The Degree of Reduction (DOR) and Degree of Metallization (DOM) estimated and presented in this article are tabulated in Table IV.
Figure 3 depicts a correlation between DOR and DOM, where a stoichiometric line (shown as dotted line) correlating DOR and DOM has also been superimposed on it. Experimental line (shown as the solid line) represents the best fit line through experimental data. A hypothetical data at (100,100) has been added to force the line to pass through the (100,100) point. It is to be noted that the line indeed very closely passes through the (100,100) point with a correlation coefficient at 92 pct. The stoichiometric line represents a situation when the sample undergoes uniform and isothermal reduction. But the reduction of pellets is non-isothermal and non-uniform across its radius. Since wustite is the lowest order among the possible iron oxide phases that contain 70 pct oxygen, stoichiometrically beyond 30 pct DOR, metallic iron will form. The experimental line is found to intercept the percent degree of reduction line much below 30 pct on extrapolation. This may be attributed to inhomogeneous reduction, i.e., when the surface of the pellet is metallized, the interior may not. Thus, metallic iron may form when the average DOR is even less than 30 pct. This clearly indicates that the reduction process in the composite pellet is non-isothermal and heat and mass transfer controlled especially in multi-layer bed RHF.
Optimum C/Fe2O3 molar ratio
Experiments are conducted using pellets containing different C/Fe2O3 molar ratios to optimize the internal carbon content of pellets for maximum reduction efficiency. C/Fe2O3 molar ratio is varied from 1.33 to maximum of 3, the stoichiometric carbon required for direct reduction of hematite to iron. The dried pellets are reduced in three-layered bed at RHF at 1523 K (1250 °C) for 20 minutes. It may be noted that the pellets stay in the preheating zones at 973 K, 1273 K (700 °C, 1000 °C) for another 10 minutes each. However, since carbon gasification is sluggish below 1273 K (1000 °C), in situ generation of CO will not be significant below 1273 K (1000 °C) leading to a poor kinetics. Besides, the pellet will take some time to preheat to 1273 K (1000 °C) when the reduction becomes significant. Therefore, it is likely that when the pellet enters the reduction zone after traversing through the preheat zone at 1273 K (1000 °C), it will attend 1273 K (1000 °C) throughout and be ready for reduction. Therefore, only partial reduction of higher oxide phases like hematite, magnetite that requires less CO might take place in the preheating/pre-reduction zone. But majority of oxygen will be removed in the reduction zone. Therefore, the credit has been given to the reduction zone. The layer-wise variation of the degree of reduction with the change in the C/Fe2O3 molar ratio in pellets is shown in Figure 4.
It is observed that the degree of reduction first increases with increase in C/Fe2O3 molar ratio and reaches its maximum at the ratio of 1.66 and then extent of reduction decreases with further increase in molar ratio beyond 1.66. Thus, there exists an optimum carbon level inside the pellet at which reduction becomes maximum and after that it decreases significantly with increase in the carbon inside the pellet. It is further interesting to note that this optimum level of carbon is much below that is required stoichiometrically to reduce hematite by direct reduction. These observations are evident in all three layers in the pellet bed. Many researchers have reported that the degree of reduction increases with the increase in internal carbon content even above the stoichiometric value in small mass systems like small single pellet, mixture, etc.[1–7] In such small mass systems, the effect of heat and mass transfer is not very significant. The present experiment demonstrates that carbon even less than the stoichiometric value for direct reduction of hematite produced higher extent of reduction in multi-layer bed RHF. Lower carbon restricts carbon gasification and promotes indirect reduction by utilizing the CO produced by direct reduction that tends to shift the final gaseous product from CO to CO, CO2 mixture and therefore making the process more carbon efficient. In such case, CO will be utilized partially and in the extreme case completely following the overall reaction when the product is CO2: Fe2O3 + 1.5C = 2Fe + 1.5 CO2 and can attain lowest stoichiometric C/Fe2O3 molar ratio of 1.5. Our experiment also indicates an optimum ratio at 1.66 which is very close to 1.5, indicating significant CO utilization could be achieved at lower carbon level in the pellet that is much below the stoichiometric carbon required for direct reduction of hematite.
Effect of coal content on the porosity of the pellets
As the carbon content is increased, the volatile matter associated with the pellet is also increased. These volatile matters get escaped at around 773 K (500 °C), which produces porosity and formation of the crack in the pellet. The porosity of reduced pellets from top layer with the variation in carbon content is shown in Figure 5.
