**2. Experimental Sample and Procedure**

To simulate the test particles, the samples were prepared using alumina powder and woody biomass powder. The woody biomass powder used in this study was commercial mangrove charcoal, which is normally used for barbecues. To simplify the experimental conditions, alumina powder was prepared as a substitute for iron ore. Alumina eliminated the effects of melt formation, reduction and re-oxidation of iron ore during coke combustion. Charcoal powder with a particle diameter of −125 μm and 125~250 μm was used in this experiment. The particle size of iron ore was to simulate the adhere powder layer, but not the coke particle. Coke powder with the same particle diameter was used to compare the results. Alumina powder with a particle diameter of −250 μm was prepared to match the particle size of the iron ore. The analysis of the results of prepared carbon material are listed in Table 1. Compared with coke, charcoal has a lower ash ratio and higher volatile matter content and therefore charcoal has a lower fixed carbon. The surface area of charcoal is higher than that of coke. This was also observed by an SEM, as shown in Figure 1. Figure 2 shows the overall view of raw materials and samples. After the alumina and coke powders were mixed, 0.5 mass % flour was mixed as a binder. The flour was just used to enable keeping the tablet shape until it was inserted to the platinum basket. It was thought that the effect of the flour could be negligible, because flour evaporates at a lower temperature—600 K—than the experimental temperature in this study. Then, the mixture was pressed into 10 mm diameter tablets by stainless dies. The height of the tablet was 10 mm with a void ratio of 35%. The void ratio was decided from the information on the tablet volume and the true density of the sample mixture. Each true density of the sample materials was measured by pycnometer. The weight ratio of coke in each sample was fixed at 20 mass %. To ensure that the volume ratio of the samples was the same as the hematite and alumina samples, 22.1 mass % coke with 77.9 mass % alumina and 20.8 mass % charcoal with 79.2 mass % alumina was also prepared.


**Figure 1.** Scanning electron micrograph of coke and charcoal (−125 μm).

**Figure 2.** Overall view of the raw materials and the sample.

The measurement of the sample weight loss during coke combustion was done by the thermobalance shown in Figure 3. The sample was placed in a platinum basket. A vertical electric resistance furnace was used to do isothermal heating. The isothermal zone was heated up to 1073 K, 1223 K, 1373 K and 1523 K. Before the combustion experiment, heat treatment of the samples was carried out at each of the given temperatures in a N2 atmosphere for 30 mins to remove water, Volatile matter (V.M.) and the binder from the samples. Then, air was passed through the reaction tube. The air flow rate was 4 NL/min. When a weight change in the sample was not observed, the experiment was terminated. It was hypothesized that coke ash did not influence the weight loss of the sample because the amount of coke in every sample stayed the same.

**Figure 3.** Schematic of the device used in the experiment.
