2020 Volume 60 Issue 12 Pages 2649-2658
The reducibility of calcio-wüstite (CW) originating from silico-ferrite of calcium and aluminum (SFCA) has been investigated from the perspectives of the morphology of SFCA and the concentration of FeO in CW. Two types of SFCA sample were prepared: columnar SFCA and acicular SFCA. The former was synthesized from reagent grade powders of Fe2O3, CaCO3, SiO2 and Al2O3, and contained columnar SFCA grains covered with slag. The latter was synthesized from iron ore and reagent grade CaCO3, and contained acicular SFCA grains, which were smaller than the columnar SFCA grains and had fine pores nearby. These samples were reduced in an XRD apparatus for high temperature use in a condition simulating a blast furnace where the oxygen partial pressure was controlled by CO–CO2 mixtures. The microstructures of the samples before and after reduction were observed by electron probe microanalysis (EPMA). XRD profiles indicated: (i) both SFCA samples were reduced to Fe via CW at 1000°C and (ii) acicular-SFCA-origin CW was reduced to Fe earlier than columnar-SFCA-origin CW, which suggests that the reducibility of acicular-SFCA-origin CW is higher than columnar-SFCA-origin CW. EMPA indicated: (i) most residual parts of acicular-SFCA-origin CW phase kept the morphologic feature of having fine pores as acicular SFCA during reduction as well and (ii) the FeO concentration in acicular-SFCA-origin CW was lower than that in columnar-SFCA-origin CW. Hence, it is concluded that the reducibility of SFCA-origin CW is dominated by the morphology of CW but not by the concentration/activity of FeO in CW.
In the current blast furnace ironmaking process, sinters are used as the main iron source. Sinters are composed of hematite (Fe2O3), magnetite (Fe3O4), calcium ferrite (mainly silico–ferrite of calcium and aluminum (SFCA)) and slag phases; among them, Fe2O3 and SFCA account for its large percentage. Sinters are normally reduced in the following two steps: (i) Fe2O3, Fe3O4 and SFCA are reduced to wüstite in the lower temperature zone and then (ii) wüstite is reduced to metallic iron (Fe) in the thermal reserve zone and in the higher temperature zone. It is worth mentioning that relevant gaseous components change depending upon the temperature zone. To provide stable operation of a blast furnace, it is very important to know characteristics of sinters.
There are various characteristics of sinters to be considered for the blast furnace operation, one of which is the reducibility of sinters;1,2,3) in particular, the reducibility of wüstite is the most important because it controls the overall reduction efficiency of a blast furnace.4,5,6) Wüstite in reduced sinters roughly falls into two types: 1) wüstite (denoted by FeO) reduced from Fe2O3 and Fe3O4 with less impurities and 2) calcio-wüstite (denoted by CW) reduced from SFCA; both FeO and CW have the rock-salt structure. There have been many studies on the reducibilities of FeO and CW in terms of chemical reaction and diffusion, and there is a consensus that the reduction of wüstite is promoted by CaO dissolved in wüstite7,8,9,10,11) from the following findings: reduction of FeO forms a dense iron layer in the outer part of the sample, which layer is a resistance to diffusion of reducing gas, whilst reduction of CW forms a porous iron layer which facilitates gas diffusion. In these previous studies used were dense bulky samples or powdery samples with the particle diameter of 50–500 μm. On the other hand, there is a report where the diameters of FeO and CW phases in reduced sinters are 30–40 μm or even smaller,12) in which case the effect of the iron layer structure on the reduction rate would be minimized.9) For example, the authors have measured the reducibilities of FeO and CW using high temperature X-ray diffraction (XRD) analysis on reduced sinter powders sieved into the particle diameter of 38–75 μm,13) and have found that FeO is more reducible than CW, which is in conflict with many previous studies but supports the results reported by Sakamoto et al.14)
Focusing on morphology, there are at least two types of CW: a) columnar CW with slag which is reduced from columnar SFCA with slag and b) acicular CW with fine pores which is reduced from acicular SFCA with fine pores.15) To examine the effect of the morphology of CW, the authors have also investigated the reducibility of acicular-SFCA-origin and columnar-SFCA-origin CW powders with 38–75 μm, which is larger than the size of CW phase, using high temperature XRD analysis, to examine the effect of the morphology of CW phase including its surrounding structure. The authors have found that acicular-SFCA-origin CW is reduced faster than columnar-SFCA-origin CW. This difference would stem from the morphology; namely, acicular-SFCA-origin CW has fine pores, which would facilitate direct contact between CW and reducing gas.15) This supposition is reasonable on the condition that gas diffusion to CW phase through its surrounding is the rate-controlling step of reduction. On the other hand, when the rate-controlling step of reduction is chemical reaction or both chemical reaction and gas diffusion, the reducibility may be worsened by the presence of gangue components such as CaO because it decreases the activity of FeO in CW.16) In actuality, CW in reduced sinters contains more than 5 mass% of CaO as solute at 1000°C.12) Thus, the reducibility of CW should be discussed in light of FeO concentration in CW as well as the morphology of CW.
