Effect of Alumina and Silica on the Reaction Kinetics of Carbon Composite Pellets at 1473 K

Abstract

The influence of compositional changes in the CaO–FeO_t–(Al₂O₃) or (SiO₂) ternary oxide system on the reaction kinetics of carbon composite pellet was investigated by Thermo-Gravimetric Analyzer (TGA) at 1473 K (1200°C). Measured CO and CO₂ gas by Infrared gas analyzer and surface area change from BET analysis assisted with the reaction kinetics of carbon composite pellet. As a result, alumina increased the reduction rate of iron oxide by increasing surface area, while silica decreased reduction rate by lowering surface area of pellet samples. A modified reaction kinetic model considering both the Boudouard reaction and surface area variation was developed for a better explanation of reduction mechanisms occurring in carbon composite pellets.

1. Introduction

As high quality raw materials are being depleted worldwide, beneficiation of iron ores and agglomeration process have become more important for iron making industry. Especially, pelletizing technique attracts many investigators as it utilizes fine iron ores from mining sites as well as secondary materials generated from steel works such as mill scale and flue dust. Iron ore pellet is well known to be an excellent feed material for blast furnace operation as it provides not only enough space for gas permeability, but also a high mechanical strength for sustaining heavy burdens in blast furnaces.¹⁾ During the pelletizing process, however, a large amount of fossil fuel is supplied for induration step which consequently produces large CO₂ emissions. Along with the increased interests in environmental issues and energy usage, extensive research regarding the direct utilization of low grade fine iron ores and pulverized non-coking coal in the iron making process has been carried out.²⁾ Among several examples of direct reduction processes, rotary hearth furnace (RHF) type process such as FASTMET and ITmk3 have reached commercial success. The process operate with iron ore - coal composite pellet producing Direct Reduced Iron (DRI) or iron nugget in a very shortened period of time (approx. 10 minutes). Comparatively inexpensive construction cost and relatively easier operation from small facilities are some of the main advantages of these processes. However, due to their limited productivity and high energy consumption rate, RHF processes are often considered as recycling process for the secondary materials generated from iron and steel works.²⁾ Therefore, technological development of carbon composite pellet for mass production of iron is mandatory for the future iron making industry. Due to its advantages of fast reaction rates coming from proximity of reactants, the application of carbon composite iron ore pellet in the blast furnace have been reported as successful by several researchers.^3,4) Researchers have also studied the improvement of blast furnace energy efficiency by lowering melting temperature of reduced iron to less than 1723 K (1450°C). Matsui et al.,⁵⁾ investigated carburization effect by solid carbon and role of concentrated gangue layer formation on the reduction process. They concluded that the direct contact of reduced iron with carbon was crucial for low temperature melting and flux addition is the most effective method to remove gangue layer which blocks melting of reduced iron. Our previous work has developed a kinetic model which is suitable for self-reducing iron oxide pellet,⁶⁾ and more recently the authors have examined the role of the Boudouard reaction in iron oxide and carbon briquettes.⁷⁾ The study focused on the effect of various carbonaceous materials (coal char, coke and graphite) on the reaction process with regard to their Boudouard reaction. Although those previous studies enabled us to establish a fundamental understanding about the reaction kinetics in a carbon composite pellet, there has been limited understanding on the role of minerals from the view point of reduction. Currently, increasing amount of alumina and silica in raw materials is inevitable situation for iron makers. Thus, many studies have been conducted on the influence of high alumina slag in blast furnace operations; slag viscosities^8,9) and desulfurizing capacities of high alumina blast furnace slags.¹⁰⁾ However, only a limited research has been done on the effect of alumina and silica regarding the reaction kinetics of iron oxide by solid carbon reduction. In the present study, the influence of compositional changes in the CaO–FeO_t–(Al₂O₃) or (SiO₂) ternary oxide system on the reaction kinetics of carbon composite pellet was investigated by Thermo-Gravimetric Analyzer (TGA) at 1473 K (1200°C). Measured CO and CO₂ gas by Infrared gas analyzer and surface area change from BET analysis assisted with the reaction kinetics of carbon composite pellet. A modified kinetic model was also developed that is useful for a better understating of reduction mechanisms occurring in carbon composite pellets.

