Queensland Alumina Limited (QAL) is the world’s largest alumina refinery and is located in the town of Gladstone, Queensland. QAL face problems associated with the presence of silica (SiO2) in their Bayer liquor circuit, due to its involvement in the formation of scale. Accurate monitoring of silica concentrations is critical to predicting scaling rates and maintaining the concentration of silica within product alumina specification. Inductively Couple Plasma – Atomic Emission Spectrometry (ICP-AES) is currently used to measure the liquor silica concentrations and this study aimed to extend the applicability of the current QAL ICP Si Analysis Method to Bayer liquor concentrations not typically seen. Liquor concentrations composed of up to 10g/L silica, 500g/L caustic and 160g/L were under investigation.
The experimental methodology focused primarily on identifying and if possible minimising matrix and spectral interferences on Si, caused by the instrument operating conditions or sample matrix compositions (predominantly Na and Al). Plasma optimisation, background interference investigations, analyte line and internal standard selection, interference/trend analysis, and calibration set comparisons were all incorporated into the experimental methodology.
The results from this study have validated and supported the continued use of many aspects of the current QAL Si Analysis ICP Method, at the extended Bayer liquor concentration ranges. Major findings from the study include:
• The optimum plasma settings were RF power 1450W; Plasma flow rate 15L/min; Auxiliary and Nebuliser flow rates 0.7L/min. Settings however did not produce ideal “robust” plasma conditions, suggesting a problem with the ICP hardware.
• The Si analyte lines 251nm and 288nm are the best for Bayer liquor analysis
• Lutetium (Lu) is ideal as an internal standard for Si correction
• The 1/200 dilution ratio used for Bayer liquors is acceptable
• No significant trends or interferences were witnessed for Si measurements over the entire Bayer liquor composition ranges of interest.
• Accuracy of Si analysis was slightly reduced at low concentrations (<2g/L) Matrix matched and non-matrix matched calibration sets both work well for Si analysis. There is some ambiguity at low concentrations however.
Recommendations emerging from the findings of the experimental work have been made to QAL. The major recommendations are:
• The current QAL ICP Si Analysis Method will be applicable to Si analysis of Bayer liquor samples over the entire composition ranges of interest. The Si concentrations in the calibration set will only need to be adjusted to cover the expected Si sample concentration range.
• Report Si concentrations measured at the analyte lines 251nm and 288nm individually and not as an average. Averaging serves no beneficial purpose and may in fact mask inaccurate measurements.
• Carry out plasma optimisation tests again once the axial and radial viewing windows and mirrors have been cleaned.
• Use an In-House Reference Standard Si solution to assess ICP operating conditions and to prove reliability of Si analysis for future research projects.
The work from this project has also uncovered a number of areas that may be worth pursuing further. The major areas for future work should focus on:
• Low concentration (<2mg/L) Si analysis
• Development of the In-House Reference Standard Si solution
• Investigation of Bayer liquor stability and the effects of delayed analysis time on Si analysis