Bayesian inference: X-ray fluorescence analysis

Image

How can the constituent parts of an unidentified substance, like a meteorite, be identified? By irradiating materials with X-rays and examining the spectrum they emit, X-ray fluorescence analysis can be used to identify a variety of elements at once. Because of this, X-ray fluorescence analysis has been used to measure the concentrations of harmful heavy metals in soil. New methods that can measure large quantities or swiftly take several measurements of unidentified materials are sought because it now takes X-ray fluorescence analysis methods roughly 10 minutes to reliably identify elements. A new technique has been created by a combined research team from the Japan Atomic Energy Agency by using Bayesian estimates with X-ray fluorescence analysis. The team was able to cut the measurement time of an X-ray fluorescence spectrum in half, from seven seconds to three seconds per measurement point. This resulted in a four-second reduction in the time needed to get analysis results that were comparable to the spectra obtained after measuring a glass standard sample for an hour. For instance, depending on the sample, 10,000 measurements may be made over a short region while generating an elemental distribution. Therefore, while producing an elemental distribution, cutting the measurement time for each point by four seconds can cut the overall measurement time by 40,000 seconds, or nearly 11 hours. We have effectively combined analytical chemistry and informatics utilising X-ray fluorescence analysis and applied Bayesian inference. If it is conceivable to use this technology to find trace levels of components, more research is required. The rapid elemental analysis could be used to monitor ongoing chemical reactions as well as analyse industrial products or waste materials that are transported on conveyor belts if it can be done in a non-destructive manner without contacting the sample.