Degree Name

Doctor of Philosophy


Department of Materials Engineering


Bath smelting has recently been the focus of several large research and development programs investigating alternative ironmaking processes. Fundamental studies have contributed to the understanding of smeltingreduction processes however these have not been entirely applicable to the smelting of composite pellets. Recently, bath smelting has also been evaluated for its potential for processing metallurgical wastes, the driving force being to develop suitable alternatives to traditional disposal methods such as dumping in landfills; one such technology is the EnvlRONment process. In general, many pyrometallurgical waste processes have been developed but few have been commercialised successfully. It is evident many processes have suffered economically because of poor understanding and control of process fundamentals. Although the EnvlRONment process, which smelts composite organic and ferrous waste pellets in a single vessel reactor, had been developed to the pilot plant stage, the mechanisms of smelting-reduction were not fully understood. Pilot plant trials highlighted critical aspects of the process but did not provide sufficient information to quantify process fundamentals, clearly further investigation was warranted. Smelting, Xray, and thermogravimetric experiments provided foundation for an in-depth study of reaction kinetics and mechanisms, while also allowing other relevant issues of waste processing, using bath smelting technology, to be addressed. Supporting the experimental investigations an extensive heat transfer model was developed which was used to predict temperature profiles and smelting times. The bath smelting reactor was modelled in 2 distinct stages, a falling gas region and an in-slag smelting region, heating was determined to be controlled by internal resistance. The general smelting behaviour of composite pellets was observed during pilot plant and laboratory trials. Smelting of composite pellets, as performed in the EnvlRONment process, was simulated in the laboratory by smelting composite pellets in crucibles containing slag at 1500°C; to allow for accurate experimental interpretation, and to increase the degree of experimental control, process pellets were replicated by surrogate laboratory pellets with a simplified number of components. Cellulose was trialled as a surrogate component for organic waste. Further experiments were performed using Xray fluoroscopy and radiography techniques to observe the progress of the smelting-reduction reactions in slag at 1500°C. Examination of the smelting reaction with the Xray techniques also allowed observation of the effect of composite smelting on slag foaming. The kinetics and mechanisms of reduction of composite pellets was investigated using non-isothermal thermogravimetric techniques to 1200°C. Thermal analysis via this route was selected because the experimental techniques are the most suited means to investigate temperature dependant phenomena of self-reducing systems. Smelting experiments provided significant fundamental information relevant to an understanding of smelting reactions in composite smelting. Analysis and interpretation of the experiments corroborated with the heat transfer model predictions, which was used together with the thermogravimetric results to clearly identify reaction mechanisms and explain other process phenomena. Key process reactions in composite smelting were identified as: - thermal decomposition of volatiles - direct reduction and carburisation of iron - pellet ablation through melting and dissolution - reduction of dissolved oxides in slag. Predicted temperature profiles within the composite pellets suggest that much of the volatile matter is lost during the falling gas region by thermal decomposition. For the surrogate pellets containing cellulose upto 7 8 % of the volatile matter is lost during the falling gas region. The heat transfer model predicted sensible melting times, for a 1.5cm diameter pellet a melting time of 92s was predicted, similar times were observed in smelting trials. An important finding of the heat transfer work was that although the in-slag smelting stage was based on a fully submerged object experimental observations show that an immersion crust does not form on the pellet during its initial contact with the slag bath, and that the composite pellet is not submerged during smelting. Thermogravimetry revealed the complexity of internal reduction in iron oxidecarbon systems when high and low volatility carbon sources contribute towards reduction. The contribution to reduction from the high volatility carbon being influenced by heating rate and the presence of secondary non-volatile material, measured charring levels for cellulose varied between 13 and 2 2 %. The main steps in the thermogravimetric reaction profiles were attributed to thermal decomposition of cellulose between 280 and 320°C and stepwise reduction of hematite above 550°C. The dominant reduction mechanism in composite bath smelting was identified as internal reduction within the pellet. Approximately 9 0 % of the total reduction occurs within the pellet, the extent of which is sufficient that much of the iron is also carburised before pellet ablation is complete. The other significant finding of the study was that slag foaming was shown not to be a critical issue for process control of composite smelting.