Analysis of the State of the Art: XRF

In honor of Spectroscopy's celebration of 30 years covering the latest developments in materials analysis, we asked a panel of experts to assess the current state of the art of X-ray fluorescence and to try to predict how the technology will develop in the future.

As a fast, quasi-nondestructive analytical technique, X-ray fluorescence (XRF) spectroscopy is widely applied in industry, primarily in quality control. High-resolution versions of the technique, such as microXRF (µXRF), achieve extremely fine spatial resolution, and are useful in a wide range of scientific disciplines, from space exploration to art conservation. In celebration of this 30th year of Spectroscopy, we asked a panel of experts about the current state of XRF (focusing mainly on energy-dispersive XRF)-including recent advances, ongoing challenges, and possible future developments. This article is part of a special group of six articles covering the state of the art of key techniques, also including inductively coupled plasma–mass spectrometry (ICP-MS), laser-induced breakdown spectroscopy (LIBS), infrared (IR) spectroscopy, near-infrared (NIR) spectroscopy, and Raman spectroscopy.

The Most Important Recent Developments

We started by asking our panel what they considered the most important advance that has been made in XRF spectroscopy or its instrumentation in recent years.

Ursula Fittschen, an assistant professor of analytical chemistry at Washington State University, named two older developments, microfocusing tubes and silicon drift detectors, and one very recent one, the development of X-ray color cameras that have good spectral resolution.

Matthew Newville, a senior research associate at the University of Chicago, agreed with Fittschen about the importance of commercially available multielement silicon drift detectors. He also cited another recent advance: electronics that use field-programmable gate arrays (FPGAs) for dynamic filtering, allowing count rates well above 1 MHz per detector channel with good energy resolution.

Optics came to mind for Christina Streli and Peter Wobrauschek, both professors at Vienna University of Technology, and for George Havrilla of Los Alamos National Laboratory.

The mainstream availability of X-ray optics for both laboratory-based and synchrotron instrumentation, Havrilla said, has generated a newfound interest in XRF and its use in new and exciting applications. "These optics include the polycapillary and doubly curved crystal optics used in both laboratories and synchrotrons, as well as the highly specialized optics that have been developed for synchrotrons that enable spatial resolution to be pushed into the nanometer scale," he explained. "There are exciting things happening at submicrometer spatial resolution that open new doors in terms of understanding the cosmos as well as life here on Earth."

Jun Kawai, a professor at Kyoto University, pointed to the popularity of handheld XRF spectrometers-noting that more than 50,000 instruments had been sold by 2010-and the commercial availability of inexpensive XRF kits. "If your computer has a USB port and you have $17,000 (USD), then you can do XRF analysis," he said, referring to the kits. What has made handheld and kit instruments popular, he says, is that now their detection sensitivity is the same or sometimes even better than kilowatt water-cooled X-ray tube XRF instruments. "Low-power instruments are also now advancing in total reflection XRF, polarization XRF, and µXRF," he said.

Kenji Sakurai, a professor and group leader at the National Institute for Materials Science in Tsukuba, Japan, said important advances had been made for both small-scale and large-scale laboratories. For the former, he named battery-driven XRF, which consists of a low-power compact X-ray source, a silicon-drift detector, and a computer for quick data processing based on the fundamental parameter method as well as comparison with the database. "Now one can find many such products from several manufacturers," he said. "This is the type of XRF spectrometer that NASA has been using on Mars."

On the other hand, Sakurai pointed out, large-scale facilities (such as new synchrotrons and X-ray free electron lasers) are extremely important for future scientific developments. "The first success in lasing in an X-ray wavelength (1.5 Å) at Stanford in April 2009 was really an epoch-making milestone," he said. "In the near future, with the help of an X-ray free electron laser, we will be able to get a clear understanding of the excitation and relaxation of inner-shell electrons, which are underlying principles of XRF."

Current Challenges

We then asked our panel what key challenges or questions researchers and users of XRF currently face.

Streli, Wobrauschek, and Newville all mentioned the challenges of quantification, particularly in analyses done at the microscale. Newville mentioned the need for robust, high accuracy, and easy-to-use software for quantitative analysis. "I think that ultimately the absorption and emission cross-sections are not known to sufficient accuracy," he said.