2. Experimental Procedure

Carbon composite pellets used in the present study were spherical shape of 8 mm in diameter and 0.9 g (±0.05 g) in weight. Hematite powder (≥99%, <5 μm) and synthetic graphite flake (≥99.99%, ≤150 μm) supplied by Sigma-Aldrich Australia were dried in air at 473 K (200°C) for 12 hours to eliminate moisture. Reagent grade CaO, SiO₂ and Al₂O₃ chemicals supplied by Sigma-Aldrich Australia were used for making the synthetic oxide mixture. Powder mixtures of hematite, graphite and oxide system were agitated in a rotating drum for 24 hours to obtain homogeneity. Equal size pellets were made by hand rolling with drops of water to cause the fines to adhere to each other by liquid bridging. Prepared samples were dried at 473 K (200°C) for 24 hours in order to remove the water bond and allowing solid particles to form physical bonds through capillary and surface forces. As shown in Table 1, different amounts of alumina and silica were added to investigate the effect of additive minerals on the reduction behavior of the carbon composite pellets. Fe₂O₃/CaO ratio of samples was fixed at 5.66 to make initial calcium ferrite (CF2) composition, and total graphite content among the composite pellet was fixed at 15 wt%. The schematic apparatus of the experiment is shown in Fig. 1. Pellets were reduced in a purpose built Thermo-Gravimetric Analyzer (TGA) at 1473 K (1200°C). The furnace was resistance heated using Super-Kanthal (MoSi2) elements, producing a 125 mm hot zone. High purity nitrogen was supplied as the inert gas through the bottom of the furnace at a rate of 1 L/min using a mass flow controller. Any secondary reactions were prevented through the elimination of stagnant gases of CO and CO₂ in the pellets. After 10 minutes of gas purging for complete removal of residual gases, the furnace was lifted in 5 seconds positioning the samples at hot zone. Weight loss during a reaction was monitored and logged every 5 seconds by computer via a digital balance. Every sample was held in the furnace over 30 minutes to ensure completion of reactions. Off gases evolved from the reduction of oxides were monitored and measured by an infrared gas analyzer (ABB, Advance Optima Series AO2020). The gas concentrations of CO and CO₂ were used to assist with the determination of reaction kinetics of carbon composite pellets. In all samples, off gases (CO, CO₂) began to evolve after 10 seconds due to the time lag between the reaction and measurement by the IR gas analyzer. In this study, the effect of heat transfer was assumed to be negligible so that it could not have influenced the reduction process. X-ray diffraction analysis was carried out using PANalytical Xpert Multipurpose X-ray Diffraction System (MPD). Ground powder samples were packed into the holder of 25 mm diameter and 5 mm depth. Voltage and current of the machine were set at 45 kv and 30 mA. The XRD patterns were obtained by recording the scattering intensities with Nickel filtered Cu Kα-radiation. The scanning were recorded over an angular range from 10° to 90° using a step size of 0.026° and collected the scattering intensity for 1 s at each step. A Micromeritics TriStar 3000 analyzer was used for the measurement of surface area of reduced pellets. From the adsorption and desorption isotherms, information about the surface characteristics of the particle was obtained.

Table 1. Chemical composition of carbon composite pellets.

Samples	Carbon/Oxides	Composition of oxides				Total weight (g)
		Fe2O3	CaO	Al2O3	SiO2
1	15/85	85	15	–	–	0.935
2		80.75	14.25	5	–	0.943
3		76.5	13.5	10	–	0.925
4		72.25	12.75	15	–	0.945
5		80.75	14.25	–	5	0.934
6		76.5	13.5	–	10	0.943
7		72.25	12.75	–	15	0.941

Fig. 1.

A schematic of experimental apparatus of Thermo Gravimetric Analyzer (TGA).