Actual CW also contains gangue components such as SiO2 and Al2O3 except for CaO.12,16) According to the phase diagrams17,18,19) and the previous papers,20,21) the solubility limits of SiO2 and Al2O3 in wüstite are much lower than that of CaO. Hence, CW can be approximated to the binary system of CaO and FeO, and thereby it is assumed that the lattice constant of CW primarily depends on the concentration of CaO and also reflects the concentration of FeO. For example, Inami and Suzuki have prepared CW samples of FeO with different CaO concentrations equilibrated at 1000°C and an identical oxygen partial pressure, and measured the lattice constant of CW using XRD analysis at room temperature.22) They have reported that the lattice constant increases with an increase in CaO concentration in CW. Applying XRD for high temperature use, the concentration changes of CaO and FeO in CW can be monitored continuously. In addition, the lattice constant of CW is also dependent on Fe vacancy concentration and also on thermal expansion in case of high temperature measurement; accordingly, these effects must be considered carefully to evaluate the change in CaO concentration from the lattice constant. As a consequence, the present study aims to measure the concentration changes of CaO in columnar-SFCA-origin CW and acicular-SFCA-origin CW with different morphologies during the reduction process using high temperature XRD analysis to clarify the dominant factor affecting the reducibility of CW.
Samples used in this study were two types of synthesized samples: i) columnar SFCA covered with slag (‘Columnar SFCA’) and ii) acicular SFCA with fine pores (‘Acicular SFCA’). Table 1 gives nominal chemical compositions of each sample. The preparation method has been described in the previous reports,13,15) the outline of which is repeated as follows:
| Fe2O3 | SiO2 | CaO | Al2O3 | CaO/SiO2 | |
|---|---|---|---|---|---|
| Columnar | 65.0 | 7.8 | 23.3 | 3.9 | 3.0 |
| Acicular | 80.5 | 6.2 | 12.3 | 1.0 | 2.0 |
‘Columnar SFCA’ samples were synthesized according to the method conducted by Maeda and Ono.16) Reagent grade powders of Fe2O3, CaCO3, SiO2 and Al2O3 were weighed to the desired compositions and mixed in an alumina mortar. The mixed powders were calcined at 1000°C for 3 h, and then were crushed and mixed again. The calcining, crushing and mixing processes were repeated three times. About 12 g of the powder was then charged into an alumina crucible and melted in air at 1300°C for 0.5 h. Subsequently, the melt was cooled down to 1100°C at a cooling rate of 4.9°C/s, and then quenched into an ice-water mixture.
‘Acicular SFCA’ samples were synthesized using hematite iron ore with high SiO2 content and reagent grade CaCO3. The iron ore was crushed and sieved to obtain particles sized under 90 μm. The CaCO3 reagent was calcined at 1100°C in air for 24 h to obtain CaO. Powdery iron ore and CaO were weighed so that the CaO/SiO2 ratio was 2, and then were mixed in an alumina mortar. About 0.45 g of mixed powders was uniaxially pressed into a shape of disc using a piston cylinder die of 10 mm in diameter. The compaction pressure of 40 MPa was applied for 30 s. The prepared sample was heated in air at 1275°C for 10 min, and then were cooled down to room temperature in air.
Both SFCA samples were crushed into fine powders using an alumina mortar and a pestle, and were subjected to XRD analyses to identify major constituent phases before reduction. This XRD analysis was carried out with CoKα radiation at an accelerating voltage of 40 kV and a filament current of 250 mA. The optical system of a parallel beam was employed. The diffraction-scanning rate was 0.5°/min, and the scan range of 2θ was between 10° and 80°. Electron probe microanalysis (EPMA) was used to observe microstructures of the synthesized SFCA samples.
