2015 Volume 55 Issue 3 Pages 616-622
X-ray diffraction analysis of a converter slag was performed by focusing on the lime-phase solid solution. Solid solutions of Ca1-xFexO and Ca1-xMnxO (0<x<1) were prepared by mechanochemical processing or high-temperature solid-state reaction, and the relationships between solid solubility x and lattice parameters were determined. Crystallized lime could be distinguished from undissolved lime by the shift of the diffraction angle associated with the formation of the solid solution. The effect of slag aging treatment with water vapor was examined by comparing the solid solubility x and the amount of lime phase in the aged and non-aged slags. It was confirmed that x affects the hydration behavior of the lime phase and that crystallized lime with high x tends to remain unchanged even after aging. Moreover, about two-thirds of the lime phase was confirmed to have been converted by hydration as a result of the aging process.
In steelmaking processes that refine pig iron into tough steel, calcined lime (CaO) is added to remove unwanted components, such as Si, P, and S, and steelmaking slag is produced as a by-product. Japan produces about fifteen million tons of steelmaking slag annually. This slag consists primarily of CaO and SiO2 and is effectively utilized as the base course material of roads, aggregates for asphalt concrete, and earthwork and ground improvement materials. However, part of the CaO in the slag remains in the unreacted form, i.e. as free lime, which expands to about double volume by hydration (CaO+H2O→Ca(OH)2). Therefore, fresh slags that contain free lime must undergo aging in order to prevent damage to the road or concrete. The aging treatments for slag stabilization are performed such that hydration by rain water is promoted by leaving the slag outdoors, or treating with steam or high-pressure steam in order to accelerate the reaction.1)
In general, slags containing high amounts of free lime have higher expansibility than their counterparts, which have lower CaO contents. The expansibility, however, varies with the particle size, physical state of the free lime, and the corresponding chemical composition. Therefore, estimating the expansibility solely as a function of the free lime content is difficult. The free lime in converter slag is typically classified into two categories namely, crystallized lime and undissolved lime. The former dissolves once and is re-deposited during solidification, while the latter remains undissolved during the steel refining process. In crystallized lime, FeO, MnO and, MgO divalent metal oxides dissolve into the CaO, thereby forming a solid solution. Hydration may be inhibited depending on the concentration of divalent oxides.2,3) Therefore, an understanding of the dependence of the hydration behavior, and hence the expansibility, on the solute content and the degree of solubility of divalent metal oxides in free lime is essential to selecting an appropriate aging treatment for the slag.
Pure CaO (undissolved lime) and its solid solution (crystallized lime) counterparts, which comprise the continuous area of the phase diagram, are regarded as the mineralogical lime phases. FeO and MnO are both highly soluble in free lime and, as such, Ca1-xFexO and Ca1-xMnxO (0<x<1) were synthesized in order to study the solid solutions. The relationships between solid solubility, indicated by mole fraction x, and the lattice parameter of the lime phases were examined by means of powder X-ray diffraction (XRD). The solid solutions were synthesized by both mechanochemical processing and high-temperature solid-state reaction methods. The solubility of FeO in CaO approaches the maximization of about 10 mol% at around 1373 K and decreases with decreasing temperature. Furthermore, the formation of the ferrite Ca2Fe2O5 phase implies that the solid solution becomes metastable at high levels of solubility. Mechanochemical processing also results in the formation of metastable phases at room temperature by imparting the mechanical energy,4) that is often used for the synthesis of solid solutions or impurity doping. This method is also thought to be favorable for homogenization of the composition. In contrast, homogeneous samples are not easily obtained from solid-state reaction processing.
The converter slag was measured by XRD after the synthesized solid solutions were analyzed. The solubility of the lime phase was estimated by examining and comparing the corresponding diffraction line with the lattice parameter of the solid solution.
As a pretreatment, CaO powder (Wako Pure Chemicals, 99.9%) was calcined at 1073 K for more than 8 h in air, in order to remove Ca(OH)2.
