ratio of the total number of counts for the Ag Lα and S Kα lines acquired across an image was a good indicator to determine the origin of sulfur. When this intensity ratio was approximately constant and close to one all over the image, the presence of sulfur was deemed likely to be due to a toning process. A Ag/S ratio smaller than one, indicated that sulfur was most probably due to environmental contamination.

Abbreviations

2D/3DTwo dimensional/three dimensional
ADAnno domini
AESAtomic emission spectrometry
AFMAtomic force microscopy
ASICApplication-specific integrated circuit
BCBefore Christ
BCEBefore the common era
BFRBrominated flame retardants
CCDCharge coupled detector
CMOSComplementary metal oxide semiconductor
CRMCertified reference material
CTComputed tomography
CZTCadmium zinc telluride
DRDietary restriction
EDEnergy dispersive
EDXRFEnergy dispersive X-ray fluorescence
FF-XANESFull-field X-ray absorption near edge structure
FORSFibre optics reflectance spectrometry
FPFundamental parameter
FTIRFourier transform infrared
FWHMFull width at half maximum
GE-XRFGrazing exit X-ray fluorescence
GI-XRFGrazing incidence X-ray fluorescence
GUIGraphical user interface
HPGeHigh purity germanium
ICPInductively coupled plasma
ICP-OESInductively coupled plasma optical emission spectrometry
ICP-MSInductively coupled plasma mass spectrometry
IPIntraperitoneal
IR-LEDInfrared-light emitting diode
JGIXAJava grazing incidence X-ray analysis
KESK-Edge subtraction
LA-ICP-MSLaser ablation inductively coupled plasma mass spectrometry
LODLimit of detection
MA-XRFMacro-X-ray fluorescence
MCMonte Carlo
MiXOMiniature lightweight X-ray optics
NISTNational institute of standards and technology
NPNanoparticle
OCOvarian cancer
PCPersonal computer
PCAPrincipal component analysis
PCSPhotoelectric cross section
PIXEParticle-induced X-ray emission
PLSPartial least squares
PMMAPolymethylmethacrylate
PTBPhysikalisch-Technische Bundesanstalt
REERare earth elements
RMReference material
RSDRelative standard deviation
SAXSSmall angle X-ray scattering
SDDSilicon drift detector
SEMScanning electron microscopy
SEM-EDSScanning electron microscopy energy dispersive X-ray spectrometry
SNSubstantia nigra
SRSynchrotron radiation
SR-CTSynchrotron radiation computed tomography
SR-MA-XRFSynchrotron radiation-macro-X-ray fluorescence
SR-TXRFSynchrotron radiation total reflection X-ray fluorescence
SR-TXRF-XANESSynchrotron radiation total reflection X-ray fluorescence X-ray absorption near edge structure
SR-XFMSynchrotron radiation X-ray fluorescence microprobe
SRXRDSynchrotron radiation X-ray diffraction
SRXRFSynchrotron radiation X-ray fluorescence
STJSuperconducting tunnel junction
SXRMBSoft X-ray micro-characterisation beamline
TEMTransmission electron microscopy
TESTransition edge sensor
TOF-SIMSTime-of-flight secondary ion mass spectrometry
TXRFTotal reflection X-ray fluorescence
UV-VISUltraviolet-visible
VIS-NIRVisible near infrared spectroscopy
VIS-NIR DRSVisible near infrared diffuse reflectance spectroscopy
WDXRFWavelength dispersive X-ray fluorescence
XAFSX-ray absorption fine structure
XANESX-ray absorption near edge structure
XASX-ray absorption spectroscopy
XFMX-ray fluorescence microscopy
XPCIX-ray phase contrast imaging
XRDX-ray diffraction
XRFX-ray fluorescence
XRF-CTX-ray fluorescence computed tomography
XRRX-ray reflectometry
XSWX-ray standing waves

References

  1. M. West, A. T. Ellis, P. J. Potts, C. Streli, C. Vanhoof and P. Wobrauschek, J. Anal. At. Spectrom., 2016, 31(9), 1706–1755 RSC .
  2. A. Taylor, N. Barlow, M. P. Day, S. Hill, M. Patriarca and M. White, J. Anal. At. Spectrom., 2016, 31(3), 554–596 RSC .