Fittschen named calibration in imaging approaches, as well as accurate in situ trace elemental determination.

Martina Schmeling, an associate professor at Loyola University Chicago, noted that traditional XRF still suffers from poor detection limits for most elements, especially for trace environmental analysis. "MicroXRF helps to some extent, but probes only on a small surface," she said. "And TXRF [total reflection XRF] is limited to highly reflective surfaces."

In Havrilla's view, the key challenge for researchers is to bring advances being developed at synchrotrons into the laboratory. "We would like to be able, in the lab, to obtain submicrometer elemental images on biological specimens and discern the elemental composition and spatial distribution of elements in materials," he said. "We want to be able to understand how the elemental distribution and composition at the nanometer scale affects and defines the properties of materials at the macroscale."

Sakurai raised a key underlying concern in XRF: the issue with calibration standards. The challenge, he explained, is that XRF intensity is not independent from the matrix-it can change depending on the chemical composition of the sample. "The technique is powerful, when one can prepare calibration standards-concentration-controlled samples with the same matrix as the sample to be analyzed-as can be done for steel, ceramics, glass and many other industrial materials," he said. But that is not always possible, and preparing calibration standards is even more difficult, he said, when the system being analyzed is inhomogeneous in two or three dimensions. "To avoid the problems caused by inhomogeneity, small area analysis and the imaging will be a great help," he said.

Another important challenge for the future, Sakurai said, will be sorting out time-dependent XRF. "Not many experiments have been done on this yet," he said. "This will be a future task-probably to be taken on with some newly developed light sources."

Don Broton, a principal scientist at CTLGroup, considered the question in the context of his firm's work in engineering and construction, where many end users are not experts in spectroscopic techniques. "I expect that with advances in software, the need to know how XRF works will be replaced with instrument interfaces or 'expert systems' to guide the user on best practices," he said. "That will bridge the gap between what users know and what they need to know to produce an acceptable result."

Lora Brehm, a research scientist at Dow Chemical Company, also raised the issue of user knowledge, noting the general lack of experience with XRF in undergraduate education in the United States. "When students are not introduced to XRF during college, they sometimes overlook the technique later," she lamented. She noted, however, that this has been improving somewhat, because the availability of inexpensive portable and benchtop XRF instruments has led to increased use in undergraduate programs.

User knowledge was also on the mind of Maggi Loubser, who is a group chief chemist at the South African cement company PPC. "The technique has matured to become a push-button black box for many applications," she said. "Users often produce very bad data unknowingly because they do not understand the limitations of the technique."

The Need for Improved Instrument Performance

Our panel then weighed in on what instrument features are most in need of improvement, and which of those are likely to be addressed in coming years.

Improvements in XRF detectors topped the list, and Streli presented a wish list of such improvements. "We would like to see higher resolution in silicon drift detectors, higher detector count rate performance, greater thickness of silicon drift detectors for better efficiency of higher energies, and a large area detector," she said.

"Energy resolution for X-ray detectors will lower detection limits," added Kawai. "The sensitivity of silicon-drift and silicon-PIN detectors is not good for higher energy X-rays, above 20 keV."

"In particular, the development of a much faster, lower-noise, higher-energy-resolution 2D detector, with a smaller pixel size and a larger detection area, will be extremely important for XRF and XRF imaging," said Sakurai. "I expect to see progress in the development of a 2D X-ray detector in the coming years."

Newville added another detector feature to the wish list. "For my applications, the ability to make multielement arrays of silicon drift detectors would be very useful," he said.

Others mentioned the need for improvements in the XRF source and tubes, particularly intensive sources for laboratory equipment. For the latter, Sakurai said, the compactness of the source will be critical. "In the laboratory-based X-ray tube, we have a clear limit in the electron-beam density on the anode," he said. "Therefore, compactness can be a key to enhancing the X-ray intensity at the sample position."

Kawai considered improvements needed in X-ray tubes for portable instruments. The filament in a traditional X-ray tube uses a lot of electric power, he noted, which means the electric battery-charging interval is short. "The laser photoelectron emission type X-ray tube has reduced the power consumption of X-ray tubes for handheld instruments," he said. "Such a power-saving and stable X-ray tube is important."

