2014 Atomic Spectrometry Update – a review of advances in X-ray fluorescence spectrometry

Contents

• 1. Introduction and reviews

• 2. Instrumentation
  • 2.1 Hand-held, mobile, on-line XRF techniques and planetary exploration
  • 2.2 Laboratory instruments and excitation sources
  • 2.3 Synchrotron and large scale facilities
  • 2.4 TXRF and related techniques
  • 2.5 X-Ray detectors
  • 2.6 Quantification and data processing

• 3. Applications
  • 3.1 Geological and climate change
  • 3.2 Industrial minerals and consequences from mining activity
  • 3.3 Industrial materials and consequences of industrial activity
  • 3.4 Environmental
  • 3.5 Archaeological and cultural heritage
  • 3.6 Forensic
  • 3.7 Clinical
  • 3.8 Drugs
  • 3.9 Biological
  • 3.10 Thin films, coatings and nanomaterials
  • 3.11 Chemical state and speciation

• 4. Abbrevations

• References

---

Abstract

This review describes advances in the continued expansion of work using the XRF group of techniques published approximately between April 2013 and March 2014. Specialised laboratory instrumentation, X-ray sources, detector development and data processing continue unabated. It is remarkable how quickly hand-held XRF instrumentation has developed to the point where many examples of its use are included in the various application sections of this review, rather than in the instrumentation section. Several new beam lines and their new end-stations were described in publications from SR research centres around the world. More analysts are attracted to TXRF and related techniques with new sample preparation techniques offered for an expanding range of applications. Nanoparticles and nanomaterials feature throughout this review particularly in clinical, biological and environmental studies. A novel approach for the reuse of industrial by-products described how acid mine drainage sludge and coal fly ash facilitated the problem of the high levels of phosphate present in waste waters (cow dung) from the dairy industry. The archaeological and cultural heritage section often includes good news stories. This year we learn how an XRF mapping technique was used to show that Pablo Picasso used a popular brand of French house paint in his works of art.

---

1. Introduction and reviews

This update reviews papers published approximately between April 2013 and March 2014 and continues the series of Atomic Spectrometry Updates in X-ray Fluorescence Spectrometry¹ that should be read in conjunction with other related reviews in the series.^2–6 Of over 1200 XRF publications viewed, we offer comment on 426 to demonstrate advances in the X-ray group of techniques. The writing team has changed this year with the departure of Darius Wegrzynek. However, we are pleased that Andrew Ellis has kindly agreed to take on the extra commitment. Our thanks go to Darius for all his hard work in recent years.

Readers new to XRF analysis will benefit from a tutorial in the form of a book by Margui and Van Grieken⁷ that provides an up-to-date review of the fundamentals of the XRF group of techniques including an overview of instrumentation, sample preparation procedures and applications. More longstanding X-ray analysts will gain from the description of related techniques such as TXRF, μ-XRF, SRXRF, PIXE, SEM and EMPA. The authors described the basic components of X-ray spectrometers ranging from hand-held systems using radioisotopes or mini X-ray tubes, bench top and floor-mounted configurations to those using synchrotron radiation as the X-ray source. The characteristics of detector types include comparisons of efficiency, resolution at MnKα, count rate capability and cooling media. Detector artefacts including escape and sum peaks are described along with information on filters, secondary targets and focusing optics. The chapter on qualitative and quantitative analysis provides a valuable overview of matrix effects and methods for correction and compensation that should not challenge the reader's mathematical prowess. Options for sample preparation procedures are introduced with examples for some biological, environmental and industrial fields of application. Layouts for WDXRF and EDXRF systems are described along with applications and case studies. The increasing use of TXRF is described with its crucial sample preparation requirements to support analysis of liquids and surfaces. Microbeam configurations are described that support both 2D and 3D elemental mapping as well as information on spatial elemental distribution down to the nanometer scale. The information on synchrotron systems lists some facilities available world-wide plus more detailed information on the beamlines available at ESRF, Grenoble, France. The Fukushima Daiichi Nuclear Power Plant accident has renewed considerable public concern about the dangers posed by radioactive contamination in the environment and the related internal exposure from the contaminating radionuclides. Zheng et al.⁸ described atomic spectrometric techniques employed in radiation protection, such as accelerator mass spectrometry, ICP-MS, PIXE and SRXRF spectrometry. Applications of atomic spectrometric techniques in radio-ecological studies in several significant nuclear contamination events in Japan, studies using a suitable stable element as an analogue of long-lived radionuclides related to high-level radioactive waste disposal, and microbeam elemental analysis for estimation of internal radionuclides radiation, were reviewed to highlight the important role of atomic spectrometric techniques in radiation protection. Zhitenko⁹ reviewed methods used at analytical laboratories in industry for determining high gold contents (from several percent to 99.9%) in alloys, compounds, solutions, and artefacts, as well as the existing international and national standards published from 2001 to 2010. A study¹⁰ on high pressure (1600 kN) sample preparation methods for pressed powder pellets from rocks, soils and stream sediments may be of interest to those analysts troubled with delamination and cracking often associated with high silica content samples. From SEM observations, the authors reported that a silica crystal lattice was destroyed and that particle size and mineralogical effects were reduced leading to improved precision and accuracy of their analytical procedure. Mello et al.¹¹ reviewed analytical methods for the determination of halogens in biological samples and other related matrices such as food, drugs and plants. The availability of CRMs for evaluation of accuracy of methods for sample preparation and analysis including TXRF spectrometry was described covering changes during the last 20 years. Motellier et al.¹² used GIXRF to investigate the deposition of TiO₂ nanoparticles (NPs) in suspension with a particular interest in the influence of substrate pre-treatment. The authors emphasised that the major advantage of this technique was the possibility to analyse the particles without pre-treatment thereby avoiding the harsh acid digestion required by most conventional methods. But they also pointed out that reliable quantitative measurements required a number of precautions; particularly, the deposition process of the sample on the flat reflecting substrate should maintain uniformity over the entire surface of the deposition residue once dried. Linear calibration curves using internal standardisation were established with ionic Ti and with two different types of TiO₂ NPs. Detection limits of 18 μg L⁻¹ and 52 μg L⁻¹ at incident angles of 0.20 degrees and 0.75 degrees, respectively, were obtained. The authors found that the correlation coefficient of the fitted linear calibration was particle-size dependent, and was assigned to sampling problems due to possible incomplete dispersion of the particles in the suspension. They also observed that the measured XRF signal of the dried deposits changed within a 4 month timespan for both types of TiO₂ NPs, demonstrating the very peculiar behaviour of these particulate samples.

2. Instrumentation

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

It is remarkable how quickly hand-held XRF instrumentation has developed to the point where many examples of its use are included in the various application sections of this review, rather than in this instrumentation and technique section. Again, it is necessary to express a word of caution about the use of terminology in research papers. ‘Portable’ XRF continues to be widely used in the literature, but sometimes it is not possible to identify whether laboratory instrumentation is being used in a field laboratory or hand-held instrumentation used for the more interesting and challenging in situ measurements in the field. That said, Bosco¹³ reviewed the development of portable, hand-held XRF spectrometers with a number of leading contributors describing the development and application of various commercial instruments. Li et al.¹⁴ described the application of an improved ‘group method of data handling’ network to the portable XRF analyser, claiming that this approach in establishing a physical model between count rate and element content was better than most of the statistical methods of calculation.

An important industrial use of XRF is in the on-line analysis of raw materials, an application illustrated this year by the work of Jia et al.,¹⁵ who used a portable XRF instrument and an auxiliary distance correction module to measure the S concentration of coal. The authors investigated the relationship between XRF intensity and distance to the surface of the sample and showed (perhaps unsurprisingly) that measurement accuracy was significantly increased when a distance correction formula was applied.

Turning now to the use of XRF in the characterisation of planetary surfaces, interest continued in analysing data from the Chandrayaan-1 mission that used solar flares as the excitation source to fluoresce elements in the surface of the Moon, measurements being made remotely from an orbiting satellite. Athiray et al.¹⁶ reported XRF lines of the major rock forming elements from the lunar surface during different solar flare conditions (November 2008 to August 2009), calculating element concentrations taking into account the incident solar flux, geometry of observation and matrix effects based on a fundamental parameter approach. The validation of results was discussed using laboratory samples of known composition and errors of a few % m/m were claimed for the remote sensing data. Using data from the same mission, Weider et al.¹⁷ reported MgO/SiO₂ and Al₂O₃/SiO₂ ratios from the western part of the Oceanus Procellarum region of the Moon's near side. The authors combined data with multi-spectral reflectance data identifying low-Ti mare basalts and more feldspathic regoliths in the highland region immediately to the North of Oceanus Procellarum. The surface composition of the planet Mars continues to attract significant interest and Evans et al.¹⁸ described GeoLab-A, a prototype geological laboratory designed to support NASA's science operations for future planetary surface missions. A commercially available hand-held XRF instrument was used in the first field trials of this laboratory at a test site near Flagstaff, Arizona. X-ray fluorescence spectrometry contributed to a study by Rampe et al.,¹⁹ whose main interest was to develop and interpret thermal emission spectrometer data to obtain information on petrology and aqueous alteration processes on Mars. Results by XRF also contributed to a paper by Stivaletta et al.,²⁰ whose interest was in the identification of minerals in terrestrial evaporites relevant to the exploration of environments suitable for supporting microbial life on the surface of Mars. Two distinct Mars analogue environments were identified by Stern et al.²¹ in Svalbard, Norway, for evaluating methodologies and techniques to be used on the Mars Science Lander. An XRD/XRF instrument was deployed as part of the field campaign at these sites, although the main focus of this investigation was an evaluation of evolved-gas techniques including a Hiden evolved gas analysis-mass spectrometer and a Picarro cavity ring down spectrometer.

2.2 Laboratory instruments and excitation sources

The most active publication area during the review period was in small-spot XRF, X-ray microfluorescence (XRMF) and X-ray imaging instruments for a variety of applications and in a number of different configurations. Of particular interest was confocal XRF and a high resolution, vacuum confocal μ-XRF instrument was developed by Nakazawa and Tsuji.²² The rhodium target X-ray tube, which was operated at 50 kV, 0.5 mA and had a spot size of 50 μm, was coupled through a 100 μm thick beryllium window in the vacuum chamber to a full-lens polycapillary optic with input and output focal lengths of 30 mm and 2.5 mm respectively. The detector was a 50 mm² Vortex SDD with an energy resolution of <130 eV at 5.9 keV that was coupled to a polycapillary half-lens with an input focal length of 3 mm and an output distance of 36 mm to the detector. The confocal optics were set at 90° and the spot size on the sample was 10 μm. The instrument could be operated in a confocal mode or the detector half-lens could be removed to allow operation in a 10 μm spot XRMF mode. For elements in the Z range 13 to 20, the confocal mode provided detection limits that were up to a factor of 2 better, although for XRF lines >5 keV the confocal arrangement was much worse due to poor transmission efficiency in the half-lens to the detector. Impressive depth resolutions were measured of 56.0 μm for 1.49 keV, 18.3 μm for 8.04 keV, and 10.9 μm for 17.4 keV X-rays, which the authors claimed were the best values to date for a confocal XRF system. Interesting examples were presented of elemental depth profiles of multi-layer paint samples that were of particular importance for forensic investigations. The authors also reported results²³ on the 3D XRF mapping of an SD memory card using an earlier, air path version of this system equipped with the same confocal lens setup but with a molybdenum target X-ray tube. The authors were able to image the two main layers of the circuitry and the ability to make 2D and 3D investigations on the same sample in the same instrument were particularly valuable and impressive. The same set-up was also used to perform depth profiling of corroded painted steel samples²⁴ and the results demonstrated how effective was the confocal EDXRF technique for non-destructive failure analysis without the need for any sample preparation or sectioning. An interesting application of a confocal EDXRF system was described²⁵ for the in situ measurement of ion distribution close to the electrode surface during electrolysis of a CuCl₂ solution, opening up the possibility for spatially-resolved analysis of mass transfer in electrolytic tanks, perhaps even with some degree of temporal resolution. A polycapillary optic EDXRF confocal system was also used by Zhao et al.²⁶ who employed a surface adaptive algorithm to establish surface topography and elemental distribution of a ceramic sample. While the claim of obtaining surface topography was made, there were limitations that would readily be overcome using the widely available techniques of SPM and AFM. The use of confocal EDXRF spectrometry in which absorption corrections were made to the XRF signals was proposed²⁷ for depth profiling of multi-layered materials. While interesting for the non-destructive analysis of complete unknowns, conventional EDXRF multi-layer thin film is already well-known for this application in well-characterised thin film layered systems. This is also the case for the straightforward Ni thin film analyses reported by Peng et al.²⁸ using a confocal EDXRF system, for which a conventional thin film EDXRF approach would seem to be perfectly suitable. A slightly different setup to the aforementioned typical confocal systems was described by Dehlinger and co-workers,²⁹ who used a rhodium target X-ray tube operated at 35 kV, 0.8 mA coupled to a polycapillary lens and a 10 mm² SDD coupled to a cylindrical glass monocapillary. The incidence and takeoff angles were 30° and 90° respectively. The authors characterised the system performance using a series of detector monocapillaries that were each 50 mm long and 1 mm outer diameter with an inner radius of 5, 10, 25 or 50 μm. The geometric considerations of the setup were presented in detail and this setup enabled the authors to establish the smallest spot that could be used for measurement using this laboratory system and geometry, which was <1 μm based upon the use of a 0.5 μm detector capillary. Clearly, the use of a higher powered small spot X-ray tube would improve the signal but it is important to recognise that this geometry, with a 90° takeoff angle may be good for best 2D spatial resolution but has no capability to deliver 3D depth profiling. An impressive improvement in X-ray spot size was reported by Komatani et al.³⁰ using a novel hybrid focussing optic. The authors used a tungsten conical pinhole placed between the polycapillary lens and sample to scrape off the outer wings of the spot output from the optic. This simple, but hard to make, pinhole had entry and exit diameters of 40 and 2.5 μm respectively and the X-ray spot on the sample was measured, using a knife edge method, to be only 2.8 μm diameter. This compared favourably with the spot from a poly or monocapillary of 12 or 8 μm respectively, although the intensity from the spot was reduced to that of the monocapillary, which was a factor of 50 less than for the polycapillary. Although providing an impressively small spot size, the very short working distance forced the detector takeoff angle to be extremely low and the drop in intensity from the straightforward polycapillary optic would only be acceptable if the really small spot size was absolutely necessary and if the high absorption-enhancement effects and low emergence depth of the fluorescence radiation could be ignored or be successfully compensated. The benefits of the high incidence and takeoff angles in a commercially available μ-XRF instrument were used to advantage by Hardy & Scruggs,³¹ who measured powdered samples through various thickness sample bags. The authors used sample bags made from either 50.8 μm nylon, 50.8 μm Teflon or 127 μm Teflon, which allowed measurement in air or under partial vacuum and reported results for a uranium-rich glass and two powdered geological materials measured using excitation through a polycapillary providing a 30 μm measurement spot. It comes as no surprise that the lower energy lines were most attenuated by the thicker sample bags and that detection limits for elements with higher energy XRF lines were improved by using primary beam filtration and/or the thicker sample bags, but the paper is of interest to those needing to use sample bags to avoid contamination to or from samples and needing a non-destructive analysis method that allows analysis of samples with no treatment at all. The physics and mathematics of X-ray scattering and transport from the surface of glass capillary optics was reported in detail by Mazuritskiy et al.,^32,33 who studied the phenomena for ultra-soft X-rays in the energy range 90 to 140 eV around the Si L_2,3 absorption edge. The authors' model showed general agreement with the high-precision data collected on an SR beamline at the BESSY II facility in Berlin although some differences were found in the fine structure around the absorption edge. The data were invaluable for the optics used in the authors' ultra-soft X-ray spectroscopy but no attempt was made to extend the models to energies >1 keV, which are those of most interest for μ-XRF instrument users. An interesting design proposal was made by Cong et al.³⁴ for an XRF computed tomography (CT) system for the in vivo imaging of gold NPs in biological tissue. The system comprised a 160 kV, 19 mA X-ray tube with a spot size of 1 to 3 mm whose output was filtered by a 1.35 mm thick tin primary beam filter then collimated into a fan beam in conjunction with two linear arrays each of 10 5 × 5 mm² CdTe detectors. The two detector arrays were aligned on either side of the fan beam and mounted on a motorised semi-circular arm that allowed them to be stepped through 180°. Measurements were to be made for 10 s at each of 30 steps of the detector arrays and those data were fed into a μ-CT system to reconstruct the XRF image of Au K lines. This straightforward design used commercially available components and was claimed to offer a realistic approach to a benchtop μ-XRF-CT system for use in molecular and cellular imaging applications without the need of access to an SR source, as was prevalent in early XRF-CT studies. A very detailed study of two novel imaging geometries for μ-XRF-CT was made by Fu et al.,³⁵ who demonstrated the effectiveness of their approaches using high intensity SR radiation at a beamline at the advanced photon source (APS), Illinois, monochromatised to 10 keV. One of the new geometries involved illuminating and scanning a single slice of the object and imaging each slice's XRF emissions using a pinhole collimator and a 2D pixellated X-ray CCD with an energy resolution of around 250 eV. The other involved illuminating a single line through the object and imaging the XRF emissions using the same CCD X-ray detector and a slit collimator. Full details of the data treatment and visualisation were presented for a phantom and an osmium-stained zebra fish specimen and, based on the data and their simulations, the authors concluded that a viable benchtop system could be created using a high power X-ray tube and the proposed new geometries with the X-ray sensitive CCD. The dispersing power of a WDXRF system was used to advantage by Ohmori and co-workers³⁶ who placed a straight polycapillary lens between the sample and the LiF200 analyzer crystal and replaced the usual gas proportional counter with a 2D pixellated X-ray detector with 487 × 195 pixels. Using an acquisition time of 10 s at each 2-theta setting, the authors were able to build up an elemental map of the sample and were able to readily discriminate Ba from Ti in one sample. Providing only a few elements are selected for mapping, the sensitivity and peak resolving power was good however, in the set-up used, the spatial resolution for K lines of the elements in the Z range 22 to 30 was only in the range 80 to 1200 μm although a figure of ∼500 μm was obtained for Zr Kα. Given the relatively poor spatial resolution and the inherently sequential nature of the method it is unlikely that WDXRF imaging of this type will supplant small-spot EDXRF instruments. Collimated X-rays from a tungsten target X-ray tube heavily filtered using a 100 μm thick tantalum primary beam filter were used in a laboratory XRF imaging system³⁷ to excite Gd K XRF lines from a marker used to identify cancerous tissue. The Gd K X-rays were detected using a CdTe X-ray detector and the cancerous cells in a nude mouse could readily be mapped when the sample was moved on an X–Y stage moving at 5 mm s⁻¹ with a measurement dwell time of 0.5 s. A detailed mathematical treatment using the Sherman equations was presented by Dul and colleagues³⁸ for the correction of absorption and enhancement effects in XRF holography of a Cu₃Au single crystal. The authors showed the effectiveness of the proposed method for monochromatic excitation and suggested that the approach might also be applied to polychromatic excitation and to other holographic techniques.

The incremental improvements in X-ray sources used for the ever-growing number of hand-held XRF systems continued during the review period and the obvious drive to deliver lower weight sources was exemplified by the latest ULTRA-LITE source³⁹ claimed to weigh only 250 g. It is pleasing to note that this low weight appears to come with no compromise in flux stability, which was claimed to be repeatable within 0.5% RSD or less over a time frame of days. The target thickness is of critical importance for the transmission target X-ray tubes typically used in hand-held XRF systems in order to ensure the expected spectrum is produced. Having gone to the trouble of using MC simulations to predict the spectrum output for various target thicknesses, Zhang et al.,⁴⁰ then simplified the data by splitting the spectrum into two regions, <5 keV and 5 to 50 keV, and re-established the blindingly obvious fact that, when operated at 50 keV and with a 4 μm thick silver target, the output in the 5 to 50 keV range was much the greater. In a similar study⁴¹ using MC simulation to predict the spectrum emitted through several different thicknesses in the range 50 to 500 μm of beryllium window it was concluded that thinner was not always best in EDXRF. The authors proposed some entertaining empirical judgement factors and concluded that a beryllium window of 250 μm was optimum as it provided some inherent filtration to improve PBR and reduced the lower energy emissions that were of less interest to them. It is interesting to note two publications on field emitter sources for low power X-ray tubes. In the first paper from Iwai and co-corkers,⁴² the authors fabricated emitters from pyrolytic graphite that had various edges and juts that provided an emission current of at least 2 mA. The authors built X-ray tubes with these cold emitters and were able to operate them at 16.6 keV and 160 μA, albeit with a limited emission current stability of around 2.8% RSD over a 24 hour period. In a second publication⁴³ a field emitter was fabricated from the exotic-sounding graphene flower cloth that the authors indicated had numerous nano-protrusions formed by free-standing graphene structures. It was not made clear how reproducible such emitters were, but an X-ray tube incorporating the new emitter delivered a beam current of up to 500 μA and provided good kV stability when operated at 300 μA. The authors used the X-ray tube to successfully analyse stainless steel samples but did not report on long term stability, which was a limiting factor for earlier attempts at fabricating cold cathode field emitters from materials such as carbon nanotubes.

In an interesting theoretical and experimental paper on the geometric dependence of Compton and Rayleigh scatter in EDXRF spectrometers, Guerra et al.⁴⁴ used a poly(propylene) and an aluminium slab sample to measure the intensity of the Compton and Rayleigh scatter peaks derived from the rhodium target X-ray tube. The setup comprised a rhodium transmission target X-ray tube operated at 35 kV and 40 μA and a collimated 25 mm² SDD in an in-plane geometry typical of portable EDXRF systems and measurements covered source-detector angles over the range 35 to 140°. Having established the sample reflectivity over the angular range, the authors developed a model for the X-ray interaction volume in the angular range of interest then used an Orchard Leaves RM (NIST-1571) to establish detection limits in four specific geometries. On the basis of the modelling and experimental data, the authors established the best source-detector geometry for EDXRF spectrometry based upon the minimum detection limits for emission lines in the energy range 1 to 30 keV to be in the range 70 to 90°, which is typically the case in many commercially-available spectrometers.

2.3 Synchrotron and large scale facilities

During the current review period several new beam lines and their new end-stations with favourable features were described in publications from SR research centres around the world. Among them, was the Bionanoprobe – a hard XRF nanoprobe with cryogenic capabilities at Argonne Nat. Lab IL USA; APS, at the Australian Synchrotron; the Imaging & Med Beam Line in Clayton Australia, the microfocus XRF beamline at the Indian facility Indus – 2 in Indore, a μ-XRF beamline at the Shanghai Synchrotron Radiation Facility (SSRF) in China and a polychromatic beam end station at the Stanford Synchrotron Radiation Lightsource (SSRL), Stanford California, USA. Further, a valuable software package for large and complex data sets has been developed at the Singapore Synchrotron Light Source, Singapore. Chen et al.⁴⁵ described the features of the Bionanoprobe, a hard XRF nanoprobe, installed at the undulator beamline at sector 21 of the APS, which provided a spatial resolution of 30 nm for 2D XRF imaging with cryogenic sample environment and cryo transfer capabilities, dedicated to studying trace elements in frozen hydrated biological systems close to the “natural state”. Complete details were provided of the instrument design and motion control systems and the performance was quantified. The first results obtained with the Bionanoprobe on frozen-hydrated cells were reported. Using this new beamline, Yuan et al.⁴⁶ imaged and quantified intracellular nanoparticles directly by their elemental signatures. The new XRF spectroscopy beamline BL 16 at the Indian SR facility Indus-2 was constructed with an emphasis on environmental, archaeological, biomedical and material science applications involving heavy metal speciation and imaging. Tiwari et al.⁴⁷ described very interestingly, that the beamline offered a combination of analytical probes, e.g. elemental mapping, μ-SRXRF and TXRF spectroscopies. For X-ray optical components, the beamline comprised a pair of Kirkpatrick-Baez focusing mirrors and a double-crystal monochromator with Si(111) symmetric and asymmetric crystals built in the bending magnet beamline. Various activities were reported from the SSRF in the Peoples' Republic of China, as described in the following section. Xie et al.⁴⁸ made preliminary investigations of trace elements in pterygium by μ-SRXRF, looking in particular at the elements Fe and Zn. A randomised clinical trial was performed using serial frozen sections of pterygium and conjunctival tissues of 40 μm thickness, taken from eight patients (10 eyes) undergoing a pterygium excision combined with limbal stem cell transplantation. The samples were mounted on a 6 μm Mylar film and mapped on BL15U. The results showed that Fe and Zn were present in both pterygium and normal conjunctiva tissue, but the content of both elements was significantly lower in the normal conjunctiva tissue. The authors claimed that these results seem to be valuable in understanding the important role of Fe and Zn in the development process of pterygium. Gao et al.⁴⁹ described the high energy resolution spectrometer at SSRF. It was installed at BL14W1, the XAFS beamline, and basically consisted of three parts: a sample holder, a spherically curved silicon crystal and an avalanche photodiode detector. With this spectrometer, SRXRF, XANES, and EXAFS measurements could be performed. It was reported that an energy resolution of 3 eV for Mn compounds was obtained. In the XAFS application the Eu^III retention on manganese dioxide was studied indicating the superiority of fluorescence detection, as in the presence of the Mn matrix, both Mn–K and Eu^III–L lines could be discriminated. Deng et al.⁵⁰ reviewed the progress of XRF CT at SSRF from 2007 onwards. XRF CT allowed the non-destructive reconstruction of elemental distribution within the sample and the SSRF development aimed to achieve XRF CT microanalysis using a hard X-ray micro focusing beamline (BL15U1). A new so-called “expectation maximisation” algorithm was introduced to speed up the data acquisition process, whose main goal was the acceleration of XRF CT imaging based on fast scanning and the new algorithm. In early 2012 at Stanford, Barberie⁵¹ began the development and operation of the polychromatic beam end-station for SRXRF at SSRL in cooperation with the University of California. This end station was used as an element-specific analytical tool for a variety of environmental, metallic and mineral samples. In particular, the article focussed on the motivation for the development and specifications of this end station. Specifically the analysis of an aerosol-deposited substrate was used to demonstrate the main purpose of this line, and also as an example of bulk sample analysis, results from the analysis of Sutter's Mill meteorite were shown.

Some interesting applications of SR were published within the review period. Fuoss et al.⁵² described in situ X-ray studies of film cathodes for solid oxide fuel cells performed at the APS. Several X-ray techniques, which included X-ray reflectivity (XRR), TXRF, high resolution X-ray diffraction and ultra-small angle X-ray scattering were used to study in situ the structural and chemical changes of film cathodes during fuel cell operations. Fenning et al.⁵³ discussed the limit on gettering efficacy of a standard gettering process for precipitated iron in multi-crystalline silicon and identified guidelines based on the modelling of phosphorous diffusion gettering, for the improvement of effective minority carrier lifetime in multi-crystalline silicon solar cells. A census of iron-silicide precipitates taken by μ-SRXRF spectrometry at the APS confirmed the presence of a high density of iron-silicide both before and after phosphorous diffusion. The similar distributions of precipitated iron seen after each step in the solar cell process confirmed that the effect of standard gettering on precipitated iron was strongly limited, as predicted by simulation. The conclusion supported by the good agreement between experimental and simulated data was that the gettering kinetics were governed by not only the total iron concentration but also by the distribution of the precipitated iron. Hall⁵⁴ described the advantages of combined XRF (FXCT) and absorption tomography using SR (SRCT), in particular from the Australian source in Clayton, Victoria. Use of SR excited XRF ensured a signal that was optimally detectable, by exploiting the monochromaticity and linear polarisation of the source. If images were collected simultaneously, distinct benefits of SRCT and FXCT for biomedical experiments could be obtained. Proof of principle modelling was used to show the possibility to recover an XRF image of a point-like source from a SRCT apparatus by suitable modulation of the illuminating planar X-ray beam. A discussion of particular reconstruction methods, which might be applied by utilising both the CT and FXCT data was reported. A study of environmental monitoring was published by Savoly et al.⁵⁵ using free living nematodes as an appropriate bio-indicator. Soil-inhabiting species such as Xiphinema vuittenezi were claimed to be perfectly suited for the characterisation of soil heavy metal pollution. Starved organisms were kept in CuSO₄ and Cu(NO₃)₂ solutions with a concentration of 1 mmol dm⁻³ for 24 h. The nematodes were then mounted on a polyimide tape, frozen for 2 min in liquid N₂ and then lyophilised for 72 h. After this procedure, the Cu distribution in the nematodes was investigated by μ-SRXRF at BESSY II using two different lateral resolutions (8 μm × 8 μm and 2 μm × 2 μm). An elevated Cu and S content was found along the stylet. Additionally XANES studies were performed with the result that in all spectra Cu was present in the 2⁺ oxidation state. Luhl et al.⁵⁶ used SR from BESSY II for the non-destructive investigation of samples with constant elemental composition, but with varying chemical compounds. The technique applied was confocal XRF at certain energies near absorption edges (marker energies), where XAFS signals of the chemical compounds differ significantly. This information was used to unveil the chemical composition and its topology. A prominent application was the degradation of a homogenous material due to chemical reactions like oxidation or reduction. Depth scans performed with the confocal setup at the marker energies allowed 3D maps to be presented. Making use of the features offered by BL16 of Indus 2, Misra et al.⁵⁷ determined Th and U in their mixed oxides in pellet form using SRXRF spectrometry. Using the established calibration curves, results from SRXRF measurement deviated from the expected values in the range of 1%, which was superior compared with those results achieved in the laboratory.

Life science experiments have seen a huge increase in number and complexity of data being collected with every experiment, in particular those performed using SRXRF spectrometry. Banas et al.⁵⁸ at the Singapore Synchrotron Light Source pointed to this problem and proposed a solution by offering a software called “R environment” to extract vital information from the vast quantity of numbers. The development team who created the R platform, claimed to offer scientists an open source environment for statistical analysis designed to handle large and complex data sets.

2.4 TXRF and related techniques

In September 2013, the 15th International Conference on Total Reflection X-Ray Fluorescence Analysis and Related Methods (TXRF2013) took place in Osaka, Japan, organised by Tsuji. During the conference⁵⁹ a significant number of papers dealing with GIXRF, GEXRF techniques and their combination with absorption spectroscopy and its application to material science, specifically nanolayer materials and ultrashallow implants, were presented. The program showed new trends for further development, in comparison with standard applications of TXRF for chemical analysis. Furthermore, there are currently ISO activities (ISO/TC201/WG3) using TXRF as a tool for environmental and biological analysis, which will help to standardise the sampling, measurement and quantification procedures in TXRF spectrometry for chemical analysis. This ISO standard hopefully will be available in the near future and will emphasise the applicability of TXRF as a standardised analytical procedure. Fernandez-Ruiz reported his activity establishing a TXRF working group in Facebook,⁶⁰ which has become a useful discussion platform for TXRF related problems.

Only two of the published articles during this review period dealt with fundamentals of data evaluation specifically for TXRF techniques. The majority of papers were devoted to instrumentation and some very interesting methods of sample preparation for TXRF analysis. Moya-Riffo et al.⁶¹ developed a procedure for deconvolution of characteristic X-ray line overlaps and the determination of its confidence interval for arsenic and lead signals in TXRF spectral analysis. The mathematical procedure worked with the basic principle of “maximum likelihood” for the identification and quantification of As and Pb, which have a severe overlap of their fluorescence lines. This procedure was applied to the analysis of two or more signals that interfered with each other and for the quantification of small signals in a very noisy environment. The authors claimed that until now, a complete study of the determination of confidence intervals in samples, where one element in the overlap is in a concentration lower than the other elements, had not been undertaken. The paper provided exact values of relative concentrations for difficult cases such as the detection of traces of arsenic in the presence of large lead interfering signals and vice versa, which was proved experimentally by TXRF spectrometry. A similar problem was treated by Santibanez et al.⁶² who presented a general procedure that included external standard calibration and a deconvolution of spectral signals and results in the determination of the relative proportion and absolute quantification of overlapping elements from high accuracy TXRF experiments. The proposed deconvolution took advantage of the cross correlation technique, offering improvements in identification of simultaneous signals and in the determination of the relative elemental proportions. The external calibration was tested with certified standard samples, producing excellent results for specific heavy metals currently of interest in the health and industry sectors. A range of linear response was achieved for 200 pg to 200 ng content, which is consistent with the values obtained by other researchers. Water surface charge has been an ongoing problem for decades. Shapovalov et al.⁶³ presented an interesting approach to clarify this influence. The authors concluded that electrophoretic mobility of air bubbles in water together with the disjoining pressure between the surfaces of aqueous films suggested that the surface of water exhibited a significant negative charge. This was commonly attributed to a strong adsorption of hydroxide ions at the interface, suggesting surface depletion of hydroxide ions. Alternatively the negative charge could arise from surface contamination with trace levels of charged surfactants. The variation in the surface charge of water was tested for its dependence on pH using the Kelvin probe technique. In the interfacial layer the abundance of “reporter ions” namely Rb⁺ and Br⁻ should be affected by a charged surface, thus these ions were monitored using TXRF spectrometry. The magnitude of the surface charge was found to be below 1 electron per 500 nm², and no evidence of variations in the surface potential between pH 2–3 and pH 9–12 was detected. The authors concluded that these findings suggest that the clean water surface exhibited negligible charge in a wide range of the pH.

Surprisingly only two papers dealing with new instrumentation were published within the review period. Kunimura et al.⁶⁴ reported improvements for the established portable TXRF spectrometer with a 5 W tube operating a system for measurements under vacuum. Unsurprisingly, the new spectrum showed a significantly reduced background compared with measurements in air, offering 8 pg detection limit for Cr. Beckhoff et al.⁶⁵ presented a novel ultra-high vacuum instrument for X-ray reflectometry and spectrometry-related techniques for nano-analysis by means of synchrotron radiation. This versatile instrument included a 9-axis manipulator that allows for an independent alignment of the samples with respect to all degrees of freedom. Thus, the new instrument supported various analytical techniques based on energy dispersive X-ray detectors such as reference-free XRF analysis, TXRF and GIXRF in addition to optional XRR measurements or polarisation-dependent XAS measurement. The authors claimed that with this instrument samples up to 100 mm × 100 mm in size could be measured with respect to their mass deposition, elemental or spatial composition, or species in order to probe surface contamination, layer composition and thickness, the depth profile of matrix elements or implants, the species of nanolayers, nanoparticles or buried interfaces as well as the molecular orientation of bonds.

Interesting applications this year included Kunimura and Kawai⁶⁶ who published results obtained using the portable TXRF spectrometer described in the earlier cited paper.⁶⁴ They analysed specimens containing ng amounts of Bi, REEs, Pb and Sb, including those with the previously mentioned overlap of Pb–L and As–K lines that were successfully deconvoluted. Liu et al.,⁶⁷ also from the Kyoto team, published analytical results, using the same portable TXRF spectrometer, from solutions containing 11 elements. They studied the excitation parameters for optimum lower detection limits and reported sub ng levels for nine elements in a 10 minute measuring time. The authors claimed that the high sensitivity from such small samples demonstrated that the portable TXRF spectrometer was a valuable tool for field studies in environmental and geological investigations.