  3. E. H. Evans, J. Pisonero, C. M. M. Smith and R. N. Taylor, J. Anal. At. Spectrom., 2016, 31(5), 1057–1077 RSC .
  4. R. Clough, C. F. Harrington, S. J. Hill, Y. Madrid and J. F. Tyson, J. Anal. At. Spectrom., 2016, 31(7), 1330–1373 RSC .
  5. S. Carter, A. Fisher, R. Garcia, B. Gibson, J. Marshall and I. Whiteside, J. Anal. At. Spectrom., 2016, 31(11), 2114–2164 RSC .
  6. O. T. Butler, W. R. L. Cairns, J. M. Cookc and C. M. Davidsond, J. Anal. At. Spectrom., 2017, 32(1), 11–57 RSC .
  7. D. Guimaraes, M. L. Praamsma and P. J. Parsons, Spectrochim. Acta, Part B, 2016, 122, 192–202 CrossRef CAS .
  8. J. Jiao, X. C. Zhan, L. Zhai, X. T. Fan, H. L. Wen, J. H. Yuan, X. Liu and S. Guo, Spectrosc. Spectral Anal., 2017, 37(1), 267–272 CAS .
  9. D. E. B. Fleming, M. N. Nader, K. A. Foran, C. Groskopf, M. C. Reno, C. S. Ware, M. Tehrani, D. Guimaraes and P. J. Parsons, Appl. Radiat. Isot., 2017, 120, 1–6 CrossRef CAS PubMed .
  10. A. Bull, M. T. Brown and A. Turner, Environ. Pollut., 2017, 220, 228–233 CrossRef CAS PubMed .
  11. A. Turner and K. R. Solman, Talanta, 2016, 159, 262–271 CrossRef CAS PubMed .
  12. A. Guzzonato, F. Puype and S. J. Harrad, Chemosphere, 2016, 159, 89–95 CrossRef CAS PubMed .
  13. L. A. Kalcsits, Front. Plant Sci., 2016, 7, 8 Search PubMed .
  14. J. F. Byrnes and P. J. Bush, J. Forensic Sci., 2016, 61(4), 1041–1045 CrossRef CAS PubMed .
  15. M. Rouillon and M. P. Taylor, Environ. Pollut., 2016, 214, 255–264 CrossRef CAS PubMed .
  16. S. Imashuku, J. Takahashi, S. Kunimura and K. Wagatsuma, ISIJ Int., 2016, 56(12), 2224–2227 CrossRef CAS .
  17. D. C. Weindorf, S. Chakraborty, J. Herrero, B. Li, C. Castaneda and A. Choudhury, Eur. J. Soil Sci., 2016, 67(2), 173–183 CrossRef CAS .
  18. V. Cardelli, D. C. Weindorf, S. Chakraborty, B. Li, M. De Feudis, S. Cocco, A. Agnelli, A. Choudhury, D. P. Ray and G. Corti, Geoderma, 2017, 288, 130–142 CrossRef CAS .
  19. S. Chakraborty, D. C. Weindorf, G. J. Michaelson, C. L. Ping, A. Choudhury, T. Kandakji, A. Acree, A. Sharma and W. Dandan, Pedosphere, 2016, 26(4), 549–560 CrossRef .
  20. J. Hong, S. Romaine and X. O. T. Mi, Earth, Planets Space, 2016, 68, 1–8 CrossRef .
  21. K. E. Young, C. A. Evans, K. V. Hodges, J. E. Bleacher and T. G. Graff, Appl. Geochem., 2016, 72, 77–87 CrossRef CAS .
  22. Q. Shan, X. L. Zhang, Y. Zhang, W. B. Jia, Y. S. Ling, D. Q. Hei and S. N. Chu, Spectrosc. Lett., 2016, 49(3), 188–193 CrossRef CAS .
  23. M. Guerra, C. Ferreira, M. L. Carvalho, J. P. Santos and S. Pessanha, Spectrochim. Acta, Part B, 2016, 122, 114–117 CrossRef CAS .
  24. F. P. Romano, C. Caliri, L. Cosentino, S. Gammino, D. Mascali, L. Pappalardo, F. Rizzo, O. Scharf and H. C. Santos, Anal. Chem., 2016, 88(20), 9873–9880 CrossRef CAS PubMed .
  25. B. Laforce, B. Vermeulen, J. Garrevoet, B. Vekemans, L. Van Hoorebeke, C. Janssen and L. Vincze, Anal. Chem., 2016, 88(6), 3386–3391 CrossRef CAS PubMed .
