Abstract
Surplus coke oven gases (COGs) and low grade hematite ores are abundant in Shanxi, China. Our group proposes a new process that could simultaneously enrich CH4 from COG and produce separated magnetite from low grade hematite. In this work, low-temperature hydrogen reduction of hematite ore fines was performed in a fixed-bed reactor with a stirring apparatus, and a laboratory Davis magnetic tube was used for the magnetic separation of the resulting magnetite ore fines. The properties of the raw hematite ore, reduced products, and magnetic concentrate were analyzed and characterized by a chemical analysis method, X-ray diffraction, optical microscopy, and scanning electron microscopy. The experimental results indicated that, at temperatures lower than 400°C, the rate of reduction of the hematite ore fines was controlled by the interfacial reaction on the core surface. However, at temperatures higher than 450°C, the reaction was controlled by product layer diffusion. With increasing reduction temperature, the average utilization of hydrogen initially increased and tended to a constant value thereafter. The conversion of Fe2O3 in the hematite ore played an important role in the total iron recovery and grade of the concentrate. The grade of the concentrate decreased, whereas the total iron recovery increased with the increasing Fe2O3 conversion.
1. Introduction
Shanxi, a large coal and coke producing province in China, has a large amount of surplus coke oven gases (COGs) [1–6], which contain substantial amounts of H2, CH4, and CO. In 2014, the coke output of Shanxi was 87.22 million tons, which accounted for 18.29% of the national output. If 1t of coke can produce 430 Nm3 of coke oven gas, of which approximately 200 Nm3 is returned to a coke oven as a coking heat source, approximately 16~18 billion Nm3 of COGs is produced just in Shanxi. Meanwhile, the apparent consumption of natural gas in China was 178.6 billion Nm3 in 2014, which imports 63 billion Nm3 of natural gas (approximately 33.9% external dependency). COG methanation for CH4 enrichment can be regarded as a simple and highly efficient way of producing gas, which has attracted many specialists and scholars to research and develop the process. However, the catalysts, which are required for both the reactivity and selectivity, are sensitive to sulfide and very expensive. Can we economically obtain enriched natural gas from surplus COGs?
Due to their natural abundance, low cost, and usefulness for a variety of applications, iron ores are considered some of the most promising and important resources for the future [7, 8]. The rapid development of the steel industry has resulted in the reduction of iron ore resources in recent years; these ores have been increasingly explored and utilized. At the same time, processed iron ores may be utilized as raw materials for the desulfurization sorbents used in gas purification; iron oxide is the main component of a renewable desulfurization sorbent for high-temperature coal gas desulfurization [9–12]. In addition, this new route uses iron ore as an oxygen carrier, which transfers oxygen from the combustion air to the fuel. Direct contact between the fuel and combustion air is prevented through the use of chemical looping combustion (CLC) [13–15].
However, using the iron ores is quite difficult due to their complex structure and nonuniform crystal size. A combination of low-temperature reduction roasting and low-intensity magnetic separation is considered a promising approach for increasing the usability of these ores [16, 17]. Low-temperature reduction roasting can convert the weakly magnetic iron minerals (Fe2O3, FeCO3, and FeS) into a strongly magnetic phase (Fe3O4), which can be easily separated by a low-intensity magnetic field. In addition, magnetization, which can be readily performed at low operating costs, is used to reduce the hematite (Fe2O3) to magnetite (Fe3O4). H2, CO, CH4, and coal can be used as reducing agents [18, 19], and the corresponding reduction reactions should proceed as follows:
Compared to coal-based reduction, gas-based reduction can be performed at lower reduction temperatures [20] and typically results in higher quality concentrates, which have lower levels of carbon deposits. H2, CO, and CH4 are usually used as reducing agents in these gas-based reductions [21–23].