2.2. Reduction Experiment in High Temperature XRDThe samples prepared as above were crushed and sieved into powders with the particle diameter of 38–75 μm, and subsequently the samples were subjected to reduction experiments using a high temperature XRD apparatus. This experiment is also the same as in the previous work,13,15) the explanation of which is repeated for reference in the following. Figure 1 shows the experimental conditions of temperature and gas composition. The partial pressure of oxygen in the sample chamber was controlled using mixtures of CO–CO2, the ratio of which was adjusted depending on temperature to simulate a blast furnace condition. A sample was heated from room temperature to 1000°C with a heating rate of 5°C/min (Step I), held at 1000°C for 100 min (Step II) to simulate the condition of the thermal reserve zone, and then was heated again up to 1250°C at a heating rate of 5°C/min (Step III). The correction of the gas switching time was performed by estimating the time required for the gas supply as the dead volume divided by the gas flow rate (100 mL/min).
Experimental conditions of temperature and gas composition.
The CoKα line was used as the X-ray source. The optical system of a parallel beam was employed. The relative amounts of SFCA, Fe2O3, Fe3O4, CW, dicalcium-silicate (2CaO·SiO2 denoted by C2S) and Fe were evaluated from the peak areas around SFCA (311), Fe2O3 (104), Fe3O4 (400), CW (200), C2S (211) and Fe (100), which areas were derived from regression curves of the corresponding peaks to the Pseudo-Voigt function.15) To evaluate the change with time in the peak area more accurately, XRD analysis was made with a diffraction scan rate of 1°/min and a scan step of 0.02° in the 2θ ranges of 39.3°–41.6° and 47.2°–49.5°, respectively, for SFCA (311) and CW (200).15) In addition, the relative amount of CaO dissolved in CW was evaluated using values of 2θ for CW (200) and CW (111), which were derived from the regression curves of the peaks measured at 1000°C to the Pseudo-Voigt function. XRD analysis was made in the 2θ range of 33°–55°, covering the peaks of the main compounds, with a diffraction scan rate of 5°/min and a scan step of 0.02°.
2.3. Characterization of Reduced SamplesEPMA was applied to observe the microstructures of reduced samples and the chemical compositions of CW, for which five ‘Columnar SFCA’ samples and three ‘Acicular SFCA’ samples were employed. These samples were heated by an infrared gold image furnace up to 1000°C in the same experimental conditions as in the XRD experiments. The correction of the gas switching time was performed by estimating the time required for the gas supply as the dead volume divided by the gas flow rate (100 mL/min). The furnace was shut down in 155 min, 175 min, 195 min (in Step I), 240 min and 280 min (in Step II) after the onset of heating for ‘Columnar SFCA’ samples, and in 208 min, 220 min and 235 min (in Step II) after the onset of heating for ‘Acicular SFCA’ samples. After the furnace was switched off, the sample temperature decreased down to 500°C in 25 s and then to 150°C in another 70 s.
The chemical compositions of CW were measured by point analysis with a beam diameter of ca. 1 μm, an accelerating voltage of 15 kV and a probe current of 20 nA. For quantitative analysis, standard samples were used: Fe2O3 was provided for Fe, and a homogeneous glassy sample of 12.5(mass%)Fe2O3-41.7CaO-33.3SiO2-12.5Al2O3 for Si, Ca, Al and O.
Figure 2 shows XRD profiles measured at room temperature for a ‘Columnar SFCA’ (a) and an ‘Acicular SFCA’ (b) samples before reduction experiments.13,15) Both samples have the diffraction peaks due to SFCA (PDF database No. 46-37) and Fe2O3 (PDF database No. 01-073-2234), and the ‘Columnar SFCA’ sample also have unknown peaks at 2θ = 22.4°, 37.5°, 39.1° and 66.4°.
XRD profiles for a ‘Columnar SFCA’ (a) and an ‘Acicular SFCA’ (b) samples.
Figures 3(a) and 3(b) show typical back-scattered electron (BE) images of a ‘Columnar SFCA’ sample, respectively, taken near and away from the alumina crucible: Fig. 3(a) contains Fe2O3 and slag phases, whereas Fig. 3(b) contains columnar SFCA grains covered with slag where the size of the grains is typically ca. 20 μm in major axis and ca. 7 μm in minor axis. This difference in these constitutional phases may stem from the composition change due to alumina pick-up from the crucible. Thus, samples for reduction experiments were taken from the poriton away from the crucible.