3. Results and Discussion

3.1. Reduction Behavior of Carbon Composite Pellets

Every sample weighing between 0.90 g and 0.95 g started to react after 5 seconds when it reached to 1473 K (1200°C) at the hot zone of the furnace. The weight losses of carbon composite pellets during the reduction process are shown in Fig. 2. The samples show different weight loss behavior when they have different amount of alumina and silica. It is clearly seen in Fig. 2 that alumina contents from 5 to 15 wt% in the oxide system increase the reaction rate by accelerating reduction of hematite. However, silica contents from 5 to 15 wt% contribute to the decreased reaction rate as the graph shows slowing down behavior. While reaction of sample 1 without alumina or silica completed nearly after 350 seconds, samples from 2 to 4 with alumina showed shortened overall reaction time which was less than 300 seconds. In contrast, silica contents in carbon composite pellets adversely affected the reduction process by retarding overall reaction. The reaction of sample 5 and 6 containing 5 and 10 wt% of silica completed in 420 seconds indicating slower overall reaction. Moreover, sample 7 with 15 wt% of silica in oxides system showed far delayed reaction progress that ended after 600 seconds. The time required for complete reduction of carbon composite pellets was confirmed to be influenced by alumina and silica from Thermo-Gravimetric Analyzer (TGA) experiment. The reactions occurring between hematite and solid carbon in composite pellets are known to occur in stages.¹¹⁾ According to the publication on the solid - gas reactions,¹²⁾ it is generally agreed that the reduction of iron oxide by solid carbon occurs through the gas intermediates of CO and CO₂, and the reaction sequence is represented by   

$$F{e}_x{O}_y\left(\text{s}\right)+CO\left(g\right)=F{e}_x{O}_{y-1}\left(s\right)+C{O}_2\left(g\right)$$(13)

  

$$C{O}_2\left(\text{g}\right)+C\left(s\right)=2CO\left(g\right)$$(13)

Fig. 2.

Effect of Al₂O₃ and SiO₂ on the weight loss of carbon composite pellet during reduction at 1473 K (1200°C).

During the reduction of iron oxide, it is assumed that all Fe₂O₃ is firstly converted to FeO, (Fe₃O₄ can be an intermediate phase, but it was not observed in the present analysis due to fast reaction) followed by further reduction of FeO to Fe. Of overall reduction process, the wüstite to iron step has been shown to be the slowest¹³⁾ as wüstite needs higher reduction potential for conversion to Fe.¹⁴⁾ On the other hand, the oxidation of carbon (i.e. Boudouard reaction) is also considered to be extremely important as it provides driving force for chemical reaction to occur. In general, the overall reaction rate of hematite – solid carbon is known to be controlled by the oxidation of carbon per reaction (2) at temperatures between 1173 K (900°C) and 1473 K (1200°C).¹⁵⁾ Therefore, in the present study, it is necessary to identify the reaction kinetics of wüsite to iron step as well as Boudouard reaction to clarify the influences of alumina and silica on the reduction of carbon composite pellet. In Fig. 3, the reduction degree of the samples are shown as a function of time calculated from the following equation by assuming complete reduction of hematite into a reduced iron;   

$$F{e}_2{O}_3+3C=2Fe+3CO\text{\hspace{0.17em}.}$$(3)

  

$${\eta }_R=\frac{W_0{X}_{F{e}_2{O}_3}-W_{F{e}_x{O}_y}}{W_0{X}_{F{e}_2{O}_3}\times \frac{M_{F{e}_2{O}_3}-2M_{Fe}}{M_{F{e}_2{O}_3}}}\times 100\%$$(4)

  

$$W_{F{e}_x{O}_y}=W_f-W_C$$(5)

where W₀ is the pellet weight before reduction (g), X_Fe2O3 is hematite fraction of carbon composite pellets (wt%) and M is the molecular weight (g/mol). W_f is the pellet weight after reduction (g) which was measured by Thermo-Gravimetric Analyzer (TGA). Carbon content in pellets (W_C) was calculated from CO and CO₂ off gases measured by an infrared gas analyzer. Weight of iron oxide in the pellet (X_FexOy) was calculated from Eq. (5) and used in Eq. (4) accordingly. The overall reaction could be divided into two major parts from Fig. 3 on the basis of slope change of graphs. First, there is a period where hematite converts to wüstite. During this step the following solid - gas reaction occurs.   

$$F{e}_2{O}_3\left(s\right)+CO\left(g\right)=2FeO\left(s\right)+C{O}_2\left(g\right)$$(5)

Fig. 3.