As such, our research group has proposed an innovative process, which combines low-temperature reduction magnetization of intractable hematite with the production of substitute natural gases (SNGs) from COGs. This combined process was used to remove H2, CO, and H2S, thereby enriching the methane gas contained in the COG and converting hematite to easily separated magnetite. The ultimate aim of this work was to use low grade hematite ore and the surplus COG comprehensively to obtain enriched SNG and high grade magnetite economically. Meanwhile, the whole process is described in Figure 1.
This paper, as a preliminary study of the innovative process, mainly describes the process of low-temperature hydrogen reduction of hematite ore and magnetic separation of the resulting magnetite ore, since hydrogen, which is the main component of COG, is an efficient reducing agent. The effect of this reduction on the properties of the magnetite ore is discussed.
2. Materials and Methods
2.1. Materials
The raw hematite ore used in this work was obtained from Guangling County, Shanxi province, China. Guangling County is located northeast of the Shanxi province. Guangling hematite ore is a Shanxi-type hematite ore.
2.2. Analysis and Characterization Method
The chemical compositions of the raw hematite ore, reduced ore samples, and concentrate ore samples were analyzed according to the National Standards of China number GB 6730.5-2008 and number GB 6730.8-2008.
The crystalline phases of the aforementioned samples were determined via X-ray diffraction (XRD) (using Cu-Kα radiation, scanning rate of 8°/min, and sweeping range of 5°–85°).
The structural characteristics of the raw hematite ore were examined using an optical microscope and a scanning electron microscope (SEM).
2.3. Experimental Procedure
A total of 100 kg of the raw hematite ore sample was first crushed using a jaw crusher and then sieved to a size of 2 mm in our laboratory. One kilogram of each of the 12 samples of the crushed hematite ore was then reduced to magnetite ore, under a total volumetric gas rate of 120 L/h, using a gas mixture of 50% H2–50% N2. The reduction process was performed in a fixed-bed reactor, and a stirring apparatus was used to enable full contact between the solids and the gas. The samples were reduced for 1, 1.5, 2, and 2.5 h, at 400°C. Similarly, the samples were reduced for 0.5, 1, 1.5, and 2 h at both 450°C and 500°C.
The total Fe content (TFe) and Fe2+ content (TFeO) of each sample were determined using a chemical analysis method. The titanium trichloride-potassium dichromate titration method was used to determine the TFe of each sample; the ferric chloride-potassium dichromate titration method was used to determine the TFeO. Fe3O4 contains one Fe2+ and two Fe3+ ions. Therefore, if all the hematite in the raw ore sample is reduced to Fe3O4, then the TFe and TFeO of the reduced sample are related through by TFeO = 3/7TFe. The conversion of Fe2O3 () to Fe3O4 in each reduced sample can then be calculated by determining the TFe and TFeO of each reduced sample. was calculated from
The grain sizes of the newly generated magnetite were calculated fromwhere is the distance between the atomic layers in the crystals, is Scherer’s constant, is the wavelength of the X-ray radiation, is the diffraction angle, and is the full width at half maximum.
As for the low-intensity magnetic separation of the reduced samples, the optimal grinding fineness and magnetic field intensity should be chosen such that the optimal magnetic separation index is obtained. In this paper, reduced samples with different degrees of grinding fineness and magnetic field intensities were obtained by varying the grinding time and the working electrical current of the Davis tube magnetic separator, respectively. The experiments aimed at optimizing the grinding fineness and magnetic field intensity were performed for samples reduced at 500°C for 0.5 h or 2.0 h. The reduced samples were dry milled in a rod mill for various times and then separated by a laboratory Davis tube magnetic separator, with a working electrical current of 2 A, in order to determine the best grinding fineness. Furthermore, in order to determine the optimum magnetic field intensity, reduced samples with the best grinding fineness were separated by a laboratory Davis tube magnetic separator operating at various electrical currents.
Each of the 12 reduced samples was dry milled in a rod mill for 15 min and then separated by a laboratory Davis tube magnetic separator (model: XCGS50), operating at a working electrical current of 2 A (magnetic field intensity: 0.156 T). The grades and weights of the magnetic concentrates were determined, and the recovery of iron (), which is the index of magnetic separation, was calculated during the low-intensity magnetic separation; , the amount of iron recovered in the final concentrate, was calculated fromwhere , , , , and are the amount of iron recovered during the low-intensity magnetic separation, quality of the concentrate, iron content of the concentrate, quality of the reduced samples, and iron content of the reduced samples, respectively.