  26. K. Ricketts, C. Guazzoni, A. Castoldi and G. Royle, Nucl. Instrum. Methods Phys. Res., Sect. A, 2016, 816, 25–32 CrossRef CAS .
  27. D. Vernekohl, M. Ahmad, G. Chinn and L. Xing, Phys. Med. Biol., 2016, 61(24), 8521–8540 CrossRef PubMed .
  28. J. Yu, X. Hong, Y. J. Ma, J. B. Zhou, Y. Q. Wang, M. Wang and Y. S. Lu, Spectrosc. Spectral Anal., 2016, 36(10), 3429–3433 CAS .
  29. S. Fess, Spectroscopy, 2016, 7 Search PubMed .
  30. X. L. Li and Z. S. Yu, Appl. Radiat. Isot., 2016, 111, 45–49 CrossRef CAS PubMed .
  31. M. Santibanez, M. Vasquez, R. G. Figueroa and M. Valente, Radiat. Phys. Chem., 2016, 122, 28–34 CrossRef CAS .
  32. Y. F. Dong, R. R. Fan, D. Y. Guo, C. L. Zhang, M. Gao, J. Z. Wang, Y. Q. Liu, D. W. Zhou and H. Y. Wang, Spectrosc. Spectral Anal., 2016, 36(2), 550–554 CAS .
  33. Q. F. Xiao, A. Maclennan, Y. F. Hu, M. Hackett, P. Leinweber and T. K. Sham, J. Synchrotron Radiat., 2017, 24, 333–337 CrossRef CAS PubMed .
  34. M. W. M. Jones, N. W. Phillips, G. A. van Riessen, B. Abbey, D. J. Vine, Y. S. G. Nashed, S. T. Mudie, N. Afshar, R. Kirkham, B. Chen, E. Balaur and M. D. de Jonge, J. Synchrotron Radiat., 2016, 23, 1151–1157 Search PubMed .
  35. C. Hall, R. G. Acres, A. Winnett and F. Wang, J. Instrum., 2016, 11, 14 Search PubMed .
  36. O. Hachmoller, A. G. Buzanich, M. Aichler, M. Radtke, D. Dietrich, K. Schwamborn, L. Lutz, M. Werner, M. Sperling, A. Walch and U. Karst, Metallomics, 2016, 8(7), 648–653 RSC .
  37. A. D. Surowka, P. Wrobel, M. M. Marzec, D. Adamek and M. Szczerbowska-Boruchowska, Spectrochim. Acta, Part B, 2016, 123, 47–58 CrossRef CAS .
  38. R. Ortega, A. Carmona, S. Roudeau, L. Perrin, T. Ducic, E. Carboni, S. Bohic, P. Cloetens and P. Lingor, Mol. Neurobiol., 2016, 53(3), 1925–1934 CrossRef CAS PubMed .
  39. A. D. Surowka, M. Topperwien, M. Bernhardt, J. D. Nicolas, M. Osterhoff, T. Salditt, D. Adamek and M. Szczerbowska-Boruchowska, Talanta, 2016, 161, 368–376 CrossRef CAS PubMed .
  40. F. Bille, G. Kourousias, E. Luchinat, M. Kiskinova and A. Gianoncelli, Spectrochim. Acta, Part B, 2016, 122, 23–30 CrossRef CAS .
  41. A. Bergamaschi, K. Medjoubi, C. Messaoudi, S. Marco and A. Somogyi, J. Synchrotron Radiat., 2016, 23, 783–794 CrossRef CAS PubMed .
  42. A. Panahifar, T. M. Swanston, M. J. Pushie, G. Belev, D. Chapman, L. Weber and D. M. L. Cooper, Phys. Med. Biol., 2016, 61(13), 5077–5088 Search PubMed .
  43. M. Rauwolf, A. Turyanskaya, A. Roschger, J. Prost, R. Simon, O. Scharf, M. Radtke, T. Schoonjans, A. G. Buzanich, K. Klaushofer, P. Wobrauschek, J. G. Hofstaetter, P. Roschger and C. Streli, J. Synchrotron Radiat., 2017, 24, 307–311 CrossRef CAS PubMed .