Reduction degree of carbon composite pellets at 1473 K (1200°C).

The weight loss on this stage can be approximately calculated as 30% from a molar ratio of Eq. (5). As seen in Fig. 3, the slopes are changing at 30 percent of reduction degree on every sample that is identical to the amount of weight loss from hematite to wüstite. This initial stage of reduction is considered to have less impact on overall reaction as it occurred fast enough that did not limit reaction rate. From the structural point of view, the transformation of hematite to wüstite gives rise to lattice distortions.¹⁴⁾ In the disturbed lattice, the diffusion of iron is accelerated and a large vacancy concentration can arise where mass is lost. The creation of voids is thus facilitated in the defect structure and all these factors contribute to a considerable pore formation.¹⁴⁾ Fast reduction rate of hematite to wüstite, therefore, can be explained by physical change of iron oxides. The time required for slope change varies depending on compositions of carbon composite pellets. While it takes 150 seconds for sample 1 to be reduced to wüstite, samples with alumina require less time (120 seconds) and silica containing samples (#5 to #7) need more time (180 to 200 second) for hematite - wüstite reduction stage. These time differences in hematite - wüstite reaction due to alumina and silica contents are explained by BET surface area analysis in the section 3.3. Secondly, a steady state reaction is observed from 30 percent of reduction degree until the end of reaction for each sample. During this stage rate constants can be calculated by assuming uniform internal reduction of carbon composite pellets. The first order irreversible unimolecular model is given as   

$$-ln\left(1-\text{X}\right)=\text{kt}$$(6)

where X is reduced fraction and k is reaction rate constant. This equation is known to be valid when gas concentration is constant during the reaction, whereas gas diffusions in the pores offer little hindrance to the reduction process.¹⁶⁾ In Fig. 4, complete linear relationship between ln(1 – X) and time can be found in the second stage of reduction. The following wüstite - carbon monoxide gas reaction is considered to be dominant during this step by showing uniform internal reactions.   

$$FeO\left(s\right)+CO\left(g\right)=Fe\left(s\right)+C{O}_2\left(g\right)$$(7)

Fig. 4.

–ln(1 – X) plots of carbon composite pellets as a function of time.

Although there are discrepancies in graphs due to the gap between hematite - wüstite and wüstite - iron reduction stages, uniform internal reaction is believed to be appropriate for the reduction mechanism of carbon composite pellet. The common analytical method of iron oxide - solid carbon reaction is to plot ln(1 – X) vs. time,¹⁶⁾ and such a plot is also possible for hematite - graphite composite containing alumina and silica. The overall reduction reaction rate, however, is seen to level off at final stage of reaction which is 360 second for sample 1, 280 second for alumina containing samples (#3, #4) and 450 second for samples with silica (#5, #6). This non-linearity shown in Fig. 4 could be explained by additional physico-chemical properties of carbon composite system in the present study. First of all, the effect of changing P_CO/P_CO2 and hence the gaseous driving force, in the reaction gas was considered as a chemical influence. In addition, physical influence coming from different amount of additives (i.e. alumina and silica) was also included in modified model by adopting surface area change which was not constant during the overall reduction reaction.

