Coke Combustion Rate with the Presence of Hematite in Quasi-particles

Abstract

The coke combustion rate in the iron ore sintering process is one of the most important factors affecting the quality and productivity of iron ore sintering. The purpose of this study was to investigate the combustion rate of coke in the presence of hematite to simulate the actual conditions in a sinter machine. The experimental samples were prepared by mixing coke powder with hematite powder. Experiments were carried out under an air or N₂ atmospheres at 1073, 1173 and 1273 K. From the results of the reaction curves, the coke combustion rates were analyzed using an unreacted-core model with one reaction interface.

From the kinetic analysis, it was found that the reaction rate constant k_c (m/s) and effective diffusion coefficient D_e (m²/s) of the sample were both affected by the experimental temperature. Furthermore, the D_e value of the sample was affected by the amount of coke. However, the k_c of the sample was not affected by this.

1. Introduction

The majority of the blast furnace charge in Japan is mainly iron ore sinter. Therefore, the quality of iron ore sinter is crucial. Table 1 shows the tension strength and reducibility of the major phases in iron ore sinter.^1,2) The strength and reducibility of each phase were different. The phases formed after the sintering process can affect the quality of sintered ore. Figure 1 shows the transformation of phases in iron ore sinter.³⁾ It is clear that the phases formed are different depending on the sintering temperature profile. Therefore, to improve the quality of the iron ore sinter, it is essential to control the temperature distribution during the sintering process.

Table 1. Tension strength and reducibility of the major phases in iron ore sinter.^1,2)

	Tension strength (MPa)	Reducibility (%)
Hematite	49	50
Magnetite	58	27
Calcium ferrite	102	44

Fig. 1.

Transformation of phase in iron ore sinter.³⁾

The combustion rate of a single coke particle is normally calculated using Hottel’s equation.^4,5,6) For the combustion rate of fine coke powder in the quasi-particles, the diffusion of oxygen is determined by experiments using mixed coke powder and alumina powder that is inactive during the reaction.⁷⁾ However, during the sintering process, coke is combined with fine iron ores, such as hematite. In addition, the liquid phase formed from calcium ferrate and the reduction and reoxidation of hematite can affect the combustion rate. The interaction between oxygen diffusion and coke combustion should also be considered. Therefore, the purpose of this study was to investigate coke combustion rate in the presence of hematite in quasi-particles.

2. Experimental

2.1. Experimental Sample

As simulant quasi-particles, the experimental samples were prepared using hematite powder and coke powder. Hematite with a particle diameter of −250 μm was used in this experiment, and the particle size of hematite was used to simulate the adhered powder layer. Coke powder of the same particle diameter was used to match the size of the hematite powder. The analysis results for coke are presented in Table 2. After the hematite and coke powders were well mixed, 0.5 mass% flour was mixed into 2 g of powder sample as a binder. Then, the mixture was pressed into a 10 mm diameter tablet using stainless steel dies, and the porosity of sample was adjusted to 35% by altering the height of the tablet. The weight ratio of coke in each sample and the height of each sample are listed in Table 3.

Table 2. Proximate analysis of coke (mass%).

F.C	V.M	ash	moist
86.2	1.0	12.8	0.74

Table 3. Mixing ratio and height of samples (mass%).

	Coke	Hematite	Height (mm)
Sample H1	10.0	90.0	7.9
Sample H2	20.0	80.0	8.5
Sample H3	30.0	70.0	9.1

2.2. Experimental Method

During these experiments, reactions shown in Eqs. (1), (2), (3), (4) can take place.   

$$C\left(s\right)+O_2\left(g\right)=C{O}_2\left(g\right)$$(1)

  

$$Fe{O}_n\left(\text{s}\right)+C\left(s\right)=Fe{O}_{n-1}\left(s\right)+CO\left(g\right)$$(2)

  

$$Fe{O}_n\left(\text{s}\right)+CO\left(g\right)=Fe{O}_{n-1}\left(s\right)+C{O}_2\left(g\right)$$(3)

  

$$C\left(\text{s}\right)+C{O}_2\left(g\right)=2CO\left(g\right)$$(4)

Equation (1) shows coke combustion. Due to the presence of iron oxide, direct reduction, as shown in Eq. (2), should be considered. Subsequently, CO formed by direct reduction can react with iron oxide, as shown in Eq. (3). Subsequently, the Boudouard reaction, as shown in Eq. (4), can progress to form CO. It is necessary to identify whether Eqs. (2), (3), (4) can take place during these experiments. Therefore, experiments were carried out under N₂ atmosphere. Later experiments were carried out under an air atmosphere to investigate the combustion of coke. The sample weight loss during coke combustion was measured using a thermobalance, as shown in Fig. 2. The sample was placed into a platinum basket. A vertical electric resistance furnace was used to create the isothermal heating conditions. The isothermal zone was heated to 1073, 1173 and 1273 K. For the experiments under N₂ atmosphere, heat treatment of samples was carried out at each given temperature under N₂ atmosphere for 10 min to remove water, volatile matter (V.M.) and the binder from the sample. Then, the experiment was started, and the experiment was ended when no more weight change was observed. The flow rate of N₂ was 8.33 × 10⁻⁵ m³/s. Microscopic observations and XRD analyses were also conducted to identify the mineral phase formed. Experiments under an air atmosphere were also carried out, and will be explained later.