  44. G. V. Brown, P. Beiersdorfer, N. Hell and E. Magee, Rev. Sci. Instrum., 2016, 87(11), 3 CrossRef PubMed .
  45. D. E. Gotta and L. M. Simons, Spectrochim. Acta, Part B, 2015, 120, 9–18 CrossRef .
  46. P. Honicke, M. Kolbe and B. Beckhoff, X-Ray Spectrom., 2016, 45(4), 207–211 CrossRef .
  47. P. Honicke, M. Kolbe, M. Krumrey, R. Unterumsberger and B. Beckhoff, Spectrochim. Acta, Part B, 2016, 124, 94–98 CrossRef .
  48. R. Kumar, A. Rani, R. M. Singh, M. K. Tiwari and A. K. Singh, Radiat. Phys. Chem., 2017, 131, 79–85 CrossRef CAS .
  49. M. W. M. Jones, K. D. Elgass, M. D. Junker, M. D. de Jonge and G. A. van Riessen, Sci. Rep., 2016, 6, 8 CrossRef PubMed .
  50. A. M. Crawford, P. Kurecka, T. K. Yim, C. Kozemchak, A. Deb, L. Dostal, C. J. Sun, D. L. Brewe, R. Barrea and J. E. Penner-Hahn, J. Synchrotron Radiat., 2016, 23, 901–908 CAS .
  51. B. De Samber, M. J. Niemiec, B. Laforce, J. Garrevoet, E. Vergucht, R. De Rycke, P. Cloetens, C. F. Urban and L. Vincze, PLoS One, 2017, 11(11), 27 Search PubMed .
  52. Y. Kashiv, J. R. Austin, B. Lai, V. Rose, S. Vogt and M. El-Muayed, Sci. Rep., 2016, 6, 9 CrossRef PubMed .
  53. Q. L. Jin, T. Paunesku, B. Lai, S. C. Gleber, S. Chen, L. Finney, D. Vine, S. Vogt, G. Woloschak and C. Jacobsen, J. Microsc., 2017, 265(1), 81–93 CrossRef CAS PubMed .
  54. B. Hesse, M. Salome, H. Castillo-Michel, M. Cotte, B. Fayard, C. J. Sahle, W. De Nolf, J. Hradilova, A. Masic, B. Kanngiesser, M. Bohner, P. Varga, K. Raum and S. Schrof, Anal. Chem., 2016, 88(7), 3826–3835 CrossRef CAS PubMed .
  55. Y. Tabuchi and K. Tsuji, X-Ray Spectrom., 2016, 45(4), 197–201 CrossRef CAS .
  56. P. Y. Sharanov, N. V. Alov and Y. A. Zolotov, Dokl. Chem., 2016, 467, 91–93 CrossRef CAS .
  57. P. M. Wrobel, M. Bogovac, H. Sghaier, J. J. Leani, A. Migliori, R. Padilla-Alvarez, M. Czyzycki, J. Osan, R. B. Kaiser and A. G. Karydas, Nucl. Instrum. Methods Phys. Res., Sect. A, 2016, 833, 105–109 CrossRef CAS .
  58. U. Fittschen, A. Guilherme, S. Bottger, D. Rosenberg, M. Menzel, W. Jansen, M. Busker, Z. P. Gotlib, M. Radtke, H. Riesemeier, P. Wobrauschek and C. Streli, J. Synchrotron Radiat., 2016, 23, 820–824 CrossRef CAS PubMed .
  59. D. Ingerle, G. Pepponi, F. Meirer, P. Wobrauschek and C. Streli, Spectrochim. Acta, Part B, 2016, 118, 20–28 CrossRef CAS .
  60. M. K. Tiwari and G. Das, X-Ray Spectrom., 2016, 45(4), 212–219 CrossRef CAS .
  61. G. Das, A. Khooha, A. K. Singh, A. K. Srivastava and M. K. Tiwari, Appl. Phys. Lett., 2016, 108(26), 4 CrossRef .
  62. E. Schneck, E. Scoppola, J. Drnec, C. Mocuta, R. Felici, D. Novikov, G. Fragneto and J. Daillant, Proc. Natl. Acad. Sci. U. S. A., 2016, 113(34), 9521–9526 CrossRef CAS PubMed .
  63. A. Singh, K. Luening, S. Brennan, T. Homma, N. Kubo, S. H. Nowak and P. Pianetta, J. Synchrotron Radiat., 2017, 24, 283–287 CAS .