The internal porosity of the reduced pellet increased significantly with the increase in carbon content. The photographs of reduced pellets from the top layer with the variation in carbon content are shown in Figure 6. These figures distinctly demonstrate that cracks in the pellet increase with increase in the amount of carbon inside the pellet; the fact that is also reported in the literature.
The presence of cracks makes the pellet more open to atmosphere and allows product gases (CO and CO2) to pass away easily to the atmosphere without further participation in the reduction process, which may be attributed for the lower level of reduction at the top layer for high carbon level in the pellet. Besides, the formation of cracks disturbs the physical integrity of the pellet and subsequently hampers the internal heat transfer. Furthermore, increase in porosity reduces the thermal conductivity of the pellet significantly, thereby affecting the heat transfer in the pellet and restricting the endothermic reduction process. Besides, higher carbon promotes more endothermic carbon gasification reaction and consequently makes heat deficiency in the pellet.
From Figure 4, it is further observed that relative extent of reduction is the highest in the top layer followed by the bottom layer and it is the lowest for the middle layer in the case of pellets with carbon-to-hematite molar ratio less than 2.33. General conviction suggests that top layer will be reduced more as it directly receives the heat from the top, while the bottom will reduce least as it receives heat from top indirectly being shielded by middle and top layers. However since the bottom layer stands on conducting crucible separated by a single alumina layer, heat is also received from the bottom wall of crucible by radiation. The middle layer receives the minimum heat being shielded from the heat sources at the top and bottom. Furthermore, it is also observed that in the case of the composite pellet with carbon level greater than C/Fe2O3 of 2.33, the bottom layer showed higher degree of reduction as compared to the top layer and middle layer being the lowest as usual. Huang et al. also indicated higher extent of reduction at the lower layer compared to top layer in a tall bed hearth furnace.
Such switch from top to bottom layer for higher reduction might be attributed as follows:
As discussed in Section III–B, it is found that the top surface is cracked with high carbon-containing pellet that allows CO, CO2 gases to pass to the atmosphere easily without further participation in the reaction. This reduces the extent of reduction at the top layer.
The volatiles in coal are released at around 773 K (500 °C), while reduction of iron ore takes place at much higher temperature (say 1173 K (900 °C) and above). Therefore, unless the escape of volatile is restricted, it is more likely that the volatile will escape without participating in the reduction reactions. In a top-heated furnace, there exists an unfavorable temperature gradient for the escape of volatile generated at the bottom of the furnace that is comparatively cooler and heavier. Therefore, in the bed, the volatile generated at the bottom would like to reside for longer duration and participate in the reduction reaction that might lead to the higher extent of reduction at the bottom especially for pellets containing high carbon with comparatively lower thermal conductivity.
Another reason may be because temperature gradient within the pellet at the bottom of the bed will be higher compared to that at the top especially for high carbon-containing pellets. Thus, it might happen that when the volatile is released from the center of the pellet, the surface has attained the reduction temperature. This leads to participation of volatile in the reduction process at the bottom of the bed and consequently higher extent of reduction. Besides, the hydrogen of volatile is more efficient than CO both thermodynamically and kinetics wise, and therefore under favorable condition, volatile will react faster than carbon.
The layer-wise shrinkage of the reduced pellets for different C/Fe2O3 molar ratios is shown in Figure 7.
The maximum shrinkage is observed with the C/Fe2O3 molar ratio of 1.66 for all three layers, which is consistent with the maximum reduction obtained for this carbon level. This may be attributed for the following three reasons: (1) Higher the DOR, higher is the gas removal from the pellet that favors shrinkage. (2) At low carbon level, the conductivity of the pellets is higher compared to the pellets containing higher carbon because conductivity of coal (~0.2 W/m-K) is significantly less than hematite (~18 W/m-K). Thus, the low carbon-containing pellet will conduct heat comparatively faster and attain higher temperature quickly compared to pellets containing large amount coal. Higher temperature favors sintering and shrinkage. (3) Again at low carbon-containing pellets, formation of liquid fayalite is more that wets pores and finally shrinks during solidification, leading to shrinkage.
Phase analysis through XRD
The layer-wise phase analysis of the reduced pellets by XRD for carbon-to-hematite molar ratio of 1.33 is shown in Figure 8(a). Figures 8(b) and (c) depicts the similar figure for the carbon-to-hematite molar ratio of 1.66 and 3, respectively.
It is observed that at lower carbon content in the pellet, wustite (W) with some hercynite (h, FeAl2O4) are observed along with hematite (H), magnetite (M), fayalite (F, Fe2SiO4) in all layers in different proportions. In the case with higher carbon content (C/Fe2O3 ratio =3), there is no sign of wustite and hercynite. Also, fayalite is observed at the top layer only. It may be further noted that no residual carbon peak is observed which may be attributed due to its low amount and amorphous nature. The relative amount of various phases present in the reduced sample has been analyzed by Rietveld analysis using TOPAS software as presented in Table V.