3. Results and Discussions
3.1. Properties of the Raw Ore Samples
The mineral composition, main chemical composition, and iron distribution of the raw hematite samples are listed in Tables 1, 2, and 3, respectively; the crystalline phase is shown in Figure 2.
The Guangling hematite ore sample consisted of 58.83% Fe2O3, 19.53% SiO2, 6.75% Al2O3, 0.21% FeO, 2.46% CaO, 1.38% MgO, 0.11% Na2O, and 3.65% K2O. Harmful elements, such as phosphorus and sulfur, were present in only low quantities. This hematite ore contained a small amount of magnetite; hematite is the precious form of iron minerals. The gangue minerals were composed mainly of quartz and small amounts of mica and kaolinite. Furthermore, there was a 3.95% loss upon ignition for the Guangling hematite ore. As Figure 2 shows, the Guangling hematite ore was composed mainly of hematite and the nonmetallic mineral, quartz.
The optical micrographs in Figure 3 reveal that the Guangling hematite ore mainly had taxitic structure, disseminated structure, and similar oolitic structure; the hematite mixed with quartz or mica, hematite, quartz, and clay (mainly kaolinite) formed an oolitic-like structure, quartz formed the core of the similar oolitic structure, and hematite and clay (mainly kaolinite) formed a concentric circle in the similar oolitic structure.
(a)
(b)
(c)
The corresponding SEM images (Figure 4) confirmed that the hematite ore consisted mainly of scaly, acicular, cryptocrystalline, and metasomatic textures, with a sand consolidation structure (rarely, with an oolitic structure). The acicular and scaly hematite ores are bright white, whereas quartz shows a grey color. The disseminated structure of the hematite was very fine.
Therefore, the hematite ore used in this study was a typical complex, fine-grained, marine-sediment-refractory hematite ore.
3.2. Reducing Magnetization of the Raw Hematite Ore
3.2.1. Thermodynamics of Reducing Fe2O3 with H2
The Gibbs free energy function method was adopted for the thermodynamic calculation of the H2-reduction of Fe2O3 [24, 25]; the results are shown in Figure 5.
Three reduction products were formed, namely, Fe3O4, FeO, and Fe. Moreover, Fe2O3 reacted with H2 to form Fe3O4 at reduction temperatures of 400°C–600°C and H2 concentrations of <65 vol.%.
3.2.2. Effect of the Reduction Temperature and Time on the Conversion of Fe2O3
Short, low-temperature reduction processes can reduce the cost of magnetizing the hematite ore. The dependence of the Fe2O3 conversion in the hematite ore on the reduction time (Figure 6) was investigated at the temperatures of 400°C, 450°C, and 500°C.
As Figure 6 shows, at 400°C, the amount of Fe2O3 converted increased linearly with the increasing reduction time. At reduction temperatures of 450°C or 500°C, the amount of Fe2O3 converted increased rapidly at the beginning of the reaction and slowly thereafter. The shrinking unreacted core model stipulates that there is a sharp boundary between the reacted (magnetite: Fe3O4) and the unreacted (hematite: Fe2O3) parts of the particle [26]. The hematite is reduced to magnetite (Fe3O4) via the diffusion of hydrogen through the product layer on the boundary (reacted-unreacted interface). As reduction proceeds, the boundary eventually recedes to the center, the hematite is exhausted, and the thickness of the product layer (magnetite layer) increases. The product layer diffusion and the interfacial reaction on the surface of the core constitute the rate-limiting steps of the process [25, 26].
The interfacial reaction on the surface of the core proceeded very slowly at low reduction temperatures. In fact, the rate of the interfacial reaction is significantly lower than that of the hydrogen diffusion process. Therefore, the interfacial reaction constituted the rate-limiting step of the process. Since the reduction temperature was constant and very low, the reaction proceeded extremely slowly; the conversion of Fe2O3 was proportional to the reduction time.