Fittschen would like to see better cost-efficiency of miniaturized and camera equipment. "XRF color cameras have a bright future," she said. "Right now they are still in the prototype stage, but when larger numbers are produced, prices should go down."

Havrilla seeks greater spatial resolution without having to go to a synchrotron. "Getting into the sub-10-µm scale in the laboratory, and even below 1 µm, would open huge areas of investigation to researchers worldwide," he said. "That said, even having detectors with spectral resolution below 10 eV or better would provide unique capabilities in the laboratory, which are only available at synchrotrons currently." Such spectral resolution is being actively pursued by a number of research groups, he added. "These groups are primarily looking to answer fundamental physics questions using micro-calorimeter detectors, but a side benefit of this research is work toward creating X-ray detectors with spectral resolutions at or below 10 eV."

Schmeling agreed. "Higher-resolution laboratory-based instruments will open up the possibility of easier access and wider range of applications," she said. "Synchrotron time is very limited, but laboratory time far less so, and in many cases the high sensitivity of the synchrotron can be offset by longer counting times in the laboratory."

"It would also be nice to see beam optics in the few-nanometer range applicable to real samples," added Wobrauschek.

From her industrial perspective, Loubser sometimes finds that instrument "improvements" are not helpful. "I wish instrument manufacturers considered that better and better sensitivity has disadvantages regarding the importance of sample preparation," she said. "Larger sample anode distances [on older instruments] were much more forgiving to irregular surface areas."

Application Areas for XRF

We also asked our panelists to comment on the most important application areas for XRF spectroscopy, including the challenges involved and how XRF competes with other techniques.

Geological-related fields are a growing area for XRF application, Brehm noted. "Geologists, geological engineers, lab technicians, drill geologists, mud loggers, and geochemists are all using XRF," she said. Some examples include the use of down-hole and portable systems for mining and energy exploration and the use of XRF core scanners for chemostratigraphy studies. "Techniques like ICP-OES and atomic absorption do not lend themselves to field analysis as they require acid digestion of the sample, but XRF can be readily applied, especially with miniaturization of components."

Schmeling agreed. "The trend is clearly to portability and ease of use in the field," she said, adding that XRF may also find a place in hydraulic fracturing and gas exploration, given the robustness and ease of use of the method. "XRF has many advantages over the mass spectrometric methods-the major one being independent of carrier gases and other consumables," she said. "It is important to keep in mind that there is an XRF spectrometer on Mars, but no ICP-MS."

Fittschen looked more broadly at how XRF competes with other techniques, noting that usage depends on analyte levels and other factors. "Conventional XRF instrumentation is most attractive in all applications where parts-per-million levels are to be analyzed in refractory material," she said. But for trace elemental analysis in the parts-per-billion range, she noted, ICP-OES is the workhorse, as long as sample material is not limited and digestion is straightforward. For limited samples, micro-analyzing tools like TXRF or graphite furnace atomic absorption spectroscopy may be a better choice. "For detection at the parts-per-trillion level, one needs ICP-MS," she added.

Several respondents mentioned industrial applications for quality control, such as in iron and steel manufacturing. "The precision of steel products is quite high and wavelength dispersive XRF is needed," said Kawai. "There is one instrument that has 40-crystal spectrometers, making it possible to do quantitative analyses of 40 elements at once with a single XRF instrument."

Industrial applications of XRF go beyond quality control, Broton pointed out. "Phase identification and standardless XRF enhance the capability of manufacturing plants to rapidly assess the use of alternative materials and formulations, as well as the by-products of their respective processes," he said. "A green future will be enhanced by better characterization of these 'wastes' so more and better reuse is achieved."

Future applications include the study of cultural heritage objects, as well as environmental, medical, and other technical fields, said Streli. "The value of XRF will be seen wherever nondestructive investigation is valuable," she said.

Havrilla agreed, adding that XRF is already proving its worth in cultural heritage studies. "The rapid growth of imaging works of art using µXRF is allowing the discovery of overpainted pictures, giving new insights into the provenance of the art," he said. "These techniques are also revealing how pigments degrade, leading us to understand that some colors we see in artwork today are not the same as they were when first painted by the artist."

