Complete phase composition and simultaneous determination of bath parameters with a sample throughput time of less than a minute on a benchtop diffractometer!
The production of aluminium from purified alumina by electrolysis (Hall-Héroult process) is a very energy-consuming process. Frequent monitoring of the electrolytic bath parameters combined with short measurement times and high sample throughput is important to ensure minimum costs for smelter operations. Over the years a number of analytical techniques, including wet chemistry, X-ray diffraction (XRD) and X-ray fluorescence (XRF) have been used to monitor electrolytical bath conditions.
Wet chemistry is the most accurate and at the same time the most time-consuming method, which is therefore not suitable for routine process control in an aluminium smelter. However, it is used as a reference method. XRF analysis is faster, but its accuracy strongly depends on the sample preparation. XRD, on the other hand, enables fast and accurate feedback and is now regarded as a standard tool for process control in the alumininum industry.
Traditionally, monitoring of the electrolytic baths’ chemistry relied on calibration-based single peak methods or on more advanced full-pattern techniques (Rietveld method)1,2, providing only the information about the crystalline phases. However, depending on the cooling rate, a certain amount of alumina and Ca-containing phases can be semi-crystalline or amorphous, which complicates the analysis and reduces its accuracy.
Recently Uwe König (Global Mining Segment Manager at Malvern Panalytical) and Nicholas Norberg (Application Specialist at Malvern Panalytical) have evaluated a new approach to the process control in an aluminium smelter. They presented their results during the latest meeting of the International Committee for Study of Bauxite, Alumina & Aluminium (ICSOBA) held this year in Hamburg, Germany. The method combines the benefits of the established full-pattern Rietveld refinement with partial least squares regression (PLSR), a well-proven statistical method3,4, and it simultaneously determines the phase composition and important process parameters (e.g. exAlF3, CaF2, total Al2O3) of an electrolytic bath using the same XRD pattern measured within 48 seconds on an Aeris Metals edition benchtop diffractometer.
- R. Feret, Breakthrough in Analysis of Electrolytic Bath Using Rietveld (2008) XRD Method, Light metals, 343-346.
- R. Feret, Selected applications of Rietveld-XRD analysis for raw materials of the aluminium industry (2013) Powder. Diffr. 28 (2), 112-123.
- de Jong, SIMPLSR: an alternative approach to partial least squares regression (2013) Chemom. Intell. Lab. Syst. 18, 251-263.
- Degen et al., HighScore Suite (2014) Powder. Diffr. 29(S2) S13-S18.