The following contributions dealt with sample preparation techniques for TXRF analysis and show the vast number of possibilities to achieve optimal conditions and sometimes simplicity in the applied procedures. A tutorial review was developed by De la Calle et al.⁶⁸ pointing to the revival of TXRF. This review discussed and compared the current approaches for sample pre-treatment including in situ microdigestion, slurry preparation, acidic digestion and extraction prior to TXRF analysis. A comprehensive revision for different sample strategies was included in this review that covered the period from 2008 until 2013. Even non-conventional pre-treatment approaches, such as microflow on-line preconcentration and lab-on-a-chip were discussed. The same group⁶⁹ also proposed ultrasound-assisted extraction of plants for routine multi-elemental TXRF analysis. Five certified reference materials were measured for Ca, Cr, Cu, Fe, K, Mn, Ni, P, Pb and Zn trying different extractant media such as various acids and oxidants. A mixture of diluted HNO₃ + HCl + HF was selected as the best option for a complete extraction. Different specimen types, i.e. herbs, spices and medical plants were analysed and, using linear discriminant analysis together with the elemental concentrations, the authors were able to differentiate commercial preparations corresponding to flower, fruit and leaf. In a further publication⁷⁰ this very active group described more detailed ultrasound-assisted single extraction (UAE) tests for rapid assessment of metal extractability from soils using TXRF techniques. The outstanding results using this UAE preparation technique showed a drastically reduced time for the acetic acid extraction from 16 h to 6 min. In addition, the amount of sample and extractants were reduced as a consequence of the miniaturisation implemented. The combination of UAE and TXRF spectrometry provided a simple means for assessing the potential metal mobility and bioavailability. Romero et al.⁷¹ described an in situ ultrasound-assisted magnetite synthesis with simultaneous ion co-precipitation, applied to magnetic solid-phase extraction (MSPE) of As, Au, Bi, Cu, Ge, Hg, Re, Pb, Tl, and Zn. The MSP enriched with metal ions was directly analysed by TXRF. A precise description of the procedure to synthesise magnetite (Fe₃O₄) NPs by sono-chemical treatment was included to provide a 10 μL aliquot of the MSP deposited on the siliconised quartz glass carrier for TXRF analysis. A recovery study was carried out and showed recoveries between 89–109% with an RSD below 10% for N = 5. The authors claimed that this effective, fast and sensitive procedure could be used for field analysis. A really critical element for analysis is mercury because of its volatile character during sample preparation and measurement. Holtkamp et al.⁷² presented results from a thorough investigation of strategies to overcome the effect of mercury losses during TXRF analysis. To prevent vaporisation, oxidation with ammonium persulfate and complexation with EDTA and DMSA were investigated. Both chemicals retained the different mercury species quantitatively for several hours. Based on this approach, a method for mercury determination by TXRF in liquid samples was developed, successfully validated by ICP-OES and applied to the analysis of mercury-containing vaccine samples. Quantitative TXRF is normally no problem, as an internal standard is used for the quantification procedure. The critical point of quantification using an external standard is the morphology and position of the sample after the drying procedure on the reflector. However, in silicon wafer surface analysis by TXRF, any intentional deposition of an internal standard is forbidden, so an external standard method has to be used. A nano-litre droplet deposition unit was developed and characterised for the application of sample preparation in TXRF by Wastl et al.⁷³ from the Vienna ATI group. The droplets produced on quartz reflectors as well as on wafers could be prepared in any desired pattern and showed good reproducibility in shape, but also in the accuracy of the pipetted volume, validated by TXRF using an external standard. The obtained fluorescence intensities were found to be invariant against reflector rotation. Angle scans of the droplet residue showed the typical curves below and above the critical angle. But below the critical angle, the fluorescence signal was relatively invariant with the angle of incidence, indicating a droplet residue behaviour. The authors claimed that this technique was well suited for the production of standards for external calibration for wafer TXRF spectrometry, but moreover, ideal for samples for the quantification of aerosols collected in special patterns by an impactor on the reflector. Pashkova et al.⁷⁴ confirmed TXRF an ideal technique for liquid sample analysis, in particular water. Some factors influencing the TXRF results were considered among them: surface density of dried water residue on the carrier, the dilution ratio of high mineralised samples with ultrapure water and the solution of the detergent TRITON X-100, the salt contents, the internal standard concentration and repeated pipetting of fresh water. Self-absorption phenomena, often normally neglected, and their influence on the quantification was demonstrated by using natural waters of varying salinity as brine, fresh water and ground-water samples. Results were compared with those obtained by wet chemistry and ICP-MS. In another interesting approach by Cantaluppi et al.⁷⁵ powder samples were analysed by direct TXRF measurement. The powder samples were simply placed directly on an adhesive carbon tab, which was then put on a quartz carrier without any digestion. Angular scans to find the best glancing angle, were carried out, different standard powders were measured and matrix effects and detection limits determined, leading to a technique that showed good agreement with data obtained with the certified values. As a positive consequence it can be concluded that a large number of samples could be analysed in a short time and that this method could be very useful for ongoing environmental monitoring. The collection efficiency of the widely used vapour phase decomposition technique for metallic impurities on silicon wafer surfaces was studied by Liou et al.⁷⁶ The collection efficiency reached up to 85% and was dependent upon parameters such as scanning droplet volume and scanning arm tracking speed and spiral tracking angle of the scanning stage. Additionally information was given on the baking temperature of the droplet which should be less than 50 °C for larger than 100 μL droplets which is recommended to achieve a high collection efficiency. A similar technique was described by Takahara et al.,⁷⁷ vapour phase treatment and its application to TXRF for trace elemental analysis of silicon wafer surface This technique which is under investigation by the ISO/TC201/WG2 to improve the detection limits of TXRF. In Round Robin tests it was shown that the intensity of the signals for an intentionally contaminated wafer with 5 × 10⁹ atoms cm⁻² of Fe and 5 × 10¹⁰ atoms cm⁻² of Ni resulted in an intensity enhancement varying greatly (1.2–4.7) among the participating laboratories. The particle morphology seemed to have an impact on the vapour phase treatment efficiency, confirmed by high resolution SEM observations revealing that a number of dots with SiO₂, silicate and/or carbon gathered to form a particle and heavy metals were segregated on it. Tikhonova and Kozlov⁷⁸ determined Gd and Sm contents in metallo-fullerenes using a parallel beam, TXRF spectrometer. The authors showed that the results could be used for measuring the contents of Gd metallo-fullerenes in powder samples, several mg in weight and in liquid samples several μL in volume. In an XRF study on segregation of I and Cs in an organic solar cell, Lindemann et al.⁷⁹ described how TXRF was able to reveal that in multi-layers of the inverted organic solar cells, caesium diffused into the organic layer and iodide diffuses into the organic solar cells. Laser ablation ICP-MS, which integrated elemental concentration across the whole multi-layer structure, indicated that the Cs : I ratio remained 1 : 1 confirming that there was no loss of iodine from the sample. Iodide diffusion to the bulk ITO layer was also found in a similar prepared ITO multi-layer structure and were consistent with XPS measurements showing that the ratio Cs : I at the ITO/CsI surface exceeds 8 : 1 and rationalise this observation.

Only one paper dealing with a TXRF related technique was published. Kayser et al.⁸⁰ reported the use of a two dimensional position sensitive area detector for scanning free GEXRF. This method is the inverse of TXRF as the exciting radiation impinges at 90° to the sample surface and the fluorescence radiation is observed at angles below the critical. In this experimental setup, a fixed sample detector arrangement was used. GEXRF is used for surface-sensitive studies with nm scale accuracy in the depth direction by measuring the intensity variation on an X-ray fluorescence line with the grazing emission angle. This combination of GEXRF with the 2D position sensitive detector results in the non-sequential and simultaneous acquisition of the GEXRF profiles of different emission lines in considerably reduced times.

2.5 X-Ray detectors

The topic of pixellated X-ray detectors in a variety of materials and configurations dominated what was a sparse year for detector papers available for review. Romano and co-workers⁸¹ described a new full-field EDXRF 2D imaging system requiring no sample or detector scanning that was equipped with a 100 W tungsten target X-ray tube and a pinhole camera comprising a 70 μm diameter laser-drilled pinhole and a 1024 × 1024 pixel CCD detector. The back-illuminated CCD chip was 13.3 × 13.3 mm² and the pixel thickness was a useful 40 μm, providing an effective energy range from 0.2 to 30 keV. The energy resolution of the detector was 157 eV at 6.4 keV when the CCD was Peltier-cooled to −100 °C and operated in vacuo. The detector was operated in single photon counting mode and the single frame exposure time was set low enough that only 10–15% of the 10⁶ pixels were illuminated to minimise the detection probability of double events in a single pixel. The overall acquisition time for an image was 900 s for the typically 50 frames acquired per measurement. The coaxial arrangement of the sample, pinhole and detector allowed for a sample area of up to 3 × 3 cm² and for a magnification factor between 0.8 and 3.2, providing a spatial resolution of 190 and 92 μm respectively. The authors demonstrated the performance of the system using a variety of samples with different elements and geometries and their pinhole camera approach demonstrated what can be achieved without the need for high precision X–Y scanning or X-ray optics. This impressive performance is being further improved and is expected to achieve an energy resolution of 133 eV at 5.9 keV and a spatial resolution down to 20 μm with magnification factors from 0.2 to 6. A new, novel and interesting method⁸² was established for achieving sub-pixel spatial resolution from a pnCCD sensor with 75 μm² pixels, each 450 μm thick, operated in single counting mode. The method took into account the energy dependence of the charge cloud created by a single photon and the probability of that charge being split across adjacent pixels. The authors considered charge splitting across up to 4 pixels, yielding a total of 13 pattern types over which the charge cloud from any single photon event could be collected. Rather than use a gold mesh with 5 μm holes placed close to the surface of the pnCCD to assess the charge cloud behaviour, the authors used a simpler approach based on the assumption that on average a charge cloud created by monoenergetic photons will always produce the same probability of splitting into singles, doubles, triples and quadruples in a CCD detector of defined pixel size. In addition, the authors showed that the charge cloud landing position within a pixel could be determined using the centre of gravity technique and a rectangular charge cloud density model enabling them to achieve a spatial resolution of just 2 μm. Full details of this novel approach were provided and the method was shown to be valid over the energy range 6.4 to 17.5 keV using measurements from five different metals (Cu, Fe, Mo, Pb, and Rb) excited by white SR at the BESSY II EDR beam line in Berlin. The size of the charge clouds was established as 8 to 10 μm over this energy range for the specific geometry of this detector and its operating parameters. The 2 μm spatial resolution was confirmed using edge profile analysis of a high contrast image taken through an optical grating and the authors believed a figure of 1 μm was achievable. This new approach may be applied to other well-characterised detector systems when adequate computing power is also available. Although far from use in laboratory XRF, a paper on the limitations of dead time models for single photon counting pixellated detector systems⁸³ will likely be of interest to those working on high count rate SR experiments. An improvement by a factor of 2 in count rate and linearity was achieved by matching the SR temporal structure to the detector system dead time. A further 2.5× improvement was demonstrated through the use of a correction algorithm based upon a transfer function described by the authors. Interest in pixellated CdTe X-ray detectors continues due to their greater stopping power for higher energy X-rays and their typically higher operating temperatures. Using a 2 × 2 arrangement of four of their existing 80 × 80 pixel CdTe detector and ASIC readout modules, Wilson and co-workers⁸⁴ described an impressively large area (16 cm²) CdTe detector array that lost only 3 pixels active area, including the guard ring, at the overlap point of the individual modules. A sensor thickness of 1 mm ensured excellent stopping power, although the energy resolution of 2.0 keV at 59.9 keV for the full array was disappointing given the individual modules had delivered 800 eV. The degraded energy resolution was ascribed to the limitations of the data acquisition system rather than the detector sensors and future work will doubtless correct those shortcomings. A detailed study of the individual pixel characterisation of this type of small pixel detector module was made by Veale et al.,⁸⁵ although using a 74 × 74 pixel, 2 mm thick CdZnTe sensor array for this study. The authors coupled the sensor to their small pixel spectroscopic readout HEXITEC ASIC and initial energy calibration of each pixel/channel was carried out using flat field irradiation using an ²⁴¹Am radioisotope source. The authors employed an X-ray microbeam on beamline B16 at the Diamond Light Source Synchrotron, Didcot, UK to study the source of non-uniformities observed in the response of these pixellated CdZnTe detectors. The SR beam was monochromatised to 20 keV and collimated to a beam of 10 × 10 μm² then stepped across the detector in 25 μm increments taking a 20 s acquisition at each step. All spectra were corrected for charge sharing and the energy resolution was established to be 1.74 ± 0.39 keV at 59.9 keV, which was significantly worse than the 800 eV typically obtained for equivalent size CdTe detectors with the same readout setup. The authors also established that the total counts per pixel varied much more widely than for CdTe detectors. The authors concluded that the observed effects were equivalent to variations in the effective area (volume) of pixels and that the effects were due to non-uniform electric fields in individual pixels rather than to impurity inclusions. One of these 80 × 80 pixel CdTe X-ray detector modules was used in conjunction with a 70 kV tungsten target X-ray tube operated at a beam current of 6 mA to enable goniometer-free, angular and energy dispersive XRD⁸⁶ without the need for any moving parts. The primary beam was shaped by two 0.5 × 0.5 mm² lead pinhole collimators before reaching the sample and the detector was set such that the primary beam passed by one corner of the detector. Any detected X-ray signal was expected to come from scattering by the sample or the air in the beam path and segments of any diffraction rings would clearly show on the 2D detector array. The angular range covered by the detector was ∼0.9–15.1° and, in addition to this angular data, each pixel also provided the energy of any detected X-ray photon. This powerful yet simple combination of ED and XRD methodologies was successfully applied to the classification of explosives and inert materials and showed promise for homeland security applications. Although strictly a pixellated detector, the indium grid microgas X-ray detector described by Krieger and colleagues⁸⁷ had limited energy dispersive utility given its reported energy resolution of 5.2% at 5.9 keV and its very limited spatial resolution. That said, the authors reported that background was reduced by a factor of 120 and that they were able to discriminate a charge track from a photon conversion in their experiment. In an interesting application of a hybrid pixellated X-ray detector, Wright and co-workers⁸⁸ were able to use a laboratory X-ray source in place of the normally-used SR beam to make combined size exclusion chromatography-SAXS measurements on proteins that were sufficiently sensitive to allow radius of gyration, maximum dimensions and molecular mass values to be assigned accurately.

Given the evident maturity of silicon drift detectors (SDD) there is little to report from papers available in the review period. Commercial suppliers continue to increase the active area of single sensor devices beyond 100 mm² to a staggering 150 mm² while electronics improvements continue to increase available output count rates to several hundred thousand cps. An experimental study⁸⁹ used PIXE spectra from 16 samples with XRF peaks in the 0.27 to 21 keV energy range to establish the overall detector efficiency of a standard commercially available 30 mm² SDD equipped with a commercially available coated polymer thin window. The main energy range of interest was 0.27 to 2 keV where the detector entrance window (AP3.3) and front contact dominated the overall efficiency. A combination of manufacturer data and assumptions was used to optimise the efficiency model and a mass thickness of 9 μg cm⁻² or a linear thickness of 34 nm was established for the SiO₂ front contact layer, which was in good agreement with data provided by the supplier. The authors' model and approach is of value for SEM-EDS systems but is of less importance for EDXRF for which X-ray energies above 1 keV are of greatest interest and where the bulk silicon and beryllium entrance window thicknesses dominate the detector efficiency. Although SDDs are now very widely deployed in commercial EDXRF and SEM-EDS systems, there are still suppliers and users who prefer cryogenically-cooled Si(Li) and HPSi detectors for EDXRF applications due to their availability in the silicon thickness range of 3 to 5 mm. Such thicker detectors provide much greater stopping power for higher energy K lines of the heavier elements in the Z range 40 to 56 and even for the light REE. Currently, the majority of SDD have a silicon sensor thickness in the range 0.3 to 0.5 mm, although some attempts have been made to use 1 mm thick sensors, so it is interesting to see the work reported by Matsuura,⁹⁰ who simulated the electrical properties of a gated SDD with silicon sensor thickness in the range 0.635 to 1.5 mm. The gated SDD was simulated for an 18 mm² device and the author reported that the new design required only half the reverse bias used in Si(PIN) diode detectors. It will be interesting to see if this design leads to functioning detectors and eventual commercialisation. Despite the incorporation of a state-of-the-art 50 mm² SDD with an energy resolution of 140 eV at 5.9 keV, Scruggs⁹¹ reported the detection limit for Cd in plastic, using the K XRF line, was still better in his commercially available EDXRF small spot analyser when a much thicker 80 mm² Si(Li) detector was used even though its maximum throughput was an order of magnitude lower than the SDD. In other examples where the Si(Li) detector had to operate at higher dead time, the SDD provided slightly better detection limits, despite its lower stopping power.

Although high purity germanium (HPGe) detectors offer even higher X-ray stopping power than silicon-based detectors, they remain on the margins of EDXRF spectrometry due to their challenging manufacture and limitations in terms of their need for cryogenic cooling and spectroscopic complexity compared to silicon-based detectors. The efficiency calibration of a 10 mm thick HPGe detector was reported by Maidana et al.,⁹² who used calibrated radioisotope sources to cover the energy range 10 to 140 keV. The authors found reasonable agreement with a previous literature model and established the thickness of the dead layer to be around 2 μm in the most sensitive part of the detector crystal. Much of the challenge of this work surrounded the behaviour in the periphery of the crystal and in establishing the spectroscopic properties and efficiency of this region, which was also found to be misaligned with the can. A 1 mm diameter pencil beam produced by collimating the sources was most useful in establishing the real effective volume of the detector and this work serves as a reminder to users of such detectors to make sure the collimation is tight enough (and has enough stopping power for the X-rays impinging on it!) only to ensure that the core sensitive area is used in which the fields are consistent and give the best spectroscopic performance. Such detectors are a little like the Universe…little known and surprising at the outer edge. Reflecting the importance of collimation/shielding for large HPGe detectors, Britton et al.⁹³ used MC simulations to establish the most effective designs and materials for shielding and background reduction in higher energy measurements. Although dealing with much higher energies than encountered in typical EDXRF systems, the simulations are of general use and would have value in optimising so-called graded collimation systems where XRF and Compton scatter from materials forming the collimator shield are significant contributors to the system blank that should be minimised to yield the best detection limits. The spectroscopic characteristics of an HPGe detector were reported by Demir and colleagues,⁹⁴ who used four radioisotope sources to cover the extended energy range 13.8 to 1212 keV. The authors made measurements at four source-detector distances and presented a theoretical n efficiency model that matched the data, concluding, rather unsurprisingly, that the detector efficiency reduced with increasing gamma-ray energy.

The low number again this year of publications on cryogenic detectors places them at the margins with other detectors. Carpenter and co-corkers⁹⁵ fabricated arrays of superconducting tunnel junction (STJ) X-ray detector pixels based upon tantalum top electrodes that were thicker than previously reported, ranging in thickness from 165 nm to 500 nm, which extended the energy range up to several keV and improved detector spectroscopic purity. Arrays comprising 36 and 112 pixels were fabricated and had active areas of 1.4 and 4.5 mm² respectively. The 36 pixel array, when operated at ∼100 mK, delivered an impressive energy resolution of between 6.8 and 7.6 eV FWHM at 525 eV for 200 μm² STJ pixels and the individual pixel response was shown to be sufficiently uniform to allow scaling to large array sizes. The energy resolution was reported to be as low as 4.8 eV FWHM for the smaller pixel (138 μm²) STJs. The authors were able to simplify the system through the use of a single ground wire and a prototype 32-channel data acquisition system, both of which are likely to improve the prospects for commercial deployment of these impressive detectors systems – at least into more SR X-ray facilities. Finally, although predominantly recommended for the detection of X-ray energies up to 70 keV, an interesting composite scintillation detector⁹⁶ using both ZnSe(Al) and GSO(Ce) scintillator materials was characterised for use with X-ray tubes operated in the range 40 to 70 kV. The response of each scintillation material was measured separately, which allowed spectra from the composite detector to be used, along with the known adjustable kV of the tube, to perform some degree of dual energy measurements of value in material inspection and low energy medical radiography.

2.6 Quantification and data processing

The first step in quantification from an EDXRF spectrum is that of spectrum processing to correct spectrum background and extract net peak intensities. An interesting chemometric method was developed by Brunetti,⁹⁷ who modified a Genetic Algorithm (GA) by means of a fine-grained approach in order to, reduce the population required and speed up the conventional GA approach. The full details of the new algorithm were described, including the parameters within the pseudo-Voigt function used to model the XRF line shapes. The new GA algorithm was applied to spectra collected using a commercially available Si(PIN) diode X-ray detector from a fairly simple copper–gold alloy containing Cr and Fe and the computed XRF spectrum comprising 32 lines showed very good agreement with the experimental spectrum in the energy range 4 to 16 eV. Although the new algorithm was recommended by the author for its much greater speed than a standard GA, the computation time of some 14 minutes on a 2.53 GHz Intel Centrino 2 processor suggests it is unlikely to supplant existing spectrum processing methods used in commercially available EDXRF instruments. A reminder was provided⁹⁸ that the use of a polycapillary optic does not reduce the potential for diffraction peaks to contaminate EDXRF spectra. The authors produced a table of the primary diffraction peak positions and intensities measured from each of 48 pure element samples, of which only vanadium showed no significant diffraction peaks. The authors warned of the potential pitfalls in falsely identifying peaks because of these diffraction peaks but, sadly, provided no insight into how such peaks might be minimised or compensated for. Of interest to those working with higher energy XRF lines was the analytical model for absorbed energy distribution developed by Yun et al.,⁹⁹ Their detector response model covered the energy range 20 to 140 keV and the authors claimed it could be applied to monolithic and pixellated X-ray detectors and specifically to compound materials such as CdTe used for higher energy X-ray and XRF imaging. The new analytical method was verified using MC simulation but remains to be applied to spectra from actual detectors. A so-called simulated annealing algorithm was proposed¹⁰⁰ in order to improve XRF peak search. The algorithm incorporated an improved peak valley detection method and the solution was computed from both ends of the energy spectrum array in order to achieve a final solution. A conventional peak model comprising a Gaussian peak, a step function and an exponential tail was incorporated into a so-called statistical distribution analytic method¹⁰¹ that was applied to spectra from a Si(PIN) diode detector. This study was extended¹⁰² to include an SDD and relatively good performance was claimed for the extraction of net peak data from the 4.5 to 26 keV energy range in an EDXRF spectrum of a Chinese RM alloy sample.

In fundamental parameter (FP) calculations for modelling and correction of absorption and enhancement effects in XRF quantification, the contribution of secondary electrons to XRF intensities is an important factor, especially for the lower energy lines. Fernandez and co-workers^103,104 used the public-domain PENELOPE MC simulation code to calculate the effect of inner-shell ionisation impacts (ISII) arising from electron impacts from single and multiple scattering of electrons such as K photoelectrons, Auger electrons and Compton scatter electrons. It was demonstrated that the angular distribution of the characteristic photons due to ISII could be safely assumed as isotropic and that the source of photons from electron interactions was well represented by a point source centred at the location of the primary collision. The new FP correction was described using a parametric expression that needed 20 parameters by line (distributed along five energy regions, four parameters each) and covered the energy range 1 to 150 keV and the K, L₁, L₂ and L₃ absorption edges for elements in the Z range 11 to 92. The results of applying the new model showed the perhaps surprisingly large contribution of ISII, which in the case of Al K line XRF at high excitation energies, was up to 16× the contribution of the initial photon excitation. The authors provided a useful insight into the inner workings of the PENELOPE code as well as their own new kernel code and they recommended the use of the new model for monochromatic excitation at energies where the ISII contribution was highest. An interesting iterative approach taken by Schoonjans et al.¹⁰⁵ who combined two open-source modules; the first being PyMca for spectrum processing and the second a MC spectrum simulation program called XMI-MSIM. The PyMca module performed non-linear least squares fitting of the XRF spectrum by means of a pseudo-Voigt peak model and provided an initial estimate of element composition. The XMI-MSIM module then took the composition estimate and used the MC core to estimate XRF and scatter peak intensities. The two modules were used in an iterative manner until the net peak intensities converged. There was the need not only for a complete set of input parameters and experimental description, as would be expected, but also for the specification of balance and measured elements that the SR set-up on beamline L at HASYLAB, Hamburg could not measure due to its monochromatised 16 keV beam and use of a 1 mm thick aluminium primary beam filter. Results were presented for 3 samples, the first of which was a glass RM (NIST SRM 1412) that gave data with an accuracy of 5% relative or better for the elements Ba, Ca, Fe, K, Sr and Zn although data for Cd and Pb were respectively 26 and 17% relative low, the reasons for which were proposed as weak excitation of the L lines and inaccurate L line cross sections respectively. The second sample was a stainless steel RM (NIST SRM 1155) that gave good agreement with certificate values for most elements measured although the inability to measure Mo at 2.41% m/m and having to use it as the balance element is a major drawback – albeit only as a result of the experimental setup and not specifically of the quantitative calculations. The final sample was a nickel–silver rod, for which, rather bizarrely, the incident synchrotron radiation was tuned to 14 keV, which excited only 2 of the three Pb L edges and is far below the Ag K absorption edge. Results for Cu, Ni and Zn at concentrations of 45, 10 and 42% respectively were accurate to better than 3% relative although the result for Pb at 2% was low by 35% relative. The inaccuracy for Pb was blamed on inaccurate L line cross-sections although one can't help but wonder at the effects of using monochromatic 14 keV SR and of having a 10 × 10 μm² beam to analyse a material in which Pb is likely to be inhomogenously distributed. This new approach has merit and the code is available to others to use although there is the need for much expert input at this stage. One can only hope that subsequent work makes use of typical laboratory EDXRF equipment not analytically limited in the way that was the case for the SR source and set-up used in this report. An FP quantitation program called x2abundance, developed for the Indian Chandrayaan-1 Lunar space mission was validated¹⁰⁶ for determination of the composition of elements in the Z range 11 to 30 in lunar surface rocks. Despite the differences from the excitation spectrum expected on the lunar surface, the authors used a monochromatised 8 keV SR beam from beamline BL-16 on the Indus-2 SR source in Indore, India in conjunction with a commercially-available Si(PIN) X-ray detector with an energy resolution of 160 eV at 5.9 keV. Despite the power of the SR source, the detector was reported to limit the element range to Z greater than 13, although the broader-band excitation expected on the Lunar mission was expected to yield improved low-Z sensitivity. Net intensities needed by the authors' program were provided by a venerable (1966 vintage) program using a straightforward Gaussian peak shape for the fitting of all spectrum peaks and incorporating blank corrections for the inadequately shielded stainless steel contamination from the chamber walls. Results from the laboratory set-up were evaluated by comparing the experimental spectra to spectra generated by the GEANT4 MC code into which the compositions calculated by 2× abundance were entered. Good agreement was reported by the authors for the concentrations of 5 elements in pressed pellet samples of a synthetic lunar rock and a terrestrial rock believed to be similar in composition to the lunar highlands. Sadly, although the agreement was OK, it was poor for Al and Si and the experimental precision was dire, being only 1% m/m for the elements Al, Ca, Fe, Si and Ti at measured concentrations of 7, 7.1, 7.5, 19.5 and 1.1% m/m respectively. Given the expense of a Lunar mission, SR beamline time and of developing a new FP method it is disappointing that more effort was not put into the rather straightforward matter of acquiring better experimental data and perhaps saving beamline costs to use a state of the art detector with performance more suited for the lower Z lines that the authors were so concerned to validate their FP method with. Watanabe and colleagues¹⁰⁷ reminded WDXRF users of the obvious necessity for applying loss or gain on ignition corrections when using an FP method for quantitative calculations of the composition of fused glass bead samples. Good agreement with known composition values was reported by the authors for a number of ferroalloys, for which calibration samples are often very hard to find thereby making empirical correction methods more difficult and usually less accurate. For those seeking better values of the fundamental parameters themselves, Caussin¹⁰⁸ developed a program to highlight some areas of difference in the values of mass absorption coefficients in four commonly-used compilations while Demir and Sahin¹⁰⁹ reported measurements in an external magnetic field of parameters such as XRF cross-sections, line ratios and fluorescence yields for 25 elements in the Z range 24 to 65.

Among the many chemometric methods applied to XRF quantification, partial least squares (PLS) regression continues to be explored and an improved approach in which MC simulation was used to generate the training data sets was reported by Rakotondrajoa et al.¹¹⁰ The authors set up a system of standards containing 9 elements (Co, Cu, Fe, Mn, Mo, Sr, Ti, Zn and Zr) and the initial step involved a rough estimate of the element concentration by comparison to a single simulated calibration sample containing 11.11% m/m of all 9 elements and assumed excitation by a SR beam monochromatised to 21.7 keV. No specific spectrum processing or matrix corrections were used for this initial estimate as the PLS process took those automatically into account. An initial training set of 90 spectra, simulated with a MC code, was used to build the PLS model and make initial predictions of the composition of each sample then a second training set comprising 180 MC-simulated spectra covering the concentration ranges estimated from the first set. The estimates from this second model were within around 5% relative of the true values for all 9 elements, which led the authors to claim that PLS regression was a good alternative in XRF spectrometry quantification and much faster than classical FP methods in which many standards are typically measured. It is true that the ability to produce a great number of “calibration” samples using MC simulation for PLS regression is efficient providing sufficient computing power is available, but the 5% relative performance for just 9 elements is hardly stellar and the claimed level of accuracy is only achieved when the unknowns and training set are in the same and rather narrow concentration range – as for conventional empirical XRF quantification methods. The PLS method certainly has merits for the measurement of obscure materials for which no reference samples are available but the intended goal of quantitative analysis with no prior knowledge of the sample is not particularly well demonstrated in this paper and there is some way to go before such an approach would be competitive in, say, a handheld XRF analyser where FP methods are the norm and are very fast in use. An interesting paper¹¹¹ used Soft Clustering methods in order to classify sub-cellular organelles by means of trace element data acquired by XRF microscopy. The authors used very large data sets of trace element XRF images from an SRXRF system on beamline 2-ID-E at APS, Argonne, IL, USA that were too large to permit manual analysis. The samples were human red blood cells infected with the parasite Plasmodium falciparum, the causative agent of malaria and a total of 27 cells were scanned using the SRXRF system. Using a hard clustering, k-means algorithm the authors were unable to identify any organelles from the background while their new soft clustering technique, based upon distribution of Fe and Zn in the images, was able to divide images into regions corresponding to food vacuole, parasite, host, and background for 70% of the cases. Further improvements in the algorithm enabled a 100% success rate for the soft clustering method. The authors recommended their approach as the size and complexity of SRXRF microscopy data sets increases and identified ways in which their approach could be extended and improved using additional chemometric tools. Egan and colleagues¹¹² also used chemometric methods to reduce the size of large XRF imaging data sets and to extract element images from the data. The authors used a conventional high power tungsten target X-ray tube operated at 100 kV in conjunction with an array detector comprising 80 × 80 pixels of a CdTe sensor with energy resolution of 800 eV at 59.5 keV to directly image XRF spectra from a phantom that incorporated lead tape randomly overlaid with tin indium and molybdenum wires. The raw data set of spectra from each pixel was 23 Mbytes but when PCA was applied the data set reduced to just 274 kbytes, a reduction of 98.8%. The PCA approach identified 5 principal components and subsequent analysis of those using a k-means method to assign data into 6 clusters resulted in clear elemental images that showed the distribution of the elements in the phantom. Full details were given of the multi-variate methods used and the authors showed the approach to work for transmitted X-ray images, opening up possibilities for many applications in material analysis and also medical and security X-ray imaging.

A small number of publications on other methods of XRF quantification was produced during the review period including an interesting refinement of the Compton–Rayleigh scatter ratio method¹¹³ that was planned to be used in the Curiosity Mars rover's Mars Science Laboratory (MSL). The authors used data collected from a wide range of geological RMS on a copy of the flight model of the MSL X-ray spectrometer that employed simultaneous XRF and alpha PIXE techniques, with excitation by its six ²⁴⁴Cm sources located in a ring around a shielded central SDD. The authors extended the GUPIX code to include the two excitation modes into the new GUAPX code and established that the best fit to the Rayleigh peak was a Gaussian and for the Compton peak was a pair of Gaussian peaks. The complexity of the Compton scatter peaks from the Pu-L series lines meant that 20 Compton peaks had to be fitted, which is no small matter. The authors established the preferred peak fitting procedure and described the pitfalls of having higher concentrations of Rb and Sr in the samples, where those elements' K lines strongly overlapped a key section of the scatter spectrum, although the low concentration of those elements in Martian rocks was expected not to be a problem for MSL measurements on Mars. Based on measurements from 60 international geological RMs, the best peak fitting and mean Z estimation methodology was established and the results enabled the authors to update their MC simulation program specifically to improve its Compton and Rayleigh peak fitting performance. An interesting method combining XRF and XRD data was described by Mikhailov et al.¹¹⁴ who were able to estimate the C content of steel from the cementite (Fe₃C) concentration from the two types of spectra. The authors claimed that the results for an air-path instrument could be as good as those from a vacuum spectrometer when this approach was adopted. Although the widespread availability of a vacuum spectrometer may not be a problem, the direct determination of C in steel is certainly a sample preparation and XRF measurement challenge that could perhaps be overcome using this method. It is also worth noting the method proposed by Iwata et al.¹¹⁵ for correcting WDXRF line intensity when transforming data from 2 theta scans into energy scans but almost certainly worth avoiding the paper by Ekinci et al.¹¹⁶ who spent valuable time collecting EDXRF spectra only to come up with the conclusion that increasing the measurement time decreased the detection limit, a fact that has been well-established theoretically and experimentally for decades…

3. Applications

3.1 Geological and climate change

In a contribution to the development of XRF techniques, Plessow¹¹⁷ observed that XRF determinations of F appeared to fail systematically in yielding reliable quantitative results. The author suggested that X-ray radiation mobilised F, as well as B and Cl, causing diffusion by radiolysis towards the specimen' surface. The resulting variations in the element's concentration in the analysed layer and consequential specimen instability might preclude XRF for the reliable determination of F in many soil and rock samples. Parsons et al.¹¹⁸ investigated the use of hand-held XRF for the in situ analysis of flood plain soil for As, reporting detection limits of 6.8 mg kg⁻¹ and a precision of 14.4% RSD. The authors reported that the presence of moisture had a significant effect on signal intensity, with an As signal-loss of 37% being observed for a soil containing 20% m/m moisture relative to a dry sample. The direct determination of gypsum in soil using a portable XRF spectrometer was described by Weindorf et al.,¹¹⁹ claiming great promise in this approach, based on the multiple linear regression of laboratory results (by thermogravimetry) and Ca and S determined by PXRF. Alvarez-Vazquez et al.¹²⁰ advocated a method of sample preparation using an ultrasonic probe slurry for the rapid and sensitive analysis of sediments by TXRF spectrometry. The method was evaluated with samples from a coastal area in Northwest Spain (Ria de Arousa) and was validated using several CRMs. Cherkashina et al.¹²¹ also used TXRF to determine Ba, Cs, Pb, Rb and Sr in potassium feldspars based on 0.05 g test portions using either Se as an internal standard or the pre-determined value of Rb. Ross et al.¹²² described a multi-sensor logger for rock cores, tested on diamond drilling samples obtained from the Matagami mining camp (Canada). The instrument was capable of making the following near simultaneous measurements: (i) volumetric magnetic susceptibility, (ii) density based on gamma attenuation, (iii) elemental determination by EDXRF, (iv) VIS/NIR spectrometry for the identification of minerals and (v) line scan continuous imaging. An instrument for in situ microanalysis was described by Yang et al.¹²³ based on a low-power Mo tube, Si(PIN) detector and polycapillary coupling optics and designed for the efficient and accurate identification of minerals. Those involved in the analysis of chromium ores will welcome the preparation of four new reference materials GCr-1 to GCr-4 (ref. 124) in which WDXRF spectrometry was used to evaluate homogeneity.

The XRF technique was used in a substantial number of geochemical investigations during the current review period but there is only space to review a small number of the abstracts available, especially ones judged to demonstrate innovative applications of the technique. Thus, Knowles et al.¹²⁵ used three related SR techniques, μ-XRF spectrometry, XANES and μ-XRD, together with ion milling and TEM analysis to characterise tubular alteration features in sub-seafloor basalts from various locations and ages. They determined the oxidation state of Fe, the coordination state of Fe, Mn and Ti, and major and trace element concentrations at the μm scale. The authors concluded that microscale mineralisation processes were common and consistent throughout ocean basins and time and delineated the sequential stages of fluid-rock interaction. Vidal-Solano et al.¹²⁶ combined laboratory WDXRF with hand-held XRF instrumentation in their study of peralkaline pyroclastic flow deposits of Northwest Mexico. They reported that the analysis of glassy rhyolites provided effective field evidence for identifying and mapping the peralkaline event under investigation. Floor et al.¹²⁷ used TXRF spectrometry to study Se sorption processes in volcanic ash and its interaction with volcanic-derived acid rain; of importance because of the risk to human health given the narrow range between essential and toxic Se concentrations. They reported that the anion content of rain water had an important impact on the mobility of Se in volcanic ash samples and commended the potential of TXRF spectrometry in this and other trace element mobility applications. Another potential hazard from large explosive volcanic eruptions is the release of bromine and chlorine and the threat posed by these elements in the catalytic destruction of ozone in the stratosphere. Kutterolf et al.¹²⁸ investigated this problem using a newly optimised μ-SRXRF facility to examine melt inclusions in volcanic phenocrysts from 14 large explosive volcanic eruptions from Nicaragua, concluding that a single eruption emitted on average 27 kt of Br and that the combined Br and Cl emission have had, and still have the potential to cause substantial depletions of the ozone layer. Synchrotron techniques were also used by Mayhew et al.¹²⁹ (μ-XRF spectrometry and XANES) to investigate the production of hydrogen during the high temperature hydration of mafic and ultramafic rocks, a process that is believed to sustain microbial communities in submarine vents and terrestrial hot-springs. Their laboratory study demonstrated a strong correlation between the generation of molecular hydrogen and the presence of spinel phase-oxide minerals leading to the proposal that the low temperature production of molecular hydrogen involved the transfer of electrons between Fe^II and water absorbed onto spinel surfaces. Winkel et al.¹³⁰ studied the uptake of As by travertines (a form of limestone deposited in hot-springs) formed in a geothermally active region in the Chalkidiki peninsular (northern Greece). Their results, obtained by XAS, μ-XAS and μ-XRF instrumentation, showed that As was present as arsenate (As^V) closely associated with the calcite matrix, indicating that calcite in travernites could sequester at least 25% of aqueous As. As a consequence, travernites could immobilise a substantial proportion of As present in geothermal ground waters. Rango et al.¹³¹ were also interested in the mechanisms for the mobilisation of As, as well as other naturally occurring contaminants (B, F, Mo, U and V) from Quarternary sedimentary aquifers of the Main Ethiopian Rift, and their enrichment in ground water. As part of this study, XRF spectrometry was used to measure the composition of rhyolites (ash, tuff, pumice and ignimbrite) and the Quaternary lacustrine sediments. Their conclusions were that a combination of oxidation state, bicarbonate composition and pH ≥ 8 enhanced the mobilisation of oxyanion-forming elements from iron oxides and consequently the contamination of local ground water.