2.1.1. Ca1–xFexO(1) Mechanochemical Method
Specific ratios of FeO powder (Kojundo Chemical Lab., 99.9%) and CaO were mixed depending on the desired compositions (x). It should be noted that FeO is non-stoichiometric and, therefore for exactness, the chemical formula should be expressed as Fe1–σO. This paper will, however, use the simple FeO denotation. Each mixture of 0.3–0.6 g was placed in an 80 mL zirconia pot with 30 g of zirconia balls, each having a diameter of 10 mm. The pot was hermetically closed in an inert gas atmosphere by filling with 1 atom of pure Ar or N2 gas (>99.995%) in order to prevent oxidation of the sample. The sample was then milled at 800 rpm for 15 min using a planetary ball mill (Nagao System, PLANET M2-3F). Furthermore, in order to prevent unevenness of the component, the cap was opened and the powder that had adhered to the inside wall was removed and milled for another 15 min. XRD patterns of the materials were obtained using a powder X-ray diffractometer (MAC science, MXLabo, Cu Kα, 40 kV–30 mA, with graphite crystal monochromator), and the formation of solid solutions was verified. The lattice parameters were determined from the diffraction lines. The diffraction angle was calibrated using the angle of silicon powder (NIST SRM 640d, Standard reference material for X-ray metrology) that was determined in a separate measurement.
(2) High-temperature Solid-state Reaction
The synthesis was performed based on the method of Abbattista et al.5) Specific ratios of FeO and CaO were homogeneously mixed and the resulting mixture was pelletized. The pellet was fired in an infrared lamp heating furnace (ULVAC, MILA-3000) at 1383 K for 7 h under an inert gas atmosphere. The heat-treated material was removed from the hot furnace and rapidly quenched in liquid nitrogen. Since Ca2Fe2O5 is easily formed on the material surface by air oxidation, the material was measured by XRD after shaving off the surface with a file.
2.1.2. Ca1–xMnxOThe high-temperature solid-state synthesis method was also used to synthesize Ca1–xMnxO samples. CaO and MnO powders (Kojundo Chemical Lab., 99.9%) were mixed and subsequently pelletized. The pellet was fired at 1473 K for 2 h under an inert gas atmosphere.
2.2. Analysis of the Free Lime in Steelmaking SlagA converter slag, with a basicity (CaO/SiO2 ratio) of around 3.5, was supplied by a blast furnace steelmaker. The samples were conventionally aged at the plant by exposure to 100˚C water vapor for 2–4 d. The aged and non-aged samples were subsequently reduced (divided and extracted) and crushed to sizes below 74 μm. In addition, the corresponding chemical compositions (Table 1) were determined by X-ray fluorescence (XRF) using glass bead specimens by Dr. Inui (KOBELCO research institute) et al.
| Slag sample | T–Fe | SiO2 | CaO | Al2O3 | MgO | P2O5 | TiO2 | MnO | C | S |
|---|---|---|---|---|---|---|---|---|---|---|
| Non-aged | 16.72 | 14.61 | 48.91 | 2.19 | 2.96 | 2.574 | 0.24 | 1.86 | 0.46 | 0.075 |
| Aged | 19.51 | 12.48 | 45.08 | 2.43 | 3.40 | 2.754 | 0.17 | 2.17 | 0.80 | 0.05 |
| (mass%) | ||||||||||
In 1981, the slag committee of the Japan Iron and Steel Federation formulated a uniform method for analytical testing of free lime in steelmaking slags. This method employs extraction by ethylene glycol6) and is referred to as the EG method. It was reported that CaO and the lime phase in the slag are almost completely extracted by the EG method. Calcium in other mineralogical phases, such as calcium silicate and the calcium ferrite series are, however, barely extracted. Therefore, in this study, free lime was first analyzed by the EG method and the results were compared with those from XRD measurements.
In accordance with the uniform method, 0.50 g of each of the aged and non-aged converter slags was placed in a conical flask, and agitated with 20 mL of ethylene glycol (Wako Pure Chemicals, special grade) at 343 K for 30 min. The liquid and solid phases were separated by centrifugation at 4500 G for 30 min, and the calcium concentration in the liquid phase was determined. Ball milling was also used to extract the free lime. 0.50 g of the slag sample and 20 mL of ethylene glycol were placed in a zirconia pot with zirconia balls, and agitated at 700 rpm for 30 min in a planetary ball mill. The extraction rate was compared with that obtained using the conical flask.
