Effect of Carbon Addition on Direct Reduction Behavior of Low Quality Magnetite Ore by Reducing Gas Atmosphere

Abstract

Recently, direct reduced iron (DRI) has been highlighted as a promising iron source for electric arc furnace (EAF)-based steelmaking. The two typical production methods for DRI are gas-based reduction and reduction using carbon composite pellets. While the gas-based reduction is strongly dependent on the reliable supply of hydrocarbon fuel, reduction using ore-coal composite pellets has relatively low productivity due to solid–solid reactions. To overcome the limitations of the above two processes, and to achieve a more efficient direct reduction process of iron ore, the possibility of combining these two methods was investigated. The experiments focused on performing an initial direct reduction using ore-coal composite pellets followed by a second stage gas reduction. It was assumed that the initial reduction of the carbon composite pellets would enhance the efficiency of the subsequent reduction by gas and the total reduction efficiency. The porosity, as well as the carbon efficiency for direct reduction, were measured to determine the optimal conditions for the initial reduction, such as the size ratio of ore and coal particles. Thereafter, further reduction by the reducing gas was carried out to verify the effect of the preliminary reduction. The reduction kinetics of the reducing gas was also discussed.

1. Introduction

With the strong demand for production methods that could effectively counteract global warming and CO₂ emissions, steelmaking production using electric arc furnaces (EAFs) is getting renewed attention. In order to prepare for the anticipated drastic growth of EAF-based steelmaking industries, it is necessary to find an alternative iron source to compensate for the shortage of high-quality scrap, which is currently the most important raw material for EAF-based steelmaking [1,2,3]. Direct reduced iron (DRI), which was developed several decades ago, is a promising candidate. Recently, many studies have been conducted on the operational changes in the EAF process that would be needed to utilize DRI [4,5]. A number of steelmaking companies with access to abundant natural gas have already been using a considerable amount of DRI in their EAF process. Since more demand for DRI is expected in the near future, more efficient production technology is required.

The production processes for DRI can be largely categorized into two types: gas-based processes and carbon-based processes. Although DRI production by gas-based processes has been more dominant than carbon-based processes, a large amount of hydrocarbon fuel, such as LNG, is required for direct reduction in gas-based DRI production. The reduction efficiency may considerably decrease unless the counter-current flow of the charges and reducing gas is properly established. Therefore, the carbon-based process can be favorable in case of a short supply of LNG and low efficiency of a reactor, such as a rotary kiln.

On the other hand, the mining amount of high-quality iron ore is currently lacking and its depletion will be encountered in the near future. Thus, innovative technologies for utilizing low-quality iron ore, such as magnetite, in DRI production are strongly required. Generally, solid carbonaceous materials such as coal are combined with iron ore into pellets for more efficient solid–solid reactions. At the initial stage of reduction by the composite carbon pellets, the emission of CO gas leaves a considerable number of pores in the pellet. If reducing gas is subsequently introduced, the porosity of the carbon-based reduction may enhance the mass transfer of gases and improve the net rate of the reduction reaction in the gas-based reduction. On the other hand, the melting point of reduced iron can be decreased by carburization when excessive carbon exists in the carbon-based reduction [6]. Because molten iron may adhere to each other as well as refractories, the formation of liquid iron should be avoided in the direct reduction in rotary or shaft kilns.

Therefore, it is meaningful to seek optimized conditions for carbon-based reduction to complement gas-based reduction. In the present study, direct reduction experiments of iron ore pellets were carried out using hybrid gas-based reduction assisted by carbon-based reduction.

2. Materials and Methods

2.1. Sample Preparation

As raw materials for iron ore and coal, magnetite and anthracite, whose compositions are shown in

Table 1. Compositions of iron ore and coal used in the present study.

Magnetite ore	t-Fe	Fe3O4	Fe2O3	MgO	SiO2	CaO
	43.5%	53.7%	6.6%	15.1%	17.7%	4.3%
Anthracite	F.C.	V.M.	Ash
	58.9	6.5%	34.7%

, were chosen. Each material was pulverized in a disk mill and dried thoroughly in a box furnace at 383 K or a desiccator for more than 48 h. Afterwards, both materials were classified into several size groups using a series of sieves. To properly simulate the practical conditions of tailing magnetite fine particles, the iron ore and coal powder were classified into three size groups: 32–53 μm (S), 53–74 μm (M), and 74–105 μm (L). Using the analyses of optical microscopes, the average size of each group was estimated to be 43 μm, 64 μm, and 91 μm, respectively, and these average values were semi-quantitively verified based on the optical observation.

Because the amount of anthracite coal obtained in the 53–74 μm group was not sufficient, only the other two size groups were used for the coal.