When the reduction temperature was high, the interfacial reaction on the core surface proceeded rapidly and at a significantly higher rate than that of the hydrogen diffusion process. Diffusion into the product layer was, therefore, the rate-limiting step of the process. During the beginning of reduction, the reaction proceeded rapidly since the product layer was very thin, and the diffusion resistance of hydrogen was correspondingly small. As the reduction proceeded, the thickness of the product layer increased the diffusion resistance of hydrogen, which in turn led to a sharp decrease in the reaction rate with the increasing reduction time.
The aforementioned results show that the transition of hematite (Fe2O3) to magnetite (Fe3O4) proceeded according to a shrinking unreacted core model. When the reduction temperature was low, the reaction rate was controlled by the interfacial reaction on the surface of the core; when the reduction temperature was high, the reaction rate was controlled by diffusion into the product layer.
On average, ~18.6%, 23.1%, and 23.4% of the hydrogen were utilized during the reduction process at 400°C, 450°C, and 500°C, respectively. This indicates that the average utilization of hydrogen increased initially and tended, subsequently, to a constant value with the increasing reduction temperature.
XRD was used to characterize 12 samples, which were reduced for 1 h, 1.5 h, 2 h, and 2.5 h, at the respective reduction temperatures of 400°C, 450°C, and 500°C, as shown in Figures 7, 8, and 9. As the figures show, the intensity of the peaks corresponding to hematite and magnetite decreases and increases, respectively, with the increasing reduction time. This indicates that the rate of reaction increased with the increasing temperature. Moreover, nearly all the hematite was reduced to magnetite during the process.
3.3. Mechanism of Reduction Magnetization Process
The grain size of the newly generated magnetite is shown in Figure 9.
Figure 10 shows that the grain size of the newly generated magnetite increased initially and then decreased slightly and increased thereafter. Based on the structural characteristics of the hematite, this grain size trend also indicates that the transition of hematite (Fe2O3) to magnetite (Fe3O4) followed a shrinking unreacted core model [27–29]. This transition progressed in three steps [30].
Step 1. H2 molecules diffused through the air and were absorbed on the surface of hematite; these H2 molecules were then activated. These active H2 molecules combined with O2− from the Fe2O3 lattice, thereby generating H2O [31]. Moreover, the two electrons released, during the reaction, reduced Fe3+ to Fe2+:
Step 2. The new generation of Fe2+ combined with Fe2O3, thereby producing Fe3O4 through lattice reconstruction [31, 32]:
Step 3. During the above process, the reduction reaction extended continuously to the inner layer, and the entire hematite particle was completely reduced to magnetite [32].
Figure 11 shows the SEM images of samples reduced for 1 h and 2.5 h at 400°C as well as for 0.5 h and 2 h at 500°C.
(a)
(b)
(c)
(d)
The grain sizes of the new-generation magnetite increased with the increasing reaction time. Furthermore, the amount of Fe2O3 converted and the number of grains generated increased with the increasing reaction temperature and time.
3.4. Low-Intensity Magnetic Separation of the Reduced Samples
3.4.1. Optimization of the Grinding Fineness and Magnetic Field Intensity
To obtain a high grade concentrate with excellent Fe recovery, experiments on the grinding fineness of the ore were performed prior to magnetic separation of the reduced samples. The results obtained for the samples reduced at 500°C for 0.5 h and 2.0 h are shown in Figures 12–15.
The figures reveal an optimum grinding time of 15 min and a corresponding grinding fineness and iron recovery of −200 mesh and 70.78%, respectively. In addition, the optimum magnetic field intensity (0.156 T) for magnetic separation occurred at an electrical current of 2 A.
The grain size of the hydrogen-reduced new-generation magnetite influenced the dissociation of the core and, therefore, played an important role in the subsequent magnetic separation. Comparing Figure 9 with Figures 14 and 15 reveals that the increases in the grain size for the former and latter figures for the new-generation magnetite were favorable and unfavorable, respectively, to the dissociation of the ore.