One of the most popular applications of the XRF technique in the analysis of geological materials is in studies of climate and environmental change over geological time scales, often by the analysis of lake sediment cores and frequently in combination with other measurements that distinguish the stratigraphy. Readers are advised to consult the companion review on geological and environmental materials⁶ to appreciate the full geological significance of these contributions. In this section, a small selection is highlighted to provide examples and illustrate the role of XRF measurements in such studies. Thus, Aquit et al.¹³² used XRF core scanner-derived elemental data, combined with benthic foraminiferal counts, planktonic stratigraphy and stable isotope measurements on bulk carbonate to investigate the Late Cretaceous paleoenvironmental evolution of the Tarfaya Atlantic coast basin, Southwest Morocco. Kanamaru et al.¹³³ provided new insights into varved sediments at the Saanich Inlet (British Columbia, Canada), using an XRF core scanner capable of offering sub-mm resolution, together with SEM backscattered electron images. These data were used to identify nine sedimentary facies that demonstrated seasonal and sub-seasonal variations so contributing to an understanding of paleoclimatic variations. High resolution scanning was also used by Heymann et al.¹³⁴ to provide in situ and continuous determinations of pre-defined element suites on split core samples from Lake Stymphalia sediments (Northeast Peloponnese, Greece). Changes in elemental behaviour were related to hydrological changes in the catchment area, changes in the lake level and rates of evaporation, all linked to the development of the climate of the Eastern Mediterranean in the Late Glacial to Mid-Holocene period. In contrast to these studies, de Angelis et al.¹³⁵ used high resolution μ-SRXRF spectrometry, IC and SEM/TEM data to study bottom ice from the Dome C site in the Antarctic as part of the European Project for Ice Coring. Their interest was in acid-salt interactions, and the presence of mineral dust and organic matter, to better understand ice formation and recrystallisation processes. Turning now to a different sampling medium, Hansson et al.¹³⁶ were interested in the biogeochemical record preserved in peat, especially the links to the carbon cycle and environmental change. As part of their studies, the WDXRF technique was used to analyse samples from three peat hummock cores selected from an ombrotrophic bog in Sweden, with a view to identifying proxies that would demonstrate the degree of decomposition of the peat. Haenssler et al.¹³⁷ had the objective of distinguishing natural and human induced influences on environmental changes preserved in the Holocene sediment sequence from the Etoliko Lagoon (Greece) combining XRF spectrometry, LOI and grain size analysis of sediments with archaeological and historical records. Results demonstrated the geological evolution of the lagoon since 11 670 BP, with the history of human occupation of the area between the Late Helladic (1700–1100 BP) to the end of the Hellenistic Period (30 BP) being distinguished by a sudden increase in coarse sediments.

Regional geology can be delineated by geochemical surveys and remote sensing, illustrated by the work of Saaltink et al.,¹³⁸ whose aim was to identify the geogenic and agricultural controls on the elemental composition of European grazing and agricultural soils, using ICP-MS for trace element and XRF spectrometry for major element determinations based on the GEMAS (geochemical mapping of agricultural and grazing land soils of Europe) data set. They reported that the four prominent factors that affected soils were all geogenic: weathering, reactive iron-aluminium oxide minerals, clay minerals and carbonate minerals, with little difference between grazing and agricultural soils. Thus, geological controls were more important than anthropogenic factors, including the accumulation of phosphate fertiliser. Haest et al.¹³⁹ used hyperspectral visible-near to shortwave infrared airborne data to produce quantitative iron (hydroxy-)oxide AlOH-clay and carbonate abundance maps of the Rocklea Dome in Western Australia. The authors used portable XRF spectrometry as one of the techniques to validate and estimate the error in these airborne-derived data. Addison et al.¹⁴⁰ showed that biogenic sediment in Southeast Alaskan temperate fjords could be used as a sensitive recorder of variability in past productivity, and by inference, past climatic conditions in the high-latitude Gulf of Alaska. Their study was based on a detailed analysis of core samples that included the use of high resolution scanning XRF with conclusions supported by satellite observations from SeaWiFS (Sea-viewing Wide Field-of-view Sensor – a satellite project that provides quantitative data on global ocean bio-optical properties to the Earth science community).

Turning now to the analysis of soils and sediments, Jean-Soro et al.¹⁴¹ undertook a study of the absorption and distribution of Pt in soil and sediment components, specifically kaolinite, hematite and humic acid. Results by μ-XRF spectrometry showed that kaolinite had the highest adsorption capacity for Pt, suggesting that the element was mainly associated with clay in soil; an observation that allowed the impact of the soil or sediment matrix to be assessed on its ability to retain or promote the dispersion of Pt in an urban environment. Burak et al.¹⁴² also used μ-XRF together with XRD spectrometry and optical microscopy to understand the distribution of Pb and Zn in dolomitic and metapelitic soils of the Brazilian Central Plateau. Results of this investigation showed that poorly crystalline Mn-rich material in the <20 μm fraction of the cambisol layer encouraged strong Pb sorption. In contrast, the same fraction in the ferralsol layer contained more Al- and Fe-oxide rich microaggregates, which also encouraged strong metal retention, whilst large sesquioxide contents reduced metal mobility, so limiting the risk of toxicity when such soils were used for agriculture. A hand-held XRF instrument was used by Dahl et al.¹⁴³ to characterise the Mo content of marine sediments under euxinia conditions (i.e., precipitated under anoxic conditions in the presence of hydrogen sulfide). The authors considered that hand-held XRF spectrometry was a quick and reliable method with low analytical costs, minor sample preparation and ease of adaptability to laboratory and field use. Bianchini et al.¹⁴⁴ used new XRF data to create geochemical maps that showed the anomalously high Cr and Ni backgrounds in the fine alluvial river sediments of the River Po (Italy). These were reported to be derived from the weathering of ophiolitic rocks within the river catchment. The authors considered that the oxidation of Cr to the hexavalent form and the higher mobility of Ni could represent a geochemical risk to agricultural activities and the culture of mussels and clams in lagoons within the Po river delta. Arsenic is a redox-sensitive element often associated with iron sulfides and Neumann et al.¹⁴⁵ reported data by the μ-XRF technique on small pyrite framboids present at a high concentration of 7.5% in sediments from the Achterwasser lagoon in the estuary of the river Oder, Southwest Baltic Sea. Variable concentrations of As were found (6 to 1142 mg kg⁻¹) and XANES measurements showed that although As^I usually substituted for S in the pyrite lattice, higher valence states (As^III and As^V) could be found in samples close to the sediment/water interface. Continuing with the adsorption properties of pyrite, Curti et al.¹⁴⁶ were interested in the role of this mineral in reducing the leaching of ⁷⁹Se, a redox sensitive nuclide important in the safety assessment of radioactive waste disposal sites. They used results from μ-XRF, μ-XANES and EXAFS methodologies to show that under nearly anoxic conditions, dissolved SeO₃²⁻ and SeO₄²⁻ sorbed directly onto pyrite surfaces and were subsequently reduced to Se⁰ over an 8 month period, demonstrating the beneficial effect of pyrite in transforming mobile Se compounds into a sparingly soluble form. Zeng et al.¹⁴⁷ investigated S speciation in lake sediments from Prairie Pothole Lakes of the S-enriched wetlands in the glaciated prairie region of North America using XANES and μ-XRF mapping, identifying pyritic S, reduced organic S and oxidised S compounds. They reported that although there was no seasonal variation in total S, during the spring-summer transition, there was a decrease in reduced organic S and a concomitant increase in oxidised S, with a reversal occurring at the summer-autumn transition. The authors commented that these seasonal changes have the potential to force the cycling of other elements such as Hg in the prairie wetlands. An unusual application of XRF core scanner data was described by Fortin et al.,¹⁴⁸ who were interested in novel methods of density determination to support a study of the Laguna Potrok Aike (Patagonia, Argentina) sedimentary record. The coherent/incoherent scatter ratio from an XRF core scanner was considered to offer high resolution, reliability and the advantage of continuous measurement of the variations in density of the sediment profile. In comparison, CT scan measurements were less noisy and provided a precise, fast and cost effective methodology with both approaches offering the advantage of being non-destructive.

In the characterisation of minerals, Gorghinian et al.¹⁴⁹ investigated red gem-quality spinels from Ratnapura (Sri Lanka) using a μ-EDXRF instrument fitted with a Kumakhov polycapillary lens. They reported some variability in the contents of Cr, Fe, Ga, Ti, V and Zn, although the samples were essentially homogeneous. Following a statistical analysis of all the results, the authors proposed that the red colour of these spinel gemstones was caused by the random isovalent substitution of Al by Cr + Fe + V. Vanadium in excess contributed to the hue through Frenkel defects based on the heterovalent substitution of 2Al by Zn + V. Rodina et al.¹⁵⁰ employed the μ-EDXRF and XANES techniques to show that garnet from placer deposits of the Taman Peninsula (southern Russian) were in the form of almandine. Gu et al.¹⁵¹ described a new mineral – ferrisepiolite – discovered in the Saishitang copper skarn deposit in Qinghai Province, China. X-ray fluorescence was one of the techniques used to characterise this mineral (in addition to XRD, EPMA and wet chemistry), which was given the chemical formula: (Fe^III, Fe^II, Mg)₄((SiFe³⁺)₆O₁₅)(O, OH)₂·6H₂O. Hatipoglu et al.¹⁵² were interested in the gemmology and genesis of gem-quality, magenta coloured kammererite ((Mg,Fe,Cr)₅Al(Si₃Al)O₁₀(OH)₈) from the Kesis (Erzincan) and Kop (Erzurum) mountains, Turkey. Using XRF and XRD results, the authors proposed that this mineral was formed from the secondary components of the hydrothermal alteration of amphibole, pyroxene and biotite minerals in the surrounding peridotite and hartzburgite ultrabasic rocks and they provided more details of the properties derived from μ-Raman spectroscopy.

The XRF analysis of fossils remains a relatively unexploited field although a number of contributions are available this year for review. In an imaginatively titled paper, Edwards et al.¹⁵³ considered ‘mapping prehistoric ghosts in the synchrotron’ emphasising the potential of detailed chemical analysis to reveal insights into the composition, preservation and biochemistry of ancient life, especially using SR rapid scanning XRF spectrometry (SRS-XRF). Perhaps one of the more exciting contributions in this field was the work of Manning et al.,¹⁵⁴ who used the SRS-XRF approach, together with XANES (for S) to reveal the plumage patterns in a 150 million year old specimen of Archaeopteryx. The distribution of trace elements and organic S were interpreted by the authors to predict the complete feather pigmentation pattern. Gueriau et al.¹⁵⁵ were concerned about the major challenge that face those who wish to interpret flattened fossil remains where, because of compression, techniques such as X-ray μ-CT is generally inapplicable in characterising internal structures. However, the authors presented results using SR X-ray spectral raster scanning to discriminate skeletal remains of exceptionally well preserved fossils from the Late Cretaceous in which, as well as the alkaline elements and P, the REE distribution pattern was critical in opening up new avenues for both fossil description and in contributing to paleoenvironmental and taphonomical studies. The μ-XRF technique was used by Yoshimura et al.¹⁵⁶ to image Fe, Mg, O, P and Sr in a fossil specimen of the cold water coral Desmophyllum sp. from the surface sediments of the Northwest Pacific. The authors considered this work established this non-destructive technique as being highly suitable for pre-screening samples in cold water sclerochronology studies. Continuing with the analysis of coral, Nguyen et al.¹⁵⁷ reported that the distribution of Mg, P and S in the axial skeleton of the Japanese red coral, Paracorallium japonicum, was linked to the dark/light bands with P and S being distributed in the organic rich dark bands and Mg in the light bands of the annual growth rings. This study was based on μ-XRF and XANES measurements. Qu et al.¹⁵⁸ conducted elemental mapping of the dental enamel of Gigantopithecus blacki (n = 3) using SRXRF spectroscopy to understand the influence of dietary variation during the time of tooth eruption. To account for the effects of diagenesis on the variation of elements in these fossil teeth, the Fe and Mn elemental distribution and levels in dental enamel of G. blacki were compared with that of a single modern pig tooth and no differences were found. The variations of Ca, REE and Sr elemental distribution in the incremental lines revealed that the plant foods utilised by G. blacki from the early Pleistocene or the middle Pleistocene had varied during the formation of dental enamel, possibly caused by the change of living environment or food resources.

3.2 Industrial minerals and consequences from mining activity

The characterisation of ore deposits during the current review period was illustrated by the work of Lund et al.,¹⁵⁹ who described a practical way of quantifying minerals from chemical assays at the Malmberget iron ore operation, Sweden. Their aim was to capture geological and metallurgical processing information and create a predictive model of mineral processing characteristics using EPMA, XRF results and data on the ferro-magnetic composition to calculate modal mineralogy and ore classification. A multi-technique approach involving the use of XRF results was used by de Oliveira et al.¹⁶⁰ to develop a detailed model of the bauxite deposit of Barro Alto, Central Brazil describing the evolution of eight different facies. Pazand et al.¹⁶¹ used the XRF technique (for the majors) and ICP-MS (for the traces) to prepare a distribution map of lithogeochemical and alteration indices for the copper mineralisation in the Sonajil area, Northwest Iran. The distribution map was based on 1248 sampling squares, each 100 × 100 m² extending over this hydrothermally altered Cu mineralised area. In the study of fluid inclusions from two quartz samples of the McArthur River and Rabbit Lake unconformity-related uranium deposit (Athabasca Basin, Canada), Richard et al.¹⁶² used the SRXRF and XANES techniques to both assess the detailed chemistry of the fluids that transported the uranium and measure the speciation of uranium in these fluids. The authors identified three sources contributing to the chemistry and reported that uranium was present and had remained in the form U^VI.

In the study of other minerals of economic interest, Nzeugang et al.¹⁶³ used the XRF technique, together with XRD, FTIR spectroscopy and DTA/DTG to assess the economic potential of clayey deposits of Nanga-Eboko, Central Cameroon. The authors reported that the firing properties of bricks made from these clays was good, despite high shrinkage values, but that they are not suitable for building construction owing to their fine grain size and high plasticity. The geochemical and mineralogical properties of roofing slate deposits of the Iberian Peninsula were investigated by Cardenes et al.¹⁶⁴ using the XRF and XRD techniques to analyse samples from most of the active quarries in the region. Differences were found in major and trace element contents and these data were of value in prospecting for new sources and in architectural conservation to identify the original source of slates. Masoud et al.¹⁶⁵ used XRF spectrometry as one of the techniques to characterise the El-Tih (Egypt) kaolin deposit. The results allowed a comparison to be undertaken with world-wide kaolin occurrences and suggested the suitability of the studied kaolin for paper coating and filling and in high-grade ceramics, after the removal of free iron and titanium oxides. Claiming that Italy holds a leadership in ceramics and glass, Langella et al.¹⁶⁶ undertook a detailed study of Campanian ignimbrite (Italy) to describe the role of temperature, alkaline and alkaline-earth cations in the zeolitisation process and the consequent crystallisation of phillipsite, chabazite and analcime. As well as the interpretation of the post-depositional mineral forming processes, this study was designed to assess the potential of the deposit for industrial use.

Contamination caused by mining waste is a common problem in many parts of the world. As a contribution to this problem, Weindorf et al.¹⁶⁷ used portable XRF instrumentation to quantify metal contamination from mining/smelting operations in soils from various land use types in Zlatna, Romania. They found that the elements As, Co, Cu and Pb exceeded government action limits at over 50% of the sites investigated and commended the speed and accuracy of the portable XRF approach, compared with laboratory determinations by ICP-OES. Krishna et al.¹⁶⁸ used WDXRF spectrometry to assess the heavy metal contamination in soils around different active and abandoned chromite mines at Nuggihalli, Karnataka, India. Soil quality guideline limits were found to be exceeded for Co, Cr and Ni in the study area indicating the need for control measures and remediation. The bioavailability of Ba to plants and invertebrates in soils contaminated by barite mining activities was investigated by Lamb et al.¹⁶⁹ using XRF spectrometry to demonstrate Ba concentrations in soil in the range 0.13–29.2% m/m. The authors presented the results of phytotoxicity and body weight loss effects in selected organisms. Although black shales are a naturally occurring geological rock type, soils derived from these facies were shown by Xu et al.¹⁷⁰ to be highly enriched in the heavy metals Ba, Cu, Ni, Pb, U, V and Zn. The authors used XRF, XRD and ICP-MS techniques in this study and proposed that the soils were formed in a two stage process involving first weathering and then pedogenesis. Often, there are few good news stories in this section, but one that might inspire future contributions was published by Wang et al.¹⁷¹ who tackled the problem of the high levels of phosphate present in waste waters from the dairy industry. Their solution was to investigate the use of acid mine drainage sludge and coal fly ash to facilitate its removal, using a number of techniques, including XRF to investigate the optimum conditions for the adsorption of phosphate.

3.3 Industrial materials and consequences of industrial activity

The emerging field regarding biomass material and combustion is reflected by the number of publication in this year's review. A quantitative multi-element method using a 4 kW WDXRF spectrometer equipped with a commercial pre-calibrated method was evaluated by Andersen et al.¹⁷² for the determination of the inorganic elemental composition of biomass. Elements of interest were Al, As, Ba, Ca, Cl, Cr, Cu, Fe, K, Mg, Mn, Mo, Na, Ni, P, Pb, Rb, S, Sr, V and Zn. The relative systematic error, determined by analysing 13 CRMs of diverse vegetable/plant origin, was typically better than ±20% for elements in the range of 25 to 100 mg kg⁻¹, better than ±15% in the range of 100 to 1000 mg kg⁻¹, and better than ±10% for concentrations above 1000 mg kg⁻¹. The relative precision was better than ±5% for concentrations higher than 25 mg kg⁻¹. The choice of matrix composition in the matrix correction model and the influence of sample moisture and sample grain-size were also addressed. Haykiri-Acma et al.¹⁷³ focussed on the characterisation of the chemically isolated parts of biomass materials, i.e. the isolated holocellulose (hemicellulose + cellulose), lignin and extractives-free bulk, by XRD and XRF spectrometry, and the results were compared with those of the parent biomass species. It was found that the isolated holocelluloses and lignins still contained ash contents up to 2.2% and 4.0%, respectively and that various minerals were found to survive after chemical treatment applied during isolation. In addition, several heavy metals were also detected. These results revealed that minerals cannot be eliminated entirely because of the nature of the chemicals used, and they unavoidably remained in the isolated macromolecules. Kirnbauer et al.¹⁷⁴ analysed ash components of biomass fuel by XRF spectrometry and thermal analyses to improve the ash handling operation of dual fluidised bed gasification plants, to reduce tendencies for fouling, slagging, and bed material agglomeration. Mass balances of inorganic matter were presented, evaluating different loop configurations. They showed a flow of potassium oxide from the combustion reactor to the gasification reactor, which led to unexpected high potassium concentrations in the fly char. A reduction of ash loops reduced the amount of potassium transferred from the combustion reactor to the gasification reactor. Singh et al.¹⁷⁵ assessed the potential slagging and fouling behaviour of the ash produced from co-combustion of a South African coal with waste tyre rubber (WTR) along with pure fired WTR. Zinc oxide is incorporated during the manufacture of tyres as a compounding additive and was also present within raw WTR. The XRF and ICP-OES analyses of the resultant ashes revealed a high composition of acidic oxides. However, it should be noted that ZnO/Zn was not likely to contribute significantly to the slagging/fouling mechanisms, due to its volatile nature.

The spatial distribution and speciation of Zn in basic oxygen furnace sludge, a metallurgical residue in the steelmaking industry, was investigated by Wang et al.¹⁷⁶ using μ-SRXRF and μ-XANES spectroscopy. The XRF analysis showed that Zn was distributed in two ways. One was associated with Fe, and its distribution showed a positive correlation with that of Fe. The other was accumulated in some well-defined hot spots showing a negative correlation with that of Fe. The XANES results indicated that ZnFe₂O₄ was the main constituent in the hot spots, whereas in other cases, Zn was mainly in the form of zinc carbonate. Nagoshi et al.¹⁷⁷ used the XAFS technique in XRF mode for the quantitative analysis of steel sheets containing 32 up to 2100 mg kg⁻¹ of Nb. The concentration of Nb in the samples was determined with high accuracy from the height of the K-edge jump. The limit of detection was estimated to be down to the single digit ppm range. Liu et al.¹⁷⁸ investigated the corrosion behaviour of Ni-based alloys in molten fluoride salt using SRXRF and SRXRD spectrometry. Results showed that the main depleted alloying element of Ni-based alloys in molten fluoride salt was Cr. In addition, the results indicated that Mo could enhance the corrosion resistance in molten FLiNaK salts. Higher-content Mo and lower-content Cr in Hastelloy C-276 alloy were responsible for the better anti-corrosive performance, compared to the Inconel 600 and Hastelloy X alloys.

Determination of the level of chlorine in crude oil and petroleum products is made on a routine base to guarantee that the chlorine present does not cause damage in oil processing equipment. Doyle et al.¹⁷⁹ presented a review of the current status of spectrometric methods, especially XRF spectrometry, for this purpose. Comparison of the performance of these methods, limitations and potential new approaches to ensure the proper spectrometric determination of chlorine was considered. Gazulla et al.¹⁸⁰ recognised the strength of the WDXRF method as a rapid and accurate method for the determination of elements in different materials over a very wide range of concentrations. In liquid petroleum products, the elements Al, Ba, Ca, Cr, Cu, Fe, Mg, Mn, Ni, P, Pb, S, Si, Ti, V and Zn were validated by analysing reference materials. The limits of detection and quantification, as well as the measurement uncertainty were calculated. Wang et al.¹⁸¹ also used WDXRF spectroscopy to determine additive elements in 98 engine oil samples of different origin. Variations in the ratios of Zn/P, S/Mo and Ca/Mg illustrated the correlation between experimental data and additive contents. The experiments indicated that Ca, P, S and Zn were the main elements of engine oil, while Mg and Mo existed in high quality engine oil, and Na, Nb, Ti and W were present in individual brands of engine oil. Scale deposits are the most common and most troublesome damage problems in the oil field. In severe conditions, scale creates a significant restriction, or even a plug, in the production tubing. Candeias et al.¹⁸² conducted a study to qualify the elements present in scale samples and quantify the thickness of the scale layer using μ-SRXRF spectrometry and computed radiography techniques. The μ-SRXRF results showed that the elements found in the scale samples were Ba, Ca, Cr, Fe, S and Sr, while the computed radiography analysis showed that the thickness of the scale layer could be identified and quantified with accuracy.

To investigate the origin of plastic material causing deposition and blocking in engines and pipelines, a WDXRF method was established by Lai et al.¹⁸³ to analyse simultaneously low-Z elements (C, N, O) and Al, Fe, Si in polyamide 6. Polyamide 6 particles were firstly comminuted to a uniform powder in liquid nitrogen, and then mixed with inorganic standard powders (Al₂O₃, Fe₂O₃, Na₂SiO₃ and SiO₂) and analysed as pressed powder. The limits of detection of Al, C, Fe, N, O and Si were 6.2, 249, 1.8, 120, 101 and 3.3 mg kg⁻¹, respectively. Moreover, a good correlation was observed between the WDXRF results and measurements by ICP-OES and the element analyser technique. Domingos et al.¹⁸⁴ studied the influence of temperature on the ageing of piping made from polyamide 11 (PA-11). The content of the additive in PA-11 was monitored from EDXRF measurements where the abundance of the S line was directly related to the ageing time, agreeing with the TGA analysis results. Chemometric handling of the XRF data allowed the classification of the PA-11 samples according to the amount of additive and weight average molar mass change, predicting the ageing time, and viscosity values of PA-11. Necemer et al.¹⁸⁵ also evaluated the EDXRF technique as a fast tool to classify different types of unknown plastics. The advantages were demonstrated using some plastic artifacts archived in the Museum and the XRF results were compared with the values gathered by other commonly applied analytical techniques.

As a consequence of several European Union Directives, there is an increasing need for rapid and easy-to-handle screening tools for compliance testing. Vanhoof et al.¹⁸⁶ conducted a field trial to evaluate the capability of portable/handheld XRF systems for screening waste loads for hazardous substances as part of the incoming inspection at waste handling plants. Six different waste materials (construction waste, shredder material, contaminated soil, waste wood, Pb granulate and slag from a municipal incineration waste) were analysed by eight different XRF instruments. The results proved that the degree of heterogeneity of the material strongly influenced the dispersion of the results of repeated measurements. The results proved that for specific samples (e.g., organic matrices, elevated contaminant concentrations and high water content), an appropriate adjustment of the matrix effects and spectral overlap corrections was required to improve the quality of the XRF results. In the overall assessment, more than 80% of the obtained results fell within a predefined range of accuracy. For compliance testing of the trace elements Br, Cd, Cr, Hg and Pb with the RoHS Directives, Rackwitz et al.¹⁸⁷ positively evaluated the analytical performance of μ-XRF with SEM-EDS based on a set of ABS reference materials. Whereas the heavy metals were more favourably detected by μ-XRF spectrometry, higher sensitivities for the lighter elements could be attained by ED-EPMA. The limits of detection were estimated based on a physical model for the calculation of μ-XRF spectra and validated.

X-ray fluorescence spectrometry is a well-established technique for the characterisation of materials, most of them related to recycling processes. A quantitative method for the determination of Ag, Au, Co, Ni and Pd in scrapped printed-circuit-board ash by XRF spectrometry was developed by Hirokawa et al.¹⁸⁸ Calibration standards were prepared by adding the appropriate amounts of NiO powder and aqueous standard solutions containing Ag, Au, Co and Pd to the base mixtures of Al₂O₃, CaCO₃, CuO, Fe₂O₃ and SiO₂ as a matrix. For XRF analysis, powder samples were compacted into a hole (12.0 mm diameter and 5.0 mm height) in an acrylic plate and covered with a 6 μm thickness of polypropylene film. Matrix effects were corrected using the intensity value of Compton scattering for Pd Kα, Ag Kα, and Au Lβ₂, and that of background scattering at 35.8° (2θ) for Co K and Ni K. The limits of detection corresponding to three times the standard deviation of the blank intensity ranged from 2.5 to 45 mg kg⁻¹. The obtained concentrations of Ag, Co, Ni and Pd were approximately the same as those resulting from an AAS analysis. Roth et al.¹⁸⁹ identified noble, precious and rare metals, as well as environmentally relevant elements in 31 printed wiring boards by EDXRF spectrometry. Moreover Pb and the flame retardant indicator Br were localised by means of μ-EDXRF spectrometry. Complementary, a GC-MS procedure was developed to identify and quantify the flame retardant substances. Altogether, a wide variety of elements of concern, and with both halogenated- and phosphate-based flame retardants being found in the investigated wiring boards. Arancibia et al.¹⁹⁰ presented a study regarding the use of tailings from mining activities as raw materials for the manufacture of glass. This procedure aimed to contribute to the environmental remediation of mining areas through the vitrification of hazardous waste, as an alternative method of stabilisation. The selected granitic raw material was shown by the XRF technique to contain high levels of heavy metals. By adding CaCO₃ and Na₂CO₃, this waste material was made into a silica glass. Field portable XRF spectrometry was used by Sandhu et al.¹⁹¹ to determine metal and metalloid concentrations in domestic and imported glass beads used for highway marking. For domestic (US-manufactured) batches, average concentrations of 8 mg kg⁻¹ for As, 23 mg kg⁻¹ for Pb, and 55 mg kg⁻¹ for Sb were observed. On the other hand, imported batches were found to have averages of 485 mg kg⁻¹ for As, 97 mg kg⁻¹ for Pb, and 106 mg kg⁻¹ for Sb. Owing to the wide variability in concentrations within the imported batches, the authors thought that additional studies were warranted to evaluate the potential for leaching of metals. Gullayanon et al.¹⁹² used a portable XRF instrument as a non-destructive tool to determine rapidly the fluorochemical concentration of carpet fibres. As fluorine was not directly measurable by XRF spectrometry, fluorochemicals were combined with chemical tracers such as Rb. The tracer concentration was used as a measure for the F level on carpet fibers, which in turn represented the fluorochemical concentration. Results showed that an accuracy of 150 ± 50 mg kg⁻¹ of F could be achieved using a portable XRF instrument. Finally, the recycling of the rare earth oxide from spent rechargeable batteries using metallurgical slags was described by Tang et al.¹⁹³ After applying a high temperature process in combination with several types of separators, XRF analyses indicated that the heat-treated materials consisted mainly of nickel, rare earth and cobalt oxides. The valuable rare earth oxides were further recovered by a high-temperature slagging treatment, whereby over 98% of nickel and cobalt oxides were reduced to the metal phase while almost all rare earth oxides remained in the molten slags. The matrix of the slag phase was Re₂O₃ deficient, typically being less than 5% m/m, which indicated the need to further develop the high-temperature process of concentrating this oxide in slags.

As with many cities all over the world with active industrial developments, the city of Penang in Malaysia has been exposed to industrial pollution. The determination of trace elemental levels in hair is a well-known method for environmental exposure monitoring. Therefore, Aldroobi et al.¹⁹⁴ determined the concentration of As and Hg in the human scalp hair of 100 residents of Penang using XRF spectrometry. The results of this study were compared with the results obtained in other cities where such measurements have also been carried out. The distribution of heavy metal pollution in surface soil at an electronic-waste recycling site near Metro Manilla in the Philippines was studied by Fujimori and Takigami¹⁹⁵ using a field-portable XRF instrument. The polluting elements (Cu, Pb and Zn) were distributed discretely in surface soil and showed similar concentration gradients. The study revealed that the heavy metals remained inside the original site where they might directly influence the health of the workers rather than the surrounding population. The marine ecosystem, which forms part of the world's environment, had been subject to impact of industrialisation and urbanisation. A study was carried out by Yadav and Jha¹⁹⁶ in the creek ecosystem in Mumbai which receives the effluents from industrial and urban settlements. Trace elements were analysed in the creek water collected from each of the contributors and analysed by TXRFspectrometry. It was found that the concentration of the elements As, Ca, Co, Cu, Fe, Mn, P, Pb and Zn were higher on the industrial side of the creek whereas Cr, Ni and V concentrations were higher on the side receiving city effluents.

At the Atomic Weapons Establishment (UK) preliminary development work using WDXRF spectrometry and ICP-OES was performed by King et al.,¹⁹⁷ to measure the chemical yield for samples containing fission products in a thermal neutron irradiated uranium sample. To be able to quantify the results after the separation of radionuclides, known amounts of either radioactive tracers or inactive carrier solutions were added at the start of the separation and measured at the end of the analysis to calculate the chemical yield of the element or isotope. In addition, the XRF technique in combination with other analytical tools was routinely applied to characterise various types of catalysts, modified clay minerals including zeolites, ceramics, slags and other novel materials. As papers dealing with this topic are focused on the industrial process rather than on the XRF technology, they were not considered in this review but interested readers should consult our companion review of advances in the analysis of metals, chemicals and materials.⁵

3.4 Environmental

Synchrotron-based XRF techniques continue to feature in reports from researchers seeking to understand the spatial distribution and speciation of elements within plants. The implementation of new algorithms and detection technologies have offered a route to more accurate analyses of the complex matrices found in plant materials. Terzano et al.¹⁹⁸ imaged and quantified the distribution of iron within tomato roots using confocal μ-SRXRF spectrometry to measure Fe concentrations ranging from a few hundred μg kg⁻¹ to several hundred mg kg⁻¹ at the microscopic level without the need to cut sections. Furthermore, semi-quantitative iron distribution maps were obtained using two geometrically opposing detectors to collect simultaneously the XRF radiation emerging from both sides of an intact cucumber leaf. Lindblom et al.¹⁹⁹ studied the influence of microbial associations on Se in the roots of Astagalus and Stanleya hyperaccumulators. The main form of selenium in both species collected from naturally seleniferous soil was an organic C–Se–C compound, which the authors thought most likely to be methyl-selenocysteine. In addition, surprisingly high fractions (up to 35%) of elemental Se (Se–O) were found, a form thus far not reported in plants but commonly produced by Se-tolerant bacteria and fungi. Sedum alfredi is one of the few species known to hyperaccumulate zinc and cadmium. Lu et al.²⁰⁰ revealed enhanced remobilisation of Zn by phloem transport in this sedum from mature leaves to new growing leaves. Zeng et al.²⁰¹ used the SRXRF and XANES techniques to analyse the S concentration and speciation in mature camphor tree leaves grown in Shanghai, China to explore the relationship between the sulfur nutrient cycle in the leaves, the atmosphere and the soil. Moving from urban to suburban locations, sulfate in the camphor tree leaves decreased as the atmospheric pollution reduced, but the authors measured a dramatic increase near the seashore, where marine sulfate emissions and pollution caused by maritime activity was significant. Feng et al.²⁰² used μ-SRXRF and XANES spectrometry to investigate Pb accumulation in the roots of Typha latifolia and rhizosphere soils collected from a brownfield site in New Jersey, USA and the role of iron plaque in controlling the Pb uptake. The sequestration of lead by T. latifolia roots suggested a potential low-cost remediation method to manage Pb-contaminated sediments for brownfield remediation while performing wetland rehabilitation. Wu et al.²⁰³ studied the distribution of metals in tissue cross sections of pak choi under stress from elevated levels of chromium and lead. The SRXRF microprobe technique was able to demonstrate that the pak choi used cell walls and vacuoles to reduce the transport of these two metals through the plant, as well as to restrict transport from root to stem. Koren et al.²⁰⁴ used μ-SRXRF imaging, and Cd K-edge XAFS and Cd L-3 edge μ-XANES measurements to compare Cd uptake and plant fitness after CdCl₂ and CdSO₄ treatments to show how the Cd ligand environment in Noccaea praecox depended on the cadmium salt type and the concentration added to the nutrient solution. Song et al.²⁰⁵ were interested in the spatial distribution and speciation of copper in root tips of cucumber revealed by μ-SRXRF and μ-XANES spectrometries. The highest content of copper was found in the root cap and meristematic zone whereas low copper content was reported in the elongation and maturation zone. The results suggested that copper was chelated by sulfur ligands in the cell walls which protected protoplasm against possible damage caused by copper excess. A better understanding of the combined pollutants, mercury and selenium was offered by Zhao and colleagues,²⁰⁶ who used synchrotron mapping to show that selenium was mainly concentrated in the stele of the roots, bulbs and veins of garlic leaves. Using XANES facilities, mercury was found to decrease the content of inorganic Se mainly in the SeO₃²⁻ form in garlic while increasing the content of organic Se (mainly C–Se–C) thereby confirming the ability of mercury to mediate changes in Se species as well as reducing Se accumulation in the plant. The combination of μ-SRXRF spectrometry and open-access, online databases of plant molecular genetics enabled Punshon et al.²⁰⁷ to design three experiments where: a gene of interest had been identified but had an unknown phenotype, an unidentified gene was associated with a known phenotype and, finally, a screening approach, where both gene and phenotype were unknown. The authors described methodological details using elemental imaging to aid or accelerate gene function characterisation by narrowing down the search for candidate genes to the tissues in which elemental distributions were altered. The work used an Arabidopsis plant in an hydrated state. Another paper by the same lead author²⁰⁸ contrasted calcium localisation and speciation in leaves of the Medicago truncatula mutant cod5 with its wild-type form. Wang et al.²⁰⁹ used synchrotron-based XRF microscopy to examine metal(loids) in hydrated cowpea roots exposed to As, Cu, Hg, Mn, Ni, Se and Zn. Development of a mathematical model enabled in situ quantitative determination of root tissues. The differences in metal(loid) spatial distribution provided valuable data on physiology, including uptake and toxicity in the plant roots. French researchers²¹⁰ confirmed that μ-SRXRF mapping offered the high resolution required to track phosphorus and sulfur in single starch granules in their quest to understand the structure and properties of potato and maize.

