X-ray fluorescence (XRF) spectrometry instrumentation has undergone significant evolution in the past couple decades.
X-ray fluorescence (XRF) spectrometry instrumentation has undergone significant evolution in the past couple decades. Participants in this Technology Forum include Alexander Seyfarth of Bruker AXS, Sergey Mamedov of Horiba Jobin Yvon, Michael W. Hinds of the Royal Canadian Mint, and Didier Bonvin and Pascal Lemberge of Thermo Fisher Scientific.
With the advance of detector technology and electronics, the use of energy-dispersive X-ray fluorescence (EDXRF) (simultaneous) spectrometry including total reflection X-ray fluorescence (TXRF), handheld devices, and µXRF has widened tremendously in the past decade. In addition, its purchasing price is also favorable. Will EDXRF eventually replace wavelength-dispersive X-ray fluorescence (WDXRF) (sequential) in the near future?
Seyfarth: EDXRF will definitely widen its base and "poach" a number of installations from WDXRF, especially in market segments with lower sample throughput and low margin products such as industrial minerals. The main driver for the increase in EDXRF will be to lower the cost of the instrumentation and make it more compact.WDXRF will remain at the top of the performance scale due to its unbeatable resolution and precision: Being able to count millions of counts at one energy will not be matchable in the near future by any potential future EDXRF detector technology. For the highest precision of major components WDXRF outperforms the optical emission spectrometry (OES) spark technique as well, thus remaining the most important analytical tool for high alloy steels, Monel, Inconel, and others. The metals market is a nice example of the various technological developments.
WDXRF (simultaneous or sequential) use decreased with the OES capabilities increasing driven by the lower cost of the OES units. Then came handheld XRF, and a new market emerged with a vengeance for EDXRF-based handheld units. The EDXRF capabilities are now such that it is possible to use it for many metals application. The price differential though between OES and EDXRF is not yet disruptive enough to drive more growth for laboratory EDXRF. There is an interesting emergence of "hybrid" units in the metals field: Instead of the traditional combination of WDXRF simultaneous channels with a "scanner" or sequential crystal based goniometer, EDXRF has successfully been integrated into a multichannel system. This approach now reduces the number of "expensive" channels to the bare essentials for the alloy's precision requirements. The EDX part then deals with the "rest" still simultaneously and offers features known from handheld XRF such as fingerprinting.
On the portable–handheld side though EDXRF is the clear winner in the metals market except for the lone carbon-in-steel application, which is performed better by OES due to the small information depth for carbon by XRF.
EDXRF instrumentation using state-of-the art silicon drift detectors (SDD) are getting closer to the theoretical resolution limit of 125 eV (Fano limit) in both benchtop and handheld instrumentation uses. Increased count rate capability with constant resolution and low dead time is achieved by relentless improvement on the detector electronics. This allows applications such as limestone and cement to be done on fusion beads with precision former only known to pressed pellets.
The dominance of WDXRF in light element analysis (C-F) is now broken as well in the benchtop class: Two commercial units are currently on the market that allow the measurement of F in solids as well as the qualitative detection of carbon.EDXRF will need an even more powerful new detector to further narrow the performance gap. Let the search for the new technology start: Development of the SDD technology was relatively fast from 1992 with full commercial adaptation in 2002.
Mamedov: For some applications, EDXRF will replace WDXRF but for some not. As an example: WDXRF has much higher energy resolution and can detect elements in different oxidation states. EDXRF can detect only elements.EDXRF is very good when you want to see all elements/composition such as in glass analysis in forensics. WDXRF is very good when you want to see a low concentration of one element. An example would be sulfur in oil/diesel. WDXRF systems are able to detect a couple elements simultaneously but it would require either a couple detectors (one per each element/line) or accurate positioning of one detector.
Hinds: The advances in EDXRF technology over the past few years have been impressive. However, I think that EDXRF will not replace WDXRF for applications that require high resolution and high count rates (which leads to much lower detection limits and better precision).
Bonvin and Lemberge: No, EDXRF will not replace WDXRF in the near future for several reasons that make WDXRF the method of choice for process control applications in industries as varied as steel, aluminum, brass, bronze, glass, cement, petroleum, polymers, geochemistry and mining:
Based on the fundamental parameters of X-ray physics and the specs of an individual spectrometer, the intensity of a pure element can be derived. Thus, a sample of unknown concentrations can be determined. This is called "standardless methods" because no type standards calibration is required. Are the results calculated from "standardless methods" quantitative or qualitative?
Seyfarth: “Standardless” is a not really a correct name although marketing in the industry has embraced it: The answer to that is a little longer and will involve some formulas to prove that it is more like a universal calibration approach, which is just standardless for the user using it. Also, the word “standard” should only be used for procedures (such as ISO or EN "standards”) not reference material; thus the correct description would be “reference material free solutions,” which has no marketing sound to it.
