Coefficient corrections for portable X-ray fluorescence data of the Niton XL3t No. 97390 (coefcor I-IV) developed according to the Munich procedure

Portable X-ray fluorescence (p-XRF) devices are commonly utilized to analyze the chemical composition of various materials, such as archaeological pottery and siliceous substances. The discussion regarding the suitability of this method for such samples is ongoing, as the data are often said to be quantitatively unreliable. Nevertheless, the development of coefficient corrections (coefcors) offers a means to transparently demonstrate the quality and comparability of p-XRF data. These coefcors are established by comparing p-XRF measurements with laboratory methods derived from the same sample set. The Frankfurt Keramiklabor (Frankfurt Procedure) procedure serves as the basis for the Munich Procedure, which outlines criteria for assessing the quality of linear regressions applied to p-XRF data. The Munich Procedure provides clear benchmarks for the coefficient of determination (r²) and the relative standard error of the estimate (rSEE). Additionally, it includes a robustness test using a bootstrap method, as detailed in the R-scripts of this article. The Munich Procedure is applied to datasets generated using the Niton XL3t No. 97390 instrument used by the Dept. of Culture Sciences, Ludwig-Maximilians-Universität München (LMU), between 2017 and 2023. Measurements were conducted in the TestAllGeo mode, using an 8 mm collimator and a 300-second measurement time in air at Goethe-Universität Frankfurt. Data processing took place at LMU. Four datasets were utilized to develop four coefcors for comparing p-XRF and WD-XRF values of the Frankfurt pottery standard set (coefcor I-IV). Additionally, five coefcors were created to adjust p-XRF data from the same instrument before and after an instrument reset (coefcor ItoII, ItoIII, IItoIII, IIItoII, IVtoIII). The instrument provides analytical data, which is used to generate the dataset for developing coefficients using R-scripts. Processed data, including calculated criteria and factors (slope and intercept) for each chemical element, are presented with relevant graphical output. These coefcor factors can be utilized to correct empirical data obtained with Niton No. 97390 by selecting the appropriate coefcor based on the measurement date. This process is also outlined in the R-scripts. To assist users in applying the Munich Procedure to their own data, an R-script is provided that offers a more detailed explanation of the calculations. While the coefcors given here are specific to the instrument and therefore provide a basis for understanding the data quality of (as yet unpublished) studies using the Niton No. 97390, the method itself can be universally applied to develop coefcors for other p-XRF instruments.