Experimental Study on the Preparation of Cementing Materials by Direct Reduction Coupling of a Hematite-Carbon Base

: The reduction of iron in hematite and process coupling of cementing material generated from gangue components are explored in this paper, and a technical proposal for preparing directly reduced iron and cementing materials considering the processes of energy and material ﬂows is proposed. An experimental study preparing cementing materials, such as tricalcium silicate and dicalcium silicate, by roasting the components, was performed. In this study, hematite was used as the raw material and powdered carbon was added, as the reducing agent, with CaO; at the same time, the gangue components of iron ore were used as the principal raw materials for the process of directly reduced iron preparation by direct reduction of iron ore. The FactSage software package was used to perform thermodynamic calculations of the reduction of iron and its inﬂuence on the formation of tricalcium silicate and dicalcium silicate. The feasibility of the direct reduction of iron to elemental iron and preparation of cementing materials by roasting of gangue components under the studied thermodynamic conditions was discussed. Di ﬀ erent temperature control strategies were used to verify the reaction coupling test. The results showed that zero-valent iron could be produced by roasting and reducing hematite under certain experimental conditions, and cementing materials, such as tricalcium silicate and dicalcium silicate, could be produced simultaneously by reacting the gangue components with CaO. Fe 2 O 3 exerted an adverse e ﬀ ect on the formation of tricalcium silicate, and su ﬃ cient reduction of the iron was a precondition for the formation and stability of tricalcium silicate.


Introduction
Direct reduction iron smelting, a new ironmaking technology, has the following characteristics: no metallurgical coke, less environmental pollution, and simple protocols. It has a higher value, especially for processing low-grade, complex, symbiotic iron ore and metallurgical dust and sludge [1][2][3][4][5][6][7][8]. However, as with traditional blast furnace ironmaking technologies, the gangue components produced by the fine-grinding process of iron mining have not been applied to reasonable and high value-added utilizations as energy carriers, resulting in energy waste. Fully utilizing the component and energy-carrying properties of gangue components after fine grinding during iron mining and realizing the high value-added resource utilization of gangue components under the reduction conditions of iron are important means to achieving low-energy, efficient, green metallurgical production. Because of conditions of iron are important means to achieving low-energy, efficient, green metallurgical production. Because of the component and energy-carrying characteristics of gangue, it is possible to prepare the gangue components into Portland cement clinker at the same time as reducing the iron. Firstly, the gangue components are the same as the main raw materials of cement clinker. Secondly, the production of cement clinker involves numerous processes and technologies that are similar to direct iron reduction and extraction technologies: a rotary kiln is used as the production equipment; the main production processes are grinding, material mixing, and material roasting and grinding; the operating temperature of the equipment is around 1300 °C; and lime needs to be added [9][10][11][12][13]. It is a new approach to energy conservation, emission reduction, non-slag smelting, and high-efficiency resource utilization to realize the preparation of directly reduced iron and cementing materials through energy and material flow processes and the coupling of the generation processes of the reduced iron and cementing materials. It is also a novel, efficient method for comprehensive utilization of hematite, providing technical support for the low-cost production of directly reduced iron; efficient, comprehensive utilization of resources; energy-savings; and cost reductions.
Coupling the reduction of iron in hematite and formation of cementing materials from gangue components is explored in this study. Using hematite as the principal raw material, the FactSage software package was used to perform a thermodynamic calculation analysis of the reduction of iron and its influence on the formation of tricalcium silicate and dicalcium silicate. Experimental verification was performed.

Raw Material Analysis and Composition
The hematite used in this study was a concentrated powder supplied by a company in eastern China. For X-ray diffraction analysis, the hematite powder was ground into particles smaller than 200 mesh (74 μm). The X-ray diffractometer (X'Pert PRO, PANalytical, The Netherlands) was used, and the scan condition was 8°/min; the analysis results are shown in Figure 1. At the same time, the chemical composition of the hematite powder was analyzed by X-ray fluorescence spectroscopy (ZSX, PrimusⅡ, Rigaku, Japan), the results of which are given in Table 1. From the analysis of the X-ray diffraction (XRD) pattern, it can be seen that the main phase present in the hematite powder was Fe 2 O 3 . There were low X-ray diffraction peaks corresponding to SiO2 and Al2O3 phases; X-ray diffraction peaks arising from minerals containing elements such as Ca, Mg, Ti, Mn, S, and P were not evident.   From the analysis of the X-ray diffraction (XRD) pattern, it can be seen that the main phase present in the hematite powder was Fe 2 O 3 . There were low X-ray diffraction peaks corresponding to SiO 2 and Al 2 O 3 phases; X-ray diffraction peaks arising from minerals containing elements such as Ca, Mg, Ti, Mn, S, and P were not evident.