To determine the optimum conditions for the carbon composite as complementary pre-reduction, the carbon equivalent, which represents the molar ratio of supplied carbon from fixed carbon against oxygen to be reduced, were varied in the range of 0–1.0. A number of gas reduction experiments without carbon addition (C_Eq. = 0) were also performed for comparison. Magnetite ore and anthracite coal under various carbon equivalent conditions were carefully mixed and pressed at 3000 kg/cm² using a pellet press into a disk-type carbon composite pellet with a diameter and height of 15 mm and 10 mm, respectively.

2.2. Experimental Procedure

The reduction experiments of the prepared pellets were carried out by analyzing the composition of the off-gas. A pellet in a quartz crucible (i.d. 22 mm, depth 27 mm) was placed in a bottom-closed quartz tube (i.d. 27 mm), which had both a gas inlet (o.d. 6 mm, i.d. 4 mm) and outlet at the top. The experimental apparatus used in this study is schematically shown in Figure 1. After a few hours of purging in a quartz tube with high-purity Ar gas as a carrier gas with 1360 cm³/min, the quartz tube holding the sample was inserted in a vertical electric resistance furnace controlled at 1373 K using Kanthal heaters. From the time when the sample was inserted into the furnace, the composition (CO, CO₂, O₂) of the off-gas was analyzed using a gas analyzer (NOVA9K, MRU, Neckarsulm, Germany). In order to calculate the amount of evolved CO or/and CO₂ gases from dynamic gas compositions, as a tracer gas, O₂ gas was directly introduced into the analyzer at a known flow rate, controlled to be 240 cm³/min by a mass flow controller at room temperature.

In the case of carbon composite reduction, only Ar gas was introduced, and the concentrations of CO and CO₂ gas in the off-gas were analyzed to investigate the reduction behavior. The reduction was carried out until the emission of both CO and CO₂ gas became negligible and was completed in 50 min in most cases. The gas reduction behavior could be determined from the amount of CO₂ emitted during the introduction of the CO + Ar mixture gas into the quartz tube. Several experiments combining carbon-based and gas-based reduction were divided into two stages for the separate consideration of the two reduction mechanisms. In the first stage, CO and CO₂ gases were detected by the composite carbon. After the evolution of CO and CO₂ gases was completed, the carrier gas was switched to CO gas to simulate the gas reduction in the second stage. When CO₂ gas evolved from the gas reduction decreased to a negligible level, the quartz tube was drawn out to be air-cooled.

To investigate the effect of pore formation, the true density of each sample after reduction was measured using an envelope density analyzer (GeoPyc 1360, Micromeritics, Norcross, GA, USA), while its apparent density was approximately estimated from the dimensions to obtain the porosity, as Equation (1)

$$porosity=\left(1-\frac{apparent density}{true density}\right)\times 100 (\%)$$

(1)

The degree of reduction was determined as follows. The flow rates of the total gas from the sample could be obtained from the known flow rate of the tracer gas, O₂ gas, and the composition of O₂ gas (vol%) using a gas analyzer as shown in Figure 2. The relationship between the flow rates of the total gas and O₂ gas is expressed by Equations (2) and (3). Subsequently, the flow rates of CO and CO₂ gas from the reduction of the sample could be determined from that of the total gas, as shown below.

$$Q_{total gas}=Q_{{\mathrm{O}}_2}\times \frac{100\%}{\%vol {\mathrm{O}}_2}$$

(2)

$$Q_{\mathrm{CO}}=Q_{total gas }\times \%vol \mathrm{CO}, Q_{{\mathrm{CO}}_2}=Q_{total gas }\times \%vol {\mathrm{CO}}_2$$

(3)

Then, the integration of the flow rates of CO and CO₂ gases with time represents the total amount of each gas from the reduction of the sample. Based on the accumulated amounts of CO and CO₂ gases in each stage, the reduction degree, $Re$, for the reduction by composited carbon and CO gas can be expressed as shown in Equation (4), respectively. Similarly, the reduction degree of the entire reaction can be estimated using Equation (5).

$$R{e}_{comp.}=\frac{\mathrm{O} in \mathrm{CO} and {\mathrm{CO}}_2}{\mathrm{O} in iron ore}\times 100 R{e}_{gas}=\frac{\mathrm{O} in {\mathrm{CO}}_2}{residual \mathrm{O} in iron ore}\times 100$$

(4)

$$R{e}_{total}=\frac{\mathrm{O} in total \mathrm{CO} and {\mathrm{CO}}_2}{\mathrm{O} in iron ore}\times 100$$

(5)