The unique potential of nano-scale elemental imaging of major, minor and trace level elemental distributions within biological tissue sections of the eco-toxicological model organism Daphnia magna was demonstrated by De Samber and colleagues,²¹¹ who used the nano-SRXRF spectrometer at the European Synchrotron Radiation Facility ID22NI beamline. A highly specialised sample preparation method coupled with the high spatial resolution (∼180 nm) and high flux (6 × 10¹¹ photons s⁻¹) proved to be essential for the visualisation of trace metal distributions (Ca, Fe and Zn) at the sub-micrometer level within the branchial sacs on the thoracic appendages (epipodites) of D. magna. These were the osmo-regulatory regions where ion exchange occurred. Servin et al.²¹² studied possible pathways of TiO₂ nanoparticle (NP) transfer from soil into the food chain using the popular garden vegetable, cucumber. Using μ-SRXRF and μ-XANES spectrometries, the authors were able to show root-to-fruit translocation of TiO₂ in the cucumber without biotransformation. This suggested that TiO₂ could be introduced into the food chain with unknown consequences. Zhao et al.²¹³ were also troubled by food crops exposed to engineered nanoparticles. In their study, cucumber plants were grown to maturity in soil amended with either CeO₂ or ZnO NPs at concentrations of 0, 400 and 800 mg kg⁻¹. Chlorophyll and gas exchange were monitored, and physical markers recorded. Neither NPs were found to impact the cucumber plant growth, gas exchange nor chlorophyll content, however, at 800 mg kg⁻¹ treatment, CeO₂ NPs reduced the yield by 31.6% compared with the control. The μ-SRXRF imaging showed Ce in the leaf vein to be vascular, suggesting that Ce moved between tissues with water flow during transpiration. Rico et al.²¹⁴ were also interested in CeO₂ NPs, this time on their ability to modify the anti-oxidative stress enzyme activities and macromolecular composition in rice seedlings. Synchrotron μ-XRF spectrometry confirmed the presence of Ce in the vascular tissues of roots. Manceau et al.²¹⁵ were interested in how root cell walls accumulate metal cations both during acquisition from the environment and removal from the protoplast thereby avoiding toxicity. Their study, which combined SRXRF, XANES and EXAFS techniques at the nano, micro and bulk scales, identified how copper was bound to cell walls of intact roots of Thlaspi arvense.

Iron insufficiency in human diets is a problem in many parts of the world. Thus, Singh and colleagues²¹⁶ studied the localisation of iron and phosphorus in grain tissues of wheat genotypes with contrasting grain iron content. Using μ-SRXRF, μ-XANES and μ-PIXE configurations, they were able to reveal shifts in iron distribution from maternal to filial tissues of grains during the evolution of wheat from its wild relatives to the present day cultivated varieties. Similarly, Neal et al.²¹⁷ reported that metal distribution and complexation in mature wheat grains were different from that in the wild-type grain. This may explain why the raised levels of minerals, transported to a developing grain, accumulate within the crease region of transgenic grains. Concentrations of minerals in whole rice grains, hulls, brown rice, bran and polished rice were reported by Lu et al.²¹⁸ Using μ-SRXRF spectrometry, the in vivo mineral distribution patterns in the grains were seen to shift during the process of germination. The authors showed that half of the total Zn, two thirds of the total Fe and most of the total Ca, K and Mn were removed by the milling process, if the hull and bran were thoroughly polished. Mobilisation of the minerals from specific seed locations during germination was also element specific. It is good to know that a preparation method for a reference material has been published²¹⁹ for the determination of cadmium in rice grain by XRF spectrometry. The base white rice grain was immersed in methanol containing the appropriate amount of cadmium and then heated and cooled. The RM was packed in a polyethylene cup (32 mm internal diameter and 23 mm high) covered with a 6 μm polypropylene film ready for XRF analysis. A Cd Kα calibration curve constructed with a suite of Cd-containing rice grains showed good linearity (r = 0.996) in the range 0.50–0.98 mg kg⁻¹, whose Cd concentrations were determine by AAS using HNO₃–H₂SO₄ decomposition after the XRF measurements. The lower limit of detection for Cd was 0.13 mg kg⁻¹.

Turning now to analysis using TXRF spectrometry, Patz et al.²²⁰ measured the Mn content in pineapple fruit and pineapple juice. The Mn concentration in the whole fruit was reported to be highly variable, however, inside an individual fruit there was an increase in the Mn content from the fruit flesh to the skin of each pineapple analysed. Looking at industrially processed pineapple products; purees (n = 12) had the lowest Mn content with a median of 4.9 mg kg⁻¹ compared with pineapple juices (10.5 mg kg⁻¹, n = 17) and concentrates (15.5 mg kg⁻¹, n = 48). Depending on the concentration of the product, the tolerable daily intake of manganese could be reached with only one glass (200 ml) of pineapple juice. Boldrin et al.²²¹ evaluated the effect of different selenium application forms and sources on rice growth, grain yield and rice Se accumulation, as well as the content of B, Cu, Fe, Mg, Mn, N, P, S and Zn in rice grains. The TXRF data for Se in the rice plants showed that selenate application to soil was more effective for shoot dry matter production and grain Se accumulation than selenite. Foliar application of both selenate and selenite increased grain yield.

Biomonitoring of environmental pollution using tree rings continues to feature in the literature. Geraldo and colleagues²²² knew that changes in the environment were often recorded as impressions in the wood. Their present study examined the growth rings of Tipuana tipu, a member of the Leguminosae family that is native to Argentina and Bolivia and was introduced into Brazil as an ornamental plant. As T. Tipu is one of the most common trees in the urban landscape in Sao Paulo city, the authors believed it would provide an accurate reflection of environmental changes. Tree ring samples were collected from a variety of locations in Sao Paulo city for analysis using SRXRF spectrometry. Samples collected from the university campus showed highest toxicity, with concentrations above the tolerable limit for the elements: Cr, Cu and Pb. However, a study in Santorini, Greece²²³ warned that tree ring dating is problematic when the tree in question is an olive. The authors warned that the determination of the number of tree rings was impossible due to intra annual wood density fluctuations, variability in tree-ring boundary structure and restriction of the tree's cambial activity to shifting sectors of the circumference, causing the tree-ring sequences along radii of the same cross section to differ.

Municipal solid waste has been dumped at a site on the fringe of Kolkata, India since the middle of the 19^th century. Gupta et al.²²⁴ studied the way soil in this area had evolved over time. Samples from three different layers, from six areas were collected for analysis, firstly by a radio-isotope sourced EDXRF system followed by a commercial polarised EDXRF spectrometer. For the first instrumental configuration, a back-scatter fundamental parameter algorithm was used with readily available single element foils augmented with compound pellets as calibrants for elemental quantification. In contrast, the more sensitive polarised spectrometer employed influence coefficients which ideally required similar matrix geological soil reference materials for the calibration. In the absence of an optimum number of soil standards, the previously used back-scatter fundamental parameters proved useful in covering the dynamic concentration range of certain heavy metal analytes (Cu, Pb and Zn) in the soil samples. The NIST SRM 2586 soil contaminated with Pb was used for quality control. The authors reported potentially toxic metal concentrations at all the sampling levels (surface, middle and lower) with a general trend to higher accumulation of Sb and Pb at the lowest level. During treatment for potable use, water utilities generate arsenic-bearing ferric wastes that are subsequently dispatched to landfills. The bio-geochemical weathering of these residuals in mature landfills are known to affect the potential mobilisation of sorbed arsenic species via desorption from solids subjected to phase transformations driven by abundant organic matter and bacterial activity. Root et al.²²⁵ devised laboratory scale columns to simulate As^V retention in such landfill leachates. After 300 days, ferric sorbents were reductively dissolved. Fluorescence-X-ray absorption spectroscopy showed that As was reduced from As^V to As^III and Fe^III sorbent was transformed to siderite and green rust. The authors also demonstrated that sulfur chemistry exerts strong control over potential mobilisation of As from ferric sorbent residuals by controlling the co-precipitation of secondary arsenic and iron sulfide. Similarly, Langner et al.²²⁶ reported the spatial distribution and speciation of As in peat using μ-XRF and XAS systems. Their findings suggested an authigenic formation of realgar and arsenopyrite in strongly reducing microenvironments in the peat. Schwarz et al.²²⁷ compared three empirically based, spatially explicit predictive models of residual soil Pb concentrations in Baltimore, USA to understand the element's known variability within cities. Sampling 61 residential properties using field portable XRF spectrometry revealed that 53% had soil Pb that exceeded the US EPA reportable limit of 400 μg g⁻¹. These data were then used as the input to three models: a traditional general linear model (GLM), and two machine learning techniques; classification and regression trees (CART) and Random Forests (RF). The GLM revealed that housing age, distance to road, distance to building and the interaction between variables explained 38% variation in the XRF data. The CART model also confirmed the importance of these variables. Using the same three predictor variables, the RF model explained 42% of variation in the XRF data. The authors considered that these models were useful tools for city agencies to model public health targets.

Readers interested in recent preconcentration strategies published for XRF spectrometry will find a wealth of experience offered by Margui et al.²²⁸ for the analysis of trace and ultra-trace analytes in liquids. Procedures included microextraction, nanomaterials, filters and activated thin layers for analysis using the current range of XRF instrumentation. The advantages and limitations were discussed for each mode including large-scale instrumentation, bench-top spectrometers and hand-held systems. In recent years, interest in environmentally friendly analytical procedures complying with the rules of green chemistry has gained strength. This approach was described by Kocot and colleagues,²²⁹ who were attracted to procedures that replaced toxic reagents, minimised waste in the laboratory and saved time. They described dispersive micro solid-phase extraction using multi-walled carbon nanotubes combined with portable TXRF spectrometry for the determination of trace amounts of Cd and Pb in environmental waters (sea, river and waste water). The linear range observed for both elements was up to 50 ng mL⁻¹ with reported detection limits of 1.0 ng mL⁻¹ and 2.1 ng mL⁻¹ for Cd^II and Pb^II ions, respectively. Zawisza and Sitko²³⁰ also offered a green approach based on using chitosan for the sorption of trace levels of Co, Cu and Ni in water samples. The effect of pH and time of the chitosan activation, sorption, salt concentration, metal ion concentrations and the amount of adsorbent on the extent of adsorption were investigated. The chitosan with adsorbed metal ions was then dissolved in acetic acid. After evaporation, the formed solid film was analysed by XRF spectrometry. As the solid film met the criteria for thin samples for XRF analysis, the matrix effects were negligible. The detection limits were quoted as 7 ng mL⁻¹ for Co, 4 ng mL⁻¹ for Cu and 5 ng mL⁻¹ for Ni. A further paper by Margui and colleagues²³¹ described liquid phase extraction strategies combined with TXRF spectrometry for the determination of low amounts of inorganic antimony species in waters. The best analytical strategy for the determination of Sb^III and Sb^V in the low μg L⁻¹ range was reported to be dispersive liquid–liquid microextraction followed by TXRF analysis. The same microextraction procedure was reported by Pytlakowska and Sitko²³² for the determination of trace amounts of Cu and Zn in surface water by EDXRF spectrometry. Under optimum conditions, an enrichment factor of 250 was obtained from only 5 mL of a water sample. The calibration graph was linear from 0.02 to 0.4 μg mL⁻¹ with detection limits of 1.8 ng mL⁻¹ and 1.7 ng mL⁻¹ for Cu and Zn respectively. Kumar et al.²³³ developed a flat sheet sorbent with poly-hydroxamic acid groups anchored on the microporous structure of a poly-propylene membrane as the preconcentration stage for the determination of heavy metals from natural waters analysed by EDXRF spectrometry. Hatzistavros and Kallithrakas-Kontos²³⁴ determined mercury at trace levels by EDXRF spectrometry using cation selective membranes. The procedure was confirmed by the analysis of ASTM Type1 water for this technique at sub-μg L⁻¹ levels. Wogelius²³⁵ described the application of synchrotron-based analytical techniques (XAS, XRF and GIXRD) for the analysis of aqueous fluids prepared by adsorption and co-precipitation at mineral fluid interfaces. Two case studies dealing with major contemporary environmental problems were presented: the mobility of arsenic in ground water and the speciation and uptake of uranium within radio-active sludge.

Sponges are aquatic, predominantly marine animals, but also inhabit freshwater environments. All freshwater sponges have a skeleton composed of silicious spikules and because they are filter feeders, have a high potential for bioaccumulation of metals. De Barros et al.²³⁶ established the inorganic composition of two Amazonian species of freshwater sponges, Drulla cristata and Drulla uruguayensis using EDXRF spectrometry. The major elements in both species were Si followed by Al, however, at lower concentration levels, Cl, Cu, S and Ti were selectively accumulated only in D. cristata, suggesting its use in environmental characterisation studies. A research team in Siberia²³⁷ studied sponges from Lake Baikal using WDXRF spectrometry and published their work to develop a calibration for 19 elements of interest: Al, Ba, Br, Ca, Cl, Cu, Fe, K, Mg, Mn, Na, Ni, P, Rb, S, Si, Sr, Ti and Zn in the absence of appropriate reference materials. Synthetic calibrants were prepared by mixing plant CRMs and SiO₂ in appropriate proportions for a calibration to enable comparison of sponges cleaned from mineral particles and symbiotic organisms with unwashed sponges. Yoshimura et al.²³⁸ described element profile and the chemical environment of sulfur in giant clam shells using synchrotron techniques. The spectra of S K-edge XANES collected from bivalve shells and S-bearing organic and inorganic reference materials indicated that inorganic sulfate was present in marine aragonitic and calcitic bivalve shells. However, XANES results did not provide discrimination between organic and inorganic sulfate in freshwater aragonitic shells. Deruytter et al.²³⁹ also used SRXRF spectrometry to determine Cu concentrations in 36 individual marine mussel larvae with a spatial resolution of 10 × 10 μm². The results indicated decreasing Cu accumulation with increasing dissolved organic carbon concentrations which was explained by an increase in Cu complexation. Salinity was found to have a non-linear effect on copper. The authors thought that this could not be explained by copper speciation or competition processes and suggested a salinity-induced alteration in physiology. More researchers appreciated the combination of μ-XANES and μ-XRD mapping of the same sample as demonstrated by Brinza et al.²⁴⁰ working on the beamline 118 at Diamond Light Source, Didcot, UK. The combined approach was used to investigated both long and short-range order in calcium carbonate granules produced by the earthworm Lumbricus terrestris. Calcite and vaterite were produced by earthworms cultured in a control artificial soil; however, granules produced by earthworms cultivated in the same soil amended with 500 μg g⁻¹ Mg also contained an aragonite. Workers in South America^241,242 evaluated the contamination risk to humans living in an area of Argentina with arsenic levels above the maximum allowed limit of 10 μg L⁻¹. Samples of domestic dog hair were used as biomarkers for TXRF spectrometric analysis. The measurements showed that, independent of gender, age and breed, dog hairs from Los Alamos had significantly higher arsenic concentrations than a set of ten dogs used as a control, coming from an area free from arsenic. The specimens were prepared by microwave digestion and offered a procedure for the use of pets as biomarkers of environmental metal contamination.

The assessment of damage to indoor museum artwork, particularly by pollutants, is a concern for curators and conservators. Godoi et al.²⁴³ reported the effects of particulate matter and gaseous pollutants in museums in tropical and sub-tropical environments. Samples of PM and gases were collected at the Oscar Niemeyer Museum in Curitiba, Brazil, where large modern and contemporary works of art are displayed. Passive diffusive samplers were used for analysis of gaseous components by IC or GC-MS. Bulk PM samples were quantified by EDXRF spectrometry. The data were compared with the concentrations obtained for the same pollutants in other museums, located in places with different climates to offer conservation advice to museum managers with limited budgets. Fomba et al.²⁴⁴ offered the first long-term, time resolved, trace metal study for remote tropical northern Atlantic marine aerosols. Samples were collected at the Cape Verde Atmospheric Observatory from January 2007 to November 2011 for subsequent analysis by TXRF spectrometry. The low Pb concentration (<0.20 ng m⁻³) was reported to be independent of air mass direction and was thought to demonstrate the complete phase-out of leaded gasoline, even in African countries. Yatkin and Gerboles²⁴⁵ investigated the mass loss of water soluble ionic compounds caused during EDXRF analysis of PM10 loaded filters. The effect was gravimetrically evaluated on co-located PM10 loaded filters using an EDXRF spectrometer with the sample chamber maintained under vacuum, air and helium media and compared with ion chromatographic data. The authors showed that EDXRF measurements made under vacuum led to mass loss of water soluble ionic compounds and other volatile constituents of PM10. The use of helium and air media considerably reduced the PM mass losses. The paper revealed that different amounts of Ca²⁺, Cl⁻, Na⁺, Mg²⁺, NH₄⁺ and NO₃⁻ were lost, whereas K⁺ and SO₄²⁻ remained unchanged. The relationship between vacuum and application time was also studied for quartz and Teflon filters. The longer the application time, the higher the PM mass loss was reported. The authors also showed that mass loss of PM on quartz filters was two times higher than that from Teflon filters. Fittschen et al.²⁴⁶ determined P and other low atomic number elements in aerosol particulates from ambient air using SR-TXRF spectrometry. Because particle size was known to be strongly correlated to its origin and sedimentation, the aerosols were collected with the aid of a low pressure Berner impactor, which separated the particulates into nine size fractions ranging from 15 to 30 nm. Measurements were determined under vacuum at the FLUO beamline at the ANKA synchrotron, Karlsruhe, Germany. An excitation energy at 3.5 keV near the P K-edge but below the Ca absorption edge was chosen to avoid interferences from the P K-line with detector escape peak artefact of the Ca–K line. With a one hour aerosol collection time, the detection limits were generally 0.2 to 0.3 ng m⁻³. Other elements such as Cl, P and Si were also determined. Indresand et al.²⁴⁷ reported the preparation of sulfur reference materials that reproduce atmospheric particulate matter sample characteristics for XRF calibration. Suppliers of commercial instrumentation offer thin film standards of known elemental masses that are useful for inter-laboratory comparisons but may differ in terms of chemical composition, substrate and geometry, thereby introducing uncertainties regarding absolute accuracy of the calibration for atmospheric samples. Long term concentration trends and comparison between analytical networks demand a calibration that is accurate and precise compared with the uncertainty of the XRF measurements themselves. This section of the review closes with an unusual sample format; exhaled breath condensate, containing organic and particulate matter in suspension. Barreiros et al.²⁴⁸ prepared their samples, collected from a group of workers in a lead processing plant, by cooling exhaled air under conditions of spontaneous breathing. Two analytical techniques, TXRF and ICP-MS were chosen for their ability to offer multi-elemental data with appropriate detection limits, from small sample volumes.

3.5 Archaeological and cultural heritage

Judging by the volume of abstracts available during the present review period, the XRF technique represents the cornerstone of many archaeometric studies with the ever expanding use of hand-held instrumentation. In this respect, education is a key aspect of training the next generation of scientific archaeologists, so it is welcoming to see that XRF spectroscopy was included in a chemistry and art course at Ithaca College, New York,²⁴⁹ underscoring the complimentary nature of scientific measurements and art history in the study of art objects.

The investigation of novel applications is one way in which a subject area can expand, illustrated, this year, by a number of papers. Thus, Kakoulli et al.²⁵⁰ included SR-base fluorescence and absorption techniques to demonstrate chronic poisoning by arsenic of pre-Columbian populations that inhabited part of the Atacama Desert (Chile) during the Middle Horizontal and Late Intermediate Period (500–1450 AD). They analysed samples of ancient human hair and concluded that poisoning was most likely to have resulted from the ingestion of As-polluted water rather than by external contamination. In another study with strong forensic links, Fraser et al.²⁵¹ used the XRF and FTIR techniques to characterise an encrusted coating on an African Komo mask from the collections of the Detroit Institute of Arts. The presence of significant quantities of iron and protein indicated the presence of blood, an observation that was confirmed by the development of an in situ methylation and mass spectrometry technique. Noting that for half of humanity, rice has been the essential staple food of billions of people in Asia, Won-in et al.²⁵² used a range of techniques, including μ-beam XRF to analyse ancient black rice from Nakhon Nayok Province, Thailand. The authors reported the detection of Si, Ca, Al and other trace elements. By contrast, cattle were the basis of the most important social institutions in southern Africa and Huffman et al.²⁵³ used the XRF technique to analyse vitrified cattle dung in Iron Age byres in the Limpopo valley, identifying high levels of K₂O and CaO. Although lightning strikes could have initiated the fire that caused the vitrification process, the authors proposed that Iron Age villagers deliberately set fire to dung for hygienic and ethical reasons when their cattle died unexpectedly. The contrast in application could not be more extreme, but Malagodi et al.²⁵⁴ used a number of techniques, including EDXRF and μ-FTIR and dendrochronology, to investigate the original varnish layers and characterise the composition of decorations on a violin top plate made by Stradivari in the 17^th century. As part of this study, evidence of inappropriate restoration was revealed and replicate determinations were made on a copy of the top plate made of materials identical to the original.

Although traditional archaeological studies might be restricted to the analysis of paper, parchment and ink, photographs as well are relevant to modern ‘archaeological’ XRF studies. In the latter area, Kaplan and Stulik²⁵⁵ undertook the first scientific investigation of the image substrate of four plates brought to England in1827 by Joseph Nicephore Niepce (one of the inventors of photography). The chemical composition of all four plates, one of which was attributed by the authors to a particular pewter smith was confirmed by XRF. The ‘gum dichromate’ process was used by late 19^th century pictorialist photographers and Vila et al.²⁵⁶ prepared test samples which were analysed by the XRF technique in an attempt to develop a method to identify this process. They reported that the presence of Cr might have more complex sources so that a more discriminatory approach was required for the identification of gum dichromate photographs. The uncertain archaeological provenance of a large part of the collection of Dead Sea scroll fragments was a topic of interest to Rabin and Hahn,²⁵⁷ who used the XRF, FTIR and Raman techniques to provide new information on the production of ancient parchment towards the end of the Second Temple period. Dzinavatonga et al.²⁵⁸ determined Ca, Cu, Fe, K, Mn and S in pre- and post-1850 paper samples from the National Library of South Africa. They found that Ca concentrations decreased significantly between 1800 and 1890, coinciding with changes in paper-making technology and from their results concluded that Fe has a potential negative impact on the preservation of paper because of its relative abundance with respect to Cu, unless de-acidification is undertaken. In the characterisation of ‘The Manueline floral charter of Sintra’, a Portuguese illuminated manuscript dated to 1514, EDXRF measurements contributed to a study by Manso et al.²⁵⁹ that allowed the identification of the gall inks used in the written text, the pigments used in illuminations and letterings and the filler and binder used to produce the colouring materials and inks.

X-ray fluorescence (often in combination with Raman and optical reflectance) has become one of the standard techniques used in the characterisation of pigments in paintings and it is only possible to feature a small selection here. The use of portable laboratory and synchrotron-based instrumentation was reviewed by Janssens et al.²⁶⁰ with a particular emphasis on macroscopic to (sub-) microscopic beams for the characterisation of pigments, paint microsamples or entire paintings. The authors acknowledged that elemental determinations alone cannot elucidate the chemical transformations that occur during the natural processes of pigment alteration, but that μ-SR-based combinations of XRF spectrometry, XAS and XRD were suitable. Zielinska et al.²⁶¹ described a system capable of simultaneously imaging large area samples using a wide field, uniform excitation X-ray beam and a position sensitive ED detector. Fluorescence radiation from an area 10 × 10 cm² was projected through a pin-hole camera on to a gas electron multiplier detector of the same size, incorporating readout strips and electronics capable of achieving a FWHM resolution of 20% at 5.9 keV and a maximum count rate of 5 × 10⁶ s⁻¹. The authors reported spatial resolutions of 1 mm (FWHM) and measurement times reduced by two to three orders of magnitude compared to conventional μ-XRF techniques. Alfeld et al.²⁶² described the first commercially available XRF scanner for paintings incorporating an X-ray tube and SDD mounted on a motorised stage capable of imaging the distribution of elements (Z = 22–42, Ti–Mo) in surface and sub-surface paint layers. The scanned area was 80 × 60 cm² with a dwell time of 10 ms and a lateral resolution of less than 100 μm. In a further contribution, Alfeld and Broekaert²⁶³ reviewed techniques for the mobile depth profiling and sub-surface imaging of historic paintings, with a particular emphasis on the way the established techniques of X-ray radiography and IR reflectography have recently been extended by scanning μ-XRF analysis and optical coherence tomography. As an example of the power of these techniques, Alfeld et al.²⁶⁴ prepared a full scale mock-up of Rembrandt's ‘An old man in military costume’, which included a free impression of a hidden portrait on top of which the original was thought to have been painted. Three XRF imaging techniques were applied: a mobile X-ray tube instrument and two SR-based facilities (one having multiple SDDs and the other a Maia detector). The results indicated that an investigation of the original painting was likely to have an excellent chance of success at revealing the hidden paint layers. As a further illustration of this approach, Alfeld et al.²⁶⁵ showed how SR-based scanning macro-XRF could be used to visualise the underpainting below ‘Portrait of an old man’ by Rembrandt. Rebollo et al.²⁶⁶ employed a mobile XRF spectrometer and a transportable imaging spectroscopy device to study two 14^th Century canvasses by Lorenzo Veneziano. Identification of different materials (e.g., gypsum and azurite) and the painting technique were in good agreement with those of the attributed artist. A portable XRF analyser was used by Gebremariam et al.²⁶⁷ for what was claimed to be the first on-site analytical examination of Ethiopian murals at the 12^th Century Yemrehanna Krestos Church. Pigments identified included red and yellow ochre, minium, cinnabar, orpiment, gypsum, white lead and carbon black. In an application to more recent (19^th and 20^th Century) artists materials, Montagner et al.²⁶⁸ used the μ-FTIR, μ-EDXRF, Raman and XRD techniques to analyse ochres used by Rodrigo (1912–1997) and Pinheiro (1857–1929) provided by commercial suppliers. The authors reported that μ-EDXRF and Raman spectroscopies were the best techniques for identifying umber due to the characteristic presence of manganese. The μ-EDXRF technique also revealed the presence of significant amounts of As in all samples of sienna. Mass et al.²⁶⁹ investigated the photo-degradation of cadmium yellow paints in Henri Matisse's Le Bonheur de vivre (1905–1906), using hand-held XRF, SEM techniques and various SR approaches (μ-XANES, μ-XRF and μ-FTIR). They reported that photo-degradation caused CdS paint to discolour to CdCO₃ with oxalates and sulfates found to be concentrated at the alteration surface. Staying with a more modern theme, Casadio and Rose²⁷⁰ used a ‘high resolution nanoprobe XRF’ mapping technique to examine impurities in zinc oxide pigments in tube and enamel paints used by early 20^th century artists. The emphasis was placed on Ripolin, a popular brand of French house paint used extensively by Pablo Picasso. One of the results reported by the authors was that the zinc oxide particles only contained a small amount of iron, indicating that the highest quality zinc oxide, free of Pb and Cd, was used in Ripolin. Continuing with a domestic theme, some would consider it an overstatement to include 18^th to 19^th century Finnish wall paper in a section on works of art, but the pigments used in this material were studied by Castro et al.²⁷¹ by Raman and EDXRF spectroscopies. These researchers reported the presence of a strange iron(III) chromate yellow pigment as well as other unusual mixtures to obtain fashionable blue and green colours, involving blue verditer, ultramarine blue, Prussian blue, chrome yellow, calcite, white lead, red and yellow iron oxide, gypsum and carbon black. The toxic emerald green pigment (copper aceto-arsenite) was also detected.

The analysis by XRF of pottery sherds once again features strongly in the literature, although the contribution of XRF data has become fairly standardised in assisting in the provenance of samples. For example, Sakalis et al.²⁷² studied various types of decorated Neolithic pottery from Polyplatanos (Imathia, Greece), using multi-element μ-XRF spectrometry to provide data for PCA. Barone et al.²⁷³ were interested in fine pottery from Syracuse (Sicily), using XRF spectrometry to contribute to a reference group for ceramics produced in this Greek colony during the Roman–Hellenistic period. Multivariate statistical analysis of these chemical data permitted the identification of pottery production at Syracuse by comparison with kiln wasters and the local Plio-Pleistocene clays. The provenance and production technology of decorated Padan terra sigillata from the site of Retratto, Adria (north eastern Italy) was investigated by Maritan et al.²⁷⁴ Data from a multi-analytical approach indicated that most samples were fired at between 850–950 °C and it was possible to distinguish different clay materials used to produce plain and decorated potsherds as well as to distinguish these samples from terra sigillata produced in Magdalensberg (in modern Austria). Studies of megalithic sarcophagi potsherds from Veeranam (Tiruvannamalai, India) were reported by Ravisankar et al.,²⁷⁵ again using a multi-technique approach, again with EDXRF data contributing to a PCA analysis that demonstrated the hierarchical clustering of samples. Gascon and Garrigos.²⁷⁶ reported that very few archaeometric studies have been undertaken at Iron Age sites in Lebanon, a situation their contribution helped to rectify by reporting XRF and XRD analyses of pottery fragments from the Tyre-Al-Bass necropolis. The aim of this investigation was to assist in the characterisation of Phoenician ceramics of Tyre and differentiate different production groups. Ferreras et al.²⁷⁷ were interested in elucidating trade between Roman provinces in the Northwest Mediterranean during the Augustan period by the analysis of amphorae from two shipwrecks located near the northern coast of Catalunya (North East Spain). Results, including XRF analyses were compared with a large data base of amphorae production, allowing the identification of the specific production workshops.

The analysis of ceramics and glazes on ceramic material provides another opportunity for XRF measurements to contribute, often in combination with Raman spectroscopy. Geil et al.²⁷⁸ provided an example of such an application, demonstrating the capabilities of SRXRF imaging of painted ceramics from prehistoric cultures of the American Southwest. The authors reported that as well as distinguishing pigments that were visually similar, layers that were obscured by over-painting could also be identified, so allowing an interpretive rather than anecdotal analysis. The effects of elevated lead caused by volatilisation during firing were investigated by Forster and Grave²⁷⁹ using portable XRF for the analysis of medieval (Byzantine Cypriot) glazed ceramics from the Mediterranean region. Owing to spectral interference from Pb, the number of discriminating elements available for multivariate discrimination analysis was reduced, although portable XRF data was still considered to expand provenancing studies especially where significantly different geology was present. Mitsuji et al.²⁸⁰ used XRF spectrometry to analyse Sue ware, a type of ancient Japanese ceramic, and reported that kiln sites could be distinguished from plots of K–Ca and Rb–Sr data. Wavelength dispersive XRF spectrometry was used to ascertain the state of degradation of ceramic tiles that have been exposed to the environment, caused by microorganisms developing within the pore system, as demonstrated by Silva et al.²⁸¹ The resulting biodegradation could cause exfoliation/detachment of the decorated glaze and the authors demonstrated the benefit of this technique in contributing to a diagnosis of the state of degradation of polychrome Portuguese decorated tiles (16^th to 18^th century) prior to future decontamination using gamma radiation. The primary technique used by Casadio et al.²⁸² to characterise over-glaze enamel production in historic (18^th to 19^th century) porcelain factories in Central Europe was XRF spectroscopy. It was claimed that the results enhanced an understanding of the dating of objects based on the detection of Zn in yellow, blue and green glazes. In combination with LIBS and XRD, XRF spectrometry was used by Blagoev et al.²⁸³ to assess the chemical and phase composition of white clay decorative tiles from the medieval archaeological site of Veliki Preslav (a Bulgarian capital in the period 893–972 AD). In an interesting application by Gondai et al.,²⁸⁴ portable XRF and portable XRD techniques were used to characterise Islamic yellow opaque glazed pottery, excavated from the Raya site, Egypt. The X-ray analysis revealed the presence of the pigment Pb₂Sn_2−xSb_xO_6+x/2 in the yellow opaque glaze and complementary SEM examination revealed differences in its weathering properties, which depended on the base glaze.

Once again, the analysis of glass features in a significant number of contributions with a trend that has seen an increase in the use of both portable XRF spectrometry and more sophisticated SR instrumentation. As an example of the latter, Klysubun et al.²⁸⁵ used the SRXRF, XANES and EXAFS techniques to characterise yellow and colourless decorative glasses from the Temple of the Emerald Budda, Bangkok, Thailand to elucidate long-lost glass-making recipes and manufacturing techniques that would contribute to the future restoration of the Temple. The glasses were found to be soda lime silica glasses and EXAFS data indicated that the Fe oxidation state was 3⁺ (compatible with its use as a yellow colorant) with Mn being 2⁺ (compatible with its use as a decolourant), and Pb as 2⁺. X-ray absorption and XRF spectroscopies were used together with LA-ICP-MS by Hormes et al.²⁸⁶ to analyse medieval glass from the Cathedral in Paderborn (Germany). The authors reported that XAS results showed that as expected, most elements were present as oxides in their most stable form, the exceptions being Ti, which was present as anatase, rather that rutile (the most stable form of TiO₂₎, and Pb, which was present as a complex mixture of oxide, sulfate and other compounds. An investigation of the colouring mechanism of ancient copper-red glass from Egypt was undertaken by Kikugawa et al.²⁸⁷ using XANES to characterise the chemical form of Cu. The authors concluded that the colouring mechanism of copper-red glass changed from crystalline Cu₂O colouring to metallic-Cu nano-clustering colouring as glass production developed in ancient Egypt. The melting behaviour of natural sands found along the Gulf of Naples (Italy), of archaeological interest in the 1^st century, was the topic of interest to Montanari et al.²⁸⁸ using a multi-technique approach that involved XRF and XRD. The thermal behaviour of four sands was investigated using a heating/hot stage microscope and the authors concluded that the addition of fluxes was necessary to achieve a complete amorphous material. Greiner-Wronowa and colleagues²⁸⁹ undertook a corrosion study of Middle Age objects (metallic rings with mounted glass beads) excavated under the Main Square, Krakow (Poland), using an SEM, EDXRF and XRD analytical approach. The authors discovered that these objects were found in soil layers that differed in chemical composition and microclimate parameters which caused local differences in corrosion processes. Adb-Allah²⁹⁰ used the XRF and XRD techniques to determine the mineralogical and elemental composition of the soil deposits and encrustations on the glass surfaces of Roman fragments excavated in northern Jordan. Following a series of experiments with different cleaning agents in which a SEM was used to examine glass surfaces, it was concluded that EDTA was generally the most effective chelating agent for cleaning durable glass surfaces with piranha solution (a mixture of sulfuric acid and hydrogen peroxide) being effective for the removal of calcareous crusts.

The whole-site geochemical characterisation of archaeological sites has been used by a number of investigators in recent years, exemplified by the work of Hayes²⁹¹ who used a portable XRF spectrometer at the Reaume Fort site, Central Minnesota, USA to examine whether the variability of surface geochemistry corresponded with sub-surface features of archaeological interest. The author concluded that this was the case when sub-surface features were shallow in depth (within 5 cm of the surface) and were associated with clayey in-fillings and that the in situ PXRF approach provided excellent discrimination in the detection of features filled with other sediments. Gauss et al.²⁹² used PXRF to undertake a comprehensive survey of the Early Bronze Age settlement of Fidvar, Vrable (Slovakia) and concluded that the PXRF analyses of soils showed patterns typical of human occupation with strong positive correlations between Ca, P and Sr, some of which could be associated with archaeological structures.