  64. A. Gianoncelli, J. Bufon, M. Ahangarianabhari, M. Altissimo, P. Bellutti, G. Bertuccio, R. Borghes, S. Carrato, G. Cautero, S. Fabiani, G. Giacomini, D. Giuressi, G. Kourousias, R. H. Menk, A. Picciotto, C. Piemonte, A. Rachevski, I. Rashevskaya, A. Stolfa, A. Vacchi, G. Zampa, N. Zampa and N. Zorzi, Nucl. Instrum. Methods Phys. Res., Sect. A, 2016, 816, 113–118 CrossRef CAS .
  65. A. Rachevski, M. Ahangarianabhari, P. Bellutti, G. Bertuccio, E. Brigo, J. Bufon, S. Carrato, A. Castoldi, G. Cautero, S. Fabiani, G. Giacomini, A. Gianoncelli, D. Giuressi, C. Guazzoni, G. Kourousias, C. Liu, R. H. Menk, G. V. Montemurro, A. Picciotto, C. Piemonte, I. Rashevskaya, Y. B. Shi, A. Stolfa, A. Vacchi, G. Zampa, N. Zampa and N. Zorzi, Nucl. Instrum. Methods Phys. Res., Sect. A, 2016, 824, 452–454 CrossRef CAS .
  66. Y. Y. Li, B. Xiong and Z. Li, Nucl. Instrum. Methods Phys. Res., Sect. A, 2016, 831, 29–33 CrossRef CAS .
  67. N. Golshani, J. Derakhshandeh, C. I. M. Beenakker and R. Ishihara, Microelectron. Eng., 2016, 160, 54–62 CrossRef CAS .
  68. H. T. Philipp, M. W. Tate, P. Purohit, K. S. Shanks, J. T. Weiss and S. M. Gruner, J. Synchrotron Radiat., 2016, 23, 395–403 CrossRef CAS PubMed .
  69. U. Boesenberg, C. G. Ryan, R. Kirkham, D. P. Siddons, M. Alfeld, J. Garrevoet, T. Nunez, T. Claussen, T. Kracht and G. Falkenberg, J. Synchrotron Radiat., 2016, 23, 1550–1560 CrossRef CAS PubMed .
  70. A. Allahgholi, J. Becker, L. Bianco, A. Delfs, R. Dinapoli, G. Arino-Estrada, P. Goettlicher, H. Graafsma, D. Greiffenberg, H. Hirsemann, S. Jack, R. Klanner, A. Klyuev, H. Krueger, S. Lange, A. Marras, D. Mezza, A. Mozzanica, J. Poehlsen, S. Rah, Q. Xia, B. Schmitt, J. Schwandt, I. Sheviakov, X. Shi, S. Smoljanin, U. Trunk, J. Zhang and M. Zimmer, J. Instrum., 2016, 11, 11 Search PubMed .
  71. C. K. Egan, J. W. Scuffham, M. C. Veale, M. D. Wilson, P. Seller and R. J. Cernik, Meas. Sci. Technol., 2016, 28(1), 8 Search PubMed .
  72. D. Doering, S. Amar-Youcef, J. Baudot, M. Deveaux, W. Dulinski, M. Kachel, B. Linnik, C. Muntz, J. Stroth and C.-M. Collaboration, J. Instrum., 2016, 11, 7 Search PubMed .
  73. P. Alle, E. Wenger, S. Dahaoui, D. Schaniel and C. Lecomte, Phys. Scr., 2016, 91(6), 15 CrossRef .
  74. H. Choi, H. S. Kim, Y. S. Kim, J. H. Ha and J. S. Chai, J. Korean Phys. Soc., 2016, 69(7), 1157–1162 CrossRef CAS .
  75. J. Pekarek, V. Dedic, J. Franc, E. Belas, M. Rejhon, P. Moravec, J. Tous and J. Voltr, Sensors, 2016, 16(10), 9 CrossRef PubMed .
  76. C. H. Hu, G. Q. Zeng, L. Q. Ge, M. F. Yu, S. L. Wei, K. Q. Zhang, J. Yang and C. Chen, Spectrosc. Spectral Anal., 2017, 37(1), 262–266 CAS .
  77. S. Shiki, G. Fujii, M. Ukibe, Y. Kitajima and M. Ohkubo, J. Low Temp. Phys., 2016, 184(1–2), 206–210 CrossRef CAS .
  78. V. A. Andrianov and V. P. Gorkov, J. Low Temp. Phys., 2016, 184(1–2), 188–193 CrossRef CAS .
  79. J. J. Yen, J. M. Kreikebaum, B. A. Young, B. Cabrera, R. Moffatt, P. Redl, B. Shank, P. L. Brink, M. Cherry and A. Tomada, J. Low Temp. Phys., 2016, 184(1–2), 30–37 CrossRef CAS .
