3.1 Temperature Pattern and Physical Appearance of Reduced Briquette
As previously reported, iron nuggets were formed by reducing of ISC/coal composite pellets under an isothermal-temperature gradient profile where the initial temperature was 1000°C and the final temperature was 1380°C. An initial temperature of 1000°C was chosen based on the results of previous experiments on the reduction of titanomagnetite by coal which showed that metallic iron was formed at temperatures higher than 900°C. Other initial temperatures may have an influence on nugget formation. Therefore, the initial temperature was varied as shown in Figure 1. Pattern A was started by setting the furnace temperature at 700°C and held for 40 min at 700°C, then the furnace temperature was increased to 1380°C in 68 min with a heating rate of 10°C/min and finally held at 1380°C for 22 min. The total treatment time was 130 min. The same procedure was carried out on the other temperature patterns until the H pattern where the temperature was constant from start to the end at 1380°C. For information, the temperature intended and set in the experiments was the furnace temperature, not the temperature at the surface of the briquette with coal in alumina crucible.
Figure 2 shows the physical appearance of the reduced briquettes. The temperature pattern A to E formed iron nuggets in large numbers but with similar and smaller sizes. The F temperature pattern produced fewer nuggets, but they are larger in size. No nuggets were formed on reduced briquettes using the temperature patterns G and H. From the observation of this physical appearance, it can be concluded that the initial temperature of ISC/coal composite briquettes reduction under an isothermal-temperature gradient profile played a very important role in the formation of iron nuggets on the surface of reduced briquettes.
3.2 Nugget size and iron recovery in Nugget
The reduced briquettes were crushed using a mortar and the iron nuggets were separated from the slag as shown in Figure 3. The results of the measurement of the nugget size are shown in Figure 4a as a function of temperature pattern from two reduced briquettes for each experimental parameter. The temperature patterns G and H were not plotted in the graph because no nuggets were visible on the surface of reduced briquettes. The number of nuggets is also shown in the Figure 4a. It can be seen that the maximum nugget size of about 4 mm can be achieved using the F temperature pattern. In addition, it is also seen that only one large nugget was produced using the F temperature pattern. In the A to D temperature pattern, the average size of the nuggets was about 1 mm with a maximum size ranging from 2.2 to 2.7 mm. Approximately 90 nuggets were separated from each reduced briquette using A to D temperature pattern. These nuggets are larger than 0.3 mm. Nuggets smaller than 0.3 mm remained in the slag which can be further recovered by using for example a magnetic separator. The temperature pattern E showed a maximum nugget size of 3 mm with an average size of 1.3 mm. The size of nugget from pattern E was between the sizes of nuggets produced from temperature patterns D and F. The number of nuggets from one reduced briquette using temperature pattern E was 44. Details of the size distribution of nugget resulting from each temperature pattern are shown in Figure 4b.
The weight of the reduced briquettes, slag and iron nuggets is shown in Figure 5a. Based on these data, the iron recovery in the nuggets was calculated where the results are depicted in Figure 5b. It can be seen that iron recovery in nuggets increased from temperature pattern A with initial temperature of 700°C and achieved optimum iron recovery in temperature pattern D with initial temperature of 1000°C. Subsequently, the iron recovery decreased as shown in temperature patterns E dan F with initial temperatures of 1100°C and 1200°C, respectively.
As previously reported for solid-state reduction of ISC under isothermal conditions, at 700 or 800°C, a certain amount of magnetite in titanomagnetite was reduced to wustite but no metallic iron was formed. Although no metallic iron was formed at 700 or 800°C in the first initial stages of temperature patterns A and B, further reduction of the iron oxides in the briquettes occurred in the second stage, where heating was carried out from 700 or 800°C to 1380°C resulting in iron nuggets as shown clearly on the surface of the reduced briquettes in Figures 2a and 2b. The iron recovery in the nuggets for temperature patterns A and B was similar at about 43% (Figure 5b).
As the initial temperature increased to 900°C, metallic iron was formed with a metallization degree of about 13% attainable during the first isothermal step for duration of 40 min. Therefore, the iron recovery in nuggets increased from 43% (temperature pattern B) to 50% (temperature pattern C). At 1000°C, the metallization degree increased to about 30% during the isothermal step so that the recovery increased to 53% for the nuggets resulting from the temperature pattern D. Temperature pattern D with an initial temperature of 1000°C seems to be the optimum temperature pattern for achieving high iron recovery in the form of nuggets.
Further increase in initial temperature tended to reduce the iron recovery in the nuggets as well as the nugget formation. At an initial temperature of 1100°C, the metallization degree increased to about 50% at the first isothermal step for 40 min, but the iron recovery in the nuggets decreased. This can be caused by the formation of large amounts of metallic iron on the surface of the briquette at the first stage of isothermal which can inhibited the migration of metal from the center to the outer surface of the briquette. The results of previous study indicated that the important step for nugget formation was the second stage on the isothermal – temperature gradient profile where the temperature was increased with a certain heating rate towards 1380°C.