There has been increased interest in recent years in the application of XRF techniques to and analysis of archaeological building materials. Bonizzoni et al.²⁹³ undertook a comparison between the XRF, TXRF and PXRF techniques for the provenance and classification of archaeological bricks from a medieval monastery on the Po Valley, northern Italy. Sagin and Boke²⁹⁴ used a multi-technique approach (XRF, XRD, SEM and TGA) to examine the bricks used in the domes of 15^th century bath buildings in Turkey. Their main findings were that the bricks were of low density, high porosity and produced from raw materials containing clays of low calcium content. These bricks were susceptible to salt crystallisation and freeze–thaw cycles, issues that did not cause accelerated degradation given the local moderate climate. To support the revival of traditional skills in stuccoes and plasterwork, Salavessa et al.²⁹⁵ used a range of techniques, including XRF, to examine mortars especially to improve their use in restoration projects. Prieto-Taboada et al.²⁹⁶ advocated the need for a cross-sectional analysis to diagnose correctly the state of building materials and allow surface deposits (crusts and patinas) to be distinguished from penetrating pollution. In this way, incorrect assumptions could be avoided when assessing the state of conservation of building materials. The effect of weathering and surface contamination on building stone in the Cuzco region of Peru was evaluated by Ogburn et al.²⁹⁷ by in situ measurements using PXRF spectrometry. The building materials examined were diorite and andesite blocks (before and after cleaning) and andesite blocks in which weathered surfaces were compared with fresh surfaces of samples from two andesite quarries. Results showed that the low atomic number elements were most impacted by contamination and weathering, whereas many higher-Z elements were relatively unaffected, noting enhanced levels of Pb and Zn because of urban pollution. The state of conservation of a 15^th century Palace house in Azpeitia (northern Spain) that had been affected by infiltration of water was reported by Gomez-Laserna et al.²⁹⁸ using the portable Raman and XRF techniques. They identified nitrates and sulfates as decay compounds and revealed that materials used in earlier restorations produced new degradation processes, so emphasising the importance of the proper selection of restoration materials. Ion et al.²⁹⁹ used EDXRF as one of the techniques contributing to an examination of the chemical weathering of chalk stone materials from Basarabi churches, medieval Dobrogea, Romania. The authors identified the crystallisation and dissolution of salts and the generation of new pores in a self-accelerating process as one of the effects of the environment on this monument. In an unusual application, Holakooei et al.³⁰⁰ used WDXRF spectrometry to provenance 15^th century Persian haft rang tiles in Northeast Iran. Tiles appeared to have been manufactured from local clay materials and so could be replaced by newly made tiles.

Considering now sculptures and stone artefacts, Barbera et al.³⁰¹ reported that PXRF was a powerful technique for the study of carbonate rock artefacts in museums. The authors used the technique to analyse a large set of limestones that outcrop in the Hyblean area of Sicily to contribute to a data base of carbonate rocks that were commonly used in the past for the manufacture of sculptures. Principal component analysis was used to classify samples both between and within different limestone formations. Vanmeert and colleagues³⁰² used μ-XRF and μ-XRD instrumentation to investigate the calcium oxalate layer that is currently used to protect monumental limestone by a passivation treatment with ammonium oxalate. X-ray fluorescence contributed with other techniques to a study of the effectiveness of laser cleaning of plaster sculptures undertaken by Pelosi et al.,³⁰³ who found that laser cleaning was effective in removing the dirty layer, preserving the original finish, whereas aqueous cleaning removed the finish. Ortiz and co-workers³⁰⁴ reported on a similar approach, using μ-XRF analyses to evaluate the laser cleaning of weathered marble surfaces to remove surface deposits, iron oxide stains and even graffiti. The authors found that this approach was generally successful although some of the graffiti left yellowing or some surface erosion.

The analysis of obsidian regularly features in this review with measurement generally used to provenance samples to source and provide information about prehistoric trade activities. This year has seen an increase in the use of portable XRF instrumentation, exemplified by the contribution of Milic,³⁰⁵ who described the use of the technique in the characterisation of obsidian from central Anatolia, the Aegean and central Europe, emphasising the importance of Rb–Sr–Zr three dimensional scatter plots for source discrimination. Kellet et al.³⁰⁶ combined portable XRF spectrometry with LA-ICP-MS in a provenance study of obsidian from the Andahuaylas region of southern Peru. Frahm and Feinberg³⁰⁷ used the sourcing results for 97 obsidian artefacts from Urkesh (Tell Mozan, Syria) to contribute to an understanding of the Early to Middle Bronze Age transition (2200 BCE) in northern Mesopotamia and the influence of climatic perturbations and the collapse of regional government. In a second contribution, Frahm et al.³⁰⁸ extoled the virtue of field measurements of obsidian, synchronised to excavation activities, rather than previous post-excavation PXRF studies of museum collections. To achieve this goal, the authors developed PXRF methods based on a count time of only 10 s (rather than the 2 to 6 minutes used in previous studies) coupled to the opportunity for automated source matching. This work was undertaken at the Middle Palaeolithic cave of Lusakert 1 in Armenia. Basaltic artefacts may be provenanced in a similar way to obsidian, as demonstrated in a study by Gluhak and Rosenberg,³⁰⁹ who used XRF and LA-ICP-MS to compare measurements of six bifacial tools found at three Neolithic and Chalcolithic sites in the Jezreel Valley, Israel. After sampling and analysing Miocene magmatic rocks in the region, the successful identification of the source of these tools arose from the geochemical differences between lava outcrops in the southern Levant.

There has been substantial activity in the analysis of a range of metal artefacts over the current review year using various XRF techniques, with a notable increase in the use of portable XRF. There is space here only to cover exemplars of these contributions, including the work of Neiva et al.,³¹⁰ who selected EDXRF instrumentation to analyse ores, pig iron and steel from a 14^th century iron works in Ipero (Brazil) to investigate the presence of the deleterious elements, Ti and P. They found large differences in chemical composition between and within samples. Castelli et al.³¹¹ developed new criteria for the rapid discrimination between traditionally maintained and artificially restored Japanese swords. The latter were restored using an acid bath treatment whereas traditionally maintained swords were polished with stone, allowing the authors to detect the presence of trace amounts of Si by the XRF technique, so distinguishing the high value originals. X-ray fluorescence was one of the techniques used by Cvikel et al.³¹² to characterise a 12 pounder wrought iron cannon ball found at a 19^th century shipwreck in Akko Harbour, Israel. Results indicated a different manufacturing process from other cannon balls recovered from this site with a possible earlier production date, indicating that this artefact could even have been used as ballast. A degree of sophistication was used by Michelin et al.³¹³ in an investigation of nano-scale corrosion mechanisms of a 450 year old archaeological nail. Analysis of the Fe L-edge profile by NEXAFS showed the presence of a 100 nm layer at the metal corrosion product interface consisting of maghemite (Fe₂O₃) and magnetite (Fe₃O₄) with iron carbonates and smaller quantities of iron oxides forming the corrosion products further from this interface. The authors recognised that these results had implications for the very long term corrosion of steels, especially in the fields of cultural heritage conservation and the storage of nuclear wastes. Satovic et al.³¹⁴ assessed the application of portable XRF instrumentation in the analysis of bronzes having an artificially corroded surface, demonstrating that accurate results could be obtained for surfaces affected by corrosion layers of up to 25 μm. X-ray fluorescence, CT and X-ray photographic microscopic observation were used by Fan et al.³¹⁵ to investigate the manufacturing techniques used in the Chen Zhang Pot, an open work bronze pot with gold and silver inlay held in the Nanjing Museum, China. Investigations showed that many parts of the pot were cast separately before assembly. In a similar type of application, Mass and Matsen³¹⁶ asked the question about whether there was a role for XRF in understanding silver hollow wares of the 18^th and 19^th centuries and provided some examples using English and American silver hollow wares from the New York Historical Society. The laser cleaning of marble has already been considered but Buccolieri et al.³¹⁷ investigated the same approach in the cleaning of a bronze bell. As a contribution to this work, EDXRF spectrometry was used to assess the efficiency of the cleaning process with variations in the elements Ca, Cl, Cu, Pb, S and Sn being of particular interest. Ferretti et al.³¹⁸ evaluated the use of XRF techniques to characterise multi-layer structures, taking four medieval processional crosses from Central Italy as examples. The authors measured variations in Ag Kα/Kβ intensities and concluded that results allowed them to distinguish ancient restorations from original parts and to characterise enamels. Multi-layer artefacts were also investigated by Cesareo et al.,³¹⁹ using variations in the relevant K and Lα/β intensities of Ag, Au, and Cu measured by EDXRF to distinguish gold, silver and copper alloys, gilded copper and silver, silvered copper and tumbaga (copper or silver with a surface enrichment of gold) in pre-Columbian artefacts from the North of Peru. Portable μ-XRF instrumentation was used by Scrivano et al.³²⁰ in an experimental archaeology approach to reconstruct ancient soldering and welding processes. The authors reported the Ag, Au and Cu content within soldering or welding areas and the results were particularly applicable to alloys with high gold concentrations, such as Tartessic jewellery (700–500 BC).

The XRF technique has traditionally been used to characterise and sort coins, especially to identify the mint at which they were struck. Such an investigation was undertaken by Gorghinian et al.,³²¹ who used μ-XRF instrumentation based on polycapillary optics to characterise 477 ancient coins issued in different mints that were active during the reign of Augustus, the First Roman Emperor. These coins, made of copper, gold and silver alloys benefited from the absence of a patina owing to usage. However, when undertaking more sophisticated studies, XRF spectrometry is likely to play a more supporting role, illustrated by the work of Cutroneo et al.,³²² who reported the Ag/O atomic ratio depth profile in silver coins by LA-ICP-MS to determine the thickness of the oxidation layer. Results were confirmed using SEM and the associated XRF microprobe with the oxidation layer ranging from 25 to 250 μm thick.

3.6 Forensic

In the area of cultural heritage, Galli and Bonizzoni³²³ reviewed the role of portable XRF analysis in discovering the provenance, proving the authenticity or supporting restoration by illustrating this role with a number of case studies. Kajiya et al.³²⁴ described a systematic study that led to the identification of a forgery of a painting (Violeiro) by the artist Di Cavalcanti (1897–1976). The authors used a number of techniques to come to this conclusion, including the analysis of pigments by PXRF instrumentation. The analysis of eight cloisonné enamels on gold of Byzantine appearance, held in a number of museums in the USA and all originating from the possessions of the Russian artist Mikhail Botkin (1839–1914), were examined by Helfenstein et al.³²⁵ The authors used analytical data obtained by XRF and SEM analysis as well as art history evidence to conclude that seven of the eight pieces were 19^th or early 20^th century copies of Byzantine originals.

The analysis of glass is an important aspect of some forensic investigations and a comparison was undertaken by Trejos et al.³²⁶ of elemental determinations undertaken by sixteen forensic science laboratories, providing a direct comparison between μ-XRF spectrometry, ICP-MS (on dissolved sample) and LA-ICP-MS. The repeatability and reproducibility (between laboratories) of μ-XRF versus ICP was 11 vs. 5% and 16 vs. 10% respectively with ICP offering overall lower detection limits. In a second comparative trial, the authors reported the results using the same techniques plus ICP-OES showing that glass samples were manufactured at different plants or at the same plant but weeks or months apart and were readily differentiated by elemental composition. Ernst et al.³²⁷ demonstrated a practical method for applying the concepts of the signal-to-noise ratio, limits of detection and limits of quantitation to the XRF spectra from glass samples, generated by μ-EDXRF spectrometry in particular, to provide evidence for forensic examiners as to whether to use analytical results for elements present at concentrations that approach the EDXRF detection limit.

X-ray fluorescence spectrometry has an important role to play in criminal investigations. Regrettably, gun crime often features in this section and is illustrated this year by the report of Fonseca et al.,³²⁸ who showed that EDXRF spectrometry could advantageously be used to replace chemometric methods to estimate the “target to muzzle” distance when investigating the relative positions of shooter and victim. This study was based on measuring Ba, Cu, Pb, Sb and Zn on plain white cotton tissue targets and plotting concentration as a function of firing distance. Interestingly, Turillazzi et al.³²⁹ undertook a similar study with pig skin as the target but using ICP-OES for the measurement step. Kawai³³⁰ reviewed evidence in two reports presented in a criminal court case in which a defendant was convicted of arsenic poisoning. Measurements by SRXRF and ICP-AES contributed to these reports in fingerprinting arsenic pesticide specimens taken from the crime scene. Kawai endorsed the report that suggested that the arsenic stored in the kitchen cupboard of the alleged crime scene was significantly different from that in the paper cup used for the poisoning. In a case of poisoning by ingestion of methomyl (a carbamate pesticide), Kinoshita et al.³³¹ complemented GC-MS blood data with EDXRF measurements that confirmed the cause of death by the detection of S and Si in the stomach lining of the victim.

The XRF technique was used in a number of other forensic investigations, including the work of Bond et al.,³³² who used XRF spectrometry to measure selected trace elements (Ca, P, S and Si being the most abundant), and ICP-MS to measure lead isotope ratios, in samples of lead from 24 church roofs in Northamptonshire, England. Their aim was to combat thefts of roofing lead. They reported that although lead isotope ratios were similar for all samples, there were larger differences in the trace element composition of samples from different churches than samples from different sheets of lead at the same church. X-ray fluorescence measurements contributed to a novel study by Roberts et al.³³³ in order to estimate the temperature to which paint had been heated, in forensic investigations of the causes of a fire. Attenuated total reflectance Fourier transform infrared spectroscopy was the principal technique. Portable XRF spectrometry was used by Shimamoto et al.³³⁴ to discriminate brands of nail polish that may contribute to evidence at crime scenes. In the forensic analysis of laser print on paper, Chu et al.³³⁵ reported that portable XRF spectrometry was able to distinguish four brands of toner, if the letter size was larger than font 20, but could not identify the printing sequence in the case of overprints. However, analysis by laser microprobe was found to be more discriminatory, especially when high resolution and high sensitivity 3D elemental mapping was required. As a contribution to the construction of a nationwide forensic soil sediment data base for Japan, Nakai et al.³³⁶ used high energy SRXRF spectrometry and high resolution SRXRD to determine the heavy mineral and heavy metal compositions of stream sediments from 3024 points across Japan. The automated system at the SPring-8 SR facility allowed the collection of 100 XRF spectra and 130 diffraction patterns per day. Results showed that XRF and XRD data collected from sediments of Shizuoka Prefecture closely reflect the geological and geographical signature of the sediment samples, which could be used for provenancing of soil collected as evidence from a crime scene.

3.7 Clinical

The use of the XRF technique in clinical applications, and especially in combination with synchrotron radiation, is a growing field as reflected by the number of contributions in this year's review. Hummer and Rompel³³⁷ presented a tutorial review introducing the use of XAS and μ-SRXRF techniques into the field of inorganic medical chemistry. The results obtained for Au, Co, Ga, Pt and Ru compounds within the last few years were presented. Moreover, in a review by Pascolo et al.³³⁸ the advantages of SR based techniques in pediatric research were shown. These authors highlighted the applicability of advanced X-ray microscopy techniques that offer exceptional spatial and quantitative resolution in elemental detection, that are of importance for understanding diseases where mismetabolism of metals that are either physiologically important (i.e. Cu, Zn) or outright toxic (i.e. Pb), underlies pathogenesis.

The elemental concentration and distribution in human prostate tumour cells was studied by Leitao et al. In a first contribution,³³⁹ the distribution of Cu, Fe and Zn in cellular spheroid derived human prostate tumour cells was investigated. The SRXRF measurements were performed in a standard geometry of 45° incidence, excited with a white beam collimated using a 20 μm diameter optical capillary in the XRF beam line at the SLNL (Campinas, Brazil). The analysed spheroids showed different elemental distributions of Cu and Fe, and the Zn concentrations being higher in the central regions. In a second contribution³⁴⁰ SR-TXRF spectrometry was used to determine the elemental concentration for Ca, Cu, Fe, K, P, Rb, S and Zn in prostate tissues affected by cancer, Benign Prostate Hyperplasia (an illness prevailing in men above the age of 50, close to 90% after the age of 80) and normal tissue. By using the Mann–Whitney U test, it was observed that the concentration for almost all elements was significantly different between the groups studied. Zaichick and Zaichick³⁴¹ performed quantitative morphometric and EDXRF analyses to clarify age-related histological and Zn content changes in non-hyperplastic adult prostate glands. The prostates were obtained from the autopsies of 63 subjects aged 21–70 years who died mainly from trauma. An interpretation of the results allowed the possible explanation of both relevant prostatic aging mechanisms and the effects of dietary interventions using supplementary Zn.

The concentrations of bio-metals are crucial for the homeostasis of human health and have significantly different concentrations when comparing human cancer tissue with normal tissue. Silva et al.³⁴² investigated the elements Ca, Cu, Fe and Zn in neoplastic and normal breast tissue using different XRF techniques and immunohistochemical assays to allow a better metabolic understanding of trace elements in breast cancer. The elemental concentrations were determined by EDXRF spectrometry and correlated with the spatial distributions of the trace elements obtained by a μ-XRF system. The results revealed that the expression of the trace elements Fe, Cu and vascular endothelial growth factor were related, indicating that higher levels of these elements could be associated with the angiogenic process in breast cancer. In another Canadian contribution³⁴³ the μ-SRXRF technique was used to determine the localisation and the relative concentrations of Ca, Cu, Fe and Zn in 128 invasive ductal breast cancer (IDC) samples and normal surrounding breast tissue. Statistical analysis revealed a significant increase in the levels of Ca, Cu, Fe and Zn concentrations by 85%, 23%, 20% and 117%, respectively, in IDC tissue when compared to the normal breast tissue. In the classification process of human breast tissues, Farquharson and co-workers^344,345 showed that X-ray data, including the measurement of bio-metal levels and scattering characteristics, could be considered as a possible discriminating variable. Levels of Br, Ca, Cu, Fe, K, Rb and Zn were evaluated using XRF spectrometry and results for tumour breast tissue were compared with normal surrounding breast tissue. The authors claimed that the strength of their classification process was the combined use of several measurements instead of only one source of data. The results of classifying unknown tissue samples were presented using two-class and three-class models that help to reveal the importance of sample histology in studies involving breast cancers. These researchers³⁴⁶ also conducted a similar study for the identification and differentiation of secondary colorectal cancer in human liver tissue using the XRF technique, coherent scatter spectroscopy and multivariate analysis. Moreover, Laursen et al.³⁴⁷ examined the association between content of the elements Br, Cl, Cu, Fe, S and Zn, as determined by XRF spectrometry, in normal and cirrhotic liver tissue from Danes and Greenlandic Inuit by dual hierarchical clustering analysis. Gaal et al.³⁴⁸ used TXRF spectrometry to demonstrate that di-2-pyridylketone-4,4,-dimethyl-3-thiosemicarbazone (Dp44mT) is a potential candidate in chelation therapy as an iron chelator. A combined treatment of each 2 μM easily available Fe^II, Cu^II and Zn^II and 5 μM Dp44mT on eight different cancer cell lines resulted in a 10–40 fold increase in the intracellular Cu content compared to the control samples. The uptake of Cu and Cu-cytotoxicity strictly depended on the Cu concentration of the culture medium, interestingly even low concentrations of 0.1 μM of Dp44mT could transport high amounts of copper inside the cell and the copper accumulation and toxicity through the chelator could hardly be influenced by iron. The results showed that Dp44mT in combination with copper was highly toxic in vitro.

Visualisation of elemental distribution in biological tissues is gaining importance in many disciplines of biological, forensic, and medical research. Figueroa et al.³⁴⁹ developed a robotic system to image rapidly the chemistry of large specimens at ppm concentration levels. Multiple images of the distribution of elements could be obtained on surfaces of 100 × 100 mm² and a spatial resolution of up to 0.2 mm² per pixel, with a spectral capture time up to 1 ms per point. This system has proven to be highly efficient for the XRF mapping of elements in large biological samples, achieving comparable results to those obtained by μ-SRXRF spectrometry. However, the new imaging system enabled the XRF scanning to be achieved in a few minutes, whereas μ-SRXRF measurements required several hours. Poitry-Yamate et al.³⁵⁰ explored the potential of a low-energy SRXRF technique to image the stable isotope of F in phosphorylated 18-fluorodeoxy-D-glucose (FDG) at 1 μm² spatial resolution in 3 μm thick brain slices. The excitation-dependent fluorescent F signal at 676 eV varied linearly with FDG concentration between 0.5 and 10 mM, whereas the endogenous F background signal was undetectable in brain tissue. The outcome of their mapping experiments demonstrated the ability to identify glucose uptake at subcellular resolution and holds remarkable potential for imaging glucose metabolism in biological tissue. Kuang et al.³⁵¹ investigated the feasibility of imaging Pt-based drug distributions in a water phantom using XRFCT. A 5 mm diameter pencil beam produced by a polychromatic X-ray source equipped with a tungsten anode was used in combination with a cadmium telluride detector. The phantom was translated and rotated relative to the stationary pencil beam in a first-generation CT geometry. The distribution of the Pt drug in the water phantom was clearly identifiable in the reconstructed XRF images, providing evidence that XRFCT is a promising approach for monitoring the spatial distribution of Pt drugs. Selenium metabolism in cancer cells was studied by Weekley et al.,³⁵² who combined the XAS and XFM techniques to link the speciation and distribution of Se to its biological activity. These techniques were preferred over the traditional hyphenated techniques that required extensive sample preparation which might affect speciation. By investigating the cellular response of E. coli to Hg²⁺ exposure, Gao et al.³⁵³ demonstrated the high sensitivity of SRXRF spectrometry in identifying metal-associated proteins compared to conventional proteomic approaches. Escherchia coli was cultured in the LB medium containing HgCl₂ and/or selenomethionine. After 2D gel electrophoresis, the distribution of Hg in the gel was measured with SRXRF spectrometry at the 4W1B, Beijing Synchrotron Radiation Facility. The proteins with differential expression and those containing Hg were identified with ESI-MS/MS and peptide mass fingerprinting analysis. The results showed that Hg²⁺ could inhibit the growth of E. coli, while a supplement of selenomethionine could shorten the lag time induced by Hg²⁺, indicating an antagonistic effect of selenomethionine against Hg²⁺ toxicity. Robison et al.³⁵⁴ continued their research regarding the Mn distribution in brains, related to Parkinson's disease. Using XRF imaging, the distribution of Mn in the hippocampal formation was measured for a rodent model of chronic Mn exposure and quantitatively compared with distributions of other biologically relevant metals such as Fe and Zn. Zhang et al.³⁵⁵ examined the biodistribution and pulmonary toxicity of nanoscale titanium dioxide (nanoTiO₂) in mice at the elemental level by SRXRF spectrometry. Notably, nanoTiO₂ particles were mainly retained in lungs after intratracheal instillation, slowly clearing but were still present after 3 months. NanoTiO₂ was shown to interfere with the natural distribution of Ca, Cu, Fe, K, and Zn in lungs.

Gold nanoparticles (GNPs) may be used as a contrast agent to identify tumour location and can be modified to target and image the biological characteristics of specific tumours. Ricketts et al.³⁵⁶ provided evidence that the high detecting sensitivity of GNPs and greater depth imaging could be achieved using two newly developed XRF systems. The first system consisted of a 10 mm² SDD coupled to a slightly focusing polycapillary optic, which allowed 2D energy resolved imaging in step and scan mode. The system had a sensitivity to GNP concentrations as low as 1 ppm and GNP concentrations different by a factor of 5 could be resolved, offering the potential to distinguish tumour from non-tumour. The second system was designed to avoid slow step and scan image acquisition; the feasibility of exciting the whole specimen with a wide beam and detecting fluorescent X-rays with a pixellated controlled drift energy resolving detector without scanning was investigated. A parallel polycapillary optic coupled to the detector was successfully used to ascertain the position where fluorescence was emitted. The tissue penetration of the technique was demonstrated to be sufficient for near-surface small-animal studies, and for 3D imaging in vitro cellular constructs. Wu et al.³⁵⁷ used the XRF technique to quantify GNPs of 1.9 and 15 nm individually and within chicken breast tissue samples. An optimal combination of excitation and emission filters was proposed to segregate the fluorescence spectra at 66.99 and 68.80 keV from the background scattering. A roadmap method was developed to subtract the scattered radiation (acquired before the insertion of GNP solutions) from the signal radiation acquired after the GNP solutions were inserted. The methods effectively minimised the background scattering in the spectrum measurements, showing linear relationships between GNP solutions from 0.1 to 10% m/m and from 0.1 to 1.0% m/m inside a chicken breast tissue sample. The feasibility of using EDXRF spectrometry for a rapid quantitative analysis of Au in tumour tissue was investigated by Magana et al.,³⁵⁸ who analysed sample biopsies in as little as 8 hours with minimal sample preparation. Samples were prepared by drying a 35 μL aliquot of tissue dissolved in KOH in a standard EDXRF cup on a Prolene® support, producing a thin sample. Calibration curves of XRF peak intensity (cps per mA) to the Au concentration (0–50 ppm) were prepared for liver, tumour, and a mix of tissue types. A linear regression demonstrated an R² correlation of 0.93. Analysis of experimental samples showed that Au accumulation in tumours (from 5.8 to 41.3 ppm) was in agreement with previous studies, where samples were analysed using INAA or ICP-MS. Mahdavi et al.³⁵⁹ designed a system to analyse Au and other heavy elements in internal organs using in vivo XRF analysis. Using Monte Carlo to simulate a source of ⁹⁹Tc to excite Au X-rays in kidney tissue, the changes in K XRF response due to variations in tissue thickness overlying the kidney at the measurement site were investigated. The results showed that a Au concentration between 3 and 10 μg g⁻¹ kidney tissue could be detected for a distance between the skin and the kidney surface of 20 to 60 mm. Finally, Kempson et al.³⁶⁰ studied the fate of intravenously administered GNPs, in the hair follicles of mice over a two week period using XRF spectrometry, whereby confocal microscopy was implemented for the 3D reconstruction of nanoparticle distributions.

Trace elements may contribute to the formation of calculi or stones or be involved in the etiopathogenesis of stone diseases (kidney, gall, ureteral, bladder). The composition and spatial distribution of elements from the inner nucleus to outer crust of a human cardiac calculus were investigated by Cheng et al.,³⁶¹ combining EDXRF and SEM-EDS measurements. Twenty-five elements were identifiable from 18 positions on the cardiac calculus by EDXRF measurements, in which the highest concentrations of toxic trace elements (Al, Au, Hg, Ni, Pb, Pt, Si, Sn, W) and higher levels of essential trace elements (Ca, Cr, P and Sr) were detected. A moderate positive was reported in Pearson's correlation coefficient between trace element concentrations of Ca, Mg or P and location differences from centre to periphery in the cardiac calculus were observed. A positive correlation was also found for Ca/Zn and Ca/Cu, indicating the gradual increase of Ca concentration from inner nucleus to outer crust of cardiac calculus. The SEM-EDS analysis revealed that the calculus was predominantly composed of calcium hydroxyapatite and cholesterol. Gallstones from four patients from Europe and particularly from England (including samples from a mother and a daughter) and Greece, were analysed by Athanasiadou et al.,³⁶² using a multi-technique approach including XRF spectrometry. The XRF results revealed that Ca was the dominant non-organic metal in all gallstones (up to about 1.95% m/m) together with Cu, Fe, Ni and Pb (up to about 19 mg kg⁻¹ for each metal). Gallstones in patients from England contained additional Mn (up to about 87 mg kg⁻¹) and Zn (up to about 6 mg kg⁻¹), while the sample from the mother contained negligible Zn and Mn, compared to that of her daughter, but significant As (about 4.5 mg kg⁻¹). All cholesterol gallstones examined were well enriched in potentially toxic metals (Pb, as well as Ni in one case) and metalloids (As also in one case) as compared to the global average. Ball et al.³⁶³ used SRXRF imaging and XAS to study the distribution of Ga in bladder tissue following oral administration of gallium maltolate during urinary tract infection. The median concentration of Ga in homogenised bladder tissue from infected mice was 1.93 μg g⁻¹ after daily administration of gallium maltolate for 5 days. The results of this study indicated that gallium maltolate might be a suitable candidate for further development as a novel antimicrobial therapy for urinary tract infections caused by uropathogenic E. coli.

In vivo monitoring of trace and bio-metals in skin samples is normally quantified using phantoms that assume a constant elemental distribution within the skin. Layered calibration skin phantoms could potentially improve the reliability by better representing the actual in vivo distribution. Desouza et al.³⁶⁴ investigated the distribution of Ca, Fe and Zn in prepared human skin samples taken from a number of locations on the body. Slices (orientation running from the skin surface into the dermis) were extracted from 18 formalin-fixed necropsy samples and scanned using the μ-XRF setup at the VESPERS beamline (Canadian Light Source). Elemental surface maps were produced using a 6 × 6 μm² beam in steps of 10 μm. Statistically significant differences (p < 0.01) were noted between the epidermal and dermal layers of skin for the elements examined, demonstrating the ability to clearly distinguish elemental content in each layer. These results indicated that the appropriate depth distribution should be taken into account when using phantoms to quantify elemental levels measured in the skin. Figueroa et al.³⁶⁵ presented a complete study of the dose distribution resulting from an in vivo XRF scanning for quantifying biohazard materials on human hands. Absorbed dose was estimated using dosimetric models specifically developed to this aim. In addition, complete dose distributions were obtained by full radiation transport calculations based on stochastic MC techniques. A dedicated subroutine was developed using the PENELOPE 2008 main code integrated with dedicated programs for the 3D visualisation of dose distribution. The obtained results showed very good agreement between approximate analytical models and full descriptions by means of MC simulations. Soares et al.³⁶⁶ performed SR-TXRF measurements to determine the concentrations of trace elements in four skin lesions related to skin disorder research. The concentrations of Ca, Cu, Fe, K, P, S, Rb and Zn were evaluated in 62 pairs of lesions and healthy samples, each one having been collected from the same patient. The results revealed significant differences of Ca, Cu, Fe, K and P levels as well as a common trend in their variations between lesion and control samples among the skin diseases.

Selenium, an essential trace element with potential anti-atherogenic and antioxidant effects, was determined by XRF spectrometry in 81 hair samples of hyperlipidemic patients and compared with the data obtained from 43 healthy volunteers.³⁶⁷ Hair selenium levels were found to be significantly higher in hyperlipidemic patients compared with volunteers with normal lipid levels. These findings warrant further investigations to study the impact of selenium supplementation on the incidence of cardiovascular events. Liu et al.³⁶⁸ used a confocal μ-XRF system based on a polycapillary X-ray lens and a conventional X-ray source to scan elemental distributions along a single hair. A polycapillary focusing X-ray lens with a large gain in power density was used to decrease the requirement for a high power X-ray source normally required in this confocal technology, The system made no use of confocal depth profiling but used the high sensitivity of the setup to perform trace element measurements on this laboratory system.

A calibration method was developed by Fleming et al.,³⁶⁹ to assess the elemental concentrations of Mn and Zn in nail clippings using a portable XRF device. In two separate sets of phantom clippings, the Mn and Zn concentrations were varied from 10 to 50 μg g⁻¹, in increments of 10 μg g⁻¹. Additionally, for each concentration, five distinct masses of sample were prepared ranging from 20 to 100 mg, in increments of 20 mg. Linear calibrations were set up using the K characteristic X-ray lines. An empirical function was fitted to the slope–mass relationship and compared with those obtained previously from other elements. Gherase et al.³⁷⁰ investigated the As distribution in nail clippings from three healthy human subjects using the μ-beam experimental setup of the hard X-ray μ-analysis beamline from the Canadian Light Source synchrotron. Fingernail and toenail clippings were embedded in polyester resin and cut in cross-sectional slices with an average thickness of 270 μm. A calibration line for the As Kα peak was produced from cross sections of nail phantom clippings with concentrations of As ranging from 0 to 20 μg g⁻¹, in increments of 5 μg g⁻¹. The energy of the X-ray beam was set at 13 keV for optimal excitation of As and the beam size was 28 × 10 μm.² Each sample was analysed using a point-by-point scanning technique in a 45° beam-sample and 90° beam-detector geometry, using a step size of 50 μm in both the horizontal and vertical directions for all samples. The distribution maps showed that small regions (<0.1 mm²) with higher As concentrations (>1 μg g⁻¹) were located predominantly in the ventral and dorsal layers of the nail. The results were in agreement with findings reported in a recent study and could be linked to nail histology and keratin structure.

The evaluation of an EDXRF imaging system equipped with a detector based on a micro patterned gas detector, named the Micro-Hole and Strip Plate, was conducted by Silva et al.,³⁷¹ who measured the diffusion of the major constituents, Hg and Zn, of a dental amalgam through human tooth structures. The full field of view system, with an active area of 28 × 28 mm² presented some important features for EDXRF imaging applications, such as a position resolution of less than 125 μm, an intrinsic energy resolution of about 14% FWHM for 5.9 keV X-rays, and a counting rate capability of 0.5 MHz. The distribution pattern of the elements Hg and Zn suggested diffusion of these elements from the amalgam to teeth tissues. Soares et al.³⁷² studied the effect of Er:YAG laser etching on dentin by μ-EDXRF spectrometry and SEM. The results of treated tooth samples showed a significant reduction of the Ca and P content after acid etching (control group) and an increase after laser irradiation with 220 mJ. Moreover, the μ-EDXRF mapping illustrated that acid etching created a homogeneous distribution of inorganic elements over dentin, while Er:YAG laser etching (220 mJ) produced an irregular elemental distribution and changed the stoichiometric proportions of hydroxyapatite, as showed by an increase in the mineral content.

The distribution and accumulation of Pb, Sr and Zn in structural units of human bone tissue was investigated by Pemmer et al.,³⁷³ using μ-SRXRF spectrometry in combination with quantitative backscattered electron (BE) imaging. Fourteen human bone samples from individuals with osteoporotic femoral neck fractures as well as from healthy individuals were analysed. Fluorescence intensity maps were matched with BE images and correlated with Ca content. The results showed that Pb and Zn had significantly increased levels in the cement lines of all samples compared to the surrounding mineralised bone matrix, whereas Pb and Sr levels were found to be correlated with the degree of mineralisation. Interestingly, Zn intensities had no correlation with Ca levels. A similar analytical approach was applied to study the differential accumulation of Pb and Zn in the double-tidemarks of articular cartilage, revealing various timescales for the accumulation of these elements.³⁷⁴ Dessombz et al.³⁷⁵ examined the presence of Zn²⁺ in osteoarthritis cartilage and meniscus from a patient undergoing a total knee joint replacement. At the mesoscopic scale, Ca²⁺ and Zn²⁺ were mapped by μ-XRF, μ-XANES and μ-XRD techniques to determine the spatial distribution of these elements in cartilage, to identify the Zn species, and to determine the chemical nature of the calcification, respectively. In their research related to osteogenesis, de Abreu et al.³⁷⁶ showed the possibilities of the μ-EDXRF technique to evaluate qualitatively and semi-quantitatively biochemical changes in the composition of bone tissue during the repair process in rats after laser therapy and pharmacotherapy. The benefits of two complementary methods for elemental bioimaging, μ-EDXRF spectrometry and LA-ICP-MS, were demonstrated by Blaske et al.³⁷⁷ by analysing a nanosilver-coated bone implant. Challenges caused by the physically inhomogeneous materials including bone and soft tissues were addressed by polymer embedding. Using μ-EDXRF spectrometry, fast sample mapping was achieved, detecting Ti and V signals from the metal implant as well as Ca and P signals representing hard bone tissue and S distribution representing soft tissues. Only by use of LA-ICP-MS, could the required high sensitivity and low detection limits for the determination of Ag be obtained. Ballarre et al.³⁷⁸ evaluated the bone quality around hybrid bioactive silica-based coated stainless steel implants with μ-Raman, μ-XRF and XAS techniques. Martin et al.³⁷⁹ applied SRXRF spectrometry to map the metal distribution in selected bone fragments representative of the remains associated with the Franklin expedition. In addition, LA-ICP-MS using a 25 μm diameter circular spot was employed to compare the Pb isotope distributions in small regions within the bone fragments. The XRF mapping showed Pb to be widely distributed in the bone while the Pb isotope ratios obtained by LA-ICP-MS of small areas representative of bone with different Pb exchange rates did not show statistically significant differences. These results were inconsistent with the hypothesis that the principle source of Pb in the remains of the expedition personnel should have come from faulty solder seals in tinned meat. Querido and Farina³⁸⁰ proved that the strontium ranelate treatment increased the formation of bone-like mineralised nodules in osteoblast cell cultures using attenuated total reflection FTIR spectroscopy, while EDXRF and μ-SRXRF measurements showed that Sr was incorporated into the intact nodules. Finally, Da Silva et al.³⁸¹ described the preparation of a chemically pure hydroxyapatite phantom material, of known composition and stoichiometry, for the calibration of in vivo XRF systems to quantify Pb and Sr in bone as a replacement for plaster of Paris. The issue with contamination by the analyte was resolved by preparing pure Ca(OH)₂ by hydroxide precipitation, which was found to bring Pb and Sr levels to <0.3 and <0.7 μg g⁻¹ Ca, respectively. The final crystal structure of the material was found to be similar to that of the bone mineral component of NIST SRM 1486 (bone meal), as determined by powder XRD spectrometry.