From Table 1 (values of the analysis of the fluorescence spectrum), it can be seen that the main phase present in the hematite powder was Fe 2 O 3 , and that the total iron content was 65.72%. There were small amounts of SiO 2 and Al 2 O 3 impurities as well; the SiO 2 content was 3.37% and the Al 2 O 3 content was 1.27%. There were a few other elements, such as Ca, Mg, Ti, Mn, S, and P. The process coupling of iron reduction and cementing materials generated by reaction of SiO 2 and CaO was calculated via thermodynamic calculation analysis, in accordance with the composition of the hematite powder. Thermodynamic calculation analysis was not conducted for the other elements, due to their low concentrations.
Based on the above analysis, the raw materials were apportioned before the experiment. The apportioning included the following two considerations: (1) the reducing agent added should completely reduce the iron; (2) the amount of CaO added should ensure sufficient combination of the SiO 2 and Al 2 O 3 for the generation of tricalcium silicate and tricalcium aluminate. The raw material apportioning calculated for the reduction roasting of hematite is shown in Table 2. All raw materials were ground into particles smaller than 200 mesh and apportioned according to the calculations, and reactant samples were prepared by sufficient mixing. The thermodynamic calculation analyses were conducted for the ratios of raw materials in the samples.

Thermodynamic Calculation Analysis
During the reduction roasting of the samples, the possible chemical reactions included the reduction of iron oxide; generation of calcium ferrite by the combination of iron oxide and CaO; generation of tricalcium silicate (alite), C 3 S (3CaO·SiO 2 ), and dicalcium silicate (belite), C 2 S (2CaO·SiO 2 ), by the combination of CaO and SiO 2 ; and generation of C 3 S by the combination of CaO and C 2 S. The equations for these reactions are shown in Equations (1)- (5). The reaction module of the FactSage software package (CRCT, Montreal, Canada; GTT-Technologies, Germany) was used to calculate the standard Gibbs free energies for all the possible reactions of the main phases in the reduction roasting process between 500 and 2000 • C at 10 • C intervals. Figure 2 shows the temperature (T) and standard Gibbs free energies (∆G) of the reactions.   Figure 2a shows that when the temperature was between 500 and 1450 • C, Equations (2)-(4) will proceed spontaneously, and tricalcium silicate, dicalcium silicate, and Ca 2 Fe 2 O 5 will be generated. The starting temperature for the reduction of iron in Equation (1) was 653 • C. However, when the temperature was below 780 • C, CaO and Fe 2 O 3 were consumed, producing calcium ferrite through Equation (2); as this was more favorable than the reduction of iron through Equation (1), it was not conducive to the reduction of iron. When the temperature was above 780 • C, Equation (1) occurred before Equation (2), reducing Fe 2 O 3 to iron. Therefore, a temperature above 780 • C is required for the reduction of Fe 2 O 3 .
By comparing the temperature versus standard Gibbs free energy curves for Equations (3) and (4), shown in Figure 2a, it can be seen that Equation (3) was more favorable than Equation (4) at temperatures below 1300 • C, and dicalcium silicate is easily generated. The formation of tricalcium silicate is more favorable when the temperature is above 1300 • C. It can also be seen from Figure 2b that Equation (5) only proceeds spontaneously when the temperature is between 1300 and 1800 • C, and tricalcium silicate can be generated through the combination of the dicalcium silicate and calcium oxide formed in the previous reactions. Figure 2a,b show that to generate cementing materials comprising predominantly tricalcium silicate, the temperature should be between 1300 and 1800 • C.