Havrilla also pointed to recent advances in optical tweezers, which use XRF to enable the mechanical manipulation of single cells, in turn enabling elemental imaging of cell contents in vivo. "This can lead us to new understandings of biological mechanisms," he said.

Sakurai also sees developments in analyzing chemical states by XRF, including important work being done at synchrotrons using X-ray absorption fine structure (XAFS) and X-ray absorption near-edge structure (XANES) with XRF detection. "I believe that chemical state analysis by XRF will open new opportunities in science and many engineering fields."

Growth in µXRF

We then asked our panelists if they are seeing growth in the use of µXRF, which has a spatial resolution with a diameter as small as 10 µm. Conventional XRF, by comparison, has a spatial resolution ranging in diameter from several hundred micrometers up to several millimeters.

Wobrauschek sees a demand for such high spatial resolution for various applications, such as for the elemental distribution in cells. "Future development in XRF will go in the direction of nano-analysis," he predicted.

"Fifteen years ago, µXRF was only possible at a synchrotron facility; today, laboratory µXRF is possible and more stable, and can be done using an X-ray tube of just a few watts," said Kawai. "This progression can been seen in two books, written 15 years apart [1,2]."

Brehm said she sees industry use of µXRF increasing, because µXRF fills a niche between electron beam technologies and conventional bulk XRF. "If the high resolution capability of scanning electron microscopy is not required, such as for small samples, inclusions, or unknown particles, the analysis can be done much faster with a µXRF instrument because of reduced sample preparation time," she said. "It is a very nice tool for troubleshooting production problems."

The advent of new commercial instrumentation is opening new applications areas for µXRF, noted Havrilla. "Some commercial laboratory instruments can push the X-ray spot size down to 10 µm," he said. "This is phenomenal. These advances will only increase the application of µXRF over the next 10 years."

Sakurai agrees that the current spatial resolution of current laboratory µXRF instruments, at nearly 10 µm, is impressive, while also noting that at third-generation synchrotrons, a 0.1–0.2 µm beam is now standard. At the SPring-8 facility, in Harima, Japan, there is a dedicated beamline for studying a variety of living cells by XRF with 0.05-µm resolution, he adds.

Fittschen noted that X-ray color cameras are full-field µXRF equipment. "I see a growing market here, because they allow for static measurement, require almost no adjustment, and provide an immediate image," she said. "They are ideally suited for nontarget screening in research, but applications in the recycling industry or in processing lines can be envisioned."

While excited about the power of µXRF, panelists also noted some challenges and cautions. "The primary issue for µXRF," Havrilla said, "is the avalanche of data that accompanies elemental maps taken over tens to hundreds of micrometers with a nominal spatial resolution of 10–50 µm."

Although forecasting that the spatial resolution of µXRF will continue to improve, for both ordinary labs and synchrotrons, Sakurai notes that such improvements will raise other concerns. "Then we will need to take care of the radiation damage," he said. "Also we may need some new theory to do reliable quantitative analysis."

Kawai, for his part, is not that convinced about the usefulness of µXRF. "If you have experience of using SEM-EDX [scanning electron microscopy coupled with energy-dispersive X-ray spectroscopy], it is easier and more informative to observe micro areas by secondary electron images as well as elemental X-ray mapping, " he said. "SEM-EDX is superior to µXRF for practical use."

Further Development of µXRF Instrumentation

So, we asked, is there still a need for further development of µXRF instrumentation? If so, in what ways?

Fittschen sees a need to develop full-field µXRF using array detectors. "This is just in the early stages," she said. "Sources, optics, and detectors for such setups need to be developed and optimized." Newville agreed. "Detector arrays with higher count rate are a continual need for µXRF and µXANES," he said.

Newville cited the need for better software. "XRF analysis software for larger, faster datasets needs to be more robust and user-friendly, and to be better coupled with other analytical methods such as XRD, XAFS, and ICP-MS," he said, referring to X-ray diffraction, X-ray absorption fine structure, and inductively coupled plasma–mass spectrometry.