A few applications considered the analysis of fluid samples using the XRF technique. Canellas et al.³⁸² determined the low Z elements Ca, Cl, K, Na, P, and S in the human serum of patients with a blood disorder by SR-TXRF spectrometry and observed a significant difference (α= 0.05) compared with the control group. The measurements were performed at the XRF beamline at the Brazilian National Synchrotron Light Laboratory (Campinas, Sao Paulo) using a monochromatic beam with a maximum energy of 20 keV and an Ultra-LEGe detector with a resolution of 148 eV at 5.9 keV. Luo et al.³⁸³ provided a new strategy for the interference-free, simple and rapid evaluation of the concentration of Ischemia-modified albumin (IMA), a biomarker of ischemic heart disease. By combining quantum dot-coupled sandwich immunoassay to quantify total human serum albumin (HSA) with XRF spectroscopy whereby the XRF intensity of Co^II bonded to intact HSA was measured, the IMA concentration could be determined within 30 minutes by calculating the difference between total HSA and intact HSA. For the determination of inorganic elements in whole blood samples from humans and laboratory animals, EDXRF spectrometry, applying the fundamental parameters method, was used by Redigolo et al.³⁸⁴ Peripheral blood samples were collected and, before coagulation, 100 μL of sample were deposited onto a Whatman No. 41 filter paper and dried, using an infrared spotlight. A comparative study between EDXRF and INAA data was carried out and the results for both techniques were statistically identical (α= 0.05). The results contributed to the establishment of reference interval values for Ca, Cl, Ci, K, Mg, Na, P, S and Zn in the healthy Brazilian population and the referred laboratory animal species. Aranda et al.³⁸⁵ developed a solid-phase preconcentration method using Aliquat 336 (tri-octylmethylammonium chloride)-activated carbon for the determination of Hg^II in drinking water. The XRF measurements were performed directly on the solid. A 1700-fold enrichment factor was obtained, enabling the determination of Hg^II in drinking water at μg L⁻¹ levels. The described methodology showed excellent reproducibility, accuracy, and detection limits with improvements achieved by eliminating the elution step of the analyte, replacing those methods based on cold vapor generation, reducing reagent consumption, and handling of samples.

3.8 Drugs

The control of metal catalyst residues in different stages of the manufacturing processes of active pharmaceutical ingredients (APIs) and, especially, in the final product is crucial. Margui et al.³⁸⁶ developed a method using a benchtop TXRF spectrometer for the rapid and simple determination of some platinum group metal (PGM) catalyst impurities (Ir, Pd, Pt and Rh) in different types of API samples. An evaluation of different sample treatments (dissolution and digestion of the solid pharmaceutical samples) was carried out and the developed methodologies were validated according to the analytical parameters to be considered and acceptance criteria for PGM determination as defined in the United States Pharmacopeia. Limits of quantification obtained for PGM metals were in the range of 2 to 4 mg kg⁻¹, which were satisfactory according to current legislation.

The US Food and Drug Administration limits for mercury in cosmetic products amounts to 1 mg kg⁻¹. As mercury is a well-documented melanotoxin added to cosmetic skin-lightening products, Hamann et al.³⁸⁷ screened 549 products, manufactured in 32 countries and available to US consumers, for Hg contents above 200 mg kg⁻¹ by using a low-cost portable XRF spectrometer. Of the 549 tested products, 6.0% (n = 33) contained Hg above 1000 mg kg⁻¹, of which 15 samples even had an excess of 10 000 mg kg⁻¹. This study confirmed the national and global presence of Hg in skin-lightening products which can cause many dermatologic, renal and neurological problems.

A rapid and straightforward method for the quantification of the total Ag content in nano-based commercially available liquid dietary supplements was reported by Sanchez-Pomales et al.,³⁸⁸ using a portable XRF analyser. Figures of merit were evaluated by analysing a series of AgNO₃ standards, resulting in a limit of detection of 3 mg kg⁻¹, a limit of quantitation of 10 mg kg⁻¹ and a broad linear range from the limit of detection up to 1% m/m. Comparative analysis between portable XRF spectroscopy and ICP-MS resulted in a good correlation with average percentage differences under 15%. Furthermore, accurate quantification of Ag in the presence of high concentrations of potential spectral interferences was also demonstrated. Researchers from Tanzania³⁸⁹ determined essential and toxic elements with the EDXRF technique in 100 clay soils commonly consumed by pregnant women in their country. The essential elements Cu, Fe, Mn, Se and Zn, and the toxic elements As, Co, Ni, Pb, Th and U were detected in concentrations above WHO permissible limits in some of the samples. These findings underlined that the habit of eating soil was exposing the pregnant mothers and their children to metal toxicity which was detrimental to their health.

3.9 Biological

Although the list of biological publications dedicated to the use of XRF technology was limited in this year's review, readers are reminded of our companion review on Clinical and biological materials, foods and beverages⁴ for additional applications. Nevertheless some interesting applications are worthwhile to highlight in this section. Abuillan et al.³⁹⁰ conducted systematic studies on an antisepsis peptide drug (Pep 19-2.5) and fish protamine to demonstrate the unique combination of the XRR and GIXRF techniques as powerful tools to identify quantitatively different modes of interaction between complex and realistic biomembrane models and membrane-active molecules. Analysis of rodent brains with μ-XRF spectrometry combined with immunohistochemistry allowed Pushkar et al.³⁹¹ to demonstrate that local Cu concentrations were thousands of times higher in the glia of the subventricular zone (SVZ) than in other cells. The Cu K-edge XANES analysis was consistent with Cu being bound as a multi-metallic Cu–S cluster similar to the one present in Cu-metallothionein. Analysis of age-related changes showed that the Cu content in astrocytes of the SVZ increased fourfold from 3 weeks to 9 months, while the Cu concentration in other brain areas remained essentially constant. Finally, Xing et al.³⁹² studied the hierarchical structure and biomineralisation in cricket teeth using SRXRF spectrometry in combination with XRD spectroscopy, small angle X-ray scattering techniques and SEM. The results showed that Zn was the main heavy element. The surface of the cricket tooth contained crystalline ZnFe₂(AsO₄)₂(OH)₂·H₂O, while the interior had a ZnCl₂ compound.

3.10 Thin films, coatings and nanomaterials

The trend of the last year for a significant number of papers dealing with thin films for solar cells continued within the current review period. Streeck et al.³⁹³ used GIXRF for the non-destructive depth profiling of In and Ga in Cu(In,Ga)Se₂ absorber films. The authors explained that the development of highly efficient thin film solar cells involved band gap engineering by tuning their elemental composition with depth. GIXRF analysis using monochromatic SR and well-characterised instrumentation was suitable for non-destructive and reference-free compositional depth profiling of thin films. The authors were able to show that double Ga gradients in Cu(In_1−x,Ga_x)Se₂ could be resolved using the GIXRF technique. The detection and role of trace impurities in high-performance organic solar cells was discussed by Nikiforov et al.³⁹⁴ The authors argued that trace impurities in organic solar cells were known to deleteriously affect device performance. Until now it was not possible to identify the character of impurities or their concentration in organic photovoltaic active layer blends. Using SRXRF spectrometry it was possible to detect and quantify trace concentrations of metal impurities in high-performance bulk hetero-junction blends. After introducing artificially known quantities of additional catalyst into polymer/fullerene blends, the authors identified both the threshold concentration at which performance degraded and the mechanism for that degradation. The knowledge of a target impurity concentration and the technique to measure accurately its presence should help to implement the material‘s preparation processes to achieve consistent, high performance in organic solar cells. Pan et al.³⁹⁵ described the structural analysis of Cu(In_1−xGa_x)Se₂ multi-layer thin film solar cells. The complex structure of CuInGaSe (CIGS) fabricated by a two-step process (the deposition step and the salinisation or co-evaporation method) was analysed by several methods, such as RBS and ICP-OES. The latter showed that diffusion appeared at the interface between both CdS and CIGS, and Mo and CIGS. The authors used XRF spectrometry to indicate that CIGS thin film presented the highest efficiency when the content ratio of In and Ga atoms was 0.7 : 0.3. The evolution of iron-silicide precipitates in multi-crystalline silicon during solar cell processing was analysed by Schon et al.³⁹⁶ The authors reported simulations with a 2D model of the precipitation of iron during the multi-crystalline ingot crystallisation process and the redistribution of iron during subsequent phosphorus diffusion gettering. They compared the simulated size distribution of the precipitates with the XRF measurements of iron precipitates along a grain boundary and found good agreement. Additionally, the authors demonstrated that the measured decrease in the line density and the increase of the mean size of the iron precipitates after the phosphorus diffusion gettering could be reproduced by their simulations. The size and spatial distribution of iron precipitates affected the kinetics of iron redistribution during the solar cell process and, ultimately, the recombination activity of the precipitated iron.

A considerable number of publications within the review period dealt with nano-particles. Li and Chan³⁹⁷ reported the use of PtFeNi tri-metallic alloy NPs as an electro-catalyst for oxygen reduction reaction in proton exchange membrane fuel cells with ultra-low Pt loading. The core–shell structured PtFeNi alloy composed of a multi-layer Pt-rich shell and a Pt-poor core was finally produced and characterised by various methods including SEM and EDXRF spectrometry. The SEM investigations of the electrode surface surprisingly confirmed that the carbon blacks were covered with an extremely high density of nano-catalysts, while EDXRF spectroscopy confirmed that these catalysts were in fact PtFeNi. The electro-catalyst based electrodes prepared from the PtFeNi ternary alloy were used as the cathodes in fuel cells and offered good performance and stability with a Pt loading of only 0.05 mg cm⁻². Gold-NP loaded cells were studied by Astolfo et al.³⁹⁸ using X-ray based techniques for cell-tracking applications with single-cell sensitivity. Complementary high-resolution imaging techniques using different length scales were applied to elucidate the cellular loading of gold NPs and its impact on the long term, and high-resolution following the behaviour of the cells utilising X-ray technology. The authors emphasised that although the method was demonstrated for malignant cell lines, the results could be applied to non-malignant cell lines as well. In particular the accumulation of the gold marker in each cell had been assessed quantitatively by SEM, 2D XRF imaging techniques and X-ray CT with micrometer and sub-micrometer resolution. The authors further determined the 3D distribution of the incorporated NPs, which were sequestered in lysosomes as a permanent marker. The latter allowed elucidation of cell size and the gold partition during mitosis, which subsequently enabled the authors to define the optimal instrument settings for a compact μ-CT system to visualise Au loaded cells. Zhang et al.³⁹⁹ studied the preparation of silver/bacterial cellulose composite membranes and its subsequent antimicrobial activity. The XRF spectrum showed that Ag was incorporated into the bacterial cellulose. Both antimicrobial activity and biocompatibility of the Ag/bacterial cellulose nano-composites were evaluated and demonstrated good antimicrobial activities against E. coli and S. aureus. Investigation of the interaction of magnetic NPs with U87MG cells using μ-SRXRF spectrometry were reported by Gianoncelli et al.⁴⁰⁰ The authors claimed that μ-SRXRF spectrometry provided unique information that had pushed the frontiers of biological research, particularly when investigating intracellular mechanisms. An investigation was carried out into the distribution and potential toxicity of Fe₂O₃ and CoFe₂O₄ NPs in U87MG glioblastoma–astrocytoma cells exposed for 24 h to NP concentrations ranging from 5 to 250 μg mL⁻¹ to monitor both morphological and chemical changes. The SRXRF maps complemented with XRM absorption and phase contrast images revealed different intracellular distribution patterns for the two types of NPs allowing different mechanisms of toxicity to be deduced. Novak et al.⁴⁰¹ reported in vivo experimental evidence for the toxicity of Co²⁺ ions dissolved from CoFe₂O₄ NPs by the digestive juices of a model organism. PIXE and low energy SRXRF spectrometries were used to study the internal cell distribution of Co and Fe. The Co²⁺ ions were found to be more toxic than nanoparticles and were found to accumulate in the hepatopancreas, in comparison with Fe²⁺ ions and CoFe₂O₄ NPs, which were not retained in vivo. Direct L-shell XRF imaging of GNPs was demonstrated by Manohar et al.⁴⁰² who used a benchtop X-ray imaging system equipped with a polychromatic X-ray tube and a Si(PIN) detector. Water-filled cylinders of 12 mm diameter tubes containing GNP concentrations of 20, 10, 5, 0.5, 0.05, 0.005, and 0 mg cm⁻³ served as calibration phantoms. An imaging phantom was created using the same cylindrical tube but filled with tissue-equivalent gel containing structures mimicking a GNP-loaded blood vessel and an approximately 1 cm³ tumour. Phantoms were irradiated by a 3 mm diameter pencil-beam from the X-ray tube excited at 62 kV with a 1 mm aluminium filter. Measurements of the calibration phantoms showed a linear relationship between the corrected XRF signal and GNP mass per imaged voxel (0.0173 cm³). The detection limit was estimated to be 0.35 μg per imaged voxel. Fiedler et al.⁴⁰³ reported the use of EDXRF spectrometry for the direct liquid analysis, under an atmosphere of helium or air, of Au, Pd, Pt and Rh nano-particulate catalysts stabilised by imidazolium propane sulfonate-based zwitterionic surfactants. This non-destructive approach offered the possibility of reusing the samples. The signals from Au, Pd, Pt and Rh samples in the presence of imidazolium propane sulfonate-based zwitterionic surfactants before and after formation of NPs were found to be essentially identical and detection limits were reported to be in the range of 0.4–3 mg L⁻¹. Motellier et al.¹² used GIXRF to investigate the deposition of TiO₂ NPs in suspension with a particular interest in the influence of substrate pretreatment. The authors emphasised that the major advantages of this technique was the possibility to analyse the particles without pre-treatment avoiding the harsh acid digestion required by most conventional methods. However, they also pointed out that reliable quantitative measurements required a number of precautions; particularly that the deposition process of the sample on the flat reflecting substrate must maintain uniformity over the entire surface of the deposition residue once dried. Linear calibration curves using internal standardisation were established with ionic Ti and with two different types of TiO₂ NPs. Detection limits of 18 μg L⁻¹ and 52 μg L⁻¹ at incident angles of 0.20° and 0.75°, respectively, were obtained. The authors found that correlation coefficient of the fitted linear calibration was particle-size dependent, which was assigned to sampling problems due to possible incomplete dispersion of the particles in the suspension. They also observed that the measured XRF signal of the dried deposits changed within a 4 month timespan for both types of TiO₂ NPs, demonstrating the very peculiar behaviour of these particulate samples.

Zawisza and Sitko⁴⁰⁴ reported the use of electrochemically assisted sorption on oxidised multi-walled carbon nanotubes (MWCNT) for the preconcentration of Co, Cr, Cu, Mn, Ni, and Zn from water samples. The proposed method was based on the application of an electric field to enhance the sorption process on oxidised MWCNT and was used to preconcentrate trace elements from water of pH = 4, which were analysed using an EDXRF spectrometer. The various parameters including pH of the solution, amounts of oxidised MWCNTs, sample volume and the influence of voltage as well as the time of assisted electric field on the sorption process were investigated for the optimisation of the analytical procedure. The authors reported that under optimised conditions the proposed preconcentration method offered a high recovery between 92 and 99% and good precision of sample preparation within 3.5–7%. The excellent detection limits for direct EDXRF measurement for the filtered MWCNT were between 1 and 5 ng mL^−1. The same authors⁴⁰⁵ investigated the preconcentration of trace amounts of Co, Fe, Ni and Zn from aqueous samples by means of microelectrodeposition from an ionic liquid (1-butyl-3-methylimidazolium hexafluorophosphate). The preconcentrated metals were determined using EDXRF spectrometry and, under the optimised conditions, the detection limits were between 2 and 6 μg L⁻¹. The method was applied successfully to water analysis. Bussy et al.⁴⁰⁶ used μ-SRXRF mapping to establish the intracellular fate of CNTs inside murine macrophages and determined the pH-dependent detachment of iron catalyst NPs. They were able to evaluate the health impact of CNTs as previously there were limited data available on the fate of CNT inside macrophages. The authors claimed that their results, while obtained with pristine single walled CNT, could likely be extended to other catalyst-containing nanomaterials and to open new ways for the interpretation and understanding of CNT toxicity. Margui et al.⁴⁰⁷ investigated the analytical possibilities of different XRF systems for the determination of trace elements in aqueous samples pre-concentrated with CNTs. Improved instrumental sensitivity and detection limits were reported for multi-element determination of Cd, Co, Cr, Cu, Fe, Ga, Mn, Ni, Pb, Se, V and Zn in liquid samples by using different XRF configurations: a benchtop EDXRF spectrometer, a benchtop polarised EDXRF spectrometer and a WDXRF spectrometer. After the extraction step, the aqueous sample was filtered and CNTs with the absorbed elements on the filter paper, were directly analysed by EDXRF spectrometry. The detection limits in all cases were in the low ng mL⁻¹ range. Nevertheless, the results clearly indicated the benefits, in terms of sensitivity, of using polarised X-ray sources with different secondary targets in comparison to conventional EDXRF systems, in particular if Cd determination was required.

Interest in analysis of nano-layers continued: Kayser et al.⁴⁰⁸ applied high-energy-resolution grazing emission X-ray fluorescence (GEXRF) spectrometry to the characterisation of thin aluminium films on silicon wafers with nominal thicknesses in the range of 1 to 150 nm. By recording the XRF signal for different shallow emission angles, the deposited aluminium surface layers on different samples could be characterised in terms of layer thickness, layer density and surface roughness. The advantages offered by SR and the employed WD detection setup were described. Leani et al.⁴⁰⁹ reported depth profiling nano-analysis of chemical state using resonant Raman spectroscopy at grazing incidence conditions. Both TXRF and X-ray Raman scattering techniques were combined to discriminate chemical environments in depth-profiling studies using an ED system. The authors claimed that this allowed, for the first time, to resolve oxidation state on surface nano-layers with a low-resolution system. Samples of pure copper and iron oxidised in tap water and salty water, respectively, were studied in the Brazilian synchrotron facility using monochromatic radiation and an EDXRF setup. The measurements were carried out in total reflection geometry with incident energy lower but close to the K absorption edge of the respective elements. The results showed the presence of very thin oxide layers, usually not observable with conventional geometries. The setup also permitted identification of the specific compound present at a particular depth in the sample with nano-meter or even sub-nano-meter resolution using TXRF.

Mashin et al.⁴¹⁰ studied the effect of the fluorescence from bulk substrates on the fluorescent intensity of vanadium nanofilms of various thicknesses and examined V Kα fluorescence enhancement by emission from the substrate. The authors calculated correction coefficients that took into account film-substrate inter-element effects for the analysis of real nanofilm structures.

Several applications involving the measurement of nanomaterials were reported in the review period: Zohoori and Karimi⁴¹¹ investigated the effect of nano SrTiO₃ supporting nano-TiO₂ on self-cleaning of cotton fabric. The photocatalytic activity of nano-strontium titanate mixed with various concentrations of nano-titania was examined under UV irradiation. The amount of nano-titania and nano-strontium titanate particles loaded on cotton fabrics was investigated using XRF spectrometry and crystallinity of coatings by XRD. Adding nano-strontium titanate to nano-titania showed the most promising photocatalytic activity toward dye degradation. El-Atwani et al.⁴¹² reported the effect of silicide formation on ion-induced nano-patterning of silicon with various ultra-thin metal coatings. Silicon substrates coated with 10 nm nickel, iron, and copper were irradiated with 200 eV Argon ions at normal incidence. Real time grazing incidence small angle X-ray scattering and XRF measurements were performed during this irradiation process. The real time measurements revealed threshold conditions for nano-patterning of silicon. The authors found that silicide formation lead to nano-structure formation and the weakening of the ion-induced mass redistribution. Kloust et al.⁴¹³ investigated in situ functionalisation and plasma electrolytic oxidation coating of iron oxide nanocrystals using seeded emulsion polymerisation. With the simple addition of functional surfactants, functional monomers or functional polymerisable linkers during the seeded emulsion polymerisation process, a broad range of in situ functionalised polymer-coated iron oxide nanocrystals was obtained. The authors concluded that this TXRF methodology would encourage synthesis and optimisation of a broad scope of nanocrystals.

Gottardi et al.⁴¹⁴ reported the effect of different levels of Nd³⁺ atoms incorporated in the microstructure and chemical structure of RF sputtered zinc oxide thin films. Undoped and Nd-doped zinc oxides films were deposited by RF co-sputtering from pure ZnO and metallic neodymium targets in argon plasma onto silicon, quartz and glass substrates. The Nd concentration in the zinc oxide host matrix was varied in the range 0–26 atom %. A comprehensive characterisation of the film properties was performed by XPS and ICP-OES, XRF, XRD and SEM techniques. At low Nd atomic concentration (Nd/Zn < 0.07) Nd atoms were shown to be successfully incorporated into the zinc oxide matrix, whose crystalline structure was preserved.

Cesareo et al.⁴¹⁵ developed a reconstruction algorithm for the characterisation of multi-layered samples based on measuring Kα/Kβ or Lα/Lβ XRF intensity ratios by EDXRF spectrometry. Their results showed that a correct calculation of the peak ratio of elements from XRF spectra, could provide important information for assessing the exact location of each layer and for calculating its thickness. This methodological approach might have important applications not only in materials science but also when dealing with the conservation and restoration of multi-layered cultural heritage objects where the use of a non-destructive techniques is often of paramount importance in achieving the best results. Ricou and Wood⁴¹⁶ used XRF analysis to determine the composition of fluoropolymer-based clear and white painted coatings. The authors described a new method for the quantification of organic and inorganic phase fractions for such fluoropolymer/acrylic-blended paints on aluminium panels using WDXRF spectrometry. The method also measured silica and titanium dioxide levels by comparing both WDXRF and XPS data. Not surprisingly, for clear coat samples, WDXRF provided far superior quantitative results compared with XPS.

3.11 Chemical state and speciation

Within the review period a significant number of papers dealing with chemical speciation using XAS applied to the area of energy research were published. Chemical speciation at buried interfaces in high-temperature processed polycrystalline silicon thin-film solar cells on ZnO:Al was described by Becker et al.⁴¹⁷ The presence of a ZnO:Al layer in the solar cell stack was found to limit the polysilicon solar cell performance with open circuit voltages, only below 390 mV (compared with 435 mV without ZnO film), even if a silicon nitride diffusion barrier was included. A considerable amount of diffused Zn inside the silicon was observed by means of GIXRF spectrometry; a depth-resolving analysis. Temperatures above 1000 °C were found to promote the formation of new chemical compounds within about 10 nm of the interface. The authors claimed that these results gave valuable insights about the temperature-limitations of Si/ZnO thin-film solar cell fabrication and the formation of high-mobility ZnO-layers by thermal annealing. Plutonium uranium mixed oxides (MOX) are widely used in nuclear reactor fuels, prompting Degueldre et al.⁴¹⁸ to characterise americium in these MOX fuels by μ-XRF spectroscopy and μ-XAS. The chemical bonds, valences and stoichiometry of Am (∼0.66% m/m) were determined from experimental data gained for the irradiated fuel material examined in its peripheral zone (rim) of the fuel. Also, the same group⁴¹⁹ analysed curium in plutonium MOX. The authors studied the local occurrence, speciation and next-neighbour environment of Cm in the (Pu,U)O₂ lattice within an irradiated MOX sample by means of μ-XRF spectroscopy and μ-XAFS. Menzel et al.⁴²⁰ used confocal μ-XRF-XAS for surface and depth-profiling of lithium-ion battery cathodes at different cycle states. The cathode material, LiNi_0.5Mn_1.5O₄ for lithium-ion batteries, was studied at the Mn–K edge and the Ni–K edge. The authors claimed this technique allowed a non-destructive, spatially resolved (3D) investigation of the oxidation states of surface areas and, to some extent, of deeper layers of the electrode. With this approach the degradation of Mn³⁺ to Mn⁴⁺ at the surface of the electrode could be directly shown. Bowden et al.⁴²¹ performed XAS and μ-XRF measurements on nickel cathodes from sodium-beta-alumina batteries and investigated Fe–Ni–Cl chemical associations. Sections of Na–Al–NiCl₂ cathodes from sodium-beta-alumina ZEBRA batteries were characterised with XRF mapping, and XANES measurements to probe the microstructure, elemental correlation, and chemical speciation after voltage cycling. Cycling was performed under identical load conditions at either 240° or 280 °C operating temperature and subsequently quenched in either the charged or discharged state. An FeS additive, introduced during battery synthesis, was found to be present as either Fe metal or an Fe^II chloride in all cathode samples. XRF mapping revealed an operating temperature and charge-state dependent spatial correlation between Cl, Fe and Ni concentrations. XANES measurements indicated that both Fe and Ni were chemically reactive and shifted between metallic and chloride phases in the charged and discharged states, respectively. However the percentage of chemically active Fe and Ni was significantly less in the cell operated at the lower temperature. The authors point out that low cost semi-conductor photocatalysts that can efficiently harvest solar energy to generate hydrogen from water or biofuels will be critical to future hydrogen economies. Pomiro et al.^422,423 determined the oxidation state of Mn and V in Mn_2−xV_1+xO_{4(x=0, 1/3 and 1)} magneto-resistive spinel using XANES and high-resolution Kβ XRF spectra. The authors noted that such determinations were extremely difficult by conventional spectroscopic methods, but their X-ray approach succeeded. They showed good agreement with detection by XAS spectra analysis and claimed that the obtained results might represent valuable and useful data for chemical state in characterising spinel-type oxides materials.

Only three papers dealing with material science have been found to be of specific interest. Radisavljevic et al.⁴²⁴ reported structural aspects of changes induced in lead telluride by doping with gallium, indium and manganese. EXAFS measurements were performed at Te-, Mn-, In- and Ga–K absorption edges at different temperatures and were complemented with XRD, flame AAS and XRF spectroscopy in order to derive complete information about elemental concentration, crystal structure, local environment around constitutive and impurity atoms (including their displacements from the regular lattice positions), local and long-range ordering and the overall influence of doping on the host crystal structure. The authors claimed that the obtained results represented an important step towards understanding the structural aspects of doping of lead telluride-based semi-conductors with Mn and group III elements and their connection to electronic and optical properties important for their applications. Yamamoto et al.⁴²⁵ analysed the coordination environment of aluminium species in zeolites and amorphous silica-alumina by XAS and X-ray emission spectroscopy. The coordination numbers of Al species in zeolites and amorphous silica-alumina samples were investigated by XANES, WDXRF and extended X-ray emission fine structure (EXEFS) spectroscopies. These spectra of zeolite samples were identical to those of four-coordinated reference compounds and were distinguishable from those for six-coordinated samples. Malherbe and Claverie⁴²⁶ performed Cr speciation in solids using WDXRF of the Cr Kβ lines by looking at the differences in the Kβ transition profiles between Cr⁰, Cr^III and Cr^VI compounds. Three different approaches were compared to determine the Cr^VI fraction of known mixtures: relative height and peak fitting using calibration mixtures, PLS of pure compounds, and PCR of pure compounds. The accuracy of these methods was found to be about the same with an average relative error in the range of 15%.

4. Abbrevations

2D/3D	2 dimensional/3 dimensional
AAS	Atomic absorption spectrometry
ABS	Acrylonitrile–butadiene–styrene
AD	Anno domini
AFM	Atomic force microscopy
API	Active pharmaceutical ingredients
APS	Advanced photon source
ASIC	Application-specific integrated circuit
ASTM	American Society for Testing and Materials
BC	Before Christ
BCE	Before common era
BE	Backscattered electron
BP	Before present
CART	Classification and regression trees
CCD	Charge coupled detector
CdTe	Cadmium telluride
CIGS	Copper indium gallium selenide
CNT	Carbon nanotube
CRM	Certified reference material
CT	Computer tomography
DMSA	Dimercaptosuccinic acid
DTA	Differential thermal analysis
DTG	Differential thermogravimetry
ED	Energy dispersive
ED-EPMA	Energy dispersive-electron probe microanalysis
EDTA	Ethylenediaminetetraacetic acid
EDXRF	Energy dispersive X-ray fluorescence
EPA	Environmental Protection Agency
EPMA	Electron probe microanalysis
ESI-MS/MS	Electrospray ionisation mass spectrometry/mass spectrometry
EXAFS	Extended X-ray absorption fine structure
EXEFS	Extended X-ray emission fine structure
FDG	Fluorodeoxy-D-glucose
FP	Fundamental parameter
FTIR	Fourier transform infrared
FWHM	Full width at half maximum
FXCT	Fluorescence X-ray computed tomography
GA	Genetic algorithm
GC-MS	Gas chromatography-mass spectrometry
GEMAS	Geochemical mapping of agricultural and grazing land soils of Europe
GEXRF	Grazing exit X-ray fluorescence
GIXRD	Grazing incidence X-ray diffraction
GIXRF	Grazing incidence X-ray fluorescence
GLM	General linear model
GNP	Gold nanoparticle
GSO	Gd2SiO5
HPGe	High purity germanium
HPSi	High purity silicon
HSA	Human serum albumin
IAEA	International Atomic Energy Agency
ICP	Inductively coupled plasma
ICP-OES	Inductively coupled plasma optical emission spectrometry
ICP-MS	Inductively coupled plasma mass spectrometry
IC	Ion chromatography
IDC	Invasive ductal breast cancer
IMA	Ischemia-modified albumin
INAA	Instrumental neutron activation analysis
IR	Infrared
ISII	Inner-shell ionisation impacts
ISO	International Organization for Standardization
ITO	indium tin oxide
LA-ICP-MS	Laser ablation inductively coupled plasma mass spectrometry
LB	Lysogeny broth
LEGe	Low energy germanium
LIBS	Laser induced breakdown spectrometry
LOI	Loss on ignition
MC	Monte Carlo
MOX	Mixed oxide
MSL	Mars Science Laboratory
MSPE	Magnetic solid phase extraction
MWCNT	Multi-walled carbon nanotubes
NASA	National Aeronautics and Space Administration
NEXAFS	Near edge X-ray absorption fine structure
NIST	National Institute of Standards and Technology
NP	Nanoparticle
PA	Polyamide
PBR	Peak background ratio
PCA	Principal component analysis
PCR	Principal component regression
PGM	Platinum group metals
PIXE	Particle-induced X-ray emission
PLS	Partial least squares
PM	Particulate matter
PXRF	Portable X-ray fluorescence
RoHS	Restriction of hazardous substances
RBS	Rutherford backscattering spectrometry
REE	Rare earth elements
RF	Radio frequency
RF	Random forests
RM	Reference material
RSD	Relative standard deviation
SD	Secure digital
SDD	Silicon drift detector
SAXS	Small angle X-ray scattering
SEM	Scanning electron microscopy
SEM-EDS	Scanning electron microscopy-energy dispersive X-ray spectrometry
Si(PIN)	Silicon PIN detector device
SLNL	Synchrotron Light National Laboratory
SPM	Scanning probe microscopy
SRCT	Synchrotron radiation computed tomography
SRM	Standard reference material
SR	Synchrotron radiation
XRM	X-ray microscopy
SRS-XRF	Synchrotron radiation rapid scanning-X-ray fluorescence
SR-TXRF	Synchrotron radiation-total reflection X-ray fluorescence
SRXRD	Synchrotron radiation X-ray diffraction
SRXRF	Synchrotron radiation X-ray fluorescence
SSRF	Shanghai Synchrotron Radiation Facility
SSRL	Stanford Synchrotron Radiation Lightsource
STJ	Superconducting tunnel junction
SVZ	Subventricular zone
TC	Technical committee
TEM	Transmission electron microscope
TGA	Thermogravimetric analysis
TXRF	Total reflection X-ray fluorescence
UAE	Ultrasound-assisted single extraction
UK	United Kingdom
US	United States
USA	United States of America
VIS/NIR	Visible/near-infrared spectroscopy
WDXRF	Wavelength dispersive X-ray fluorescence
WG	Working group
WHO	World Health Organisation
WTR	Waste tyre rubber
XAFS	X-ray absorption fine structure
XANES	X-ray absorption near edge structure
XAS	X-ray absorption spectroscopy
XFM	X-ray fluorescence microscopy
XPS	X-ray photoelectron spectroscopy
XRD	X-ray diffraction
XRF	X-ray fluorescence
XRFCT	X-ray fluorescence computed tomography
XRF-XAS	X-ray fluorescence-X-ray absorption spectroscopy
XRMF	X-ray microfluorescence
XRR	X-ray reflectometry
YAG	Yttrium aluminium garnet
Z	Atomic number
ZEBRA	Zeolite Battery Research Africa