Typical BE images of a ‘Columnar SFCA’ sample near (a) and away from alumina crucible (b).
Figure 4 shows a typical BE image of an ‘Acicular SFCA’ sample, where Fe2O3 and SFCA grains are observed with many fine pores. The columnar SFCA phase precipitated from liquid slag could not be observed in the samples. This SFCA grain is in an acicular (or plate-like) shape and is considered to be SFCA but not SFCA-I crystallographically. The size of the acicular SFCA grains is typically ca. 10 μm in major axis and ca. 0.5 μm in minor axis, which are smaller than those of the columnar SFCA grains. Hida et al.23) have reported that acicular shaped SFCA is generated through the reaction between solid hematite and CaO–Fe2O3 melt with high concentraion of CaO and low concentrations of SiO2 and Al2O3, which situation is also reflected in Fig. 4.
Typical BE image of an ‘Acicular SFCA’ sample.
Figures 5 and 6 show typical XRD profiles of a ‘Columnar SFCA’ and an ‘Acicular SFCA’ samples during the reduction reaction for Step I (Figs. 5(a) and 6(a)), Step II (Figs. 5(b) and 6(b)) and Step III (Figs. 5(c) and 6(c)), where temperature in Figs. 5(a), 5(c), 6(a) and 6(c) represents one at which each scan was started for Steps I and III, and time in Figs. 5(b) and 6(b) represents the total time after heating was started. Figure 7 shows the changes with reduction time in the relative amounts of SFCA, Fe2O3, Fe3O4, CW, C2S and Fe phases in a ‘Columnar SFCA’ (a) and an ‘Acicular SFCA’ (b) samples, estimated from Figs. 5 and 6, together with the temperature history. Here, the relative amount of CW phase includes FeO phase originating from Fe2O3. On the other hand, as for the ‘Acicular SFCA’ sample, the effect of FeO phase on the relative amount of CW has been eliminated after 205 min as described in the dscussion section 4.1. First, see Fig. 7(a) for ‘Columnar SFCA’. The Fe2O3 phase disappears around 100 min, at which the Fe3O4 phase starts to appear and then disappears around 160 min. At the same time, the columnar SFCA starts to disappear; instead, the CW peaks supperimposed by FeO peaks and C2S peaks strat to appear. The SFCA would primarily produce CW although there may be some contribution from Fe2O3-origin Fe3O4, and C2S would be originated from SFCA. The CW would not be directly produced from SFCA but through Fe3O416,24) and/or other intermediate compounds.25) However, such peaks have not been observed around 160 min probably because the stable resion for Fe3O4 and/or other intermediate compounds is too short on the present experimental condition.15) The reduction process from SFCA to CW should be studied more in the future. Afterwards, the columnar-SFCA-origin CW peaks supperimposed by FeO peaks start to disappear around 240 min; complementarily, the Fe phase starts to appear.15) The ‘Acicular SFCA’ sample in Fig. 7(b) basically shows the same behavior as the ‘Columnar SFCA’ sample. However, the acicular-SFCA-origin CW phases start to disappear to produce the Fe phase around 220 min, which is earlier than for the ‘Columnar SFCA’ sample.15) Thus, this suggests that acicular-SFCA-origin CW is more reducible than columnar-SFCA-origin CW. This difference in reducibility is discussed later.
Typical XRD profiles of a ‘Columnar SFCA’ sample during the reduction reaction for Step I (a), Step II (b) and Step III (c), where temperature on the right or the left side of Figs. 5(a) and 5(c) represents one at which each scan was started for Steps I and III, and time on the left side of Fig. 5(b) represents the total time after heating was started.
Typical XRD profiles of an ‘Acicular SFCA’ sample during the reduction reaction for Step I (a), Step II (b) and Step III (c), where temperature on the right or the left side of Figs. 5(a) and 5(c) represents one at which each scan was started for Steps I and III, and time on the left side of Fig. 5(b) represents the total time after heating was started.
Changes with time in the relative amounts of SFCA, Fe2O3, Fe3O4, CW, C2S and Fe phases for a ‘Columnar SFCA’ (a) and an ‘Acicular SFCA’ (b) samples together with the temperature history.