"Reference-free or standard-less quantitative analysis software is needed," added Kawai.

Havrilla feels that µXRF would benefit from the development of imaging detectors that offer the inverse of the current state-of-the-art. "Imaging detectors offer the use of global illumination–excitation of the sample, while the detector provides the spatial resolution," he explained. "This will generate faster elemental images."

Brehm sees confocal µXRF as a potential growth area. "Availability of this type of instrument through a commercial vendor is needed," she said. "Development of software for quantitative elemental analysis in the confocal configuration is needed as well."

Havrilla agreed. "Confocal µXRF is still a research curiosity, yet it is the only elemental analysis technique capable of providing nondestructive 3D elemental imaging," he said. "Commercial development of a confocal µXRF would offer significant advantages over conventional 3D elemental analysis instrumentation."

Calibration Standards for µXRF

The lack of adequate and readily available calibration standards is a problem for µXRF, however, and presents an ongoing challenge. "Suitable standards are urgently required!" declared Wobrauschek.

The problem, Fittschen explained, is that references are usually homogeneous in nature whereas samples interesting to µXRF usually are not. "A reference needs to account for these inhomogeneities, which is a challenge," she said. "A combined approach using structured references and numerical models should be optimal from my point of view."

Havrilla agreed. "Metals and alloys that appear homogeneous at the bulk level are quickly seen as heterogeneous at the micrometer scale," he said. "For µXRF to move into the mainstream analytical laboratory, quantitative capabilities need to be developed, and that requires micrometer-scale homogeneous calibration standards for composition as well as spatial distribution."

Synchrotron-Radiation µXRF and µXANES

When the power of a synchrotron beam is added to µXRF and µXANES spectroscopy, the spatial resolution becomes much smaller-as low as 0.1–0.2 µm-and new types of analysis become possible. We asked our panelists to comment on current state, and usage, of these synchrotron-based techniques.

Fittschen pointed out that because of rapid, high performance scanning detectors like the Maia detector-an advanced system designed specifically for scanning XRF microprobe applications-2D µXANES of minor compounds over large areas (such as 4 mm²) can be accomplished with submicrometer spatial resolution in as little as 6 h. "And thanks to improved optics, spatial resolution of around 30 nm can be achieved at third-generation synchrotrons," she said.

Using synchrotrons for µXRF and µXANES allows analysis truly at the micrometer and even 10-nm scale, said Newville, who runs a synchrotron beamline at the University of Chicago. In addition, he noted, one can easily tune the X-ray energy, which makes it possible to selectively turn on and off fluorescence lines from particular elements. "Because the incident fluxes are so high, analyzable µXRF spectra can be measured in well under 1 s, allowing mapping of elemental composition with micrometer resolution and for low concentrations," he said. "This enables one to measure metals composition and correlations in heterogeneous materials like soils and zoned mineral phases, archeological specimens, and also cellular-level spatial sensitivity in plants and animal tissues."

"On top of the elemental features obtained by µXRF, µXANES offers the functional aspects by providing not only the element, but the oxidation state of the element as well," adds Havrilla.

In addition to providing gains in spatial resolution, these techniques have led to major advances in imaging, Sakurai pointed out. "Now XAFS and XANES images can be taken with a resolution of 1000 x 1000 pixels," he said (3,4). The third advance is in terms of light elements, he added. "Many interesting trials to see XAFS [analysis] of light elements such as C and O have been recently reported," he said.

Kawai has been developing and using XANES methods for 30 years, but he has a tempered perspective. "If you can use a synchrotron radiation microbeam, micro-X-ray diffraction is more useful," he said. "Because XRD is possible for all elements, but XANES is beamline dependent, so only one element absorption edge is possible at a time." He added that µXANES is useful for catalyst analysis, but that infrared or Raman spectroscopy is sometimes more informative.

Future Development of Synchrotron-Radiation µXRF and µXANES

In the future, our panelists foresee that synchrotron-radiation µXRF and µXANES will continue to develop.

"We're trying to go faster, and to be able to easily couple µXRF, µXANES, and in some cases µXRD to characterize samples as fully as possible," said Newville. "In addition, X-ray beams are getting brighter and focusing technologies are getting better so that it is becoming possible to do these routinely at a resolution of 100 nm or smaller."

