Quantitative verification of 1:35 diluted fused glass disks with 10 mg sample sizes for the wavelength-dispersive X-ray fluorescence analysis of the whole-rock major elements of precious geological specimens

Abstract

Herein, an improved wavelength-dispersive X-ray fluorescence (WD-XRF) analytical technique for determining the major elements (including SiO2, TiO2, Al2O3, TFe2O3, MnO, MgO, CaO, Na2O, K2O, and P2O5) in precious geological samples, particularly extraterrestrial basaltic rock samples, was developed. For WD-XRF analysis, undersized (approximately 11 mm in diameter) glass disks were prepared by fusing 10 mg of a powdered sample with 350 mg of lithium borate flux at a sample-to-flux ratio of 1:35. Specialized PtAu crucibles and molds were crafted for glass disk preparation to facilitate automatic fluxer mounting. This procedure enabled the automatic preparation of undersized glass disks, with repeatability between sample batches. Calibration curves were constructed for 30 certified reference materials (CRMs), including felsic, intermediate, mafic, and ultramafic igneous rocks. Regular-sized CRM disks with large sample sizes were prepared at the same sample-to-flux ratio as the test samples to avoid issues caused by CRM heterogeneity. The influence of various factors, including glass disk thickness and homogeneity, releasing agent amount, and interference between lines (Br Lα with Al Kα, Rh Lγ2 and Rh Lγ3 with K Kα, and K Kβ with Ca Kα), on analytical accuracy was investigated. Two homogeneous synthetic reference glass samples (ARM-2 and ARM-3) were pulverized and used to verify the analytical results for small samples using the WD-XRF method, which could eliminate the uncertainty caused by sample homogeneity. Subsequently, three CRMs (JB-1b, JA-3, and JR-1) were used to validate the accuracy and precision of the method. Finally, two lunar meteorite samples (NWA4898 and NWA4734) were assayed by the proposed method, and comparison with inductively coupled plasma optical emission spectroscopy results confirmed the reliability of the method. The proposed small sample protocol, which involves the destructive pretreatment of precious and limited extraterrestrial basaltic samples, has great application potential for precious samples such as lunar soil samples.