3.4.2. Effect of Reduction on the Magnetic Separation
The relationship between the grade of the concentrate and the conversion of Fe2O3 (in the hematite ore) at a given reduction temperature was determined by analyzing the total iron content; the results are shown in Figure 16.
As Figure 16 shows, the grade of the concentrate decreased sharply with the increasing amount of converted Fe2O3. This indicates that the Fe2O3 conversion played an important role in determining the grade of the concentrate.
According to the shrinking unreacted core model, the reduction of hematite (Fe2O3) to magnetite (Fe3O4) proceeded from the outer surface to the center of the hematite ore particle. The product layer (magnetite layer) thickened with the progression of reduction, and the unreacted gangue in the hematite ore particle was inevitably encapsulated by the new-generation magnetite (Fe3O4). This enclosure hindered the dissociation of the gangue and the newly generated magnetite (Fe3O4), and, as a result, the grade of the magnetic concentrate decreased with the increasing amount of converted Fe2O3.
The dependence of the iron recovery on the conversion of Fe2O3 at specific reduction temperatures was determined by analyzing the total iron content in both the hematite and concentrate ores; the results are shown in Figure 17.
As Figure 17 shows, the total iron recovery increased significantly with the increasing amount of converted Fe2O3. This indicates that the Fe2O3 conversion played an important role in the recovery of iron.
The amount of hematite reduced to magnetite increased with the increasing amount of converted Fe2O3 and resulted, in turn, in increased total iron recovery. Therefore, the conversion of Fe2O3 in the hematite ore had a significant effect on the total iron recovery and concentrate grade. In fact, the concentrate grade decreased, whereas the total iron recovery increased, with the increasing Fe2O3 conversion.
The grade of the concentrate reduced at 450°C for 30 min and milled in a rod mill for 15 min could be improved (to grade: 56.99%, iron recovery rate: 61.93%) by performing a simple magnetic separation process, using a working electrical current of 2 A (magnetic field intensity: 0.156 T); this process could also be used to improve the iron recovery rate of the concentrate (to grade: 52.06%, iron recovery rate: 90.5%), which was reduced at 400°C for 150 min and rod-milled for 15 min.
4. Conclusions
We conclude the following:(1)One kilogram of hematite (−2 mm) was reduced under a total volumetric gas rate of 120 L/h (gas mixture composition: 50% H2–50% N2). At reduction temperatures lower than 400°C, the rate of the reduction reaction of the hematite ore fines was controlled by the interfacial reaction on the core surface. However, at temperatures higher than 450°C, the reaction rate was controlled by the product layer diffusion. With an increasing reduction temperature, the average utilization of hydrogen increased initially and tended to a constant value thereafter.(2)The grade of the concentrate decreased and the total iron recovery increased with increasing Fe2O3 conversion. The grade of the concentrate, which was reduced at 450°C for 30 min and rod-milled for 15 min, was improved (to grade: 56.99%, iron recovery rate: 61.93%) through a simple magnetic separation process using a working electrical current of 2 A (magnetic field intensity: 0.156 T); this process was also used to improve the iron recovery rate (to grade: 52.06%, iron recovery rate: 90.5%) of the concentrate, which was reduced at 400°C for 150 min and rod-milled for 15 min.(3)During the reduction magnetization, the grain size of the new-generation magnetite increased initially and then decreases slightly and increased thereafter. The former and latter increases in the grain size were favorable and unfavorable, respectively, to the dissociation of the hematite ore. In addition, the transition of hematite (Fe2O3) to magnetite (Fe3O4) followed a shrinking unreacted core model.
Competing Interests
The authors declare that they have no conflict of interests.
Acknowledgments
The authors thank the colleagues from Taiyuan University of Technology for their assistance during this work. This work was sponsored by the Taiyuan Green Coke Clean Energy Co., Ltd. (China).