Sakurai notes that many users of µXRF will need more information than what can be obtained through single-point analysis. As a result, he foresees the technique being combined with scanning, 1D line profiling, or 2D imaging.

Havrilla also foresees that these synchrotron techniques will increase in importance as the number of beamlines increases and the effort to make these measurements becomes more turnkey, opening up these techniques to a wider set of users.

If these techniques do become very popular, Sakurai said, long measuring times at synchrotrons will be a problem. "Then it will be necessary to develop some kind of extremely quick scanning method to see quite a wide area, or to use some other imaging method such as projection-type XRF imaging."

Sakurai also provides two cautions. "One should be careful about the possibility of radiation damage effects in µXRF and imaging," he said. And for accurate quantitative analysis, he believes new theory will be required. "The present Sherman's equation does not consider inhomogeneity in the same depth-plane," he said.

Total Reflection XRF

TXRF is another type of XRF spectroscopy that has been gaining interest and adoption in recent years, particularly for certain industrial applications, such as microelectronics manufacturing, where its ability to conduct micro and trace analysis is useful. In TXRF, a small amount of sample-prepared as a thin film or layer-is placed on a flat, reflective support. X-rays hit the sample at a glancing angle (<0.1°), exciting atoms in the top layers of the material, and the fluorescence is detected by a silicon-lithium or a silicon drift detector. Matrix effects are negligible in TXRF, making quantification possible. We asked our panel about the current role of TXRF, and how they think this technique, and its use, will develop in the coming years.

TXRF is a very attractive technique that allows for trace elemental analysis in microscopic amounts of sample, said Fittschen, adding that for many elements it has detection limits that are comparable to those of ICP-OES without needing high purity argon. It also allows for a quick qualitative overview of elements present in a sample. "TXRF is an ideal, cost-efficient method for a small laboratory where nontargeted screening is required," she said.

"TXRF applications have been growing steadily, in particular in the environmental and biological sectors," said Schmeling. "The big advantage of TXRF to its main competitor, ICP-MS, is its small sample volume and its ability to analyze both solid and liquid samples." As a result, she believes that in areas with very limited sample amounts such as biological specimens, TXRF applications will continue to grow.

Sakurai agreed. "TXRF can be used for trace element analysis, and usually provides a detection limit that is 2–3 orders of magnitude better than ordinary XRF," he said. "And by using low-power white X-rays in very efficient ways, or by using efficient and therefore wider-band monochromators, you can get even better performance, but will still encounter some inherent limits."

Loubser finds the technique very useful in her work. "TXRF is an invaluable tool where samples are difficult to prepare as fused beads or pressed pellets, such as with platinum concentrate slurries," she said. "And the detection limits approach those of ICP-MS."

The main problem with TXRF is that not many instrument models are available, said Fittschen. "There is demand for both cost-efficient 'simple' TXRF instrumentation as well as sophisticated, research-orientated TXRF instrumentation," she said.

There is a lot of potential for further applications of TXRF, especially using transportable systems, Wobrauschek noted, particularly given the easy sample preparation and the method's usefulness for screening. "Straight TXRF enables spatially resolved elemental determination on any smooth surface," he said. "A further development of TXRF-grazing incidence XRF, or GIXRF-will allow progress in nanometer-layer characterization required by nanotechnology."

For TXRF use to expand, commercial laboratory-based instrumentation is needed, Havrilla says, along with a well-developed sample preparation methodology to enable easy calibration and sample analyses. "What is needed is a robust calibration and sample preparation system that can produce good precision and accuracy," he said. "This would enable TXRF to take off in terms of applications."

Portable and Handheld XRF

Our panel also commented on trends in portable and handheld XRF instrumentation.

"Portable XRF instrumentation has enabled completely new applications in the field, as well as in production lines in industries," noted Sakurai. Streli agreed. "Handheld instruments are the fastest expanding market for XRF spectrometers," she said. "These systems expand the overall use of XRF."

"Portability is a great asset and will set XRF from its competitor ICP-MS or ICP-OES apart," added Schmeling. "No gases or liquids have to be used to run the analysis, thus any field analysis, especially in oil, gas, and natural resources exploration, is promising." In addition, she said, testing for heavy metals in imported jewelry and trinkets is likely to grow.