  80. S. E. Busch, W. S. Yoon, J. S. Adams, C. N. Bailey, S. R. Bandler, J. A. Chervenak, M. E. Eckart, A. J. Ewin, F. M. Finkbeiner, R. L. Kelley, C. A. Kilbourne, S. J. Lee, J. P. Porst, F. S. Porter, J. E. Sadleir, S. J. Smith and M. Sultana, J. Low Temp. Phys., 2016, 184(1–2), 23–29 CrossRef CAS .
  81. V. A. Andrianov, S. Kraft-Bermuth and P. Scholz, J. Low Temp. Phys., 2016, 184(1–2), 238–243 CrossRef CAS .
  82. J. Rowley, L. Pei, R. C. Davis, R. R. Vanfleet, S. Liddiard, M. Harker and J. Abbott, J. Vac. Sci. Technol., A, 2016, 34(5), 5 Search PubMed .
  83. F. K. Zhao, J. Wang and A. M. Wang, Appl. Spectrosc., 2016, 70(12), 1994–2004 CrossRef CAS PubMed .
  84. C. Guo, W. C. Lai, Y. Hu, J. C. Huang, J. Zhai, F. Cheng, G. X. Wang, G. Q. Zeng and Q. X. Zhang, Spectrosc. Spectral Anal., 2016, 36(4), 1235–1239 CAS .
  85. J. L. Campbell, C. M. Heirwegh and B. Ganly, Nucl. Instrum. Methods Phys. Res., Sect. B, 2016, 383, 143–151 CrossRef CAS .
  86. S. P. Durdagi, Microchem. J., 2017, 130, 27–32 CrossRef CAS .
  87. M. Jabua, D. Gotta, T. Strauch, C. Weidemann, B. Fricke and K. Rashid, Spectrochim. Acta, Part B, 2016, 121, 11–17 CrossRef CAS .
  88. R. D. Perez, C. Sosa, V. Sbarato, J. Leani and H. J. Sanchez, Spectrochim. Acta, Part B, 2015, 117, 23–28 CrossRef .
  89. F. L. Melquiades and E. L. Thomaz, J. Environ. Qual., 2016, 45(3), 1104–1109 CrossRef CAS PubMed .
  90. M. Alfeld, M. Wahabzada, C. Bauckhage, K. Kersting, G. Wellenreuther, P. Barriobero-Vila, G. Requena, U. Boesenberg and G. Falkenberg, J. Synchrotron Radiat., 2016, 23, 579–589 CAS .
  91. S. H. Jiang, B. A. Wei, P. Feng, P. He, L. Z. Deng and M. Y. Chen, J. Med. Imaging Health Inform., 2016, 6(7), 1720–1726 CrossRef .
  92. K. Janssens, G. Van der Snickt, F. Vanmeert, S. Legrand, G. Nuyts, M. Alfeld, L. Monico, W. Anaf, W. De Nolf, M. Vermeulen, J. Verbeeck and K. De Wael, Top. Curr. Chem., 2016, 374(6), 52 CrossRef PubMed .
  93. B. Brunetti, C. Miliani, F. Rosi, B. Doherty, L. Monico, A. Romani and A. Sgamellotti, Top. Curr. Chem., 2016, 374(1), 35 CrossRef PubMed .
  94. R. Debastiani, R. Simon, D. Batchelor, G. Dellagustin, T. Baumbach and M. Fiederle, Microchem. J., 2016, 126, 438–445 CrossRef CAS .
  95. R. Debastiani, R. Simon, J. Goettlicher, S. Heissler, R. Steininger, D. Batchelor, M. Fiederle and T. Baumbach, Appl. Phys. A: Mater. Sci. Process., 2016, 122(10), 12 CrossRef .
  96. I. Garofano, J. L. Perez-Rodriguez, M. D. Robador and A. Duran, J. Cult. Herit., 2016, 22, 1028–1039 CrossRef .
  97. M. Kim, J. Lee, J. M. Doh, H. Ahn, H. D. Kim, Y. M. Yang and Y. Lee, Surf. Interface Anal., 2016, 48(7), 409–414 CrossRef CAS .