In order to investigate the phase composition of the reactants at high temperatures, when the iron was insufficiently reduced, the phase module of the FactSage software package was used to calculate and analyze the equilibrium phase diagram of the Fe 2 O 3 -CaO-SiO 2 ternary system at 1350 • C. This is shown in Figure 3.
Metals 2020, 10, x FOR PEER REVIEW 4 of 9 Figure 2a shows that when the temperature was between 500 and 1450 °C, Equations (2)-(4) will proceed spontaneously, and tricalcium silicate, dicalcium silicate, and Ca2Fe2O5 will be generated. The starting temperature for the reduction of iron in Equation (1) was 653 °C. However, when the temperature was below 780 °C, CaO and Fe 2 O 3 were consumed, producing calcium ferrite through Equation (2); as this was more favorable than the reduction of iron through Equation (1), it was not conducive to the reduction of iron. When the temperature was above 780 °C, Equation (1) occurred before Equation (2), reducing Fe2O3 to iron. Therefore, a temperature above 780 °C is required for the reduction of Fe2O3.
By comparing the temperature versus standard Gibbs free energy curves for Equations (3) and (4), shown in Figure 2a, it can be seen that Equation (3) was more favorable than Equation (4) at temperatures below 1300 °C, and dicalcium silicate is easily generated. The formation of tricalcium silicate is more favorable when the temperature is above 1300 °C. It can also be seen from Figure 2b that Equation (5) only proceeds spontaneously when the temperature is between 1300 and 1800 °C, and tricalcium silicate can be generated through the combination of the dicalcium silicate and calcium oxide formed in the previous reactions. Figure 2a,b show that to generate cementing materials comprising predominantly tricalcium silicate, the temperature should be between 1300 and 1800 °C.
In order to investigate the phase composition of the reactants at high temperatures, when the iron was insufficiently reduced, the phase module of the FactSage software package was used to calculate and analyze the equilibrium phase diagram of the Fe2O3-CaO-SiO2 ternary system at 1350 °C. This is shown in Figure 3.  Figure 3 shows that at 1350 °C, Fe2O3 reacts with CaO, generating Ca2Fe2O5, and coexists with C2S and C3S in the high-CaO content regions (regions 2 and 3 in Figure 3); here, Ca2Fe2O5 becomes an impurity in the cementing materials. When Fe2O3 has low concentrations or is absent (region 1 in Figure 3), the main phases present in the system are C3S and excess free CaO. Therefore, Fe2O3 should  Figure 3 shows that at 1350 • C, Fe 2 O 3 reacts with CaO, generating Ca 2 Fe 2 O 5 , and coexists with C 2 S and C 3 S in the high-CaO content regions (regions 2 and 3 in Figure 3); here, Ca 2 Fe 2 O 5 becomes an impurity in the cementing materials. When Fe 2 O 3 has low concentrations or is absent (region 1 in Figure 3), the main phases present in the system are C 3 S and excess free CaO. Therefore, Fe 2 O 3 should be sufficiently reduced to iron, by the roasting process, to prevent the formation of Ca 2 Fe 2 O 5 . In the regions of high Fe 2 O 3 and SiO 2 content (regions 4, 5, and 6 in Figure 3), there are low melting point liquid phases containing three components-namely, Fe 2 O 3 , CaO, and SiO 2 -and C 3 S is not generated.