References

1. M. West, A. T. Ellis, P. J. Potts, C. Streli, C. Vanhoof, D. Wegrzynek and P. Wobrauschek, J. Anal. At. Spectrom., 2013, 28(10), 1544–1590.
2. E. H. Evans, M. Horstwood, J. Pisonero and C. M. M. Smith, J. Anal. At. Spectrom., 2013, 28(6), 779–800.
3. R. Clough, C. F. Harrington, S. J. Hill, Y. Madrid and J. F. Tyson, J. Anal. At. Spectrom., 2013, 28(8), 1153–1195.
4. A. Taylor, M. P. Day, S. Hill, J. Marshall, M. Patriarca and M. White, J. Anal. At. Spectrom., 2013, 28(4), 425–459.
5. S. Carter, A. S. Fisher, M. W. Hinds, S. Lancaster and J. Marshall, J. Anal. At. Spectrom., 2013, 28(12), 1814–1869.
6. O. T. Butler, W. R. L. Cairns, J. M. Cook and C. M. Davidson, J. Anal. At. Spectrom., 2014, 29(1), 17–50.
7. E. Margui and R. Van Grieken, X-ray Fluorescence Spectrometry and Related Techniques, 2013, ISBN-13: 978-1-60650-391-1.
8. J. Zheng, K. Tagami, S. Homma-Takeda and W. T. Bu, J. Anal. At. Spectrom., 2013, 28(11), 1676–1699.
9. L. P. Zhitenko, Inorg. Mater., 2013, 49(14), 1294–1302.
10. Q. Zhang, Z. S. Yu, X. L. Li and G. H. Li, Spectrosc. Spectr. Anal., 2013, 33(12), 3402–3407.
11. P. A. Mello, J. S. Barin, F. A. Duarte, C. A. Bizzi, L. O. Diehl, E. I. Muller and E. M. M. Flores, Anal. Bioanal. Chem., 2013, 405(24), 7615–7642.
12. S. Motellier, S. Derrough, D. Locatelli, M. Amdaoud and K. Lhaute, Spectrochim. Acta, Part B, 2013, 88, 1–9.
13. G. L. Bosco, TrAC, Trends Anal. Chem., 2013, 45, 121–134.
14. F. Li, L. Q. Ge, Y. Y. Luo, Q. X. Zhang and Y. Gu, Spectrosc. Spectr. Anal., 2013, 33(6), 1711–1713.
15. W. B. Jia, Y. Zhang, C. G. Gu, D. Q. Hei, Y. S. Ling and Q. Shan, Sci. China: Technol. Sci., 2014, 57(1), 39–43.
16. P. S. Athiray, S. Narendranath, P. Sreekumar, S. K. Dash and B. R. S. Babu, Planet. Space Sci., 2013, 75, 188–194.
17. S. Z. Weider, K. H. Joy, I. A. Crawford, B. J. Kellett, B. M. Swinyard and C. J. Howe, Icarus, 2014, 229, 254–262.
18. C. A. Evans, M. J. Calaway, M. S. Bell and K. E. Young, Acta Astronaut., 2013, 90(2), 289–300.
19. E. B. Rampe, M. D. Kraft and T. G. Sharp, Icarus, 2013, 225(1), 749–762.
20. N. Stivaletta, F. Dellisanti, M. D'Elia, S. Fonti and F. Mancarella, Icarus, 2013, 224(1), 86–96.
21. J. C. Stern, A. C. McAdam, I. L. Ten Kate, D. L. Bish, D. F. Blake, R. V. Morris, R. Bowden, M. L. Fogel, M. Glamoclija, P. R. Mahaffy, A. Steele and H. E. F. Amundsen, Icarus, 2013, 224(2), 297–308.
22. T. Nakazawa and K. Tsuji, X-Ray Spectrom., 2013, 42(5), 374–379.
23. T. Nakazawa and K. Tsuji, X-Ray Spectrom., 2013, 42(3), 123–127.
24. K. Nakano, K. Akioka, T. Doi, M. Arai, H. Takabe and K. Tsuji, ISIJ Int., 2013, 53(11), 1953–1957.
25. S. Peng, Z. G. Liu, T. X. Sun, Y. Z. Ma and X. L. Ding, Anal. Chem., 2014, 86(1), 362–366.
26. G. C. Zhao, T. X. Sun, Z. G. Liu, H. Yuan, Y. D. Li, H. H. Liu, W. G. Zhao, R. X. Zhang, Q. Min and S. Peng, Nucl. Instrum. Methods Phys. Res., Sect. A, 2013, 721, 73–75.
27. P. Wrobel and M. Czyzycki, Talanta, 2013, 113, 62–67.
28. S. Peng, Z. G. Liu, T. X. Sun, Y. D. Li, H. H. Liu, W. G. Zhao, G. C. Zhao, X. Y. Lin, P. Luo, Q. L. Pan and X. L. Ding, Spectrosc. Spectr. Anal., 2013, 33(8), 2223–2226.
29. M. Dehlinger, C. Fauquet, S. Lavandier, O. Aumporn, F. Jandard, V. Arkadiev, A. Bjeoumikhov and D. Tonneau, Nanoscale Res. Lett., 2013, 8, 6.
30. S. Komatani, K. Nakamachi, K. Nakano, S. Ohzawa, H. Uchihara, A. Bando and K. Tsuji, Nucl. Instrum. Methods Phys. Res., Sect. B, 2013, 309, 260–263.
31. J. Hardy and B. Scruggs, Spectroscopy, 2013, 8–9.
32. M. I. Mazuritskiy, A. M. Lerer, A. A. Novakovich and R. V. Vedrinskii, Jetp Lett., 2013, 98(3), 130–133.
33. M. I. Mazuritskiy, S. B. Dabagov, K. Dziedzic-Kocurek and A. Marcelli, Nucl. Instrum. Methods Phys. Res., Sect. B, 2013, 309, 240–243.
34. W. X. Cong, H. O. Shen, G. Cao, H. Liu and G. Wang, J. X-Ray Sci. Technol., 2013, 21(1), 1–8.
35. G. Fu, L. J. Meng, P. Eng, M. Newville, P. Vargas and P. La Riviere, Med. Phys., 2013, 40(6), 11.
36. T. Ohmori, S. Kato, M. Doi, T. Shoji and K. Tsuji, Spectrochim. Acta, Part B, 2013, 83–84, 56–60.
37. Y. Yanbe, E. Sato, H. Chiba, T. Maeda, R. Matsushita, Y. Oda, O. Hagiwara, H. Matsukiyo, A. Osawa, T. Enomoto, M. Watanabe, S. Kusachi, S. Sato and A. Ogawa, Jpn. J. Appl. Phys., 2013, 52(9), 4.
38. D. T. Dul, K. M. Dabrowski and P. Korecki, Europhys. Lett., 2013, 104(6), 6.
39. S. Cornaby, Spectroscopy, 2013, 11.
40. Q. X. Zhang, L. Q. Ge, Y. Gu, G. Q. Zeng, Q. Yang and Y. Y. Luo, Spectrosc. Spectr. Anal., 2013, 33(8), 2231–2234.
41. Y. Gu, S. Q. Xiong, L. Q. Ge, Z. G. Fan, Q. X. Zhang and Z. Y. Zhu, Spectrosc. Spectr. Anal., 2014, 34(1), 252–256.
42. Y. Iwai, T. Koike, Y. Hayama, A. Jouzuka, T. Nakamura, Y. Onizuka, M. Miyoshi and H. Mimura, J. Vac. Sci. Technol., B: Microelectron. Nanometer Struct.--Process., Meas., Phenom., 2013, 31(2), 4.
43. Y. Iwai, K. Muramatsu, S. Tsuboi, A. Jyouzuka, T. Nakamura, Y. Onizuka and H. Mimura, Appl. Phys. Express, 2013, 6(10), 3.
44. M. Guerra, M. Manso, S. Pessanha, S. Longelin and M. L. Carvalho, X-Ray Spectrom., 2013, 42(5), 402–407.
45. S. Chen, J. Deng, Y. Yuan, C. Flachenecker, R. Mak, B. Hornberger, Q. Jin, D. Shu, B. Lai, J. Maser, C. Roehrig, T. Paunesku, S. C. Gleber, D. J. Vine, L. Finney, J. VonOsinski, M. Bolbat, I. Spink, Z. Chen, J. Steele, D. Trapp, J. Irwin, M. Feser, E. Snyder, K. Brister, C. Jacobsen, G. Woloschak and S. Vogt, J. Synchrotron Radiat., 2014, 21, 66–75.
46. Y. Yuan, S. Chen, T. Paunesku, S. C. Gleber, W. C. Liu, C. B. Doty, R. Mak, J. J. Deng, Q. L. Jin, B. Lai, K. Brister, C. Flachenecker, C. Jacobsen, S. Vogt and G. E. Woloschak, ACS Nano, 2013, 7(12), 10502–10517.
47. M. K. Tiwari, P. Gupta, A. K. Sinha, S. R. Kane, A. K. Singh, S. R. Garg, C. K. Garg, G. S. Lodha and S. K. Deb, J. Synchrotron Radiat., 2013, 20, 386–389.
48. Q. Xie, L. Peng, F. Cai, A. G. Li and K. Yang, J. Instrum., 2013, 8, 11.
49. X. Gao, S. Q. Gu, Q. Gao, Y. Zou, Z. Jiang, S. Zhang, X. J. Wei, H. S. Yu, G. D. Sheng, P. Q. Duan and Y. Y. Huang, X-Ray Spectrom., 2013, 42(6), 502–507.
50. B. A. Deng, Q. Yang, G. H. Du, Y. J. Tong, H. L. Xie and T. Q. Xiao, Nucl. Instrum. Methods Phys. Res., Sect. B, 2013, 305, 5–8.
51. S. R. Barberie, T. A. Cahill, C. F. Cahill, T. M. Cahill, C. R. Iceman and D. E. Barnes, Nucl. Instrum. Methods Phys. Res., Sect. A, 2013, 729, 930–933.
52. P. Fuoss, K. C. Chang and H. You, J. Electron Spectrosc. Relat. Phenom., 2013, 190, 75–83.
53. D. P. Fenning, J. Hofstetter, M. I. Bertoni, G. Coletti, B. Lai, C. del Canizo and T. Buonassisi, J. Appl. Phys., 2013, 113(4), 12.
54. C. Hall, J. Instrum., 2013, 8, 14.
55. Z. Savoly, G. Pepponi, P. I. Nagy, C. Streli, G. Buzanich, E. Chinea-Cano and G. Zaray, X-Ray Spectrom., 2013, 42(4), 321–329.
56. L. Luhl, I. Mantouvalou, I. Schaumann, C. Vogt and B. Kanngiesser, Anal. Chem., 2013, 85(7), 3682–3689.
57. N. L. Misra, M. K. Tiwari, S. S. Kumar, S. Dhara, A. K. Singh, G. S. Lodha, S. K. Deb, P. D. Gupta and S. K. Aggarwal, X-Ray Spectrom., 2013, 42(1), 4–7.
58. K. Banas, A. M. Banas, M. Gajda, W. M. Kwiatek, B. Pawlicki and M. B. H. Breese, Radiat. Phys. Chem., 2013, 93, 82–86.
59. 15th International Conference on TXRF Analysis and Related Methods, http://www.a-chem.eng.osaka-cu.ac.jp/txrf2013/.
60. https://www.facebook.com/TXRFSpectrometry .
61. A. Moya-Riffo, L. Bennun, V. Sanhueza and M. Santibanez, X-Ray Spectrom., 2013, 42(2), 93–99.
62. M. Santibanez, L. Bennun and L. M. Marco-Parra, X-Ray Spectrom., 2013, 42(6), 442–449.
63. V. L. Shapovalov, H. Mohwald, O. V. Konovalov and V. Knecht, Phys. Chem. Chem. Phys., 2013, 15(33), 13991–13998.
64. S. Kunimura, S. Kudo, H. Nagai, Y. Nakajima and H. Ohmori, Rev. Sci. Instrum., 2013, 84(4), 3.
65. J. Lubeck, B. Beckhoff, R. Fliegauf, I. Holfelder, P. Hönicke, M. Müller, B. Pollakowski, F. Reinhardt and J. Weser, Rev. Sci. Instrum., 2013, 84, 045106.
66. S. Kunimura and J. Kawai, X-Ray Spectrom., 2013, 42(3), 171–173.
67. Y. Liu, S. Imashuku and J. Kawai, Anal. Sci., 2013, 29(8), 793–797.
68. I. De La Calle, N. Cabaleiro, V. Romero, I. Lavilla and C. Bendicho, Spectrochim. Acta, Part B, 2013, 90, 23–54.
69. I. De La Calle, M. Costas, N. Cabaleiro, I. Lavilla and C. Bendicho, Food Chem., 2013, 138(1), 234–241.
70. I. De La Calle, N. Cabaleiro, I. Lavilla and C. Bendicho, J. Hazard. Mater., 2013, 260, 202–209.
71. V. Romero, I. Costas-Mora, I. Lavilla and C. Bendicho, J. Anal. At. Spectrom., 2013, 28(6), 923–933.
72. M. Holtkamp, T. Elseberg, C. A. Wehe, M. Sperling and U. Karst, J. Anal. At. Spectrom., 2013, 28(5), 719–723.
73. A. Wastl, F. Stadlbauer, J. Prost, C. Horntrich, P. Kregsamer, P. Wobrauschek and C. Streli, Spectrochim. Acta, Part B, 2013, 82, 71–75.
74. G. V. Pashkova, A. G. Revenko and A. L. Finkelshtein, X-Ray Spectrom., 2013, 42(6), 524–530.
75. C. Cantaluppi, M. Natali, F. Ceccotto and A. Fasson, X-Ray Spectrom., 2013, 42(4), 213–219.
76. B. W. Liou, J. Comput. Theor. Nanosci., 2013, 10(5), 1072–1079.
77. H. Takahara, Y. Mori, H. Shibata, A. Shimazaki, M. B. Shabani, M. Yamagami, N. Yabumoto, K. Nishihagi and Y. Gohshi, Spectrochim. Acta, Part B, 2013, 90, 72–82.
78. A. E. Tikhonova and V. S. Kozlov, PhSS, 2014, 56(1), 17–19.
79. W. R. Lindemann, T. Xiao, W. J. Wang, J. E. Berry, N. A. Anderson, R. S. Houk, R. Shinar, J. Shinar and D. Vaknin, Org. Electron., 2013, 14(12), 3190–3194.
80. Y. Kayser, J. Szlachetko and J. Sa, Rev. Sci. Instrum., 2013, 84(12), 11.
81. F. P. Romano, C. Altana, L. Cosentino, L. Celona, S. Gammino, D. Mascali, L. Pappalardo and F. Rizzo, Spectrochim. Acta, Part B, 2013, 86, 60–65.
82. A. Abboud, S. Send, N. Pashniak, W. Leitenberger, S. Ihle, M. Huth, R. Hartmann, L. Struder and U. Pietsch, J. Instrum., 2013, 8, 17.
83. B. A. Sobott, C. Broennimann, B. Schmitt, P. Trueb, M. Schneebeli, V. Lee, D. J. Peake, S. Elbracht-Leong, A. Schubert, N. Kirby, M. J. Boland, C. T. Chantler, Z. Barnea and R. P. Rassool, J. Synchrotron Radiat., 2013, 20, 347–354.
84. M. D. Wilson, S. J. Bell, R. J. Cernik, C. Christodoulou, C. K. Egan, D. O'Flynn, S. Jacques, S. Pani, J. Scuffham, P. Seller, P. J. Sellin, R. Speller and M. C. Veale, IEEE Trans. Nucl. Sci., 2013, 60(2), 1197–1200.
85. M. C. Veale, S. J. Bell, D. D. Duarte, A. Schneider, P. Seller, M. D. Wilson, V. Kachkanov and K. J. S. Sawhney, Nucl. Instrum. Methods Phys. Res., Sect. A, 2013, 729, 265–272.
86. D. O'Flynn, C. B. Reid, C. Christodoulou, M. D. Wilson, M. C. Veale, P. Seller, D. Hills, H. Desai, B. Wong and R. Speller, J. Instrum., 2013, 8, 16.
87. C. Krieger, J. Kaminski and K. Desch, Nucl. Instrum. Methods Phys. Res., Sect. A, 2013, 729, 905–909.
88. G. S. A. Wright, H. C. Lee, C. Schulze-Briese, J. G. Grossmann, R. W. Strange and S. S. Hasnain, J. Synchrotron Radiat., 2013, 20, 383–385.
89. S. Limandri, G. Bernardi and S. Suarez, X-Ray Spectrom., 2013, 42(6), 487–492.
90. H. Matsuura, Jpn. J. Appl. Phys., 2013, 52(2), 5.
91. B. Scruggs, Spectroscopy, 2013, 8–9.
92. N. L. Maidana, V. R. Vanin, V. Jahnke, J. M. Fernandez-Varea, M. N. Martins and L. Brualla, Nucl. Instrum. Methods Phys. Res., Sect. A, 2013, 729, 371–380.
93. R. Britton, J. L. Burnett, A. V. Davies and P. H. Regan, J. Radioanal. Nucl. Chem., 2013, 298(3), 1491–1499.
94. D. Demir, M. Eroglu and A. Tursucu, J. Instrum., 2013, 8, 12.
95. M. H. Carpenter, S. Friedrich, J. A. Hall, J. Harris, W. K. Warburton and R. Cantor, IEEE Trans. Appl. Supercond., 2013, 23(3), 4.
96. V. A. Litichevskyi, A. D. Opolonin, S. N. Galkin, A. I. Lalaiants and E. F. Voronkin, Instrum. Exp. Tech., 2013, 56(4), 436–443.
97. A. Brunetti, Comput. Phys. Commun., 2013, 184(3), 573–578.
98. A. V. Zablotskii, A. A. Viryus, O. I. Lyamina, A. Y. Kuzin, T. A. Kupriyanova, P. A. Todua and M. N. Filippov, Meas. Tech., 2013, 56(6), 625–629.
99. S. Yun, H. K. Kim, H. Youn, J. Tanguay and I. A. Cunningham, IEEE Trans. Med. Imaging, 2013, 32(10), 1819–1828.
100. G. Q. Zeng, L. Q. Ge, Y. Y. Luo, Q. X. Zhang, F. Cheng and G. X. Wang, Spectrosc. Spectr. Anal., 2013, 33(9), 2583–2585.
101. Z. Li, X. G. Tuo, J. B. Yang, M. Z. Liu, Y. Cheng and L. Wang, Chin. Phys. C, 2013, 37(1), 5.
102. Z. Li, X. G. Tuo, R. Shi and J. B. Yang, Sci. China: Technol. Sci., 2014, 57(1), 19–24.
103. J. E. Fernandez, V. Scot, L. Verardi and F. Salvat, X-Ray Spectrom., 2013, 42(4), 189–196.
104. J. E. Fernandez, V. Scot, L. Verardi and F. Salvat, Radiat. Phys. Chem., 2014, 95, 22–25.
105. T. Schoonjans, V. A. Sole, L. Vincze, M. S. del Rio, K. Appel and C. Ferrero, Spectrochim. Acta, Part B, 2013, 82, 36–41.
106. P. S. Athiray, M. Sudhakar, M. K. Tiwari, S. Narendranath, G. S. Lodha, S. K. Deb, P. Sreekumar and S. K. Dash, P&SS, 2013, 89, 183–187.
107. M. Watanabe, H. Inoue, Y. Yamada, M. Feeney, L. Oelofse and Y. Kataoka, Powder Diffr., 2013, 28(2), 132–136.
108. P. Caussin, Powder Diffr., 2013, 28(2), 90–94.
109. D. Demir and Y. Sahin, Radiat. Phys. Chem., 2013, 85, 64–69.
110. A. Rakotondrajoa, G. Buzanich, M. Radtke, U. Reinholz, H. Riesemeier, L. Vincze and R. Raboanary, X-Ray Spectrom., 2013, 42(4), 183–188.
111. J. Ward, R. Marvin, T. O'Halloran, C. Jacobsen and S. Vogt, Microsc. Microanal., 2013, 19(5), 1281–1289.
112. C. K. Egan, S. D. M. Jacques and R. J. Cernik, X-Ray Spectrom., 2013, 42(3), 151–157.
113. J. L. Campbell, G. M. Perrett, J. A. Maxwell, E. Nield, R. Gellert, P. L. King, M. Lee, J. M. O'Meara and I. Pradler, Nucl. Instrum. Methods Phys. Res., Sect. B, 2013, 302, 24–31.
114. I. F. Mikhailov, A. A. Baturin, A. I. Mikhailov and L. P. Fomina, Probl. At. Sci. Technol., 2013,(2), 172–176.
115. A. Iwata, K. Yuge and J. Kawai, X-Ray Spectrom., 2013, 42(1), 16–18.
116. R. Ekinci, N. Ekinci and E. Senemtasi, Asian J. Chem., 2013, 25(5), 2557–2560.
117. A. Plessow, X-Ray Spectrom., 2013, 42(1), 19–32.
118. C. Parsons, E. M. Grabulosa, E. Pili, G. H. Floor, G. Roman-Ross and L. Charlet, J. Hazard. Mater., 2013, 262, 1213–1222.
119. D. C. Weindorf, J. Herrero, C. Castaneda, N. Bakr and S. Swanhart, Soil Sci. Soc. Am. J., 2013, 77(6), 2071–2077.
120. M. A. Alvarez-Vazquez, C. Bendicho and R. Prego, Microchem. J., 2014, 112, 172–180.
121. T. Y. Cherkashina, S. V. Panteeva, A. L. Finkelshtein and V. M. Makagon, X-Ray Spectrom., 2013, 42(4), 207–212.
122. P. S. Ross, A. Bourke and B. Fresia, Ore Geol. Rev., 2013, 53, 93–111.
123. H. Yang, L. Q. Ge, Y. Gu, Q. X. Zhang and S. Q. Xiong, Spectrosc. Spectr. Anal., 2013, 33(11), 3143–3147.
124. Z. Z. Cheng, H. K. Huang, M. Liu, T. X. Gu, W. D. Yan and M. C. Yan, Geostand. Geoanal. Res., 2013, 37(1), 95–101.
125. E. Knowles, H. Staudigel and A. Templeton, Earth Planet. Sci. Lett., 2013, 374, 239–250.
126. J. R. Vidal-Solano, R. L. S. Cruz, O. Zamora, A. Mendoza-Cordova and J. M. Stock, J. Iber. Geol., 2013, 39(1), 121–130.
127. G. H. Floor, E. Margui, M. Hidalgo, I. Queralt, P. Kregsamer, C. Streli and G. Roman-Ross, Chem. Geol., 2013, 352, 19–26.
128. S. Kutterolf, T. H. Hansteen, K. Appel, A. Freundt, K. Kruger, W. Perez and H. Wehrmann, Geology, 2013, 41(6), 707–710.
129. L. E. Mayhew, E. T. Ellison, T. M. McCollom, T. P. Trainor and A. S. Templeton, Nat. Geosci., 2013, 6(6), 478–484.
130. L. H. E. Winkel, B. Casentini, F. Bardelli, A. Voegelin, N. P. Nikolaidis and L. Charlet, Geochim. Cosmochim. Acta, 2013, 106, 99–110.
131. T. Rango, A. Vengosh, G. Dwyer and G. Bianchini, Water Res., 2013, 47(15), 5801–5818.
132. M. Aquit, W. Kuhnt, A. Holbourn, E. Chellai, K. Stattegger, O. Kluth and H. Jabour, Cretaceous Res., 2013, 45, 288–305.
133. K. Kanamaru, P. Francus, R. Francois, M. Besonen and C. Laj, GFF, 2013, 135(3–4), 316–339.
134. C. Heymann, O. Nelle, W. Dorfler, H. Zagana, N. Nowaczyk, J. B. Xue and I. Unkel, Quat. Int., 2013, 302, 42–60.
135. M. de Angelis, J. L. Tison, M. C. Morel-Fourcade and J. Susini, QSRv, 2013, 78, 248–265.
136. S. V. Hansson, J. Rydberg, M. Kylander, K. Gallagher and R. Bindler, Holocene, 2013, 23(12), 1666–1671.
137. E. Haenssler, M. J. Nadeau, A. Vott and I. Unkel, Quat. Int., 2013, 308, 89–104.
138. R. Saaltink, J. Griffioen, G. Mol, M. Birke and G. P. Team, J. Soils Sediments, 2014, 14(1), 121–137.
139. M. Haest, T. Cudahy, A. Rodger, C. Laukamp, E. Martens and M. Caccetta, Remote Sensing of Environment, 2013, 129, 17–31.
140. J. A. Addison, B. P. Finney, J. M. Jaeger, J. S. Stoner, R. D. Norris and A. Hangsterfer, J. Geophys. Res.: Oceans, 2013, 118(7), 3444–3461.
141. L. Jean-Soro, A. Oleron-Hamdous, B. Bechet and M. Legret, J. Soils Sediments, 2013, 13(3), 569–574.
142. D. L. Burak, F. van Oort, T. Becquer, E. Foy and M. P. F. Fontes, Eur. J. Soil Sci., 2013, 64(1), 131–144.
143. T. W. Dahl, M. Ruhl, E. U. Hammarlund, D. E. Canfield, M. T. Rosing and C. J. Bjerrum, Chem. Geol., 2013, 360, 241–251.
144. G. Bianchini, D. Di Giuseppe, C. Natali and L. Beccaluva, Ofioliti, 2013, 38(1), 1–14.
145. T. Neumann, F. Scholz, U. Kramar, M. Ostermaier, N. Rausch and Z. Berner, Sedimentology, 2013, 60(6), 1389–1404.
146. E. Curti, L. Aimoz and A. Kitamura, J. Radioanal. Nucl. Chem., 2013, 295(3), 1655–1665.
147. T. Zeng, W. A. Arnold and B. M. Toner, Environ. Sci. Technol., 2013, 47(3), 1287–1296.
148. D. Fortin, P. Francus, A. C. Gebhardt, A. Hahn, P. Kliem, A. Lise-Pronovost, R. Roychowdhury, J. Labrie, G. St-Onge and P. S. Team, Quat. Sci. Rev., 2013, 71, 147–153.
149. A. Gorghinian, A. Mottana, A. Rossi, F. M. Oltean, A. Esposito and A. Marcelli, Rendiconti Lincei-Scienze Fisiche E Naturali, 2013, 24(2), 127–140.
150. I. S. Rodina, A. N. Kravtsova, A. V. Soldatov, G. E. Yalovega, Y. V. Popov and N. I. Boyko, Opt. Spectrosc., 2013, 115(6), 858–862.
151. X. P. Gu, X. D. Xie, X. B. Wu, G. C. Zhu, J. Q. Lai, H. Kenich and J. W. Huang, Eur. J. Mineral., 2013, 25(2), 177–186.
152. M. Hatipoglu, M. B. Oguzer and H. B. Buzlu, J. Afr. Earth Sci., 2013, 84, 20–35.
153. N. P. Edwards, R. A. Wogelius, U. Bergmann, P. Larson, W. I. Sellers and P. L. Manning, Appl. Phys. A: Mater. Sci. Process., 2013, 111(1), 147–155.
154. P. L. Manning, N. P. Edwards, R. A. Wogelius, U. Bergmann, H. E. Barden, P. L. Larson, D. Schwarz-Wings, V. M. Egerton, D. Sokaras, R. A. Mori and W. I. Sellers, J. Anal. At. Spectrom., 2013, 28(7), 1024–1030.
155. P. Gueriau, C. Mocuta, D. B. Dutheil, S. X. Cohen, D. Thiaudiere, S. Charbonnier, G. Clement, L. Bertrand and O. T. Consortium, PLoS One, 2014, 9(1), 9.
156. T. Yoshimura, A. Suzuki, Y. Tamenori and H. Kawahata, Geo-Mar. Lett., 2014, 34(1), 1–9.
157. L. T. Nguyen, M. A. Rahman, T. Maki, Y. Tamenori, T. Yoshimura, A. Suzuki, N. Iwasaki and H. Hasegawa, Geochim. Cosmochim. Acta, 2014, 127, 1–9.
158. Y. T. Qu, C. Z. Jin, Y. Q. Zhang, Y. W. Hu, X. Shang and C. S. Wang, Appl. Phys. A: Mater. Sci. Process., 2013, 111(1), 75–82.
159. C. Lund, P. Lamberg and T. Lindberg, Miner. Eng., 2013, 49, 7–16.
160. F. S. de Oliveira, A. Varajao, C. A. C. Varajao, B. Boulange and C. C. V. Soares, Catena, 2013, 105, 29–39.
161. K. Pazand, A. Hezarkhani and M. Ataei, Arab. J. Geosci., 2013, 6(5), 1447–1456.
162. A. Richard, J. Cauzid, M. Cathelineau, M. C. Boiron, J. Mercadier and M. Cuney, Geofluids, 2013, 13(2), 101–111.
163. A. N. Nzeugang, R. M. Eko, N. Fagel, V. K. Kabeyene, A. Njoya, A. B. Madi, J. R. Mache and U. M. Chinje, Clay Min., 2013, 48(4), 655–662.
164. V. Cardenes, A. Rubio-Ordonez, C. Monterroso and L. Calleja, Chem Erde-Geochem., 2013, 73(3), 373–382.
165. A. A. Masoud, G. Christidis and K. Koike, Clay Min., 2013, 48(1), 1–20.
166. A. Langella, D. L. Bish, P. Cappelletti, G. Cerri, A. Colella, R. de Gennaro, S. F. Graziano, A. Perrotta, C. Scarpati and M. de Gennaro, Appl. Clay Sci., 2013, 72, 55–73.
167. D. C. Weindorf, L. Paulette and T. Man, Environ. Pollut., 2013, 182, 92–100.
168. A. K. Krishna, K. R. Mohan, N. N. Murthy, V. Periasamy, G. Bipinkumar, K. Manohar and S. S. Rao, Environ. Earth Sci., 2013, 70(2), 699–708.
169. D. T. Lamb, V. P. Matanitobua, T. Palanisami, M. Megharaj and R. Naidu, Environ. Sci. Technol., 2013, 47(9), 4670–4676.
170. J. Z. Xu, B. Peng, C. X. Yu, G. Yang, X. Y. Tang, C. Y. Tan, S. R. Xie, X. L. Tu, Z. C. Bao, M. J. Quan and M. Xiao, Environ. Earth Sci., 2013, 70(1), 175–190.
171. Y. R. Wang, D. C. W. Tsang, W. E. Olds and P. A. Weber, Environ. Technol., 2013, 34(24), 3177–3182.
172. L. K. Andersen, T. J. Morgan, A. K. Boulamanti, P. Alvarez, S. V. Vassilev and D. Baxter, Energy Fuels, 2013, 27(12), 7439–7454.
173. H. Haykiri-Acma, S. Yaman, M. Alkan and S. Kucukbayrak, Energy Conv. Manag., 2014, 77, 221–226.
174. F. Kirnbauer, M. Koch, R. Koch, C. Aichernig and H. Hofbauer, Energy Fuels, 2013, 27(6), 3316–3331.
175. S. Singh, W. Nimmo and P. T. Williams, Fuel, 2013, 111, 269–279.
176. L. H. Wang, X. M. Lu and Y. Y. Huang, X-Ray Spectrom., 2013, 42(6), 423–428.
177. M. Nagoshi, T. Aoyama, Y. Tanaka, T. Ishida, S. Kinoshiro and K. Kobayashi, ISIJ Int., 2013, 53(12), 2197–2200.
178. M. Liu, J. Y. Zheng, Y. L. Lu, Z. J. Li, Y. Zou, X. H. Yu and X. T. Zhou, J. Nucl. Mater., 2013, 440(1–3), 124–128.
179. A. Doyle, A. Saavedra, M. L. B. Tristao, L. A. N. Mendes and R. Q. Aucelio, Spectrochim. Acta, Part B, 2013, 86, 102–107.
180. M. F. Gazulla, M. Orduna, S. Vicente and M. Rodrigo, Fuel, 2013, 108, 247–253.
181. H. Wang, Z. X. Lin and B. L. Wu, Spectrosc. Spectr. Anal., 2013, 33(9), 2579–2582.
182. J. P. Candeias, D. F. de Oliveira, M. J. dos Anjos and R. T. Lopes, Radiat. Phys. Chem., 2014, 95, 408–411.
183. M. S. Lai, L. W. Xiang, J. M. Lin and H. F. Li, Sci. China: Chem., 2013, 56(8), 1164–1170.
184. E. Domingos, T. M. C. Pereira, P. R. Filgueiras, M. Bueno, E. V. R. de Castro, R. C. L. Guimaraes, G. L. de Sena, W. F. C. Rocha and W. Romao, X-Ray Spectrom., 2013, 42(2), 79–86.
185. M. Necemer, P. Kump, P. Sket, J. Plavec, J. Grdadolnik and M. Zvanut, Acta Chim. Slov., 2013, 60(3), 701–705.
186. C. Vanhoof, K. A. Holschbach-Bussian, B. M. Bussian, R. Cleven and K. Furtmann, X-Ray Spectrom., 2013, 42(4), 224–231.
187. V. Rackwitz, M. Ostermann, U. Panne and V. D. Hodoroaba, J. Anal. At. Spectrom., 2013, 28(9), 1466–1474.
188. Y. Hirokawa, Y. Shibata, T. Konya, Y. Koike and T. Nakamura, X-Ray Spectrom., 2013, 42(3), 134–140.
189. T. Roth, M. Wolf, M. Pohlein and R. van Eldik, Anal. Bioanal. Chem., 2013, 405(23), 7215–7229.
190. J. R. H. Arancibia, P. Alfonso, M. Garcia-Valles, S. Martinez, D. Parcerisa, C. Canet and Y. F. M. Romero, Boletin De La Sociedad Espanola De Ceramica Y Vidrio, 2013, 52(3), 143–150.
191. N. K. Sandhu, L. Axe, P. K. Ndiba and K. Jahan, Environ. Eng. Sci., 2013, 30(7), 387–392.
192. R. Gullayanon, T. E. Michaels and M. A. Rudat, X-Ray Spectrom., 2013, 42(4), 232–236.
193. K. Tang, A. Ciftja, C. van der Eijk, S. Wilson and G. Tranell, J. Min. Metall., Sect. B, 2013, 49(2), 233–236.
194. K. S. A. Aldroobi, A. Shukri, S. Bauk, E. M. A. Munem and A. M. A. Abuarra, Radiat. Phys. Chem., 2013, 91, 9–14.
195. T. Fujimori and H. Takigami, Environ. Geochem. Health, 2014, 36(1), 159–168.
196. V. B. Yadav and S. K. Jha, J. Radioanal. Nucl. Chem., 2013, 295(3), 1759–1762.
197. A. R. King, A. Davies, N. Saint and N. Pockett, J. Radioanal. Nucl. Chem., 2013, 296(2), 1143–1147.
198. R. Terzano, M. Alfeld, K. Janssens, B. Vekemans, T. Schoonjans, L. Vincze, N. Tomasi, R. Pinton and S. Cesco, Anal. Bioanal. Chem., 2013, 405(10), 3341–3350.
199. S. D. Lindblom, J. R. Valdez-Barillas, S. C. Fakra, M. A. Marcus, A. L. Wangeline and E. A. H. Pilon-Smits, Environ. Exp. Bot., 2013, 88, 33–42.
200. L. L. Lu, S. K. Tian, J. Zhang, X. Yang, J. M. Labavitch, S. M. Webb, M. Latimer and P. H. Brown, New Phytol., 2013, 198(3), 721–731.
201. J. R. Zeng, G. L. Zhang, L. M. Bao, S. L. Long, M. G. Tan, Y. Li, C. Y. Ma and Y. D. Zhao, J. Environ. Sci., 2013, 25(3), 605–612.
202. H. Feng, Y. Qian, F. J. Gallagher, M. Y. Wu, W. G. Zhang, L. Z. Yu, Q. Z. Zhu, K. W. Zhang, C. J. Liu and R. Tappero, Environ. Sci. Pollut. Res., 2013, 20(6), 3743–3750.
203. Z. P. Wu, K. McGrouther, D. L. Chen, W. D. Wu and H. L. Wang, J. Agric. Food Chem., 2013, 61(20), 4715–4722.
204. S. Koren, I. Arcon, P. Kump, M. Necemer and K. Vogel-Mikus, Plant Soil, 2013, 370(1–2), 125–148.
205. J. Song, Y. Q. Yang, S. H. Zhu, G. C. Chen, X. F. Yuan, T. T. Liu, X. H. Yu and J. Y. Shi, Biol. Plant., 2013, 57(3), 581–586.
206. J. T. Zhao, Y. Hu, Y. X. Gao, Y. F. Li, B. Li, Y. X. Dong and Z. F. Chai, Metallomics, 2013, 5(7), 896–903.
207. T. Punshon, F. K. Ricachenevsky, M. N. Hindt, A. L. Socha and H. Zuber, Metallomics, 2013, 5(9), 1133–1145.
208. T. Punshon, R. Tappero, F. K. Ricachenevsky, K. Hirschi and P. A. Nakata, Plant J., 2013, 76(4), 627–633.
209. P. Wang, N. W. Menzies, E. Lombi, B. A. McKenna, M. D. de Jonge, E. Donner, F. P. C. Blamey, C. G. Ryan, D. J. Paterson, D. L. Howard, S. A. James and P. M. Kopittke, Sci. Total Environ., 2013, 463, 131–139.
210. A. Buleon, M. Cotte, J. L. Putaux, C. d'Hulst and J. Susini, Biochim. Biophys. Acta, Gen. Subj., 2014, 1840(1), 113–119.
211. B. De Samber, K. A. C. De Schamphelaere, C. R. Janssen, B. Vekemans, R. De Rycke, G. Martinez-Criado, R. Tucoulou, P. Cloetens and L. Vincze, Anal. Bioanal. Chem., 2013, 405(18), 6061–6068.
212. A. D. Servin, M. I. Morales, H. Castillo-Michel, J. A. Hemandez-Viezcas, B. Munoz, L. J. Zhao, J. E. Nunez, J. R. Peralta-Videa and J. L. Gardea-Torresdey, Environ. Sci. Technol., 2013, 47(20), 11592–11598.
213. L. J. Zhao, Y. P. Sun, J. A. Hernandez-Viezcas, A. D. Servin, J. Hong, G. H. Niu, J. R. Peralta-Videa, M. Duarte-Gardea and J. L. Gardea-Torresdey, J. Agric. Food Chem., 2013, 61(49), 11945–11951.
214. C. M. Rico, M. I. Morales, R. McCreary, H. Castillo-Michel, A. C. Barrios, J. Hong, A. Tafoya, W. Y. Lee, A. Varela-Ramirez, J. R. Perata-Videa and J. L. Gardea-Torresdey, Environ. Sci. Technol., 2013, 47(24), 14110–14118.
215. A. Manceau, A. Simionovici, M. Lanson, J. Perrin, R. Tucoulou, S. Bohic, S. C. Fakra, M. A. Marcus, J. P. Bedell and K. L. Nagy, Metallomics, 2013, 5(12), 1674–1684.
216. S. P. Singh, K. Vogel-Mikus, I. Arcon, P. Vavpetic, L. Jeromel, P. Pelicon, J. Kumar and R. Tuli, J. Exp. Bot., 2013, 64(11), 3249–3260.