  98. M. Eveno, E. Mysak, K. Muller, G. Bastian, N. Pincas and I. Reiche, Stud. Conserv., 2016, 61(2), 102–112 CrossRef .
  99. K. Nakano, A. Tabe, S. Shimoyama and K. Tsuji, Microchem. J., 2016, 126, 496–500 CrossRef CAS .
  100. K. Janssens, G. Van der Snickt, M. Alfeld, P. Noble, A. van Loon, J. Delaney, D. Conover, J. Zeibel and J. Dik, Microchem. J., 2016, 126, 515–523 CrossRef CAS .
  101. K. Janssens, S. Legrand, G. Van der Snickt and F. Vanmeert, Elements, 2016, 12(1), 39–44 CrossRef CAS .
  102. S. Mosca, R. Alberti, T. Frizzi, A. Nevin, G. Valentini and D. Comelli, Appl. Phys. A: Mater. Sci. Process., 2016, 122(9), 10 CrossRef .
  103. M. Vermeulen, G. Nuyts, J. Sanyova, A. Vila, D. Buti, J. P. Suuronen and K. Janssens, J. Anal. At. Spectrom., 2016, 31(9), 1913–1921 RSC .
  104. I. Reiche, M. Eveno, K. Muller, T. Calligaro, L. Pichon, E. Laval, E. Mysak and B. Mottin, Appl. Phys. A: Mater. Sci. Process., 2016, 122(11), 13 CrossRef .
  105. D. Thurrowgood, D. Paterson, M. D. de Jonge, R. Kirkham, S. Thurrowgood and D. L. Howard, Sci. Rep., 2016, 6, 10 CrossRef PubMed .
  106. C. Ruberto, A. Mazzinghi, M. Massi, L. Castelli, C. Czelusniak, L. Palla, N. Gelli, M. Betuzzi, A. Impallaria, R. Brancaccio, E. Peccenini and M. Raffaelli, Microchem. J., 2016, 126, 63–69 CrossRef CAS .
  107. L. Bonizzoni, S. Bruni, A. Galli, M. Gargano, V. Guglielmi, N. Ludwig, L. Lodi and M. Martini, Microchem. J., 2016, 126, 172–180 CrossRef CAS .
  108. R. S. Popelka-Filcoff, C. E. Lenehan, E. Lombi, E. Donner, D. L. Howard, M. D. de Jonge, D. Paterson, K. Walshe and A. Pring, Analyst, 2016, 141(12), 3657–3667 RSC .
  109. M. A. Garrote, M. D. Robador and J. L. Perez-Rodriguez, Spectrochim. Acta, Part A, 2017, 173, 279–291 CrossRef CAS PubMed .
  110. E. Vataj, E. Hobdari, S. Rohrs, P. Vandenabele and N. Civici, Appl. Phys. A: Mater. Sci. Process., 2017, 123(1), 15 CrossRef .
  111. R. Alberti, V. Crupi, R. Frontoni, G. Galli, M. F. La Russa, M. Licchelli, D. Majolino, M. Malagodi, B. Rossi, S. A. Ruffolo and V. Venuti, J. Anal. At. Spectrom., 2017, 32(1), 117–129 RSC .
  112. A. Coccato, B. Vekemans, L. Vincze, L. Moens and P. Vandenabeele, Appl. Phys. A: Mater. Sci. Process., 2016, 122(12), 10 CrossRef .
  113. I. Marcaida, M. Maguregui, H. Morillas, C. Garcia-Florentino, U. Knuutinen, J. A. Carrero, S. F. O. de Vallejuelo, A. P. Marti, K. Castro and J. M. Madariaga, Anal. Chem., 2016, 88(12), 6395–6402 CrossRef CAS PubMed .
  114. M. Guerra, M. L. Carvalho, A. Le Gac, M. Manso, C. Mortari, S. Longelin and S. Pessanha, J. Appl. Phys., 2016, 119(10), 7 CrossRef .
  115. K. V. Manukyan, B. J. Guerin, E. J. Stech, A. Aprahamian, M. Wiescher, D. T. Gura and Z. D. Schultz, Anal. Methods, 2016, 8(42), 7696–7701 RSC .
  116. E. Jadot, N. Schiavon and M. Manso, Spectrochim. Acta, Part B, 2015, 119, 10–16 CrossRef .
  117. R. Cesareo, R. F. Jordan, A. Fernandez, A. Bustamante, J. Fabian, S. D. Zambrano, S. Azeredo, R. T. Lopes, G. M. Ingo, C. Riccucci, G. Di Carlo and G. E. Gigante, X-Ray Spectrom., 2016, 45(3), 138–154 CrossRef CAS .