Figure 8 shows BE images of ‘Columnar SFCA’ samples after reduction where the samples were quenched at (a) 175 min in Step I, and (b) 240 min and (c) 280 min in Step II. The particle sample in Fig. 8(a) consists of columnar-SFCA-origin CW and slag, whilst the sample in Fig. 8(b) contains Fe as well, which seems to form around the perimeter of the CW. At 280 min in Step II, as shown in Fig. 8(c), the sample contains the ‘silicate’ phase in addition to columnar-SFCA-origin CW, slag and Fe. The ‘silicate’ should originate from gangue minerals contained in SFCA initially. Comparison with Fig. 7(a) suggests that the ‘silicate’ phase corresponds to C2S although the ‘silicate’ area was too small to quantitatively analyze the chemical compositions by EPMA. In addition, the ‘silicate’ phase is not clearly seen in Figs. 8(a) and 8(b), probably because the average atomic number (12.3) of C2S is close to that (17) of FeO, cmpared with that (26) of Fe. In BE images, the brighter area corresponds to the larger average atomic number of that area relative to adjacent areas. The precipitation of C2S at the reduction of SFCA was also observed by Maeda and Ono.16) With respect to Fig. 8(c), it is worth noting that the residual columnar-SFCA-origin CW seems to be covered with Fe and ‘silicate’.
BE images of ‘Columnar SFCA’ samples after reduction where the samples were quenched at (a) 175 min in Step I, and (b) 240 min and (c) 280 min in Step II.
Figure 9 shows BE images of ‘Acicular SFCA’ samples after reduction where the samples were quenched at (a) 208 min, (b) 220 min and (c) 235 min in Step II. The particle sample in Fig. 9(a) consists of CW and FeO phases, which have been reduced from acicular SFCA and Fe2O3, respectively. In contrast, the sample in Fig. 9(b) also contains Fe and ‘silicate’ phases in addition to acicular-SFCA-origin CW and FeO. Comparison with Fig. 7(b) suggests that the ‘silicate’ phase corresponds to C2S. The acicular-SFCA-origin CW and FeO phases have been reduced to Fe to some extent at 220 min. It is worth noting that most residual parts of acicular-SFCA-origin CW phase still keep the morphologic feature of acicular SFCA and have fine pores nearby although the ‘silicate’ phase penetrates some pores. At 235 min in Step II, as shown in Fig. 9(c), the sample contains acicular-SFCA-origin CW, Fe and ‘silicate’ phases only, and most of acicular-SFCA-origin CW and FeO phases are reduced to Fe. The ‘silicate’ phase is not clearly seen in Fig. 9(a) for the same reason as stated previously.
BE images of ‘Acicular SFCA’ samples after reduction where the samples were quenched at (a) 208 min, (b) 220 min and (c) 235 min in Step II.
Figure 10 shows the chemical compositions (mass%) of FeO, SiO2, CaO and Al2O3 in the CW phase for five ‘Columnar SFCA’ (a) and three ‘Acicular SFCA’ (b) samples, respectively, against heating time, where the error bars represent the standard deviations of five experimental values for each sample. With respect to ‘Columnar SFCA’ samples, the detected values showing exceptionally high gangue mineral contents, which may be due to ‘silicate’ phases, have been eliminated from the data in Fig. 10. It can be seen from Fig. 10(a) that CaO, SiO2 and Al2O3 concentrations in the columnar-SFCA-origin CW phase are almost constant independently of time, 4.4 mass%, 2.1 mass% and 0.3 mass%, respectively. As shown in Fig. 10(b), in contrast, CaO and SiO2 concentrations in the acicular-SFCA-origin CW phase increase from 7.2 mass% to 9.8 mass% and from 3.5 mass% to 5.0 mass% with time, respectively; complementarily, the FeO concentration slightly decreases. The concentration values of CaO and SiO2 may be overestimated by the presence of the ‘silicate’ phase around acicular-SFCA-origin CW; even so it would be possible that the CaO concentration in acicular-SFCA-origin CW increases with time since the solubility limit of CaO in iron-saturated CW is around 15 mass% according to the CaO–FeOx binary phase diagram.18) With respect to Fig. 10(b), it is worth noting that the FeO concentration in acicular-SFCA-origin CW is lower by about 5 mass% than in columnar-SFCA-origin CW although the Fe2O3 concentration in the ‘Acicular SFCA’ sample was higher in the nominal chemical composition given in Table 1.
