An improved PD-AsLS method for baseline estimation in EDXRF analysis.

Area of Science:

• Analytical Chemistry
• Spectroscopy

Background:

• Baseline correction is crucial in energy-dispersive X-ray fluorescence (EDXRF) analysis.
• Existing methods like AsLS, airPLS, and arPLS have parametric sensitivity issues.
• X-ray energy spectra possess unique statistical characteristics that necessitate tailored baseline correction.

Purpose of the Study:

• To propose a novel asymmetrically reweighted penalized least squares method based on the Poisson distribution (PD-AsLS) for automatic X-ray spectra baseline correction.
• To evaluate the accuracy and reliability of the PD-AsLS method.
• To compare PD-AsLS with existing classical methods.

Main Methods:

• Monte Carlo (MC) simulation to generate background spectra.
• PD-AsLS algorithm to estimate the baseline of simulated background spectra.
• Calculation of relative and absolute errors, correlation coefficient (COR), and root mean square error (RMSE) for accuracy assessment.
• Experimental validation using repeated measurements and comparison with arPLS, airPLS, and AsLS.

Main Results:

• PD-AsLS achieved a COR above 0.95 and RMSE less than 6 in simulations.
• Experimental results demonstrated PD-AsLS's stability, with a lower coefficient of variation (CV) than measured spectra.
• PD-AsLS showed superior baseline estimation compared to arPLS, without overestimation.

Conclusions:

• PD-AsLS reliably estimates X-ray spectra baselines.
• The method effectively suppresses statistical fluctuations in spectra.
• PD-AsLS offers an improved approach for baseline correction in EDXRF analysis.