3.2. Influence of Boudouard Reaction

In the present iron oxide – solid carbon composite system, reaction proceeds between graphite flake and hematite powder generating CO and CO₂ gases. There are two combined reactions; one is iron oxide reduction and the other is oxidation of carbon. Boudouard reaction in Eq. (2) play an extremely important role in overall reduction process as it provides reducing gas (CO) for the iron oxide reduction. The concentration of CO gas (namely driving force for chemical reaction) should be constant for the sake of uniform internal reduction model.¹⁶⁾ However, as the concentrations of CO and CO₂ gases are changing within composite pellet during the reaction, shortage or oversupply of CO gases are highly expected in the practical experiments. The rising discrepancies between uniform internal reduction model and the real situation of solid carbon reduction could be minimized by analysis of CO and CO₂ in off-gases. In the previous studies,^6,7) Moon and Sahajwalla developed a rate equation, namely uniform conversion model which included both terms of iron oxide reduction and the carbon oxidation. Analysing the reaction off-gas provided an effective and dynamic method to assist in the determination of the reaction controlling the self-reduction process. The oxygen potential of the reaction off gas can be calculated in terms of P_CO/P_CO2 ratio. To determine which of the reactions are controlling, it is necessary to identify the reaction at equilibrium. A given reaction is at equilibrium when the off gas P_CO/P_CO2 concentration falls on the equilibrium P_CO/P_CO2 value for that reaction.⁶⁾ As shown in Fig. 5, each equilibrium ln(P_CO/P_CO2) value at 1473 K (1200°C) for Fe₂O₃–FeO equilibrium, FeO–Fe equilibrium and Boudouard reaction are calculated by assuming that P_CO + P_CO2 = 1. The results of ln(P_CO/P_CO2) are plotted as a function time that indicated the equilibrium state at each reduction stage. During the first 50 seconds, ln(P_CO/P_CO2) value approaches hematite - wüstite equilibrium implying solid – gas reaction of Eq. (5). Between 50 and 150 seconds, all graphs are approaching wüstite – iron equilibrium as Eq. (7). The time required for initial hematite to wüstite reduction is confirmed to be similar to the weight loss results from Thermo Gravimetric Analysis (TGA). From 150 seconds of reaction, the complimentary wüstite - iron and Boudouard reaction is taken to be the controlling reaction. The off-gas P_CO/P_CO2 concentration indicates that the wüstite - iron reaction is at equilibrium, and it can be inferred that the complimentary Boudouard reaction is the controlling reaction. However, there are no significant differences in ln(P_CO/P_CO2) variations when comparing each sample. This implies that alumina and silica contents included in carbon composite pellets do not significantly affect Boudouard reaction. The model developed for uniform conversion of carbon composite briquette is given as⁶⁾   

$$-\frac{ln\left(1-X\right)}{\left(P_{C{O}_2}-P_{C{O}_2}^{eq.}\right)}=\text{k}\cdot \text{t}$$(8)

where X is reduced fraction and k is reaction rate constant. $\left(P_{C{O}_2}-P_{C{O}_2}^{eq.}\right)$ is the Boudouard reaction driving force, P_CO2 is concentration of evolved CO₂ gas and $P_{C{O}_2}^{eq.}$ is equilibrium CO₂ concentration of Boudouard reaction. In the present experimental condition, P_CO2 was measured from infrared gas analysis and $P_{C{O}_2}^{eq.}$ was calculated as 8.4388 × 10^–4 (mol/cm³) at 1473 K (1200°C) by assuming P_CO + P_CO2 = 1. In Fig. 6, the uniform conversion model is applied to the present data, and the results were plotted along with the reaction off gas as a function of time. It is shown in the figure that plotted lines are considerably straightened compared to those leveled off lines in Fig. 4. The reaction model (8) validated that its rate was heavily influenced by the P_CO/P_CO2 levels within the composite during the course of conversion. Thus formerly developed model seems to be more appropriate to explain reaction kinetics of current carbon composite pellets. However, the uniform conversion model⁶⁾ is not fully applicable for the kinetic analysis of the present pellet samples as the graphs show non-linear relationships. The reaction rate of carbon composite pellet containing alumina or silica is not solely dependent on chemical influences, but also largely affected by physical influence coming from morphological changes during reduction process. Surface area variations of samples, therefore, were investigated by BET analysis which is shown to have a huge impact on reaction kinetics of hematite - solid carbon reaction.¹⁶⁾

Fig. 5.

ln(P_CO/P_CO2) plots from off-gases as a function of time.

Fig. 6.

–ln(1 – X)/ $\left(P_{C{O}_2}-P_{C{O}_2}^{eq.}\right)$ plots from off-gases as a function of time.