In addition to the reduction degree, as an index of residual carbon, the carbon efficiency for the reduction, ${\epsilon }_C$, could be further obtained for the whole reaction, as shown in Equation (6)

$${\epsilon }_C=\frac{\mathrm{C} in total \mathrm{CO} and {\mathrm{CO}}_2}{\mathrm{C} in anthracite+\mathrm{C} in introduced \mathrm{CO}}\times 100$$

(6)

3. Results and Discussion

3.1. Effect of Particle Size in Carbon-Based Reduction

Figure 3 shows the changes in the degree of reduction and carbon efficiency with different particle sizes. As the sizes of iron ore and coal particles decreased, the reduction reaction by carbon appeared to be promoted, which may be attributed to the larger surface area of the smaller particle size. This phenomenon indicates that as the number of contacts between magnetite ore and coal increases, the direct reduction reaction is more likely to occur than in the case with larger particle size.

It could be further verified by estimating the contacts between different kinds of particles, that is, iron ore and coal. A few decades ago, Norio and Tanaka developed a simple and effective expression as shown in Equation (7), based on a simple packing model, for the number of contacts between randomly mixed solid particles of different sizes [7].

$$N_T=\frac{N_D·n_D+N_d·n_d}{2}, N_D=16\left(1-{\epsilon }_A\right)\frac{4{\left(\frac{D+\overline{x}}{2\overline{x}}\right)}^2{\left(\frac{D+\overline{x}}{2\overline{x}}\right)}^2}{3+\frac{\overline{x^2}}{{\left(\overline{x}\right)}^2}}\times r_d$$

(7)

Here, N_T and N_i stand for total number of contacts and the number of particles surrounding a particle i, while n_i and r_i the number and the fraction of a particle i, respectively. D, d, and $\overline{x}$ denote the diameters of larger and smaller particle and the average diameter, respectively. Moreover, ε_A refers to the surface porosity, which can be deduced from the volume porosity.

Using their model, the number of contacts between ore and coal particles was estimated to sharply increase when either the ore or coal particle size decreased, as shown in Figure 4. This phenomenon indicates that as the particle size decreases, the number of contacts between magnetite ore and anthracite increases, and a direct reduction reaction is more likely to occur than in the case with larger particle size. However, in Figure 5, it is shown that the relationship between the degree of reduction and the total number of contacts varies with the coal size.

The influence of the particle sizes of the iron ore and the carbon composite on the degree of reduction needs to be further clarified.

Recalling Figure 3, it can be also seen that the effect of the ore particle size on the reduction seems to be more significant than that of the coal particle size. Since the reduction characteristics can be also evaluated based on the porosity created by CO and CO₂ gas evolution, the porosity of the reduced pellets was investigated. The effect of the size of iron ore and coal particles on the porosity of the reduced pellets is presented in Figure 6. As the ore particle size increased, the fraction of pores in the reduced pellets decreased. Contrary to the ore particle size, the coal particle size has a negligible influence on the porosity, which is essential for enhancing the subsequent reduction by gas. The pores can be formed by the evolution of CO and CO₂ gas from the reduction of the ore. Whereas carbon sources for the gases are supplied from not only coal particles but also CO gas, only ore particles are able to provide oxygen for CO and CO₂ gas formation. Therefore, pore formation appeared to be more dependent on the surface area of the ore particles than on that of the coal particles.

3.2. Optimization of Carbon-Based Reduction Prior to Gas Reduction

To determine the optimum conditions for pre-reduction by composited carbon for better subsequent gas reduction, experiments for carbon-based and gas-based combined reduction were conducted using the smallest groups of ore and coal particles. Carbon with a certain equivalent was composited into iron ore pellets, and the reduction of the pellet by CO gas was initiated after the completion of the composited carbon reaction. Similar to the previous section, the degree of reduction was separately evaluated for the carbon-based and gas-based cases, as shown in Figure 7. Although the reduction mechanism shifts from gas-based reduction to carbon-based reduction as the carbon equivalent increases, which is not much different from expected, the overall degree of reduction adding up both the gas-based and carbon-based reduction did not change significantly.

Instead of the degree of reduction, the carbon efficiency could be used to establish the optimal conditions for maximizing the reduction efficiency by CO gas. Based on the amount of the emitted CO₂ gas, the carbon efficiency of CO gas was estimated by varying the carbon equivalent. Figure 8 shows the influence of the carbon equivalent on the carbon efficiency of CO gas. For comparison, the efficiency of CO gas in the gas-only reduction (C_Eq. = 0) was calculated by integrating the CO₂ gas emission from the point at which the degree of reduction by composited carbon in each carbon equivalent is equal to the degree of reduction during the gas reduction without carbon. In gas-only reduction without carbon, the carbon efficiency gradually decreases as the reduction progresses. On the other hand, the carbon efficiency of CO gas after the preliminary reduction was improved compared to the gas-only reduction. This improvement becomes evident when the carbon equivalent is greater than 0.2, possibly because of the formation of additional pores by the preliminary reduction.

