Thesis (MTech (Chemical Engineering))--Cape Peninsula University of Technology, 2006 / Better metallurgical accounting practices are achieved through implementation of robust and accurate analytical techniques, in particular the sample preparation techniques involved, as well as characterization of all possible errors associated with such techniques and those associated with the characteristics of the materials to be analyzed. As a contribution to the AMIRA P754 project which aims at developing standards and tools for metal balancing, reconciliation and reporting from mine resource to final product, this thesis presents best practices for the characterization of ferrochrome smelter slag. The characterization ofthe slag is also essential for process control, and thus its implications to the performance of the smelting process are also discussed in this thesis. Slag samples from a ferrochrome smelter were analyzed using an XRF powder pellet analytical technique in contrast to the rep technique used at the plant laboratory, to determine their composition. This was done to evaluate the possibility of using XRF as an alternative technique to improve the turnaround times at the plant laboratory. It was found that the XRF analysis of the composite slag by pressed powder pellets could be performed rapidly, but because of the grain size effects resulted from the entrained alloy particles which could not pulverize well, the results were not accurate. This then led to the introduction of an extra stage (oxidation of the entrained alloy particles) in the sample preparation procedure, which prolonged the turnaround time by about 16 hours. However, an important achievement was that the entrained alloy particles could be oxidized into their oxides to obtain homogeneous samples and therefore representative sub-samples. This is an important improvement in the practices at the plant laboratory which involved excluding some of the entrained alloy particles and thus analyzing biased, non-representative samples. A rapid estimation of the percentage of the entrained alloy in slag was proven to be possible using a simple model (which relates density and the mass fraction of the entrained alloy), in conjunction with a density measuring device (a pycnometer). This ability serves as a significant improvement in the mass balancing and metallurgical accounting practices at ferrochrome producers. This thesis also describes the manner in which heterogeneity can be quantified as well as the errors associated with heterogeneity and their implications to the analysis results. Heterogeneity in ferrochrome slag has been found to depend significantly on the amount of alloy entrainment in slag, and therefore the slag viscosity, tapping temperature and chemical composition. One can therefore conclude that the sampling constant and therefore sample mass is dependent on furnace operating conditions.