X-ray Fluorescence Analysis of Feldspars and Silicate Glass: Effects of Melting Time on Fused Bead Consistency and Volatilisation

Abstract

Reproducible preparation of lithium tetraborate fused beads for XRF analysis of glass and mineral samples is of paramount importance for analytical repeatability. However, as with all glass melting processes, losses due to volatilisation must be taken into account and their effects are not negligible. Here the effects of fused bead melting time have been studied for four Certified Reference Materials (CRM’s: three feldspars, one silicate glass), in terms of their effects on analytical variability and volatilisation losses arising from fused bead preparation. At melting temperatures of 1065 °C, and for feldspar samples, fused bead melting times shorter than approximately 25 min generally gave rise to a greater deviation of the XRF-analysed composition from the certified composition. This variation might be due to incomplete fusion and/or fused bead inhomogeneity but further research is needed. In contrast, the shortest fused bead melting time for the silicate glass CRM gave an XRF-analysed composition closer to the certified values than longer melting times. This may suggest a faster rate of glass-in-glass dissolution and homogenization during fused bead preparation. For all samples, longer melting times gave rise to greater volatilisation losses (including sulphates and halides) during fusion. This was demonstrated by a linear relationship between SO3 mass loss and time1/2, as predicted by a simple diffusion-based model. Iodine volatilisation displays a more complex relationship, suggestive of diffusion plus additional mechanisms. This conclusion may have implications for vitrification of iodine-bearing radioactive wastes. Our research demonstrates that the nature of the sample material impacts on the most appropriate fusion times. For feldspars no less than ~25 min and no more than ~60 min of fusion at 1065 °C, using Li2B4O7 as the fusion medium and in the context of feldspar samples and the automatic fusion equipment used here, strikes an acceptable (albeit non-ideal) balance between the competing factors of fused bead quality, analytical consistency and mitigating volatilisation losses. Conversely, for the silicate glass sample, shorter fusion times of less than ~30 min under the same conditions provided more accurate analyses whilst limiting volatile losses.

Alternative raw material research for decarbonization of UK glass manufacture

“…The XRF data were analyzed using a modified version of the OXI program. 62 A total carbon elemental assay was conducted using a LECO CS844ES combustion furnace. The analysis was carried out by an ISO 17025 (UKAS) accredited testing facility, and the instrument was calibrated using appropriate certified reference materials.…”

“…Uncertainties associated with the XRF analysis are estimated at ±2 % of the measured concentrations. The XRF data were analyzed using a modified version of the OXI program 62 …”

Open-cell mullite ceramic foams derived from porous geopolymer precursors with tailored porosity

“…A continuous weight loss trend was observed for the control sample without the ion-exchange treatment above 600 . The weight loss phenomenon could be ℃ explained by the volatilization of low-melting-point metal oxides [90][91][92][93] and their compounds [93][94][95] at high temperatures. This loss was also detected for the starting materials (bauxite [96] and FA [13]), and the porous structure for the samples without the ion exchange sintered at 1300 could not be maintained ℃ due to the formation of a glass phase (vitrification).…”

Biomass ashes as potential raw materials for mineral wool manufacture: initial studies of glass structure and chemistry

“…The XRF spectra were analysed using a version of the OXI programme, 39 a Wide Range Oxide XRF programme developed and modified in-house. Details of the included oxides and calibration ranges can be found in Bell et al 40 Uncertainties associated with each measurement were estimated to be related to the maximum of the relative standard deviation of triplicate measurements of a standard reference material analysed using the same instrument and software. 40 Based on these values, oxides present in amounts > 10 wt.% have a Relative Standard Deviation (RSD) of ~0•2 %, oxides present in amounts between 1 and 10 wt.…”

“…Details of the included oxides and calibration ranges can be found in Bell et al 40 Uncertainties associated with each measurement were estimated to be related to the maximum of the relative standard deviation of triplicate measurements of a standard reference material analysed using the same instrument and software. 40 Based on these values, oxides present in amounts > 10 wt.% have a Relative Standard Deviation (RSD) of ~0•2 %, oxides present in amounts between 1 and 10 wt. % have an RSD of ~ 1 %, oxides between 0•1 and 1 wt.% have an RSD of 5 %, and oxides present below 0•1 wt% have an RSD of 10 %.…”