"As I mentioned before, XRF usage is being expanded not only by handheld instruments, but also by cheap XRF kits," said Kawai. "With these low-power kits, users can arrange their spectrometers to meet the needs of their particular type of analysis."

The panelists also agreed that portable instruments will not replace laboratory systems. "The sensitivity, stability, and quantitative capabilities of the laboratory instrument cannot be replaced," said Havrilla.

Havrilla warned that handhelds still require a solid background in the technique. "Handheld and similar portable XRF instrumentation has been touted as 'simple to use' but this does not mean the user should not know anything about XRF," he cautioned. "To the contrary, the user should have more than a working knowledge of XRF to fully capitalize on the power of the handheld instruments being sold today."

Loubser agreed. "As much as the 'black box' capability of these systems is expanding the overall use of XRF, that very capability also holds the biggest danger," she said. "Aim-and-shoot cowboy spectroscopists doing quantification on inhomogenous or uneven samples often produce dubious results that damage the reputation of the technique."

The next stage of the use of mobile XRF instrument should be discussed in the context of other sophisticated technologies, said Sakurai. "These include robots and compact helicopters that carry XRF machines, manless measurements, and very fast automatic sample-changing systems for the laboratories, hospitals, and industry, for testing huge number of samples in a short time," he said.

Combining XRF with Other Techniques

We asked our panel how they are seeing other techniques combined with XRF, and for what application areas.

"There is combined SEM–µXRF instrumentation that offers the benefits of both excitation characteristics," Fittschen noted. "And given that XRF is quasi-nondestructive, a number of combinations can be envisioned using XRF, such as in prescreening methods."

"XRD is being combined with XRF using a microbeam for material characterization," said Streli. "Also, GIXRF is being used in combination with X-ray reflectometry (XRR) for thin-layer and implant characterization, and small-angle X-ray scattering (SAXS) is being used in combination with XRF."

"Along with µXANES and µXRD, we occasionally use optical fluorescence, say with biological samples tagged with fluorescent proteins," said Newville.

In industry, the combination of XRD and XRF is changing the production landscape of cement, noted Broton. "XRF is useful for combining raw materials and optimizing mix constituents, whereas XRD is good for confirming final product quality," he said. "The combination of instruments will likely be installed in all cement manufacturing facilities in the future." The power industry is also recognizing the benefits of this technology combination, he added.

Loubser is excited about developments in this combination. "The XRD–XRF combination is improving with each supplier launch of a new model!" she declared. The integration of XRF and XRD into a single system, with the same operating software, is a huge bonus, she added. "This is opening an entire new field not just for mineral beneficiation and cement manufacture but for many other industries," she said.

Havrilla says that Raman and infrared spectroscopy are two common instrumental methods typically combined with XRF, but it is not easy. "Full integration of the elemental information obtained by XRF and the molecular information from the Raman or IR instrumentation requires a huge investment in time and effort," he said. The great advantage of doing so, he said, is that it provides information about chemical phases, not just about elemental or molecular composition. "This integration generates a comprehensive characterization of the material, with the added bonus of being nondestructive, such that the sample can be further analyzed, say using mass spectrometry."

Advanced Computation, Modeling, and Data Analysis

We then asked our panel about the importance of advanced computation, modeling, and data analysis in XRF spectroscopy.

Several commented on improvements in the fundamental parameters approaches and databases, for XRF and for XRF imaging. The fundamental parameters database is now being revised through a joint effort of international standards-setting bodies, including PTB in Germany, NIST in the United States, LNE in France, NIMS in Japan, and others. "The computational method for fundamental parameters has already been completed, but database revisions will improve the accuracy of quantitative analyses," Kawai said.

"By combining modeling with experimental measurements, improvements in both the hardware and computational model generates advances in quantitative measurements of material composition, which is the ultimate goal for analytical scientists," said Havrilla.

"I would believe that XRF has good potential for determining chemical composition of completely unknown samples without any calibration standards," said Sakurai. "However, for this, we need to improve the performance of the fundamental parameter method. We need to establish some good reliable ways to avoid getting wrong results with some satisfactory fitting."