Chemical compositions (mass%) of FeO, SiO2, CaO and Al2O3 for a ‘Columnar SFCA’ in Step I and Step II (a) and for an ‘Acicular SFCA’ (b) samples in Step II against heating time.
The concentration changes of CaO and FeO in CW can be monitored continuously from 2θ values of CW peaks using high temperature XRD analysis based on the fact that the lattice constant of CW increases with an increase in CaO concentration.10,22) On the other hand, as stated in Introduction, the lattice constant of CW is also dependent on Fe vacancy concentration as well as thermal expansion. Thus, these effects must be considered to evaluate the change in CaO concentration from the lattice constant. Additionally, Fe2O3 is inevitably contained in ‘Columnar SFCA’ and ‘Acicular SFCA’ samples as shown in Fig. 7, and we cannot deny the possibility that 2θ values of CW peaks are affected by superposed FeO peaks originating from Fe2O3.13) Accordingly, now focus on the CW (200) peaks at 2θ ~ 48° in Figs. 5(b) and 6(b) (Step II) around the time at which Fe peaks appear. Step II is the isothermal heating at 1000°C; accordintly, the change in 2θ values of CW in Step II is not affected by thermal expansion of lattice constant. Figure 11 shows close-ups of XRD profiles of Figs. 5 and 6 around CW (200) peaks in Step II (isothermal heating at 1000°C), together with regression curves of the CW peaks to the Pseudo-Voigt function for a ‘Columnar SFCA’ (a) and an ‘Acicular SFCA’ (b) samples. It can be seen from Fig. 11(a) that all the experimental peaks seem to be well regressed to each curve with high symmetry, which suggests that 2θ values of CW peak and FeO peak originating from Fe2O3 are close to each other. For an ‘Acicular SFCA’ sample in Fig. 11(b), comparison between the experimental profiles and the respective regression curves indicate that there are two peaks observed in the experimental profiles until 225 min. These two peaks would correspond to acicular-SFCA-origin CW and FeO, and the peak at the lower angle is assigned to CW according to the dependence of the lattice constant of CW on CaO concentration.10,22) Inspection of Fig. 11(b) implies that the reduction of FeO proceeds faster than CW, which is in accord with the results reported by Cai et al.13) and Sakamoto et al.14) In the present study, the sizes of CW and FeO phases are ca. 38–75 μm or even smaller. Thus, the difference between reducibilities of CW and FeO may be attributed to the FeO activity rather than the structure of the reduced iron layer.9)
Close-ups of XRD profiles around CW (200) peaks in Step II (isothermal heating at 1000°C) together with fitting curves regressed to the Pseudo-Voigt function for CW peaks on a ‘Columnar SFCA’ (a) and an ‘Acicular SFCA’ (b) samples.
Figure 12 shows the 2θ values of acicular-SFCA-origin CW and FeO peaks taken from Figs. 11(b) against heating time, where the dashed lines indicate the approximate times at which the Fe phases start to appear in Fig. 7(b). The 2θ value of FeO was visually determined because the analytical deconvolution of XRD profiles could not be carried out in Fig. 11(b). Both 2θ values monotonically decrease with time until ca. 220 min, after which the 2θ value of FeO becomes roughly constant whilst the 2θ value of acicular-SFCA-origin CW continues to decrease. These decreases should be affected by the changes in Fe vacancy concentrations in CW and FeO, which are both non-stoichiometric oxides.10,26,27,28) These oxides are reduced in Step II, and the Fe vacancy concentration decreases progressively with the reduction time. It has been reported that the decrease in Fe vacancy concentration causes the increase in lattice constant,22,29,30,31,32) which would lead to the decrease in the 2θ values of CW and FeO. Around 220 min, CW and FeO would coexist with Fe, as shown in Fig. 7(b); as a result, the 2θ value of FeO would become constant, as shown in Fig. 12.
2θ values of FeO and CW peaks for an ‘Acicular SFCA’ sample in Fig. 8(b) as a function of heating time.
