% Calibration of element concentrations determined using a portable XRF % analyser. % % The calibration is based on linear regression models fitted through % semi-quantitative pXRF data and corresponding high-accuracy ICP-MS data. % Measurements were conducted on a wide range of marine sediment samples % collected in the South Indian Ocean and Tasman Sea. Each sample was % analysed both using a portable XRF analyser (Olympus Vanta Series) and an % inductively-coupled plasma mass spectrometer (Thermo Fischer Scientific % Element 2). The calibration data is stored in the "CalData" folder. In % order to improve the accuracy of the calibration (especially of elements % present in low quantities in the samples the calibration is currently % based on) the calibration data sets should be extended with corresponding % pXRF and ICP-MS data of different sediment types. % % As pXRF and ICP-MS target slightly different sets of elements, two % different calibrations are being performed. % % Element-specific calibration: % For each element that was analysed both by pXRF and ICP-MS, an element- % specific calibration (ESC) line is being determined by fitting a linear % regression model through the corresponding pXRF and ICP-MS element % concentrations. The pXRF concentrations are then corrected for slope (m) % and y-intercept (b) of the corresponding ESC. % % General calibration: % Where an element was analysed only by pXRF, a general calibration (GC) is % applied. The GC line is based on a linear regression model fitted through % the entire data set of corresponding pXRF and ICP-MS concentrations (i.e. % an average of all ESCs). The pXRF concentrations are then corrected for % slope and y-intercept of the GC. % % PXRF analyses are conducted in two different modes, "geoChem" and "Soil", % each targeting slightly different sets of elements. Therefore, % ESC and GC are calculated separately for each pXRF mode. % % How to use the script: % - Replace input data (pXRF concentrations) in the "input" folder. Make % sure to leave the structure of the excel file as is. Replace only the % the data (concentrations and, optionally, the sample-IDs, remove or add % rows if necessary). Make sure, that the pasted concentrations align % with the concentrations of the input file (row 1). % - Set Matlab working directory to corresponding path of the "XRFcal" % folder % - Run script by typing "XRFcal" into the Command Window (now input % variables required % - When prompted, select pXRF data to be calibrated (geoChem, Soil or % both) % - Calibration output will be stored in a date/time subfolder within the % the "output" folder. % % Calibration Info: % GC and ESC coefficients (slope, y-intercept and R2 error) can be found % in the "Calibration_Info" subfolder (CalibrationInfo.txt). % "CalibrationFigure_GEOCHEM" and "CalibrationFigure_SOIL" display % regression plots and lines (ICP-MS vs. pXRF concentrations; green plots - % R2>.9, orange - .5<R2<.9, red - R2<.5).
JakeWeis/XRFcal
Folders and files
| Name | Name | Last commit date | ||
|---|---|---|---|---|