Broton agreed. "Standardless compositional analysis has proven invaluable in investigations already," he said. "The more accurate the calculated values become, the more it will be used in industry."

In addition, Broton sees the value in software advances that help nonexpert users. "Computer programs that essentially evaluate and recommend instrument setup conditions for a wide variety of material types with little operator input will allow multiple users the ability to analyze virtually any material that will fit in the instrument," he said.

Time-Resolved Imaging Using New X-ray Cameras

We also asked our panelist to comment on a very recent development, time-resolved imaging using new X-ray cameras.

"Time-resolved imaging enables new applications," said Streli. "Ultra-short pulses in the femtosecond range allow fundamental research of atomic and subatomic structures."

The full-field imaging X-ray cameras are very exciting, agreed Havrilla. "While they are still experimental, they are becoming more reliable and will eventually become useful in providing elemental imaging in the laboratory." In parallel, he added, the development of microcalorimeter detectors offering sub-10-eV spectral resolution will enable high-resolution µXRF and µXANES in the laboratory. "Both of these detector advances will lead to exciting times for the future of XRF."

Sakurai, who succeeded in making an XRF movie many years ago (5,6), believes the value of applying time-resolution to X-ray techniques is to see how materials are synthesized or how other processes actually occur. "Our XRF movie shows the dendritic growth of metals during electric deposition," he explained. "Such measurements should be useful to see, for example, the reaction of the electrodes in batteries."

Summary

Significant developments have been made in XRF technology and instrumentation in the past 10–15 years, including notable improvements in microfocusing tubes, silicon-drift detectors, electronics, and optics. Advances in portable and handheld devices have made the technique practical for a wide variety of applications in industry. On the other extreme, powerful versions of the technique such as µXRF and µXANES have brought spatial resolution down to 10 µm in the laboratory and 0.1–0.2 at synchrotrons, making ultrafine measurements possible in a wide range of materials-from artwork to space exploration equipment to plant and tissue cells-providing information not only about elemental composition but oxidation states as well. XRF imaging has also advanced. TXRF, meanwhile, has become a very attractive technique for trace elemental analysis in microscopic amounts of sample.

In all fields, of course, everyone wants everything to be better, faster, and cheaper, and for XRF, that means ongoing improvements in detectors, sources, and tubes, and software, and the desire to achieve synchrotron-like resolution in the laboratory. Future developments will also surely include advances in time-resolved XRF, confocal µXRF, and imaging detectors. XRF is increasingly being combined with other techniques, including XRD, SEM, and Raman and IR spectroscopy, for increased power and versatility.

Challenges still remain, however, including the need for calibration standards, and the corresponding difficulty using XRF for quantification, particularly in microanalysis. There is also a great need for improvements in the fundamental parameters approach and databases. Given the developments that have been made in the last few decades, it seems likely that advances in all of these aspects of XRF will continue apace.

Acknowledgments

I would like to express my thanks to Lora Brehm for all of her help suggesting and recruiting panelists and for providing feedback on questions. I would also like to thank Stefan Vogt of Argonne National Laboratory and Matt Newville for suggesting additional questions, and to Tim Elam of Washington University, for recommending Stefan Vogt, Matt Newville, and Lora Brehm to me.

References

(1) K.H.A. Janssens, F.C.V. Adams, and A. Rindby, Eds., Microscopic X-ray Fluorescence Analysis (Wiley, Hoboken, New Jersey, 2000).

(2) M. Haschke, Laboratory Micro-X-Ray Fluorescence Spectroscopy, Instrumentation and Applications (Springer, New York, 2014).

(3) K. Sakurai and M. Mizusawa, Nanotechnol. 15, S428 (2004).

(4) M. Mizusawa and K. Sakurai, J. Synchrotron Rad. 11, 209–213 (2004).

(5) K. Sakurai and H. Eba, Anal. Chem. 75, 355 (2003).

(6) AIP Conference Proceedings 705 (Synchrotron Radiation Instrumentation 2003, San Francisco, USA) 889–892 (2004).

Laura Bush is the editorial director of Spectroscopy.