In order to study the interaction between the reduced iron and the CaO and SiO 2 in the system, the FactSage software package was used to calculate and analyze the phase diagram of the Fe-CaO-SiO 2 ternary system, in which the iron accounts for 50% of the total mass. These results are shown in Figure 4.   Figure 4 shows that tricalcium silicate is an unstable compound. Dicalcium silicate can form stable tricalcium silicate through combination with calcium oxide (regions 2, 3, and 4 in Figure 4) only within a temperature range of 1300-1800 °C; when the temperature is below 1300 °C or above 1800 °C (regions 1 and 5 in Figure 4), C3S decomposes into C2S and free CaO-this is in line with the thermodynamic calculation results given in Figure 2b. At the same time, it can also be seen that the reduced iron had no obvious effect on the CaO-SiO2 system, meaning that iron can coexist with the CaO-SiO2 system. When the concentration of CaO was such that CaO/SiO2 < 1.40 and the temperature was between 1300 and 1800 °C, there were no C2S or C3S phases present in the system, and CaO mainly formed CaSiO3 and Ca3Si2O7 by combining with SiO2. When CaO/SiO2 was between 1.40 and 1.86, C2S was present in the system (regions 7 and 8 in Figure 4). When CaO/SiO2 was between 1.86 and 2.79, C2S coexisted with C3S (regions 3 and 4 in Figure 4). When CaO/SiO2 exceeded 2.79, the main phases in the system were C3S, free CaO, and coexistent iron (region 2 in Figure 4). Thus, CaO/SiO2 = 1.40 is the composition point for Ca3Si2O7, CaO/SiO2 = 1.86 is the composition point for C2S, and CaO/SiO2 = 2.79 is the composition point for C3S. Therefore, when cementing materials with a tricalcium silicate main phase are being prepared, the masses of CaO and SiO2 used in the blending process should have a ratio of CaO/SiO2 = 2.79, and the roasting temperature should be between 1300 and 1800 °C.
The equipment module of the FactSage software package was used to perform a theoretical analysis of changes in the reactant phase composition during the temperature ramping, with hematite as the raw material, powdered carbon as the reducing agent, and CaO as the additive, under a protective CO atmosphere. The ratio of the reactants used was as shown in Table 2. The phase change of the system between 500 and 1500 °C was analyzed, and the temperature difference was 50 °C. The Origin software package was used to prepare the plot. The theoretical reaction products, as determined by the calculations, are shown in Figure 5.  Figure 4 shows that tricalcium silicate is an unstable compound. Dicalcium silicate can form stable tricalcium silicate through combination with calcium oxide (regions 2, 3, and 4 in Figure 4) only within a temperature range of 1300-1800 • C; when the temperature is below 1300 • C or above 1800 • C (regions 1 and 5 in Figure 4), C 3 S decomposes into C 2 S and free CaO-this is in line with the thermodynamic calculation results given in Figure 2b. At the same time, it can also be seen that the reduced iron had no obvious effect on the CaO-SiO 2 system, meaning that iron can coexist with the CaO-SiO 2 system. When the concentration of CaO was such that CaO/SiO 2 < 1.40 and the temperature was between 1300 and 1800 • C, there were no C 2 S or C 3 S phases present in the system, and CaO mainly formed CaSiO 3 and Ca 3 Si 2 O 7 by combining with SiO 2 . When CaO/SiO 2 was between 1.40 and 1.86, C 2 S was present in the system (regions 7 and 8 in Figure 4). When CaO/SiO 2 was between 1.86 and 2.79, C 2 S coexisted with C 3 S (regions 3 and 4 in Figure 4). When CaO/SiO 2 exceeded 2.79, the main phases in the system were C 3 S, free CaO, and coexistent iron (region 2 in Figure 4). Thus, CaO/SiO 2 = 1.40 is the composition point for Ca 3 Si 2 O 7 , CaO/SiO 2 = 1.86 is the composition point for C 2 S, and CaO/SiO 2 = 2.79 is the composition point for C 3 S. Therefore, when cementing materials with a tricalcium silicate main phase are being prepared, the masses of CaO and SiO 2 used in the blending process should have a ratio of CaO/SiO 2 = 2.79, and the roasting temperature should be between 1300 and 1800 • C.
The equipment module of the FactSage software package was used to perform a theoretical analysis of changes in the reactant phase composition during the temperature ramping, with hematite as the raw material, powdered carbon as the reducing agent, and CaO as the additive, under a protective CO atmosphere. The ratio of the reactants used was as shown in Table 2. The phase change of the system between 500 and 1500 • C was analyzed, and the temperature difference was 50 • C. The Origin software package was used to prepare the plot. The theoretical reaction products, as determined by the calculations, are shown in Figure 5. From the calculation and analysis results shown in Figure 5, it can be seen that the reactions occur in the low-temperature region (<653 °C). The generated phases included Ca2Al2SiO7, CaCO3, Ca2FeSi2O7, and a small amount of reduced iron, and there are large amounts of the reactants C and Fe2O3. When the temperature exceeds 653 °C, Fe2O3 is largely reduced by C to iron, and most of the reactants C and Fe2O3 are consumed. When the temperature is 673 °C, C2S is produced and the intermediate product, Ca2FeSi2O7, is consumed. When the temperature is 787 °C, CaCO3 completely decomposes into CaO and CO2. When the temperature is between 787 and 1300 °C, the main phases are Fe, CaO, C2S, and Ca3Al2O6. When the temperature exceeds 1300 °C, CaO combines with C2S to form C3S; the other phases do not change, and the final products, under high-temperature reaction conditions, are iron, C3S, and Ca3Al2O6.