  118. R. O. Vidigal, P. Valerio, M. F. Araujo, A. M. M. Soares and R. Mataloto, X-Ray Spectrom., 2016, 45(1), 63–68 CrossRef .
  119. F. Lopes, F. L. Melquiades, C. R. Appoloni, R. Cesareo, M. Rizzutto and T. F. Silva, X-Ray Spectrom., 2016, 45(6), 344–351 CrossRef CAS .
  120. D. Comelli, M. D’Orazio, L. Folco, M. El-Halwagy, T. Frizzi, R. Alberti, V. Capogrosso, A. Elnaggar, H. Hassan, A. Nevin, F. Porcelli, M. G. Rashed and G. Valentini, Meteorit. Planet. Sci., 2016, 51(7), 1301–1309 CrossRef CAS .
  121. M. Macovei and G. C. Popescu, Carpath. J. Earth Env., 2016, 11(1), 65–78 Search PubMed .
  122. E. Figueiredo, M. A. S. Pereira, F. Lopes, J. G. Marques, J. P. Santos, M. F. Araujo, R. J. C. Silva and J. C. Senna-Martinez, Spectrochim. Acta, Part B, 2016, 122, 15–22 CrossRef CAS .
  123. A. Brunetti, J. Fabian, C. W. La Torre and N. Schiavon, Appl. Phys. A: Mater. Sci. Process., 2016, 122(6), 8 CrossRef .
  124. A. Brunetti, A. Depalmas, F. di Gennaro, A. Serges and N. Schiavon, Spectrochim. Acta, Part B, 2016, 121, 18–21 CrossRef CAS .
  125. N. Schiavon, A. de Palmas, C. Bulla, G. Piga and A. Brunetti, Spectrochim. Acta, Part B, 2016, 123, 42–46 CrossRef CAS .
  126. S. Scrivano, I. Ortega-Feliu, B. Gomez-Tubio, F. J. Ager, M. L. de la Bandera, M. A. Respaldiza and M. A. Ontalba-Salamanca, Radiat. Phys. Chem., 2017, 130, 133–141 CrossRef CAS .
  127. L. Troalen and M. F. Guerra, Appl. Phys. A: Mater. Sci. Process., 2016, 122(3), 12 CrossRef .
  128. E. S. Blakelock, Archaeometry, 2016, 58(6), 912–929 CrossRef CAS .
  129. L. T. Yi, Z. G. Liu, K. Wang, X. Lin, M. Chen, S. Q. Peng, K. Yang and J. B. Wang, Appl. Phys. A: Mater. Sci. Process., 2016, 122(4), 7 CrossRef .
  130. Y. Q. Li, Y. M. Yang, J. Zhu, X. G. Zhang, S. Jiang, Z. X. Zhang, Z. Q. Yao and G. Solbrekken, Ceram. Int., 2016, 42(7), 8495–8500 CrossRef CAS .
  131. D. Guimaraes, A. A. Dias, M. Carvalho, M. L. Carvalho, J. P. Santos, F. R. Henriques, F. Curate and S. Pessanha, Talanta, 2016, 155, 107–115 CrossRef CAS PubMed .
  132. S. Choudhury, T. Swanston, T. L. Varney, D. M. L. Cooper, G. N. George, I. J. Pickering, V. Grimes, B. Bewer and I. Coulthard, Archaeometry, 2016, 58, 207–217 CrossRef CAS .
  133. H. Morillas, J. Garcia-Galan, M. Maguregui, C. Garcia-Florentino, I. Marcaida, J. A. Carrero and J. M. Madariaga, Microchem. J., 2016, 128, 288–296 CrossRef CAS .
  134. L. E. Karran and B. J. Colston, J. Cult. Herit., 2016, 18, 290–298 CrossRef .
  135. M. Carl and M. L. Young, Microchem. J., 2016, 126, 307–315 CrossRef CAS .
  136. C. Sessa, L. Barro, S. A. Centeno, H. Bagan, F. Caro and J. F. Garcia, X-Ray Spectrom., 2016, 45(3), 176–184 CrossRef CAS .

Footnote

Review coordinator.

This journal is © The Royal Society of Chemistry 2017