217. A. L. Neal, K. Geraki, S. Borg, P. Quinn, J. F. Mosselmans, H. Brinch-Pedersen and P. R. Shewry, J. Biol. Inorg. Chem., 2013, 18(5), 557–570.
218. L. L. Lu, S. K. Tian, H. B. Liao, J. Zhang, X. E. Yang, J. M. Labavitch and W. R. Chen, PLoS One, 2013, 8(2), 9.
219. T. Inui, S. Nakatao, Y. Koike, M. Kitano and T. Nakamura, Bunseki Kagaku, 2013, 62(10), 925–930.
220. C. D. Patz, C. Cescutti, H. Dietrich and W. Andlauer, Dtsch. Lebensm.-Rundsch., 2013, 109(6), 315–319.
221. P. F. Boldrin, V. Faquin, S. J. Ramos, K. V. F. Boldrin, F. W. Avila and L. R. G. Guilherme, J. Food Compos. Anal., 2013, 31(2), 238–244.
222. S. M. Geraldo, F. B. Canteras and S. Moreira, Radiat. Phys. Chem., 2014, 95, 346–348.
223. P. Cherubini, T. Humbel, H. Beeckman, H. Gartner, D. Mannes, C. Pearson, W. Schoch, R. Tognetti and S. Lev-Yadun, PLoS One, 2013, 8(1), 5.
224. D. Gupta, S. Roy, R. Ghosh and A. K. Mitra, X-Ray Spectrom., 2013, 42(4), 268–275.
225. R. A. Root, S. Fathordoobadi, F. Alday, W. Ela and J. Chorover, Environ. Sci. Technol., 2013, 47(22), 12992–13000.
226. P. Langner, C. Mikutta, E. Suess, M. A. Marcus and R. Kretzschmar, Environ. Sci. Technol., 2013, 47(17), 9706–9714.
227. K. Schwarz, K. C. Weathers, S. T. A. Pickett, R. G. Lathrop, R. V. Pouyat and M. L. Cadenasso, Environ. Geochem. Health, 2013, 35(4), 495–510.
228. E. Margui, B. Zawisza and R. Sitko, TrAC, Trends Anal. Chem., 2014, 53, 73–83.
229. K. Kocot, B. Zawisza, E. Margui, I. Queralt, M. Hidalgo and R. Sitko, J. Anal. At. Spectrom., 2013, 28(5), 736–742.
230. B. Zawisza and R. Sitko, Appl. Spectrosc., 2013, 67(5), 536–541.
231. E. Margui, M. Sague, I. Queralt and M. Hidalgo, Anal. Chim. Acta, 2013, 786, 8–15.
232. K. Pytlakowska and R. Sitko, Anal. Methods, 2013, 5(21), 6192–6199.
233. S. A. Kumar, S. P. Pandey, N. Thakur, H. Parab, R. N. Shinde, A. K. Pandey, D. N. Wagh, S. D. Kumar and A. V. R. Reddy, J. Hazard. Mater., 2013, 262, 265–273.
234. V. S. Hatzistavros and N. G. Kallithrakas-Kontos, Anal. Chim. Acta, 2014, 809, 25–29.
235. R. A. Wogelius, Cryst. Res. Technol., 2013, 48(10), 877–902.
236. I. B. de Barros, E. S. G. dos Santos, D. E. D. Gomes, C. Volkmer-Ribeiro, C. C. Silva and V. F. da Veiga, X-Ray Spectrom., 2013, 42(2), 59–62.
237. E. V. Chuparina, L. P. Paradina and V. A. Trunova, X-Ray Spectrom., 2013, 42(5), 388–393.
238. T. Yoshimura, Y. Tamenori, A. Suzuki, R. Nakashima, N. Iwasaki, H. Hasegawa and H. Kawahata, Chem. Geol., 2013, 352, 170–175.
239. D. Deruytter, J. Garrevoet, M. B. Vandegehuchte, E. Vergucht, B. De Samber, B. Vekemans, K. Appel, G. Falkenberg, K. Delbeke, R. Blust, K. A. C. De Schamphelaere, L. Vincze and C. R. Janssen, Environ. Sci. Technol., 2014, 48(1), 698–705.
240. L. Brinza, P. F. Schofield, M. E. Hodson, S. Weller, K. Ignatyev, K. Geraki, P. D. Quinn and J. F. W. Mosselmans, J. Synchrotron Radiat., 2014, 21, 235–241.
241. C. Vazquez, O. Palacios, S. Boeykens and L. M. M. Parra, X-Ray Spectrom., 2013, 42(4), 220–223.
242. M. C. R. Castro, V. Andreano, G. Custo and C. Vazquez, Microchem. J., 2013, 110, 402–406.
243. R. H. M. Godoi, B. H. B. Carneiro, S. L. Paralovo, V. P. Campos, T. M. Tavares, H. Evangelista, R. Van Grieken and A. F. L. Godoi, Sci. Total Environ., 2013, 452, 314–320.
244. K. W. Fomba, K. Muller, D. van Pinxteren and H. Herrmann, Atmos. Chem. Phys., 2013, 13(9), 4801–4814.
245. S. Yatkin and M. Gerboles, Atmos. Environ., 2013, 73, 159–168.
246. U. E. A. Fittschen, C. Streli, F. Meirer and M. Alfeld, X-Ray Spectrom., 2013, 42(5), 368–373.
247. H. Indresand, W. H. White, K. Trzepla and A. M. Dillner, X-Ray Spectrom., 2013, 42(5), 359–367.
248. M. A. Barreiros, T. Pinheiro, P. M. Felix, C. Franco, M. Santos, F. Araujo, M. C. Freitas and S. M. Almeida, J. Radioanal. Nucl. Chem., 2013, 297(3), 377–382.
249. G. Wells and M. Haaf, J. Chem. Educ., 2013, 90(12), 1616–1621.
250. I. Kakoulli, S. V. Prikhodko, C. Fischer, M. Cilluffo, M. Uribe, H. A. Bechtel, S. C. Fakra and M. A. Marcus, Anal. Chem., 2014, 86(1), 521–526.
251. D. Fraser, C. S. DeRoo, R. B. Cody and R. A. Armitage, Analyst, 2013, 138(16), 4470–4474.
252. K. Won-in, T. Sako, C. Thongleurm, S. Intarasiri, U. Tippawan, T. Kamwanna, W. Pattanasiriwisana, S. Tancharakorn, N. Kamonsutthipaijit and P. Dararutana, J. Radioanal. Nucl. Chem., 2013, 297(2), 285–290.
253. T. N. Huffman, M. Elburg and M. Watkeys, J. Archaeol. Sci., 2013, 40(10), 3553–3560.
254. M. Malagodi, C. Canevari, L. Bonizzoni, A. Galli, F. Maspero and M. Martini, Appl. Phys. A: Mater. Sci. Process., 2013, 112(2), 225–234.
255. A. Kaplan and D. C. Stulik, Imaging Sci. J., 2013, 61(8), 629–646.
256. A. Vila, S. A. Centeno, L. Barro and N. W. Kennedy, Stud. Conserv., 2013, 58(3), 176–188.
257. I. Rabin and O. Hahn, Anal. Methods, 2013, 5(18), 4648–4654.
258. K. Dzinavatonga, T. R. Medupe, L. C. Prinsloo and E. E. Ebenso, Asian J. Chem., 2013, 25(16), 9384–9386.
259. M. Manso, A. Le Gac, S. Longelin, S. Pessanha, J. C. Frade, M. Guerra, A. J. Candeias and M. L. Carvalho, Spectrochim. Acta, Part A, 2013, 105, 288–296.
260. K. Janssens, M. Alfeld, G. Van der Snickt, W. De Nolf, F. Vanmeert, M. Radepont, L. Monico, J. Dik, M. Cotte, G. Falkenberg, C. Miliani and B. G. Brunetti, in Annual Review of Analytical Chemistry, ed. R. G. Cooks and J. E. Pemberton, Annual Reviews, Palo Alto, 2013, vol. 6, pp. 399–425.
261. A. Zielinska, W. Dabrowski, T. Fiutowski, B. Mindur, P. Wiacek and P. Wrobel, J. Instrum., 2013, 8, 15.
262. M. Alfeld, J. V. Pedroso, M. V. Hommes, G. Van der Snickt, G. Tauber, J. Blaas, M. Haschke, K. Erler, J. Dik and K. Janssens, J. Anal. At. Spectrom., 2013, 28(5), 760–767.
263. M. Alfeld and J. A. C. Broekaert, Spectrochim. Acta, Part B, 2013, 88, 211–230.
264. M. Alfeld, W. De Nolf, S. Cagno, K. Appel, D. P. Siddons, A. Kuczewski, K. Janssens, J. Dik, K. Trentelman, M. Walton and A. Sartorius, J. Anal. At. Spectrom., 2013, 28(1), 40–51.
265. M. Alfeld, D. P. Siddons, K. Janssens, J. Dik, A. Woll, R. Kirkham and E. van de Wetering, Appl. Phys. A: Mater. Sci. Process., 2013, 111(1), 157–164.
266. E. Rebollo, L. Nodari, U. Russo, R. Bertoncello, C. Scardellato, F. Romano, F. Ratti and L. Poletto, J. Cult. Herit., 2013, 14(3), E153–E160.
267. K. F. Gebremariam, L. Kvittingen and F. G. Banica, X-Ray Spectrom., 2013, 42(6), 462–469.
268. C. Montagner, D. Sanches, J. Pedroso, M. J. Melo and M. Vilarigues, Spectrochim. Acta, Part A, 2013, 103, 409–416.
269. J. L. Mass, R. Opila, B. Buckley, M. Cotte, J. Church and A. Mehta, Appl. Phys. A: Mater. Sci. Process., 2013, 111(1), 59–68.
270. F. Casadio and V. Rose, Appl. Phys. A: Mater. Sci. Process., 2013, 111(1), 1–8.
271. K. Castro, U. Knuutinen, S. F. O. de Vallejuelo, M. Irazola and J. M. Madariaga, Spectrochim. Acta, Part A, 2013, 106, 104–109.
272. A. J. Sakalis, N. A. Kazakis, N. Merousis and N. C. Tsirliganis, J. Cult. Herit., 2013, 14(6), 485–498.
273. G. Barone, P. Mazzoleni, A. Aquilia and G. Barbera, Archa, 2014, 56(1), 70–87.
274. L. Maritan, M. Secco, C. Mazzoli, V. Mantovani and J. Bonetto, Appl. Clay Sci., 2013, 82, 62–69.
275. R. Ravisankar, G. R. Annamalai, A. Naseerutheen, A. Chandrasekaran, M. V. R. Prasad, K. K. Satpathy and C. Maheswaran, Spectrochim. Acta, Part A, 2013, 115, 845–853.
276. E. M. Gascon and J. B. I. Garrigos, Appl. Clay Sci., 2013, 82, 79–85.
277. V. M. Ferreras, C. Capelli, R. Cabella and X. N. Prieto, Appl. Clay Sci., 2013, 82, 70–78.
278. E. C. Geil, S. A. LeBlanc, D. S. Dale and R. E. Thorne, J. Archaeol. Sci., 2013, 40(12), 4780–4784.
279. N. Forster and P. Grave, X-Ray Spectrom., 2013, 42(6), 480–486.
280. T. Mitsuji, S. Nakazono and H. Hirakawa, Bunseki Kagaku, 2013, 62(2), 73–87.
281. T. P. Silva, M. O. Figueiredo and M. I. Prudencio, Appl. Clay Sci., 2013, 82, 101–105.
282. F. Casadio, A. Bezur, K. Domoney, K. Eremin, L. Lee, J. L. Mass, A. Shortland and N. Zumbulyadis, Stud. Conserv., 2012, 57, 61–72.
283. K. Blagoev, M. Grozeva, G. Malcheva and S. Neykova, Spectrochim. Acta, Part B, 2013, 79–80, 39–43.
284. H. Gondai, Y. Shindo, M. Kawatoko and I. Nakai, Bunseki Kagaku, 2013, 62(2), 143–154.
285. W. Klysubun, B. Ravel, P. Klysubun, P. Sombunchoo and W. Deenan, Appl. Phys. A: Mater. Sci. Process., 2013, 111(3), 775–782.
286. J. Hormes, A. Roy, G. L. Bovenkamp, K. Simon, C. Y. Kim, N. Borste and S. Gai, Appl. Phys. A: Mater. Sci. Process., 2013, 111(1), 91–97.
287. T. Kikugawa, Y. Abe, A. Nakamura and I. Nakai, Bunseki Kagaku, 2014, 63(1), 31–40.
288. F. Montanari, C. Boschetti, P. Miselli, M. Hanuskova, P. Baraldi and C. Leonelli, Ceram.-Silik., 2013, 57(3), 258–264.
289. E. Greiner-Wronowa, D. Zabiegaj and P. Piccardo, Appl. Phys. A: Mater. Sci. Process., 2013, 113(4), 999–1008.
290. R. Abd-Allah, J. Cult. Herit., 2013, 14(2), 97–108.
291. K. Hayes, J. Archaeol. Sci., 2013, 40(8), 3193–3211.
292. R. K. Gauss, J. Batora, E. Nowaczinski, K. Rassmann and G. Schukraft, J. Archaeol. Sci., 2013, 40(7), 2942–2960.
293. L. Bonizzoni, A. Galli, M. Gondola and M. Martini, X-Ray Spectrom., 2013, 42(4), 262–267.
294. E. U. Sagin and H. Boke, J. Cult. Herit., 2013, 14(3), E73–E76.
295. E. Salavessa, S. Jalali, L. M. O. Sousa, L. Fernandes and A. M. Duarte, Constr. Build. Mater., 2013, 48, 858–867.
296. N. Prieto-Taboada, O. Gomez-Laserna, I. Martinez-Arkarazo, M. A. Olazabal and J. M. Madariaga, Anal. Chem., 2013, 85(20), 9501–9507.
297. D. Ogburn, B. Sillar and J. C. Sierra, J. Archaeol. Sci., 2013, 40(4), 1823–1837.
298. O. Gomez-Laserna, M. A. Olazabal, H. Morillas, N. Prieto-Taboada, I. Martinez-Arkarazo, G. Arana and J. M. Madariaga, J. Raman Spectrosc., 2013, 44(9), 1277–1284.
299. R. M. Ion, I. R. Bunghez, S. F. Pop, R. C. Fierascu, M. L. Ion and M. Leahu, Metal. Int., 2013, 18, 89–93.
300. P. Holakooei, F. C. Petrucci, R. Tassinari and C. Vaccaro, X-Ray Spectrom., 2013, 42(2), 105–115.
301. G. Barbera, G. Barone, V. Crupi, F. Longo, D. Majolino, P. Mazzoleni and V. Venuti, X-Ray Spectrom., 2013, 42(1), 8–15.
302. F. Vanmeert, D. Mudronja, S. Fazinic, K. Janssens and D. Tibljas, X-Ray Spectrom., 2013, 42(4), 256–261.
303. C. Pelosi, D. Fodaro, L. Sforzini, A. R. Rubino and A. Falqui, Opt. Spectrosc., 2013, 114(6), 917–928.
304. P. Ortiz, V. Antunez, R. Ortiz, J. M. Martin, M. A. Gomez, A. R. Hortal and B. Martinez-Haya, Appl. Surf. Sci., 2013, 283, 193–201.
305. M. Milic, J. Archaeol. Sci., 2014, 41, 285–296.
306. L. C. Kellett, M. Golitko and B. S. Bauer, J. Archaeol. Sci., 2013, 40(4), 1890–1902.
307. E. Frahm and J. M. Feinberg, J. Archaeol. Sci., 2013, 40(4), 1866–1878.
308. E. Frahm, B. A. Schmidt, B. Gasparyan, B. Yeritsyan, S. Karapetian, K. Meliksetian and D. S. Adler, J. Archaeol. Sci., 2014, 41, 333–348.
309. T. M. Gluhak and D. Rosenberg, J. Archaeol. Sci., 2013, 40(3), 1611–1622.
310. A. C. Neiva, H. P. F. Pinto and F. J. G. Landgraf, Radiat. Phys. Chem., 2014, 95, 368–372.
311. L. Castelli, L. Giuntini, F. Taccetti, E. Barzagli, F. Civita, C. Czelusniak, M. E. Fedi, N. Gelli, F. Grazzi, A. Mazzinghi, L. Palla, F. P. Romano and P. A. Mandao, X-Ray Spectrom., 2013, 42(6), 537–540.
312. D. Cvikel, D. Ashkenazi, A. Stern and Y. Kahanov, Mater. Charact., 2013, 83, 198–211.
313. A. Michelin, E. Drouet, E. Foy, J. J. Dynes, D. Neff and P. Dillmann, J. Anal. At. Spectrom., 2013, 28(1), 59–66.
314. D. Satovic, V. Desnica and S. Fazinic, Spectrochim. Acta, Part B, 2013, 89, 7–13.
315. T. F. Fan, L. Wan and H. Zhang, China Foundry, 2013, 10(1), 1–6.
316. J. L. Mass and C. R. Matsen, Stud. Conserv., 2012, 57, 191–198.
317. G. Buccolieri, V. Nassisi, A. Buccolieri, F. Vona and A. Castellano, Appl. Surf. Sci., 2013, 272, 55–58.
318. M. Ferretti, C. Polese and C. R. Garcia, Spectrochim. Acta, Part BSpectrochim. Acta, Part B, 2013, 83–84, 21–27.
319. R. Cesareo, A. D. Bustamante, J. S. Fabian, S. D. Zambrano, W. Alva, L. Z. Chero, M. C. D. Espinoza, R. R. Rodriguez, M. F. Seclen, F. V. Gutierrez, E. B. Levano, J. A. Gonzales, M. A. Rizzutto, E. Poli, C. Calza, M. Dos Anjos, R. P. De Freitas, R. T. Lopes, C. Elera, I. Shimada, V. Curay, M. G. Castillo, G. E. Gigante, G. M. Ingo, F. Lopes, U. Holmquist and D. Diestra, Appl. Phys. A: Mater. Sci. Process., 2013, 113(4), 889–903.
320. S. Scrivano, B. Gomez-Tubio, I. Ortega-Feliu, F. J. Ager, A. I. Moreno-Suarez, M. A. Respaldiza, M. L. de la Bandera and A. Marmolejo, X-Ray Spectrom., 2013, 42(4), 251–255.
321. A. Gorghinian, A. Esposito, M. Ferretti and F. Catalli, Nucl. Instrum. Methods Phys. Res., Sect. B, 2013, 309, 268–271.
322. M. Cutroneo, L. Torrisi, F. Caridi, R. Sayed, C. Gentile, G. Mondio, T. Serafino and E. D. Castrizio, Appl. Surf. Sci., 2013, 272, 25–29.
323. A. Galli and L. Bonizzoni, X-Ray Spectrom., 2014, 43(1), 22–28.
324. E. A. M. Kajiya, P. Campos, M. A. Rizzutto, C. R. Appoloni and F. Lopes, Radiat. Phys. Chem., 2014, 95, 373–377.
325. E. Helfenstein, K. Eremin, R. Newman, G. Gates, T. Drayman-Weisser, C. S. DeRoo, P. Klausmeyer and I. Freestone, Stud. Conserv., 2012, 57, 147–156.
326. T. Trejos, R. Koons, S. Becker, T. Berman, J. Buscaglia, M. Duecking, T. Eckert-Lumsdon, T. Ernst, C. Hanlon, A. Heydon, K. Mooney, R. Nelson, K. Olsson, C. Palenik, E. C. Pollock, D. Rudell, S. Ryland, A. Tarifa, M. Valadez, P. Weis and J. Almirall, Anal. Bioanal. Chem., 2013, 405(16), 5393–5409.
327. T. Ernst, T. Berman, J. Buscaglia, T. Eckert-Lumsdon, C. Hanlon, K. Olsson, C. Palenik, S. Ryland, T. Trejos, M. Valadez and J. R. Almirall, X-Ray Spectrom., 2014, 43(1), 13–21.
328. J. F. Fonseca, M. M. Cruz and M. L. Carvalho, X-Ray Spectrom., 2014, 43(1), 49–55.
329. E. Turillazzi, G. P. Di Peri, A. Nieddu, S. Bello, F. Monaci, M. Neri, C. Pomara, R. Rabozzi, I. Riezzo and V. Fineschi, Forensic Sci. Int., 2013, 231(1–3), 142–149.
330. J. Kawai, X-Ray Spectrom., 2014, 43(1), 2–12.
331. H. Kinoshita, N. Tanaka, M. Jamal, M. Kumihashi, R. Okuzono, K. Tsutsui and K. Ameno, Forensic Sci. Int., 2013, 227(1–3), 103–105.
332. J. W. Bond, S. V. Hainsworth and T. L. Lau, J. Forensic Sci., 2013, 58(4), 1003–1007.
333. K. Roberts, M. J. Almond and J. W. Bond, J. Forensic Sci., 2013, 58(2), 495–499.
334. G. G. Shimamoto, J. Terra and M. Bueno, J. Braz. Chem. Soc., 2013, 24(5), 731–735.
335. P. C. Chu, B. Y. Cai, Y. K. Tsoi, R. Yuen, K. S. Y. Leung and N. H. Cheung, Anal. Chem., 2013, 85(9), 4311–4315.
336. I. Nakai, S. Furuya, W. Bong, Y. Abe, K. Osaka, T. Matsumoto, M. Itou, A. Ohta and T. Ninomiya, X-Ray Spectrom., 2014, 43(1), 38–48.
337. A. A. Hummer and A. Rompel, Metallomics, 2013, 5(6), 597–614.
338. L. Pascolo, F. Esteve, C. Rizzardi, S. James and R. H. Menk, Curr. Med. Chem., 2013, 20(17), 2157–2175.
339. R. G. Leitao, C. A. N. Santos, A. Palumbo, P. Souza, G. R. Pereira, M. J. Anjos, L. E. Nasciutti and R. T. Lopes, IEEE Trans. Nucl. Sci., 2013, 60(2), 758–762.
340. R. G. Leitao, A. Palumbo, P. Souza, G. R. Pereira, C. G. L. Canellas, M. J. Anjos, L. E. Nasciutti and R. T. Lopes, Radiat. Phys. Chem., 2014, 95, 62–64.
341. V. Zaichick and S. Zaichick, Age, 2014, 36(1), 167–181.
342. M. P. Silva, D. M. Silva, A. L. C. Conceicao, A. Ribeiro-Silva and M. E. Poletti, X-Ray Spectrom., 2013, 42(4), 303–311.
343. A. Al-Ebraheem, K. Geraki, R. Leek, A. L. Harris and M. J. Farquharson, X-Ray Spectrom., 2013, 42(4), 330–336.
344. M. J. Farquharson, A. Al-Ebraheem, S. Cornacchi, G. Gohla and P. Lovrics, X-Ray Spectrom., 2013, 42(5), 349–358.
345. A. Mersov, G. Mersov, A. Al-Ebraheem, S. Cornacchi, G. Gohla, P. Lovrics and M. J. Farquharson, Radiat. Phys. Chem., 2014, 95, 210–213.
346. S. Darvish-Molla, A. Al-Ebraheem and M. J. Farquharson, ApSpe, 2014, 68(1), 79–87.
347. J. Laursen, N. Milman, N. Pind, H. Pedersen and G. Mulvad, J. Trace Elem. Med. Biol., 2014, 28(1), 50–55.
348. A. Gaal, G. Orgovan, Z. Polgari, A. Reti, V. G. Mihucz, S. Bosze, N. Szoboszlai and C. Streli, J. Inorg. Biochem., 2014, 130, 52–58.
349. R. G. Figueroa, E. Lozano and G. Bongiovanni, Revista Mexicana De Fisica, 2013, 59(4), 292–295.
350. C. Poitry-Yamate, A. Gianoncelli, B. Kaulich, G. Kourousias, A. W. Magill, M. Lepore, V. Gajdosik and R. Gruetter, J. Neurosci. Res., 2013, 91(8), 1050–1058.
351. Y. Kuang, G. Pratx, M. Bazalova, J. G. Qian, B. W. Meng and L. Xing, Med. Phys., 2013, 40(3), 7.
352. C. M. Weekley, J. B. Aitken, L. Finney, S. Vogt, P. K. Witting and H. H. Harris, Nutrients, 2013, 5(5), 1734–1756.
353. Y. X. Gao, X. M. Peng, J. C. Zhang, J. T. Zhao, Y. Y. Li, Y. F. Li, B. Li, Y. Hu and Z. F. Chai, Metallomics, 2013, 5(7), 913–919.
354. G. Robison, T. Zakharova, S. Fu, W. D. Jiang, R. Fulper, R. Barrea, W. Zheng and Y. Pushkar, Metallomics, 2013, 5(11), 1554–1565.
355. J. C. Zhang, B. Li, Y. Zhang, A. G. Li, X. H. Yu, Q. Huang, C. H. Fana and X. Q. Cai, Analyst, 2013, 138(21), 6511–6516.
356. K. Ricketts, C. Guazzoni, A. Castoldi, A. P. Gibson and G. J. Royle, Phys. Med. Biol., 2013, 58(21), 7841–7855.
357. D. Wu, Y. H. Li, M. D. Wong and H. Liu, Med. Phys., 2013, 40(5), 10.
358. I. B. Magana, P. Adhikari, M. C. Smalley, S. E. Eklund and D. P. O'Neal, Anal. Methods, 2013, 5(12), 3148–3151.
359. N. Mahdavi, M. Shamsaei, M. Shafaei and A. Rabiei, Radiat. Phys. Chem., 2013, 91, 40–43.
360. I. M. Kempson, C. C. Chien, C. Y. Chung, Y. Hwu, D. Paterson, M. D. de Jonge and D. L. Howard, Adv. Healthcare Mater., 2012, 1(6), 736–741.
361. C. L. Cheng, H. H. Chang, P. J. Huang, Y. T. Chu and S. Y. Lin, Biol. Trace Elem. Res., 2013, 152(1), 143–151.
362. D. Athanasiadou, A. Godelitsas, D. Sokaras, A. G. Karydas, E. Dotsika, C. Potamitis, M. Zervou, S. Xanthos, E. Chatzitheodoridis, H. C. Gooi and U. Becker, J. Trace Elem. Med. Biol., 2013, 27(2), 79–84.
363. K. R. Ball, F. Sampieri, M. Chirino, D. L. Hamilton, R. I. R. Blyth, T. K. Sham, P. M. Dowling and J. Thompson, Antimicrob. Agents Chemother., 2013, 57(11), 5197–5201.
364. E. D. Desouza, I. Abu Atiya, A. Al-Ebraheem, B. C. Wainman, D. E. B. Fleming, F. E. McNeill and M. J. Farquharson, Appl. Radiat. Isot., 2013, 77, 68–75.
365. R. G. Figueroa, E. Lozano and M. Valente, Revista Mexicana De Fisica, 2013, 59(4), 339–342.
366. J. Soares, C. G. L. Canellas, M. J. Anjos and R. T. Lopes, Radiat. Phys. Chem., 2014, 95, 317–319.
367. P. Fulop, I. Seres, Z. Jenei, I. Juhasz and G. Paragh, J. Cell. Mol. Med., 2013, 17(3), 350–355.
368. H. H. Liu, Z. G. Liu, T. X. Sun, S. Peng, W. G. Zhao, W. Y. Sun, Y. D. Li, X. Y. Lin, G. C. Zhao, P. Luo and X. L. Ding, Spectrosc. Spectr. Anal., 2013, 33(11), 3153–3156.
369. D. E. B. Fleming, M. R. Gherase and M. Anthonisen, X-Ray Spectrom., 2013, 42(4), 299–302.
370. M. R. Gherase, E. D. Desouza, M. J. Farquharson, F. E. McNeill, C. Y. Kim and D. E. B. Fleming, Physiological Measurement, 2013, 34(9), 1163–1177.
371. A. L. M. Silva, R. Figueroa, A. Jaramillo, M. L. Carvalho and J. Veloso, Spectrochim. Acta, Part B, 2013, 86, 115–122.
372. L. E. S. Soares, O. C. L. Martin, L. T. Moriyama, C. Kurachi and A. A. Martin, J. Biomed. Opt., 2013, 18(6), 8.
373. B. Pemmer, A. Roschger, A. Wastl, J. G. Hofstaetter, P. Wobrauschek, R. Simon, H. W. Thaler, P. Roschger, K. Klaushofer and C. Streli, Bone, 2013, 57(1), 184–193.
374. A. Roschger, J. G. Hofstaetter, B. Pemmer, N. Zoeger, P. Wobrauschek, G. Falkenberg, R. Simon, A. Berzlanovich, H. W. Thaler, P. Roschger, K. Klaushofer and C. Streli, Osteoarthritis Cartilage, 2013, 21(11), 1707–1715.
375. A. Dessombz, C. Nguyen, H. K. Ea, S. Rouziere, E. Foy, D. Hannouche, S. Reguer, F. E. Picca, D. Thiaudiere, F. Liote, M. Daudon and D. Bazin, J. Trace Elem. Med. Biol., 2013, 27(4), 326–333.
376. G. M. A. de Abreu, A. M. D. Santo, A. A. Martin and E. A. L. Arisawa, Photomed. Laser Surg., 2013, 31(8), 378–385.
377. F. Blaske, O. Reifschneider, G. Gosheger, C. A. Wehe, M. Sperling, U. Karst, G. Hauschild and S. Holl, Anal. Chem., 2014, 86(1), 615–620.
378. J. Ballarre, P. M. Desimone, M. Chorro, M. Baca, J. C. Orellano and S. M. Cere, J. Struct. Biol., 2013, 184(2), 164–172.
379. R. R. Martin, S. Naftel, S. Macfie, K. Jones and A. Nelson, Appl. Phys. A: Mater. Sci. Process., 2013, 111(1), 23–29.
380. W. Querido and M. Farina, Cell Tissue Res., 2013, 354(2), 573–580.
381. E. Da Silva, B. Kirkham, D. V. Heyd and A. Pejovic-Milic, Anal. Chem., 2013, 85(19), 9189–9195.
382. C. G. L. Canellas, S. M. F. Carvalho, A. V. B. Bellido, R. G. Leitao, M. J. Anjos and R. T. Lopes, X-Ray Spectrom., 2013, 42(4), 312–315.
383. Y. Luo, C. M. Wang, T. L. Jiang, B. Zhang, J. F. Huang, P. Liao and W. L. Fu, Biosens. Bioelectron., 2014, 51, 136–142.
384. M. M. Redigolo, R. O. Aguiar, C. B. Zamboni and I. M. Sato, J. Radioanal. Nucl. Chem., 2013, 297(3), 463–467.
385. P. R. Aranda, L. Colombo, E. Perino, I. E. De Vito and J. Raba, X-Ray Spectrom., 2013, 42(2), 100–104.
386. E. Margui, I. Queralt and M. Hidalgo, Spectrochim. Acta, Part B, 2013, 86, 50–54.
387. C. R. Hamann, W. Boonchai, L. P. Wen, E. N. Sakanashi, C. Y. Chu, K. Hamann, C. P. Hamann, K. Sinniah and D. Hamann, J. Am. Acad. Dermatol., 2014, 70(2), 281–287.
388. G. Sanchez-Pomales, T. K. Mudalige, J. H. Lim and S. W. Linder, J. Agric. Food Chem., 2013, 61(30), 7250–7257.
389. D. Mwalongo and N. K. Mohammed, Radiat. Phys. Chem., 2013, 91, 15–18.
390. W. Abuillan, E. Schneck, A. Korner, K. Brandenburg, T. Gutsmann, T. Gill, A. Vorobiev, O. Konovalov and M. Tanaka, Phys. Rev. E: Stat., Nonlinear, Soft Matter Phys., 2013, 88(1), 11.
391. Y. Pushkar, G. Robison, B. Sullivan, S. X. Fu, M. Kohne, W. D. Jiang, S. Rohr, B. Lai, M. A. Marcus, T. Zakharova and W. Zheng, Aging Cell, 2013, 12(5), 823–832.
392. X. Q. Xing, Y. Gong, Q. Cai, G. Mo, R. Du, Z. J. Chen and Z. H. Wu, Chin. Phys. C, 2013, 37(2), 6.
393. C. Streeck, S. Brunken, M. Gerlach, C. Herzog, P. Honicke, C. A. Kaufmann, J. Lubeck, B. Pollakowski, R. Unterumsberger, A. Weber, B. Beckhoff, B. Kanngiesser, H. W. Schock and R. Mainz, Appl. Phys. Lett., 2013, 103(11), 5.
394. M. P. Nikiforov, B. Lai, W. Chen, S. Chen, R. D. Schaller, J. Strzalka, J. Maser and S. B. Darling, Energy Environ. Sci., 2013, 6(5), 1513–1520.
395. H. P. Pan, L. K. Bo, T. W. Huang, Y. Zhang, T. Yu and S. D. Yao, Acta Phys. Sin., 2012, 61(22), 8.
396. J. Schon, A. Haarahiltunen, H. Savin, D. P. Fenning, T. Buonassisi, W. Warta and M. C. Schubert, IEEE J. Photovolt., 2013, 3(1), 131–137.
397. B. S. Li and S. H. Chan, Int. J. Hydrogen Energy, 2013, 38(8), 3338–3345.
398. A. Astolfo, F. Arfelli, E. Schultke, S. James, L. Mancini and R. H. Menk, Nanoscale, 2013, 5(8), 3337–3345.
399. X. J. Zhang, Y. Fang and W. B. Chen, Synth. React. Inorg., Met.-Org., Nano-Met. Chem., 2013, 43(7), 907–913.
400. A. Gianoncelli, P. Marmorato, J. Ponti, L. Pascolo, B. Kaulich, C. Uboldi, F. Rossi, D. Makovec, M. Kiskinova and G. Ceccone, X-Ray Spectrom., 2013, 42(4), 316–320.
401. S. Novak, D. Drobne, M. Goobic, J. Zupanc, T. Romih, A. Gianoncelli, M. Kiskinova, B. Kaulich, P. Pelicon, P. Vavpetic, L. Jeromel, N. Ogrinc and D. Makovec, Environ. Sci. Technol., 2013, 47(10), 5400–5408.
402. N. Manohar, F. J. Reynoso and S. H. Cho, Med. Phys., 2013, 40(8), 6.
403. H. D. Fiedler, E. E. Drinkel, B. Orzechovicz, E. C. Leopoldino, F. D. Souza, G. I. Almerindo, C. Perdona and F. Nomet, Anal. Chem., 2013, 85(21), 10142–10148.
404. B. Zawisza and R. Sitko, Analyst, 2013, 138(8), 2470–2476.
405. B. Zawisza and R. Sitko, Spectrochim. Acta, Part B, 2013, 82, 60–64.
406. C. Bussy, E. Paineau, J. Cambedouzou, N. Brun, C. Mory, B. Fayard, M. Salome, M. Pinault, M. Huard, E. Belade, L. Armand, J. Boczkowski, P. Launois and S. Lanone, Part. Fibre Toxicol., 2013, 10, 12.
407. E. Margui, B. Zawisza, R. Skorek, T. Theato, I. Queralt, M. Hidalgo and R. Sitko, Spectrochim. Acta, Part B, 2013, 88, 192–197.
408. Y. Kayser, J. Szlachetko, D. Banas, W. Cao, J. C. Dousse, J. Hoszowska, A. Kubala-Kukus and M. Pajek, Spectrochim. Acta, Part B, 2013, 88, 136–149.
409. J. J. Leani, H. J. Sanchez, R. D. Perez and C. Perez, Anal. Chem., 2013, 85(15), 7069–7075.
410. N. I. Mashin, E. A. Chernyaeva, A. N. Tumanova and A. A. Ershov, Inorg. Mater., 2013, 49(12), 1194–1198.
411. S. Zohoori and L. Karimi, Fibers Polym., 2013, 14(6), 996–1000.
412. O. El-Atwani, S. Gonderman, A. DeMasi, A. Suslova, J. Fowler, M. El-Atwani, K. Ludwig and J. P. Allain, J. Appl. Phys., 2013, 113(12), 9.
413. H. Kloust, C. Schmidtke, A. Feld, T. Schotten, R. Eggers, U. E. A. Fittschen, F. Schulz, E. Poselt, J. Ostermann, N. G. Bastus and H. Weller, Langmuir, 2013, 29(15), 4915–4921.
414. G. Gottardi, R. Pandiyan, V. Micheli, G. Pepponi, S. Gennaro, R. Bartali and N. Laidani, Mater. Sci. Eng., B, 2013, 178(9), 609–616.
415. R. Cesareo, J. T. de Assis, C. Roldan, A. D. Bustamante, A. Brunetti and N. Schiavon, Nucl. Instrum. Methods Phys. Res., Sect. B, 2013, 312, 15–22.
416. P. Ricou and K. Wood, X-Ray Spectrom., 2013, 42(6), 429–436.
417. C. Becker, M. Pagels, C. Zachaus, B. Pollakowski, B. Beckhoff, B. Kanngiesser and B. Rech, J. Appl. Phys., 2013, 113(4), 7.
418. C. Degueldre, C. Cozzo, M. Martin, D. Grolimund and C. Mieszczynski, J. Solid State Chem., 2013, 202, 315–319.
419. C. Degueldre, C. Borca and C. Cozzo, Talanta, 2013, 115, 986–991.
420. M. Menzel, A. Schlifke, M. Falk, J. Janek, M. Froba and U. E. A. Fittschen, Spectrochim. Acta, Part B, 2013, 85, 62–70.
421. M. E. Bowden, K. J. Alvine, J. L. Fulton, J. R. Lemmon, X. Lu, B. J. Webb-Robertson, S. M. Heald, M. Balasubramanian, D. R. Mortensen, G. T. Seidler and N. J. Hess, J. Power Sources, 2014, 247, 517–526.
422. F. Pomiro, S. Ceppi, J. M. De Paoli, R. D. Sanchez, A. Mesquita, G. Tirao and E. V. P. Miner, J. Solid State Chem., 2013, 205, 57–63.
423. S. Ceppi, A. Mesquita, F. Pomiro, E. V. P. Miner and G. Tirao, J. Phys. Chem. Solids, 2014, 75(3), 366–373.
424. I. Radisavljevic, N. Ivanovic, N. Novakovic, N. Romcevic, M. Mitric, V. Andric and H. E. Mahnke, J. Mater. Sci., 2013, 48(23), 8084–8100.
425. T. Yamamoto, T. Kudo and J. Kawai, Microporous Mesoporous Mater., 2013, 182, 239–243.
426. J. Malherbe and F. Claverie, Anal. Chim. Acta, 2013, 773, 37–44.

---

Footnote

† Review coordinator.

---