Based on the above analysis, the preparation of elemental iron, C3S, C2S, and other cementing materials by reduction roasting of hematite, after apportioning the reactants according to the thermomechanical analysis, is feasible. At the same time, in order to ensure sufficient reduction of the iron and effective formation of the cementing materials (such as C3S and C2S), the reaction temperature needs to be controlled in stages.
In order to prevent Fe2O3 and CaO combining to form Ca2Fe2O5, which affects the reduction of iron, the reaction temperature, for Fe2O3, should be between 780 and 1220 °C. This temperature should be maintained for a sufficient time to fully reduce the iron and prevent formation of the low melting point liquid phase (the melting point of fayalite is 1220 °C) from Fe2O3 and SiO2. By this stage, C2S has been generated by the combination of CaO and SiO2. To achieve sufficient reduction of the iron, the temperature should be raised for a second time, and C3S generated through the reaction of C2S with CaO (1300-1800 °C). Finally, the samples should be cooled by fast cooling. During fast cooling, the decomposition rate of tricalcium silicate is slow enough that it can exist as a metastable state at ambient temperature.

Analysis of Experimental Results
In this experiment, to verify the effects of the different reaction conditions on the results of the thermodynamic calculations, three different control processes were designed, as follows: (1) 8 g of sample was placed in a corundum crucible in an airtight, box-type resistance furnace. Without a From the calculation and analysis results shown in Figure 5, it can be seen that the reactions occur in the low-temperature region (<653 • C). The generated phases included Ca 2 Al 2 SiO 7 , CaCO 3 , Ca 2 FeSi 2 O 7 , and a small amount of reduced iron, and there are large amounts of the reactants C and Fe 2 O 3 . When the temperature exceeds 653 • C, Fe 2 O 3 is largely reduced by C to iron, and most of the reactants C and Fe 2 O 3 are consumed. When the temperature is 673 • C, C 2 S is produced and the intermediate product, Ca 2 FeSi 2 O 7 , is consumed. When the temperature is 787 • C, CaCO 3 completely decomposes into CaO and CO 2 . When the temperature is between 787 and 1300 • C, the main phases are Fe, CaO, C 2 S, and Ca 3 Al 2 O 6 . When the temperature exceeds 1300 • C, CaO combines with C 2 S to form C 3 S; the other phases do not change, and the final products, under high-temperature reaction conditions, are iron, C 3 S, and Ca 3 Al 2 O 6 .
Based on the above analysis, the preparation of elemental iron, C 3 S, C 2 S, and other cementing materials by reduction roasting of hematite, after apportioning the reactants according to the thermomechanical analysis, is feasible. At the same time, in order to ensure sufficient reduction of the iron and effective formation of the cementing materials (such as C 3 S and C 2 S), the reaction temperature needs to be controlled in stages.
In order to prevent Fe 2 O 3 and CaO combining to form Ca 2 Fe 2 O 5 , which affects the reduction of iron, the reaction temperature, for Fe 2 O 3 , should be between 780 and 1220 • C. This temperature should be maintained for a sufficient time to fully reduce the iron and prevent formation of the low melting point liquid phase (the melting point of fayalite is 1220 • C) from Fe 2 O 3 and SiO 2 . By this stage, C 2 S has been generated by the combination of CaO and SiO 2 . To achieve sufficient reduction of the iron, the temperature should be raised for a second time, and C 3 S generated through the reaction of C 2 S with CaO (1300-1800 • C). Finally, the samples should be cooled by fast cooling. During fast cooling, the decomposition rate of tricalcium silicate is slow enough that it can exist as a metastable state at ambient temperature.