Further increasing the initial temperature to 1200°C reduced the porosity on the briquette surface in addition to increasing the metallization degree so that the formation of nuggets became more difficult. At initial temperatures of 1300°C and 1380°C, no nugget was formed although the metallization degree throughout the briquettes increased. Besides the initial temperature, the heating rate can also affect the formation of iron nuggets.
Geng et al. investigated the reduction of ISC pellets by embedding in coal as a reducing agent under isothermal conditions in the temperature range of 1100 to 1300°C. The reduced pellets were crushed and ground to less than 0.043 mm and the metallic iron was recovered by magnetic separator technique. It was reported that the optimum temperature was 1200°C to achieve high iron recovery and high iron grade. In the temperature range of 1250 to 1300°C, the semi molten phase was formed which prevented the reducing gas from the embedded coal entering the center of pellet [17, 21].
As mentioned earlier, the nuggets which were recovered have the size of larger than 0.3 mm. The nuggets that were less than 0.3 mm in size cannot be recovered by hand sorting where the other techniques muss be applied. The slag was analyzed by XRD where the results are shown in Figure 6 for temperature patterns A, D and F. It can be seen that the metallic iron was still dominant in the slag. Titanium was presented as titanium oxide (TiO) and armalcolite ((Mg0.5Fe0.5)Ti2O5).
The Indonesian ISC was used as experimental material by previous researchers. Gao et al.  reported that the phases in an ISC mixture with 25% coal reduced at 1250°C were metallic iron, ilmenite, and armalcolite. A similar experiment was carried out by Geng et al.  and found that the phases in pellets reduced by embedding in the coal were metallic iron, ilmenite and pseudobrookite (FeTi2O5). Recently, Zhao et al.[21, 23] reported that the phases in the reduced mixture consisting of ISC and 25% coal at 1200-1300°C were metallic iron, ilmenite, and anosovite ((Fe,Mg)Ti2O5). From the experiments mentioned above, the iron nuggets were not formed, the metallic iron was still attached to other oxides. Therefore, a series of crushing and grinding treatments are required to liberate the metallic iron.
3.3 Microstructure and SEM-EDS Observation
The surface of briquette that was reduced under temperature pattern A was analyzed by SEM-EDS as shown in Figure 7. Many nuggets with a size of less than 20 µm, called micro nuggets, were visible on the surface of briquette. These micro nuggets have not been agglomerated to form larger nuggets. These micro nuggets were impossible to separate from the slag by hand sorting and remained in the slag as shown by XRD pattern in Figure 6 where the metallic iron was the dominant phase. Other separation technique is required, e.g., a magnetic separator to recover these micro nuggets. In addition, a nugget with a size of 0.33 mm was clearly visible on the surface of briquette. This nugget size may still be difficult to separate from the slag manually by hand sorting as shown in Figure 4a for the minimum size of the nuggets that was counted for iron recovery. On the surface of this nugget, “flower”-like motifs with light gray color revealed. Inside this nugget, the shape like micro nuggets can be seen clearly where the micro nuggets were covered by a transparent layer. A similar phenomenon was observed in other nuggets as shown in Figure 8 where the agglomerated micro nuggets were covered by a transparent layer forming a large nugget size.
SEM-EDS analysis on the surface of nuggets which were produced from temperature pattern A, D and F are shown in Figure 9. From Figure 9a, it is clear that the formation of “flower”-like motifs on the surface of nuggets in Figure 7 was caused by the presence of manganese together with sulphur, and iron form a sulphide phase which had a lower melting point. Manganese came from ISC as listed in Table 2 and sulfur from coal as the sulfur content in the ISC is very low (0.005%). The combination of iron, manganese and sulfur may form FeS⋅2MnS with has a melting point of 1128°C that solidified last on the surface of nuggets. The other phases are oxides containing aluminum, magnesium, and iron. Figure 9b and 9c show the oxides were formed on the surface of nugget containing iron, titanium, aluminum, and oxygen.
The microstructure images from optical microscopy of the bisection of nuggets resulting from temperature patterns A, D and F are shown in Figure 11a – 11c while of the polished surface of reduced briquette is shown in Figure 11d. Grain boundaries are clearly visible in Figure 11a with a magnification of 200 ⋅ which may be formed from the agglomeration of small metal particles with one another. It was believed that the metal was not completely melted, the metal phase was semi liquid. During agglomeration, the unreduced oxides were trapped between one metal particle and another. Grain boundaries and trapped oxides as well as sulfide precipitations at the grain boundaries can be seen clearly in Figure 11c. Figure 11d shows metallic iron formed on the surface of the reduced briquettes. To recover the metallic iron with a magnetic separator, a comminution step consisting of crushing and grinding is required to liberate the metallic iron particles from the unreduced oxides. This technique was reported by pervious researchers. [13, 17, 21, 23]