XRF analysis was used to quantify the amount of calcium in the ethylene glycol liquid phase. Ethylene glycol solutions of ~3000 ppm CaO were prepared as standard solutions and put in sample cells for the liquid. XRF intensities were measured using XRF analysis equipment, ED-10 (Techno X Co., Ltd., W target, 7.2 keV–2.0 mA). Ca Kα intensities were collected by 100-s integrations and a calibration curve was obtained. Furthermore, the XRF intensities of the ethylene glycol samples were measured after extraction from the slag and the calcium concentrations were determined from the calibration curve. The amount of free lime in the slag was also calculated.
2.2.2. Free Lime Analysis by XRDAged and non-aged converter slag samples were measured by XRD and the constituent crystalline phases were identified. The measurements were repeated ten times and added up at an approximate diffraction angle of 2θ=37–38˚ where the 200 reflection, the most intense line of the lime phase, is observed. Moreover, the lattice parameters a of the lime phase were calculated from the angles of the 200 reflection and the values were compared to those of the solid solutions synthesized above.
XRD patterns of the Ca1–xFexO obtained by mechanochemical synthesis are shown in Fig. 1. Here, the solid solubility x corresponds to the composition of the raw powder mixture used in the synthesis. The patterns of the reagent materials, i.e., CaO and FeO are also shown. The diffraction lines have been shifted and the shifts vary based on the composition, i.e. the value of x, for each sample. The shifts in the diffraction lines confirm, therefore, that solid solutions were formed. The lime and wüstite phases were both observed for a composition of x=0.20. According to the pseudo binary phase diagram of CaO–FeO,7,8) the solid solubility limit of FeO in the lime phase is 12 mol% at 1373 K. Mechanochemical synthesis should also result in the formation of non-equilibrium phases which are not shown in the phase diagram. These phases form as a result of the localized high energy that is generated by friction and compression, which arise during the pulverization of solid matter. Non-equilibrium phases were, however, not observed in the current experiment. α–Fe shown in Fig. 1 was the impurity contained in the reagent FeO, which remained after the synthesis.
X-ray diffraction patterns for Ca1–xFexO (lime and wüstite phases) obtained by mechanochemical synthesis.
XRD patterns of Ca1–xFexO synthesized by high-temperature solid-state reaction are shown in Fig. 2. Peak shifts are observed which, as in the case of the mechanochemical samples, confirm the formation of the lime-phase solid solution. Mechanochemically synthesized samples consisting of lattice distortions, structural defects, and small crystallite sizes, which originate from the mechanical stresses, give rise to broad diffraction profiles. In contrast, owing to their high crystallinity, samples synthesized by high-temperature solid-state reaction exhibit sharp diffraction profiles. Small amounts of Ca2Fe2O5 and CaCO3 impurities, which were generated immediately after removing the samples from the furnace, were also observed in the XRD pattern.
X-ray diffraction patterns for Ca1–xFexO (lime phase) obtained by high-temperature solid-state synthesis.
The cubic lattice parameter a corresponding to each solid solubility x, as determined from the XRD patterns of Figs. 1 and 2, are shown in Table 2. The former is plotted as a function of the latter in Fig. 3. Although, as previously mentioned, the full width at half maximum (FWHM) of the diffraction lines and the crystallinity differed markedly, there was not much difference between the lattice parameters of the mechanochemical and high-temperature synthesized samples.
| x of Ca1-xFexO | Lattice parameter (nm) | Synthesis method |
|---|---|---|
| 0 (CaO) | 0.4817 (1) | — |
| 0.023 | 0.4802 (2) | SS |
| 0.040 | 0.4798 (1) | SS |
| 0.06 | 0.4788 (5) | MC |
| 0.063 | 0.4791 (2) | SS |
| 0.080 | 0.4785 (1) | SS |
| 0.088 | 0.4781 (3) | SS |
| 0.10 | 0.4776 (6) | MC |
| 0.104 | 0.4774 (2) | SS |
| 0.60 | 0.4501 (2) | MC |
| 0.65 | 0.4466 (3) | MC |
| 0.70 | 0.4439 (5) | MC |
| 1 (FeO) | 0.4321 (3) | — |
Lattice parameter a as a function of the solid solubility x of Ca1–xFexO. (a) 0≤x≤1 and (b) enlargement of the lime phase.