Fig. 2.

Schematic diagram of the experimental device. (Online version in color.)

3. Results

3.1. Experiments under N₂ Atmosphere

The weight change curves at each temperature are shown in Figs. 3, 4, 5. The dashed line shows the estimated total weight of volatile matter, binder and the amount of coke at each coke content. At the reaction temperature of 1073 K no further weight change was observed after 200 s regardless of the weight ratio of coke, as shown in Fig. 3, and the weight change of all samples was at the same level. However, at 1173 and 1273 K, the weight change increased with increasing reaction time, and the coke contents in the samples became larger, and a larger weight change was obtained after the reaction as shown in Figs. 4 and 5. At 1073 K, the difference between final weight change and the dashed line was approximately 0.01 g, which is very small compared with that observed at 1173 and 1273 K. Because of the direct reduction of hematite, the increasing coke content and higher temperature may accelerate the reduction of hematite, causing a larger final weight change.

Fig. 3.

Weight change curve at 1073 K.

Fig. 4.

Weight change curve at 1173 K.

Fig. 5.

Weight change curve at 1273 K.

Figure 6 shows the microstructure of each sample after the heat treatment. s Hematite and coke phase were observed in all samples. However, a magnetite phase was observed at 1173 and 1273 K. Furthermore, the amount of magnetite phase at 1273 K appears to be higher than that at 1173 K. It is thought that the starting temperature of direct reduction may be between 1073 K and 1173 K, and with increasing reaction temperature, more magnetite can be formed by the reduction of hematite.

Fig. 6.

Microstructure of Sample H1, H2 and H3 after heat treatment. (Online version in color.)

Figures 7, 8, 9 show the XRD results of samples at 1073, 1173 and 1273 K, respectively. Peaks of hematite and coke were observed in all the samples, which is the same with the microscopic observation. At 1073 K, the magnetite peaks were found to be different from the microscopic observations. For the results at 1173 K, the phases were the same as in the microscopic observation. However, at 1273 K, peaks corresponding to FeO were also obtained in Sample H2 (20 mass% coke), and H3 (30 mass% coke) samples according to the XRD results. It is considered that a small amount of FeO can be formed during the reduction process when the amount of coke is large. From the results above, it is considered that the reactions shown in Eqs. (2), (3), (4) can hardly occur at 1073 K, and experiments under an air atmosphere were carried out as mentioned below.

Fig. 7.

X-ray diffraction pattern of Sample H1, H2 and H3 heat-treated at 1073 K. (Online version in color.)

Fig. 8.

X-ray diffraction pattern of Sample H1, H2 and H3 heat-treated at 1173 K. (Online version in color.)

Fig. 9.

X-ray diffraction pattern of Sample H1, H2 and H3 heat-treated at 1273 K. (Online version in color.)

3.2. Experiments under Air Atmosphere

For the experiments under air atmosphere, the samples were heated to 1073 K, and held for 6 min to remove the binder, and then the temperature was raised to the experimental temperature for 1 min under N₂ atmosphere. The combustion experiment was started when the N₂ atmosphere was replaced by an air atmosphere, and the experiment was finished when no more weight change was observed. The flow rate of air was 8.33 × 10⁻⁵ m³/s.

Figure 10 shows the XRD the result of heat-treated Sample H3 (30 mass% coke) which had the highest magnetite content in the experiment under N₂ atmosphere. Peaks of hematite and coke were identified, and no magnetite peaks were observed. The same tendency was also obtained for Sample H1 (10 mass% coke) and Sample H2 (20 mass% coke). It is considered that the final weight change of the sample corresponds to the combustion of coke. Therefore, the ratio of carbon removed during the experiment can be represented by the fractional reaction (F) at the reaction time which can be described by Eq. (5).   

$$F=\frac{\mathrm{\Delta }{w}_t}{W}$$(5)

where W is the total weight of coke without ash in the sample (g), and Δw_t is the weight change during the experiment (g).

Fig. 10.

X-ray diffraction pattern of Sample H3 heat-treated under Air atmosphere. (Online version in color.)