Analysis of Experimental Results
In this experiment, to verify the effects of the different reaction conditions on the results of the thermodynamic calculations, three different control processes were designed, as follows: (1) 8 g of sample was placed in a corundum crucible in an airtight, box-type resistance furnace. Without a protective atmosphere, the temperature was continuously raised from room temperature to 1450 • C at a ramp rate of 10 • C/min; the product, A, was obtained after heating for 1 h and cooling to room temperature. (2) 8 g of sample was placed in a corundum crucible and then, placed in a tube-type resistance furnace. Under a protective CO atmosphere, with a flow rate of 60 mL/h, the temperature was continuously raised from room temperature to 1000 • C and, after heating for 2 h, continuously raised to 1450 • C at a ramp rate of 10 • C/min; the product, B, was obtained after heating for 1 h at 1450 • C and cooling to room temperature.
(3) 8 g of sample was reduced and roasted, and the same atmosphere and temperature ramp to 1450 • C were used as for sample B; the difference was that, after heating for 1 h at 1450 • C, the product, C, was obtained using the fast cooling method. Finally, XRD and scanning electron microscopy-energy-dispersive X-ray spectroscopy (SEM-EDS) analyses were conducted on the samples obtained. The same diffractometer was used as for the raw material analysis; the scanning electron microscope used was a VEGA II XMU from TESCAN, Czech Republic, and the EDS was type 7718 from Oxford Instruments, UK. The analysis results are shown in Figures 6 and 7.
Metals 2020, 10, x FOR PEER REVIEW 7 of 9 temperature. (2) 8 g of sample was placed in a corundum crucible and then, placed in a tube-type resistance furnace. Under a protective CO atmosphere, with a flow rate of 60 mL/h, the temperature was continuously raised from room temperature to 1000 °C and, after heating for 2 h, continuously raised to 1450 °C at a ramp rate of 10 °C/min; the product, B, was obtained after heating for 1 h at 1450 °C and cooling to room temperature. (3) 8 g of sample was reduced and roasted, and the same atmosphere and temperature ramp to 1450 °C were used as for sample B; the difference was that, after heating for 1 h at 1450 °C, the product, C, was obtained using the fast cooling method. Finally, XRD and scanning electron microscopy-energy-dispersive X-ray spectroscopy (SEM-EDS) analyses were conducted on the samples obtained. The same diffractometer was used as for the raw material analysis; the scanning electron microscope used was a VEGA II XMU from TESCAN, Czech Republic, and the EDS was type 7718 from Oxford Instruments, UK.   The iron oxide accounted for 70.58% of the mass of the sample, and the elemental iron in the theoretical reaction product accounted for 80.27% of the total mass. Therefore, strong peaks from elemental iron and iron oxide appear in the XRD patterns of the samples after the reaction, and the X-ray diffraction peaks of the other phases are generally low.
From the results of the XRD analysis, it can be seen that, for sample A, the reduction of iron could not be ensured, due to the lack of a protective atmosphere; the iron was present in the form of Fe2O3, which formed Ca2Fe2O5 through combination with CaO. The consumption of CaO resulted in temperature. (2) 8 g of sample was placed in a corundum crucible and then, placed in a tube-type resistance furnace. Under a protective CO atmosphere, with a flow rate of 60 mL/h, the temperature was continuously raised from room temperature to 1000 °C and, after heating for 2 h, continuously raised to 1450 °C at a ramp rate of 10 °C/min; the product, B, was obtained after heating for 1 h at 1450 °C and cooling to room temperature. (3) 8 g of sample was reduced and roasted, and the same atmosphere and temperature ramp to 1450 °C were used as for sample B; the difference was that, after heating for 1 h at 1450 °C, the product, C, was obtained using the fast cooling method. Finally, XRD and scanning electron microscopy-energy-dispersive X-ray spectroscopy (SEM-EDS) analyses were conducted on the samples obtained.   The iron oxide accounted for 70.58% of the mass of the sample, and the elemental iron in the theoretical reaction product accounted for 80.27% of the total mass. Therefore, strong peaks from elemental iron and iron oxide appear in the XRD patterns of the samples after the reaction, and the X-ray diffraction peaks of the other phases are generally low.
