Effect of Coke Reactivity upon Coke Strength with Focus on Microstructure

The primary aim of this project was to examine changes in the microstructure of metallurgical coke during high-temperature reaction with CO₂.  Additionally, the project aimed to relate the changes in microstructure to the quality of coke, in terms of its reactivity and strength.  In order for the work to be relevant to industry, the reaction process mimicked, as closely as possible, the standard CSR test.  CSR-sized samples of five cokes prepared from single coals coked in a pilot coke oven were obtained. The coals were selected to cover a range of measures: rank, vitrinite and CSR. CT scanning at the Australian Synchrotron was used to produce three dimensional images of microstructure of three unreacted lumps from each of the cokes at a pixel size of 8.9 micron. The same lumps were then imaged using CT after 30, 60, 90 and 120 min reaction with carbon dioxide under simulated CSR test conditions and the images were then compared. In this way the evolution of the coke structure during reaction with CO₂ at 1100^oC could be followed. After 120 min reaction a selection of samples was tumbled in an I-drum and the remaining core of each sample was imaged to identify where the structure failed.

After reaction, some cokes showed gradients in density near the particle surface, which is attributed to gas transport (pore diffusion) being significant in controlling the reaction rate. There were some cokes where gas transport did not appear to be the limiting process; these cokes appeared to react more uniformly. With the small number of cokes so far tested it was not possible to link this behaviour back to coal properties.

A method was developed for analysing the stress distribution in particles put under load, by comparing the density of high stress points (which we call the critical stress index, CSI) in the outer shell of the coke lumps compared with the bulk. Once again different cokes showed different developments of relative CSI between the shell and the interior, suggesting that that for some cokes the loss in strength was restricted to surface regions whereas for others it wasn't. Cokes that exhibit significant gradients in coke density near the surface (due to reaction) show increased CSI at the surface and hence are expected to be more susceptible to surface damage during the tumbling process, which would decrease the CSR value for that coke. Thus, our study seems to indicate that the cokes with the best CSR have a balance between reactivity and internal pore diffusion.  

It is clear from the images that inerts play an important role in the reactivity.  However, the response of the inert-derived material to CO₂ attack varied quite markedly. Some inerts completely disappeared due to reaction, while others only partially reacted.  Some inerts reacted from the outside in (presumably due to low internal porosity), while others reacted throughout the inert.

A similar series of experiments were performed at 900^oC to see if differences due to different rate limiting processes could be detected in the images. Unfortunately the weight loss was so low (1-8%) that it was difficult to detect changes in individual lumps within the time available.