3.3. Effect of Surface Area Change on the Reaction Rate

In the previous study, surface area change was confirm to be influential on the reduction of carbon composite pellet.¹⁷⁾ According to BET results, alumina based pellet showed higher surface area compared to silica based pellets during overall reaction.¹⁷⁾ This demonstrated larger inner surface of alumina pellet that contributed faster reaction rate. Figure 7 illustrates the surface area change of carbon composite pellets at different reduction times. After 5% of reduction degree, surface areas of all samples were in between 3.81 m²/g and 4.76 m²/g. While sample 2, 3 and 4 show lower surface area of 4.38 m²/g, 3.81 m²/g and 4.18 m²/g respectively, other samples have comparatively higher surface area above 4.5 m²/g. Alumina content in the oxide system of samples 2, 3 and 4 is considered to have lowered surface area by the sintering effect at the initial stage of the reduction.¹⁸⁾ The effect of alumina and silica on the internal surface area of carbon composite pellets is more apparent after 30% of reduction degree which is about 200 seconds of reaction. Alumina containing samples show gently reducing slopes, while surface area drops far more rapidly in silica containing samples (#5, #6 and #7). Surface areas of samples with alumina remain approximately at 0.5 m²/g even after 400 seconds, which means a supply of large number of reaction sites resulting in faster hematite to wüstite and wüstite to iron reduction. The reason for this phenomenon is considered to be a formation of high melting point (1663 K (1390°C)) calcium aluminate compound (CaAl₂O₄). As seen in Al₂O₃–CaO–FeO_t ternary phase diagram,¹⁹⁾ introduction of alumina content from 5 to 15 wt% significantly increases melting point at fixed CaO / Fe₂O₃ ratio (CF2). It is considered that no mutual solubility occurs in the subsystem; Fe₂O₃–CaAl₂O₄. Thus, alumina does not form a low melting eutectic line with iron oxides producing high melting point (1663 K (1390°C)) compound instead. Phase analysis by X-ray Diffraction confirmed the formation of calcium-aluminate phase as shown in Fig. 8. After 30% of reduction degree, major peaks for CaAl₂O₄ were observed substantiating above hypothesis. In contrast, samples with silica contents (#5, #6 and #7) show rapid decrease in surface area. Sample 5 and 6 show very low surface areas of 0.06 m²/g and 0.04 m²/g respectively after 400 seconds of reaction. More significant drop in surface area was observed at sample 7 as it decreases to 0.01 m²/g only after 180 seconds. Surface area could not be measured after 600 seconds since samples (#5, #6 and #7) converted to complete dense phase. It is considered that the formation of a liquid fayalite phase generated dense internal structures. From the CaO–SiO₂–FeO_t ternary phase diagram,¹⁹⁾ increasing silica contents from 5 to 15 wt% confirmed to formulate the fayalite phase (Fe₂SiO₄) at a fixed CaO / Fe₂O₃ ratio (CF2). Due to its low melting point of 1478 K (1205°C) and low viscosity (≤10 Pa·s),²⁰⁾ fayalite lowered the surface area of pellets by increasing liquid phases in the structure. When silica is introduced in the oxide system, the fayalite phase increases the stability of oxides. Activity of iron oxide decreases due to its higher chemical bonding energy.²¹⁾ Thus, fayalite is more difficult to be reduced by carbon as given equation of   

$$\begin{array}{l}2\text{Fe}\left(\text{s}\right)+O_2\left(g\right)+Si{O}_2\left(s\right)\\ =F{e}_{2}Si{O}_4\left(s\right), \mathrm{\Delta }{G}_{T\left(K\right)}^o=-568\mathrm{\hspace{0.17em} }260+144.48T kJ/mol\end{array}$$(9)

Fig. 7.

Dependence of surface area of carbon composite pellets on time.

Fig. 8.

XRD phase analysis of sample 4 containing 15 wt% of alumina as degree of reduction.