From the results of the XRD analysis, it can be seen that, for sample A, the reduction of iron could not be ensured, due to the lack of a protective atmosphere; the iron was present in the form of Fe2O3, which formed Ca2Fe2O5 through combination with CaO. The consumption of CaO resulted in reductions in the CaO/SiO2 and CaO/Al2O3 ratios. The X-ray diffraction peaks of phases with low The iron oxide accounted for 70.58% of the mass of the sample, and the elemental iron in the theoretical reaction product accounted for 80.27% of the total mass. Therefore, strong peaks from elemental iron and iron oxide appear in the XRD patterns of the samples after the reaction, and the X-ray diffraction peaks of the other phases are generally low.
From the results of the XRD analysis, it can be seen that, for sample A, the reduction of iron could not be ensured, due to the lack of a protective atmosphere; the iron was present in the form of Fe 2 O 3 , which formed Ca 2 Fe 2 O 5 through combination with CaO. The consumption of CaO resulted in reductions in the CaO/SiO 2 and CaO/Al 2 O 3 ratios. The X-ray diffraction peaks of phases with low calcium contents, such as CaAl 4 O 7 and Ca 2 SiO 4 , were obvious in the reaction products, and diffraction peaks of SiO 2 and Ca 3 Al 2 O 6 were also present, which was consistent with the analysis results shown in Figure 3. With the protection of a reducing atmosphere, sample B was heated at 1000 • C for 2 h, ensuring the complete reduction of iron, and then, at 1450 • C for 1 h, matching the formation conditions of C 3 S; however, C 3 S decomposed into C 2 S and CaO phases, due to the use of furnace cooling. Thus, the final reaction products of sample B were phases such as elemental iron, Ca 2 SiO 4 , CaO, CaAl 4 O 7 , CaAl 2 O 4 , and Ca 3 Al 2 O 6 ; this was consistent with the analysis results shown in Figure 4. Under a protective CO atmosphere, sample C was heated according to the same protocol as sample B, to realize the reduction of iron and formation of C 3 S and C 3 A phases, and fast cooling was used to avoid large-scale decomposition of C 3 S. The final reaction products were Fe, C 3 S, and Ca 3 Al 2 O 6 , and the decomposition products were Ca 2 SiO 4 and CaAl 2 O 4 , which showed substantially weaker X-ray diffraction peaks than the Ca 2 SiO 4 in sample B. At the same time, from the results of the SEM-EDS analysis of sample C, it can be seen that the product morphology consisted of many iron grains, with different diameters and cementing materials among them; this was consistent with the analysis results of the theoretical products shown in Figure 5.

Conclusions
(1) It is feasible to prepare elemental iron, C 3 S, C 2 S, and other cementing materials that can coexist at high temperatures by reduction roasting of hematite after apportioning the reactants using thermomechanical analysis. (2) The coupling reaction temperature needs to be controlled in stages. The reduction temperature of Fe 2 O 3 should be between 780 and 1220 • C, and the formation temperature of C 3 S should be between 1300 and 1800 • C. In order to obtain a stable C 3 S phase, the fast cooling method must be used for sampling. (3) Under the protective, reducing CO atmosphere, the temperature of the first stage is 1000 • C, and the heating time is 2 h; the temperature of the second stage is 1450 • C, and the heating time is 1 h. The fast cooling method is used for sampling; the final products of the hematite reduction roasting are Fe, C 3 S, and Ca 3 Al 2 O 6 , and the decomposition products are Ca 2 SiO 4 and CaAl 2 O 4 . (4) The reduction of iron has a great influence on the formation of C 3 S. When iron has not been sufficiently reduced, it will form calcium ferrite, and other phases, through combination of the CaO and SiO 2 ; this hinders the formation of C 3 S. Sufficient reduction of the iron is a precondition for the formation of C 3 S.

Funding:
The authors wish to express their thanks to the National Natural Science Foundation of China (51874224) and Project of young talents in basic research of Natural Science in Shaanxi Province (2014JQ7282) for the financial support of this research.

Conflicts of Interest:
The authors declare no conflict of interest.