To relate the theoretical intensities calculated by fundamental parameters (FP) with the measured intensity and to determine the instrument function, XRF needs the measurement of samples with known composition. This "base" calibration can then be used within the FP application to enable standardless quantification for the user after it has been transferred from the master spectrometer to the user’s system or the master set has been measured on that system.
FP methodology works best on samples that are close to the premise of the FP method:
Mamedov: The standardless method gives quantitative results. The standardless method is routinely used in EDXRF. Actually, all vendors provide a fundamental parameter method (FPM), which is equivalent to standardless with the software. It is possible to do correction in FPM and expand the capability of the method. Using standards will give you better accuracy. For example, using a calibration curve method accuracy may be as high as 0.4%. The standardless method will give you about 3% of accuracy. For someone working with one group of materials (for example, glasses), it is possible to generate a calibration curve that will take into account background and secondary absorption. It will give much better results.
Hinds: Regarding the use of the "standardless analysis," it is firstly a qualitative technique to let the analyst know what elements are present in the sample (within the limits of XRF). It can be used as a quantitative method depending on the analytical needs. If the end user needs a ballpark value then OK or if measured value can be within 10–20% of the true value then OK. Some samples can be determined with better accuracy depending on the type of sample, number elements in the sample, the element concentrations, and the specific "standardless analysis" program used. I strongly recommend investing some time in validating a "standardless" program by analyzing known samples and comparing the values obtained from the "standardless" program.
Bonvin and Lemberge: Results from “standardless methods” are both quantitative and qualitative.
The qualitative accuracy aspect of the analysis will depend upon the instrument, parameter and conditions, counting times, and the effectiveness of the algorithm use in the measurements. The better these factors are the better the qualitative analysis will be, especially for trace elements. If an element is not seen qualitatively, then there is no chance that it will be determined quantitatively.
To classify results as more quantitative than just qualitative will depend upon the accuracy needed for the application. The results obtained by a “standardless method” using either a peak base or a less precise scan-based algorithm will rarely be as accurate as a quantitative calibration. However, the accuracy achieved may be good enough to meet the user’s needs for an application. When the measurement is done using a scan, covering elements from fluorine to uranium, the counting time per step can only be very short (for example, 0.1 s) in order to achieve the whole acquisition within 12 min. The results obtained from such short counting time will only be semiquantitative. When the acquisition of the intensities is done by counting on specific peaks representing the various elements from boron to uranium and some well chosen background positions, the counting time per each measurement can be increased to 4 s, or 12 s for the less sensitive peaks. This improves the limit of detection and precision by factor of 7 or 11. In this case we can talk about quantitative analysis especially when the samples have the correct dimensions for presentation to the instrument (30 mm diameter and 5 mm thickness) and their surface can be well prepared and they do not suffer from morphology effects (grain size effects, shadow effects, inhomogeneity, metallurgical effects, unmeasurable elements, and so forth). As standardless programs are solely based on theory, treating the sample as homogeneous down to the atomic level, infinitely thick and having a perfectly flat/smooth surface, such effects if they are present can only be eliminated empirically by adjusting the calibration using a typical standard!
New AI-Powered Raman Spectroscopy Method Enables Rapid Drug Detection in Blood
February 10th 2025Scientists from China and Finland have developed an advanced method for detecting cardiovascular drugs in blood using surface-enhanced Raman spectroscopy (SERS) and artificial intelligence (AI). This innovative approach, which employs "molecular hooks" to selectively capture drug molecules, enables rapid and precise analysis, offering a potential advance for real-time clinical diagnostics.
Detection of Microplastics in Bottled Water Using Raman Microspectroscopy
February 10th 2025Spectroscopy sat down with Oskar Hagelskjaer, Founder and CEO of Microplastic Solution, to discuss his latest study whose findings challenge EU Directive 2020/2184 regarding microplastic detection in potable water.
Best of the Week: Interview with Juergen Popp, Microplastic Detection, Machine Learning Models
February 7th 2025Top articles published this week include a video interview that explores using label-free spectroscopic techniques for tumor classification, an interview discussing how near-infrared (NIR) spectroscopy can classify different types of horsetails, and a news article about detecting colorless microplastics (MPs) using NIR spectroscopy and machine learning (ML).
Enhancing Tumor Classification with AI and Raman: A Conversation with Juergen Popp
February 7th 2025Spectroscopy sat down with Juergen Popp of the Leibniz Institute for Photonic Technology to talk about the Photonics West Conference, as well as his work using label-free spectroscopy techniques for precise tumor margin control.