Therefore, samples containing silica (#5, #6 and #7) had a liquid fayalite formation within the iron oxide and solid carbon hindering reaction. Fayalite was also confirmed by XRD analysis since major peaks was observed after 30% reduction degree as shown in Fig. 9. The results demonstrated a significant influence of oxide composition present in the carbon composite pellet. The dependence of the extent of reduction on composition is evident from BET surface area measurement for the system of varying chemistries used in carbon composite pellet. According to Turkdogan and Vinters,¹⁶⁾ for solid – gas uniform internal reduction, the integrated form of the rate equation is given as   

$$\text{ln}\left(1-X\right)=-S{\varphi }_{C{O}_2}\left(1-\theta \right)\left[P_{CO}-{\left(P_{CO}\right)}_e\right]t$$(10)

where X is the reduced fraction, S is the pore surface area of wüstite per unit mass of oxygen, (1 – θ) is the fraction of vacant sites in the adsorbed layer and ϕ_CO2 is the specific rate constant. P_CO is partial pressure of CO gas in the gas and (P_CO)_e is the corresponding equilibrium value for coexistence with iron and wüstite phases. The uniform internal reduction Eq. (10) is valid when the reduction time is independent of particle sizes. However, particle sizes of composite pellet and their surface area is not expected to remain constant during the overall reduction process. Variations of surface area during reduction can be calculated from BET-N₂ measurement as an exponential function as following equation.   

$$S=Aexp\left(-B\cdot t\right)+C$$(11)

Fig. 9.

XRD phase analysis of sample 7 containing 15 wt% of silica as degree of reduction.

Empirical constant values A, B and C obtained from each sample are listed in Table 2. The new developed model includes both the chemical influence of Boudouard reaction and the physical influence coming from the morphological change of internal structure of carbon composite pellets. A combined formula of Eqs. (8) and (11) can be given as   

$$-\frac{ln\left(1-X\right)}{S\left(P_{C{O}_2}-P_{C{O}_2}^{eq.}\right)}=k\cdot t$$(12)

where X is reduced fraction and k is reaction constant. S is surface area obtained from Eq. (11), and $\left(P_{C{O}_2}-P_{C{O}_2}^{eq.}\right)$ is the Boudouard reaction driving force from gas analysis. Data obtained from the present experiments were calculated with Eq. (12) and the results are plotted in Fig. 10 as a function of time. Although sample 5, 6 and 7 containing silica show non-linear relation with time, samples from 1 to 4 show strong linearity when plotted with time. Due to their abrupt change in surface area, Eq. (12) could not be applied to samples with silica (#5, #6 and #7). These samples containing silica generated wüstite and existing silica in pellet formed fayalite phase which provided liquid phase interface between iron oxide and solid carbon. Internal uniform reaction is considered to be largely prohibited by liquid fayalite, so the Eq. (12) is not applicable for samples containing silica. However, samples from 1 to 4 showed linear relationship with time which confirms new kinetic model described by Eq. (12). Their slowly decreasing surface area appeared to secure enough reaction sites for wüstite – iron conversion as well as Boudouard reaction, which contributed to suitability of the uniform internal reaction model. Rate constants derived from Eq. (12), listed in Table 3, shows a positive influence of alumina on reduction rate. It is apparent that faster iron oxide reduction occurs when higher amount of alumina is introduced to the composite pellet system. Consequently, alumina content in carbon composite pellet increases reaction rate by providing enough surface area, while silica hinders reduction reaction by formulating liquid fayalite phase at 1474 K (1200°C).

Table 2. Empirical constant values for S = Aexp(–B·t) + C.

Sample	A	B	C
1	5.6711	0.0071	0.1777
2	4.9569	0.0087	0.5560
3	4.1027	0.0051	0.2884
4	4.9482	0.0070	0.1838

Fig. 10.

–ln(1 – X)/S $\left(P_{C{O}_2}-P_{C{O}_2}^{eq.}\right)$ plots of carbon composite pellets as a function of time.

Table 3. Rate constants K of carbon composite pellets from Eq. (12).

Sample	Alumina contents (wt%)	Rate constant (K) (cm·mol–1·sec–1)
1	0	1014.0
2	5	1052.1
3	10	1132.4
4	15	1491.2

4. Conclusions

The effect of alumina and silica on the reaction kinetics of carbon composite pellet at 1474 K (1200°C) was investigated. Following conclusions can be made from this study.

(1) The reduction rate of iron oxide in carbon composite pellet was influenced by alumina and silica contents. Alumina increased the reaction rate by increasing surface area while silica decreased the reaction rate by lowering surface area of pellet samples.