3.3. Kinetics of Reduction of Iron Ore by CO Gas

As discussed in the previous section, the efficiency of the reduction by gases could be improved by more pore formation, which can increase the mobilities of CO and CO₂ gases. The effect of gas mobility is also meaningful in discussing the kinetics of the reduction reaction by the reducing gas. The kinetics of the reduction of iron ore by reducing gas has been intensively analyzed by many researchers [8,9,10,11]. It has been found that the reduction behavior of iron ore by carbon or carbonaceous gas is strongly dependent on process conditions such as temperature, pressure, size of the solid particles, and pore structure. Because the entire reduction process is divided into several reaction steps, the rate of the total reduction reaction should be governed by a rate-controlling step.

There are typical equations for expressing the reduction reaction rate, which have been proposed to be related to specific rate-controlling steps, i.e., Boudouard reaction, topochemical reaction, and gas diffusion. The dominant reaction mechanism of reduction by gas in the present study was determined using known rate equation functions for each determining step. As shown in Figure 9, a simple comparison of the linearity of the functions with the present results indicated that the reduction by gas is mainly governed by gas diffusion over a wide range of carbon equivalents. The rate equation function for gas diffusion can be written in a simplified form as Equation (8) [12], whose slope, K, is known to include the effective counter-current diffusivity of CO and CO₂. Thus, the reduction kinetics can be discussed using the slope against the reaction time. In particular, the initial slope in the earliest stage should be considered to properly evaluate the effect of the pores formed by preliminary reduction.

$$\left[3-2x-3{\left(1-x\right)}^{2/3}\right]=Kt+C$$

(8)

As shown in Figure 10, at a low carbon equivalent, the initial effective diffusivity of CO and CO₂ gases increases with the increment of the carbon equivalent. It can be expected that more pores with more composited carbon enabled faster diffusion of CO and CO₂ gases. When the carbon equivalent increases above 0.7, the gas mobilities of CO and CO₂ are slightly reduced, because the reduction by CO gas is limited by an insufficient amount of unreduced residual ore. Hence, it can be concluded that a carbon equivalent of 0.5~0.7, seems to be a favorable condition to maximize the efficiency of the reducing gas under the present experimental conditions, based on the reaction rate and carbon efficiency.

It should be noted that preliminary reduction by carbon and subsequent reduction by gas cannot be clearly separated under actual conditions in industrial kilns. A reducing gas atmosphere (considerable CO partial pressure) may suppress the preliminary reduction by carbon in the early stages. Experiments on the hybrid reduction of iron ore were carried out using carbon and CO gas simultaneously. Since it was not possible to determine the degree of reduction by analyzing CO and CO₂ gas in the off-gas, the metallization ratio of the pellets after interrupting the reaction at a certain time was quantified from the concentrations of metallic Fe, Fe⁺², and Fe³⁺. The metallization ratios of the pellets are shown in Figure 11. Compared to the pellets interrupted similarly during the reduction by CO gas, the metallization ratio of pellets by combined reduction is substantially higher, which means that the reduction rate of the combined reduction by carbon + CO gas is superior to that of the gas reduction only. It was verified that the present hybrid reduction can be applied to even industrial conditions using reducing gas.

Conclusively, it is expected that the consumption of hydrocarbon fuel for DRI production can be cut down because the reduction is completed in a shorter time, as shown in Figure 11. Hence, the hybrid DRI production process is possible with less CO₂ emissions. As explained in Section 3.2, moreover, the efficiency of reducing gas could be improved by the preliminary reduction using composited carbon, resulting in further reduction of CO₂ emission. The overall effect of the hybrid reduction of iron ore may be quantitatively analyzed in further study.

4. Summary

The feasibility of combining carbon composite reduction and gas reduction was experimentally investigated. It was expected that pre-reduction by the composite carbon could improve pore formation and gas reduction. First, the fundamental conditions for carbon-based reduction to form more pores were sought. When the particle size decreased, the number of contacts between the iron ore and coal particles increased, and the degree of reduction and carbon efficiency increased, leaving more pores in the pellet. It was also found that ore particle size has a greater influence on pore formation than coal particle size. In the carbon-CO gas 2-step reduction experiments with a controlled carbon equivalent, the efficiency of gas reduction was enhanced by the preliminary reduction by carbon composited with a carbon equivalent greater than 0.2. The analysis using the reduction kinetics showed that the reduction rate is dependent on the diffusivity of CO and CO₂, which can be maximized at a carbon equivalent of 0.5~0.7. Finally, the applicability of the combined reduction to industrial practice was confirmed based on a higher reaction rate.