Now focus on the consecutive decrease in the 2θ value of acicular-SFCA-origin CW after 220 min. Figure 13(a) shows changes with time in 2θ values of columnar- and acicular-SFCA-origin CW (200) peaks against reduction time together with the temperature history, and Fig. 13(b) is for columnar-SFCA-origin CW (111) peak. The arrows in these figures indicate the time at which the CW peak starts to decrease and the Fe peak starts to increase in Fig. 7. In Figs. 13(a) and 13(b), both 2θ values of columnar-SFCA-origin CW (200) and CW (111) peaks become constant around 240 min; on the contrary, in Fig. 13(a) the 2θ value of acicular-SFCA-origin CW (200) peak continues to decrease even after 220 min. After the time pointed by the arrow, both CWs are considered to coexist with Fe, and the Fe vacancy concentration should be constant. This should be true for both CWs; accordingly, the consecutive change after 220 min for the acicular-SFCA-origin CW would be relevant to the dissolution of CaO into CW, as shown in Fig. 10(b). As shown in Figs. 8 and 9, both CWs exist in contcat with the slag and/or ‘silicate’ phases; in particular, the acicular-SFCA-origin CW consists of smaller grains than the columnar-SFCA-origin CW and resultantly has larger specific surface area. As a result, gangue components such as CaO would diffuse into CW in the acicular-SFCA-origin CW more readily than in the columnar-SFCA-origin CW. However, the dissolution of CaO into CW might be associated with the reduction of CW into Fe instead of the diffusion of CaO from slag and/or ‘silicate’ phases into CW; CaO might be concentrated in the residual CW phases during the reduction process, resulting in the increase in CaO content of the acicular-SFCA-origin CW composed of smaller grains. The mechanism of the concentration changes of CaO in CW should be studied in the future. In any case, the above discussion on the XRD profiles of CW at high temperature has been made in term of Fe vacancy concentration and dissolution of CaO into CW assuming that the dissolution of SiO2 into CW does not change the lattice constanat.10) It can be considered that the concentration changes of FeO in columnar- and acicular-SFCA-origin CWs monitored by high temperature XRD analysis are consistent with the compositions measured by EPMA.
Changes with time in 2θ values of CW (200) peaks for a ‘Columnar SFCA’ and an ‘Acicular SFCA’ samples together with the temperature history (a) and of CW (111) peak of the ‘Columnar SFCA’ sample. The arrows in these figures indicate the time at which the Fe peaks appear in Fig. 7.
As shown in Fig. 7, it has been found that both a ‘Columnar SFCA’ and an ‘Acicular SFCA’ samples produced CW at roughly the same time after heating and were reduced to Fe via CW at 1000°C. It has also been found that acicular-SFCA-origin CW was reduced to Fe earlier than columnar-SFCA-origin CW, which suggests that the reducibility of acicular-SFCA-origin CW is higher than columnar-SFCA-origin CW. On the other hand, it can also be found from comparison between Figs. 10 and 13 that the FeO concentration in acicular-SFCA-origin CW is lower than that in columnar-SFCA-origin CW. This suggests that the reducibility of CW is not dominated by a thermodynamic factor such as the activity of FeO in CW. As a consequence, it is concluded that the reducibility of CW is dominated by a kinetic factor such as the morphology of CW. It is suggested that mass transfer occurs more readily in acicular-SFCA-origin CW, in Figs. 10(b) and 13(a) as well.
• The ‘Columnar SFCA’ sample contained columnar SFCA grains sized typically ca. 20 μm in major axis and ca. 7 μm in minor axis, and these particales are covered with slag.
• The ‘Acicular SFCA’ sample contained acicular SFCA grains sized typically ca. 10 μm in major axis and ca. 0.5 μm in minor axis. These grains were smaller than the columnar SFCA grains, and the acicular SFCA grains have fine pores nearby.
• From high temperature XRD analysis, it has been found that both samples produced CW at roughly the same time after heating and were reduced to Fe via CW at 1000°C. It has also been found that acicular-SFCA-origin CW was reduced to Fe earlier than columnar-SFCA-origin CW, which suggests that the reducibility of acicular-SFCA-origin CW is higher than columnar-SFCA-origin CW.
• From EMPA, it has been found that during reduction the residual parts of columnar-SFCA-origin CW seems to be covered with Fe and ‘silicate’ whilst most residual parts of acicular-SFCA-origin CW phase kept the morphologic feature of having fine pores as acicular SFCA during reduction as well. It has also been that the FeO concentration in acicular-SFCA-origin CW was lower than that in columnar-SFCA-origin CW.
As a consequence, it is concluded that the reducibility of SFCA-origin CW is dominated by the morphology of the CW but not by the concentration/activity of FeO in the CW.
