The general saying is that “pXRF instruments are precise but not accurate.” But what does that really mean?
In terms of pXRF instrumentation, precision is a gauge of how well an instrument can measure something and how repeatable that measurement is, whereas accuracy is a measure of how close the value of your result is to the true value of the sample.
To improve the accuracy of pXRF instruments we use Certified Reference Materials (CRMs), as they provide known (accurate) values to compare with our precise data. When we compare the CRM pXRF values to that of the known values, a correction factor, if necessary, can be applied to enable the pXRF data to closely correlate with the elemental composition of the certified material. Then, when that correction factor is applied to the entire dataset, we can have confidence that the values from the pXRF are a good indication of the chemical composition of our previously unknown samples. Another option is to have a custom calibration made to eliminate the need for a correction factor. The custom calibration is created with samples that have a similar chemistry and matrix to the samples being analysed, therefore producing a much closer correlation with the certified chemistry and enabling more accurate data to be collected in real-time.
Limit of detection is the lowest quantity of a substance that can be distinguished from the absence of that substance with a stated confidence level, while limit of quantification is the lowest concentration that can be quantitatively detected with a stated accuracy and precision. The detection limit of an element is determined by the length of measurement time (e.g. 1 minute per sample), the physical matrix affects (e.g. particle size and homogeneity) and chemical matrix affects (e.g. the concentration of any interfering elements). Therefore, we need to ask ourselves “at what concentration can the instrument reliably detect the presence of an element?” and “at what concentration can the instrument reliably measure the quantity of an element?”
For pXRF instruments, the detection limit provided by the instrument is typically 3 standard deviations from the mean concentration of that element. Therefore, if a pXRF reading gives a result that is less than the reported error, then the element is below the limit of detection and is not present in a detectable quantity within the sample. If a pXRF reading is more than three times the reported error, then the value is deemed quantifiable and can be used as a dependable estimation of the abundance of that element in the sample. It is important to be aware that the detection limits quoted by manufacturers for their instruments are produced by analysing a silica blank standard for a long time (usually 120 seconds) and does not provide representative detection limits for real-world samples.
Presenting the samples to the instrument in the appropriate conditions can be the difference between substandard and reliable results.
The small spot size (5 mm diameter) of the instrument means that only a small proportion of the sample will be analysed. By reducing the grain size of the sample, the inherent heterogeneity of the sample is decreased, and the analysed proportion of the sample becomes more representative of the material being analysed.
Infinite thickness is defined as the minimum thickness a sample must be in order to absorb all the x-rays of the primary x-ray beam emitted from an XRF instrument. If the infinite thickness is not met, then some of the X-ray are lost from the sample and result in underreporting of some elements. The infinite thickness is dependent the depth of x-ray penetration and is primarily determined by the density of the sample medium.
An important aspect of sampling for XRF analyses is to ensure samples are dry prior to testing. Wet samples often seriously interfere with readings, making results unreliable. The presence of moisture causes systematic under reporting of elements. This can be a result of the water molecules diffracting the incident energy, reducing the amount of energy that interacts with the sample and is subsequently detected by the pXRF instrument. Therefore, it is always recommended that samples are dried prior to analysis to obtain reliable results.
You must make sure that the window is clean and in contact with the sample for the full length of the analysis. The polypropylene window on the pXRF instrument is 4 µm thick to enable an abundance of primary x-ray beam to reach the sample and the florescence to then get back to the detector. Any additional barrier between the sample and the instrument e.g. plastic sample bags, attenuate the signal to and from the instrument causing systematic under reporting of element but can also cause elements that are present in the plastic to be added to the reported element concentrations.
Image reference: Gazley, M; Wellnitz, K (2019) X-ray transmission and pXRF: implications for the analysis of geological samples through different media. Geochemistry: Exploration, Environment, Analysis.19(4):343
Certified Reference Materials (CRM’s) are of great assistance to the QAQC (Quality Assurance and Quality Control) of pXRF data. They provide a certified check for data accuracy (as discussed above) as well as a quick way to check for low level contamination of the instrument. For users of portable X-ray fluorescence (pXRF) instruments, routine analysis of CRMs monitors the quality of your results during the sampling program and the baseline performance of the instrument. As mentioned previously, the use of CRMs allows the user to quantify the accuracy and precision of the results and determine whether post-processing is necessary.
For best practise we recommend:
• The use of multiple CRMs to assess and monitor the performance of your instrument,
• That the CRMs used contain the elements of interest to your project and cover the expected concentration range,
• A CRM is analysed every 20 to 30 samples,
• During sampling the result for an element(s) of interest in the CRM is monitored to maintain confidence in the results,
• The use a quartz blank to monitor for low-level contamination on the analyse window or in the detector.