Figure 11 shows the fractional reaction at each temperature for the samples. It is clear that the samples with lower coke contents have shorter reaction times. In addition, it can be considered that a higher temperature causes a shorter reaction time. The final fractional reaction of all the experiments was nearly 1, which indicates the weight change was caused by the combustion of coke. Among all the reaction curves, Sample H1 at 1273 K had a temporary fractional reaction over 1. To verify the mechanism of the combustion reaction, samples with fractional reaction of 0.2 and 0.8 were also prepared under 1073, 1173 and 1273 K, and microscopic observations of the cross-section of the samples were carried out.

Fig. 11.

Fractional reaction curves of the samples.

Figure 12 shows the microstructure of Sample H1 with a fractional reaction of 0.2. Hematite and coke phases were observed at both 1073 K and 1173 K. However, at 1273 K hematite and magnetite phases were observed, but no coke was observed. It is thought that oxidization of magnetite occurs earlier than the combustion of coke when the temperature is less than 1173 K. When the temperature is high such as 1273 K, the rate of coke combustion is faster, and both oxidization of magnetite and combustion of coke can occur. In addition, the fractional reaction for 200 s in the experiment of Sample H1 at 1273 K was over 1 because the combustion of coke and the reduction of hematite to magnetite occurred simultaneously, which caused the weight change over the fixed carbon content. Furthermore, under this condition, it is considered that the reduction rate from hematite to magnetite is greater than the oxidation rate from magnetite to hematite. In other experiments, except for Sample H1 at 1273 K, the reduction rate from hematite to magnetite was at the same level as the oxidation rate from magnetite to hematite. As a result, the fractional reaction was not affected and did not exceed 1.

Fig. 12.

Microstructure of Sample H1 with the fractional reaction of 0.2. (Online version in color.)

Figure 13 shows cross-sectional observations of the samples. It is clear that the combustion reaction was a topochemical reaction.

Fig. 13.

Cross-sectional observations of the samples with the fractional reaction of 0.8. (Online version in color.)

4. Kinetic Analysis (Unreacted Core Model for Coke)

From the fractional reaction curves obtained from the combustion experiment except for Sample H1 at 1273 K, the combustion reaction rate constant was determined using the unreacted core model. The following processes during the combustion reaction should be considered.⁷⁾

I. O₂ is transported from the gas phase to the particle surface through the gas film.

II. O₂ is transported from the particle surface to reaction interface through the iron oxide powder layer after coke combustion.

III. Combustion reaction at the reaction interface.

During these processes, Eq. (6) can be obtained.   

$$\begin{array}{l}t={\displaystyle \frac{{\rho }_{Cm}{r}_0}{C_{O_2}}}\\ \cdot \left[{\displaystyle \frac{F}{3k_f}}+{\displaystyle \frac{d_r}{12D_e}}\left\{3-3{\left(1-F\right)}^{{\displaystyle \frac{2}{3}}}-2F\right\}+{\displaystyle \frac{1}{k_C}}\left\{1-{\left(1-F\right)}^{{\displaystyle \frac{1}{3}}}\right\}\right]\end{array}$$(6)

where C_(O2) is the O₂ concentration in the gas phase (mol/m³); D_e is the effective diffusion coefficient (m²/s); F is the fractional reaction (–); k_C is the interfacial chemical reaction rate constant (m/s); r₀ is the initial radius (m); t is time (s); ρ_Cm is the carbon concentration in sample (mol/m³); d_r is the diameter of sample (m).

The gas film mass transfer coefficient k_f (m/s) can be calculated using Ranz-Marshall’s equation.⁸⁾ The value of the effective diffusion coefficient in the iron oxide layer D_e and the interfacial chemical reaction rate constant of coke k_C were obtained using mixed-control plot.

Figure 14 shows the relationship between k_c and coke ratio in the sample. It is clear that k_c increases with increasing temperature. In addition, k_c is a particular value regardless of the coke ratio at the same reaction temperature.

Fig. 14.

Relationship between reaction rate constants k_c and coke ratio in sample. (Online version in color.)

Figure 15 shows the Arrhenius plot of k_c. It is clear that the k_c is not affected by the amount of coke in samples and k_c is larger with higher temperature.

Fig. 15.

Temperature dependence of reaction rate constants k_c. (Online version in color.)

The temperature dependence of k_c is expressed as follows.   

$$k_c=3.78\text{exp}\left(-\frac{40.9\times {10}^3}{RT}\right)\mathrm{\hspace{0.17em} } \left(\text{m}/\text{s}\right)$$(7)

Figure 16 shows the relationship between D_e and the coke ratio in the sample. It is clear that when the coke content in the sample is larger, a larger D_e can be obtained.

Fig. 16.

Relationship between effective diffusivities D_e and coke ratio in the sample. (Online version in color.)