(2) The overall reaction of carbon composite pellets was largely controlled by uniform internal reaction as the plot ln(1 – X) vs. time showed linearity in every sample.

(3) This study helped with developing a modified kinetic model of carbon composite pellet given as   

$$-\frac{ln\left(1-X\right)}{S\left(P_{C{O}_2}-P_{C{O}_2}^{eq.}\right)}=k\cdot t$$

where X is reduced fraction and k is reaction constant. S is surface area obtained from Eq. (11), and $\left(P_{C{O}_2}-P_{C{O}_2}^{eq.}\right)$ is the Boudouard reaction driving force from gas analysis. The equation includes both the chemical influence of Boudouard reaction and the physical influence coming from the morphological change of internal structure of carbon composite pellets.

The study implies that high alumina iron ore system can be used for the application of carbon composite pellets as it had superior reduction behavior compared to silica based oxide systems. The result is due to the higher surface area of alumina containing pellets that contributed to faster reduction kinetics.

Acknowledgements

The authors thank to Mr. N. Saha Chaudhury of SMaRT center (Centre for Sustainable Materials Research and Technology) at University of New South Wales for experimental design.

References

• 1)   K.  Myer: Pelletizing of Iron Ores, Springer-Verlag, New York, (1980), 3.
• 2)   F.  Habashi: Handbook of Extractive Metallurgy, WILEY-VCH, Weinheim, (1997), 104.
• 3)   Y.  Ueki,  T.  Maeda,  M.  Shimiu,  Y.  Matsui and  A.  Kasai: Tetsu-to-Hagané, 89 (2003), 1205.
• 4)   H.  Ono,  K.  Tanizawa and  T.  Usui: ISIJ Int., 51 (2011), 1274.
• 5)   T.  Matsui,  N.  Ishiwata,  Y.  Hara and  K.  Takeda: ISIJ Int., 44 (2004), 2105.
• 6)   J.  Moon and  V.  Sahajwalla: ISIJ Int., 43 (2003),1136.
• 7)   J.  Moon and  V.  Sahajwalla: Metall. Mater. Trans., 37B (2006), 215.
• 8)   J. R.  Kim,  Y. S.  Lee,  D. J.  Min,  S. M.  Jung and  S. H.  Yi: ISIJ Int., 44 (2004), 1291.
• 9)   Y. S.  Lee,  D. J.  Min,  S. M.  Jung and  S. H.  Yi: ISIJ Int., 44 (2004), 1283.
• 10)   B.  Agrawal,  G. J.  Yurek and  J. F.  Elliott: Metall. Mater. Trans., 14B (1983), 221.
• 11)   R.  Haque and  H. S.  Ray: Metall. Mater. Trans., 26B (1995), 400.
• 12)   J.  Szekely and  H. Y.  Sohn: Gas - Solid Reactions. Academic Press, Inc., New York, (1976), 176.
• 13)   D.  Carvalho,  P.  Quariguasi and  J.  D’Abreu: Can. Metall. Q, 33 (1994), 217.
• 14)   A. K.  Biswas: Principles of Blast Furnace Ironmaking. Cootha Publishing House, Brisbane, (1981), 1.
• 15)   R.  Fruehan: Metall. Mater. Trans., 8B (1977), 279.
• 16)   E. T.  Turkdogan and  J.  Vinters: Metall. Mater. Trans., 3B (1972), 1561.
• 17)   H.  Park and  V.  Sahajwalla: Metall. Mater. Trans., (2013), accepted.
• 18)   D. F.  Ball,  J.  Dartnell,  J.  Davison and  A.  Grieve: Agglomeration of Iron Ores, Heinemnann Educational Books Limited, London, (1973), 95.
• 19)   Verein  Deutscher Eisenhuttenleute[21]: Slag ATLAS, Verlag Stahleisen Gmbh, Düsseldorf, (1995), 99.
• 20)   H. S.  Park and  I.  Sohn: Metall. Mater. Trans., 42B (2011), 692.
• 21)   O.  Kubaschewski: Metallurgical Thermochemistry, Materials Science and Technology, New York, (1979), 24.