In order to check the reliability of such phase quantification by Rietveld analysis, oxygen mass balance has been carried out based on the estimated phase constituents by Rietveld analysis, and the degree of reduction is calculated. The DOR calculated by Rietveld analysis closely matches with the experimental DOR with an error of ±7 pct. The oxygen balance also includes alumina and silica in the sample both before and after reduction. From Table V, it may be observed that the amounts of fayalite formed in the top layer of the reduced pellet in the case with C/Fe2O3 of 1.33 and 1.66 are 18.2 and 22.4, respectively, which are comparatively higher compared to 5.3 for the said ratio of 3. Fayalite is liquid at the reduction temperature that blocks the gas passage. It also reduces the activity of FeO in fayalite that makes it difficult to be reduced by CO. Also, there is a possibility that the carbon particles might get entrapped in the liquid slag phase and become difficult to be accessed for its gasification as well as for reduction resulting in lower DOR and DOM. It may be noted that in no case, the DOR exceeds 80 pct, which may be attributed to the liquid fayalite formation that halts the DOR beyond a certain limit. Some proper additives that restrict fayalite formation could be the scope for any future study. It is also observed that the amount of FeO formed in the bottom layer is comparatively lower as compared to top layer for the same degree of reduction. Also for the same degree of reduction, the amount of iron formed in the bottom layer is comparatively higher than that at the top layer. During cooling, since the top layer is directly exposed to the atmosphere, it gets cooled faster, whereas the lower layers cool at slower rate being shielded from open atmosphere at the top. During slow cooling, FeO decomposes as: 4FeO = Fe3O4 + Fe that results in higher iron and lower FeO content in the lower layers compared to the top layer. Effect of such decomposition is also observed in Figure 3, where the data from middle and bottom layer were found to deviate away from the stoichiometric line to the data from top layer. The increase of magnetite at lower layers might also be attributed to such decomposition.
In order to compare the process efficiency of different pellets with varying carbon contents, the following performance parameters have been calculated and compared. These parameters are the degree of reduction, metallization, carbon efficiency, and productivity. Carbon efficiency is defined as the stoichiometric carbon required to the actual carbon used in the pellet. The productivity is defined as the amount of metallic iron produced per unit area of hearth per unit time. The performance analysis with different C/Fe2O3 molar ratios is shown in Table VI.
From Table VI, it is observed that the carbon efficiency is higher for lower C/Fe2O3 ratio and is above 100 pct. 100 pct efficiency indicates that the product gas is fully CO and O/C molar ratio is 1 and zero utilization of CO that has the potential to pick up one more ore oxygen atom. Carbon efficiency above 100 pct indicates better CO utilization, higher O/C in the product gas and lower use of carbon than stoichiometric carbon required for hematite reduction by direct reduction (C/Fe2O3 = 3). It is further noted that the carbon efficiency is higher in top layer compared to lower layers for lower carbon level (C/Fe2O3 ratio <2.33) and is maximum in bottom layer for high carbon levels (C/Fe2O3 ratio >2.33), which is consistent with the observation of reduction efficiency for these carbon levels. The productivity is maximum (about 52 kg metallic iron/h-m2) for C/Fe2O3 molar ratio of 1.66 which is the optimum carbon level for the maximum degree of reduction in the present experimental condition. Productivity is also found to be higher at the bottom layer as compared to other layers for all carbon levels, which may be attributed to slow cooling and FeO decomposition to Fe compared to top layer. A productivity of 46-58 kg metallic iron/h-m2 was reported by Kakogawa Demonstration Plant (KDP) employing Fastmet process. Similarly, an optimum productivity of 36 kg metallic iron/h-m2 was reported for a pilot plant operating with Comet process. All these reported productivity data are consistent with the present experimental data.
The compressive strength of the reduced pellet is measured by using an Instron universal testing machine with a cross-head movement of 0.25 mm/minute. The variation of the compressive strength of the reduced pellets with the variation in C/Fe2O3 molar ratio layer-wise is depicted in Figure 9.
It is observed that the compressive strength is the maximum for the reduced pellets with the C/Fe2O3 molar ratio of 1.66 for all three layers and is around 3000 N/pellet. The pellet of such high strength can be effectively utilized in a blast furnace, or electric arc furnace (EAF) for subsequent reduction and melting. The peak strength of the partially reduced pellet at this ratio may be attributed to more iron content, consolidation and fewer cracks in the pellets.