The effective diffusion considering the porosity of the sample and temperature can be expressed as Eq. (8).⁹⁾   

$$D_e=D_{O_2}{\left(\frac{T}{T_0}\right)}^n\frac{\epsilon }{\tau }\hspace{1em}\left({\text{m}}^2/\text{s}\right)$$(8)

Equation (8) can be transformed into the following equation, and the temperature independence of the diffusion coefficient n can be shown.   

$$\text{ln}{D}_e=\text{ln}\left(D_{O_2}\frac{\epsilon }{\tau }\right)+n\text{ln}\frac{T}{T_0}$$(9)

Figure 17 shows the relationship between lnD_e and lnT/T₀. The value of n can be obtained as 2.61, which is a larger value than the temperature independence of the diffusion coefficient between 1.75 and 2.⁹⁾ It is considered that the larger temperature independence of the diffusion coefficient was affected by the reduction and oxidization of iron oxide during the experiment.

Fig. 17.

Temperature dependence of effective diffusivities D_e. (Online version in color.)

Figure 18 shows the relationship between the labyrinth factor τ and porosity ε. The value of the labyrinth factor obtained in this experiment is a proper value of the labyrinth factor, which normally lies between 1 and 10.

Fig. 18.

Porosity dependence of labyrinth factor τ_H.

As a result, D_e and τ can be expressed by the following equations.   

$$D_e=1.78\times {10}^{-5}{\left(\frac{T}{273.15}\right)}^{2.61}\frac{\epsilon }{\tau }\mathrm{\hspace{0.17em} } \left({\text{m}}^2/\text{s}\right)$$(10)

  

$$\tau =\frac{68.9}{\text{exp}\left(4.64\epsilon \right)}+1\mathrm{\hspace{0.17em} } (-)$$(11)

Figures 15 and 17 show the comparison of k_c and D_e obtained in this study and in previous research,⁷⁾ which combined 30 mass% coke with alumina to calculate the combustion rate of coke. It is clear that the k_c of coke combined with hematite is larger than that of coke combined with alumina, although the molar heat capacity of hematite is 1.1 times larger than that of alumina. The sample with alumina requires more heat at higher temperatures because the alumina content is 1.6 times larger than that of hematite when the weights of hematite and alumina are the same. In other words, the sample where coke was combined with hematite could reach a higher temperature during the experiment, and as a result, the value of k_c for the sample with hematite was larger. Moreover, the oxidization and reduction of iron oxide may affect the value of k_c. On the other hand, the D_e values in both studies were the same. The results show that, the reaction of hematite hardly affects the diffusion of oxygen.

5. Conclusions

This study investigated the coke combustion rate in the presence of hematite in quasi-particles. The following conclusions were drawn:

(1) Both k_c and D_e are affected by the temperature.

(2) The amount of the coke in sample can affect the D_e value. However, k_c is not affected by the amount of coke in the sample.

(3) The interfacial chemical reaction rate coefficient and effective diffusion coefficient of the experimental data were calculated as follows:   

$$k_c=3.78\text{exp}\left(-\frac{40.9\times {10}^3}{RT}\right)\mathrm{\hspace{0.17em} } \left(\text{m}/\text{s}\right)$$

  

$$D_e=1.78\times {10}^{-5}{\left(\frac{T}{273.15}\right)}^{2.61}\frac{\epsilon }{\tau }\mathrm{\hspace{0.17em} } \left({\text{m}}^2/\text{s}\right)$$

References

• 1)   N.  Oyama,  K.  Nushiro,  Y.  Konishi,  K.  Igawa and  K.  Sorimachi: Tetsu-to-Hagané, 82 (1996), 719 (in Japanese).
• 2)   T.  Iwami: Study about Low Carbon Sintering Process, Research Institute of Low Carbon Sinter Technology, The Iron and Steel Institute of Japan, Tokyo, (2011), 11 (in Japanese).
• 3)   M.  Sasaki and  Y.  Hida: Tetsu-to-Hagané, 68 (1982), 563 (in Japanese).
• 4)   C. M.  Tu,  H.  Davis and  H. C.  Hottel: Ind. Eng. Chem., 26 (1934), 749.
• 5)   H.  Davis and  H. C.  Hottel: Ind. Eng. Chem., 26 (1934), 889.
• 6)   A. S.  Parker and  H. C.  Hottel: Ind. Eng. Chem., 28 (1936), 1334.
• 7)   K.  Ohno,  K.  Noda,  K.  Nishioka,  T.  Maeda and  M.  Shimizu: ISIJ Int., 53 (2013), 1588. https://doi.org/10.2355/isijinternational.53.1588
• 8)   W. E.  Ranz and  W. R.  Marshall: Chem. Eng. Prog., 48 (1952), 173.
• 9)   K.  Sato: Butsuseiteisuusuizanhou (Estimation of Physical Property Value), Maruzen Publishing Co., Ltd., Tokyo, (1975), 129 (in Japanese).