The nonstoichiometry (FeδO, δ<1) of the FeO raw material was estimated to be Fe0.973O from the observed lattice parameter a=0.4321 nm and the theoretical value of a = 0.3856+0.0478δ.9) If Fe defects were directly introduced during the formation of the solid solution, then the lattice parameter will vary with the ratio of defects. Consideration of nonstoichiometry is, therefore, important for the wüstite phase. However, Fe occupies only about 10% of the cation sties in the lime phase resulting, therefore, in a low ratio of defects. If approximately 3% of the Fe-defects in the FeO reagent transferred to the solid solution in the same proportion, then the defects in the cation site would be 0.3%. The defects result in an approximately 1 × 10−4 nm change in lattice parameter. This change falls within the error of the measurement and implies that the nonstoichiometry is small and has negligible influence on the lattice parameter of the lime phase solid solution. In Fig. 3, the lattice parameter decreases linearly with increasing solid solubility, which reflects the fact that the ionic radius of Fe2+ is smaller than that of Ca2+. The lattice parameter changes discontinuously at the boundary between the lime and wüstite phases, although it exhibits high linearity within the lime phase; in other words, it obeys Vegard’s law. Equation (1), obtained by the method of least squares, describes the aforementioned behavior of the lattice parameter.
| (1) |
For samples with x ≥ 0.60, diffraction lines of Ca(OH)2 were not detected in the XRD pattern of the mechanochemically synthesized solid solution, which was exposed to water vapor at 353 K for 3 h; this result means that hydrolysis was completely prevented. At x=0.10, diffraction lines of Ca(OH)2 and its carbonation product, CaCO3 were observed. It is apparent, however, that the hydration reaction was prevented to some extent since the diffraction intensities were weak compared to those at x=0. At x=0.06, the solid solution was completely hydrated and the diffraction lines from the lime phase were, accordingly, not detected.
3.1.4. Ca1–xMnxO Obtained by High-temperature Solid-state ReactionFigure. 4 shows the lattice parameter, a, plotted as a function of the solid solubility, x, for Ca1–xMnxO synthesized by high-temperature solid-state reaction. The lattice parameter a decreases linearly with increasing solid solubility x, since the Mn2+ ions are smaller than their Ca2+ counterparts. All proportional solid solutions that obey Vegard’s law were formed and, as such, Eq. (2), which expresses the linear dependence of the lattice parameter (a) on the solid solubility (x) was obtained.
| (2) |
Lattice parameter a as a function of the solid solubility x of Ca1–xMnxO.
Like FeO, MnO tends to be nonstoichiometric. However, nonstoichometry is not considered here because, as discussed in 3.1.3, free lime rich in CaO has a low density of defects. Ca1–xMnxO has been recorded in crystal structure databases. For the range of compositions considered, the lattice parameters determined in this work are in excellent agreement with those reported by Jay et al.10)
3.1.5. Condition for the Formation of Substitutional Solid SolutionThe starting components of the pseudo binary systems, i.e. CaO, FeO, and MnO all have the cubic rock salt-type structure, and, therefore, Ca1–xFexO and Ca1–xMnxO are substitutional solid solutions. The Hume Rothery rule for substitutional alloys states that substitutional solid solutions are easily formed when the ionic diameter of the solute and solvent ions differ by no more than 15%. However, the solid solutions in the present study are composed of oxides, and, hence, an exact determination of the ionic size is difficult. Nevertheless, the maximum difference in size will, to some extent, exceed 15%. The ionic radii for a coordination number of 6 were cited from the table of R. D. Shannon11) (Fe2+:0.061 nm, Ca2+:0.100 nm, Mn2+:0.083 nm) and the difference was calculated using Eq. (3). The resultant value for the CaO–FeO system, CaO is the solvent, was found to be 39%.
| (3) |
This value far exceeds 15%, which implies that the solid solution region is narrow. This, in turn, means that the solid solution is formed in a limited composition range at high temperatures, where an increase in the entropy of mixing effectively cancels the large positive enthalpy.
The lattice parameters in Fig. 3 change discontinuously between the lime and wüstite phases. The linear changes do not converge even when the lattice compression in the wüstite phase is considered from the viewpoint of non-stoichiometric defect structure. Lime and wüstite phases have different bonding manners and never mix freely with each other. On the other hand, since the difference between the ionic radii of Ca2+ and Mn2+ is smaller, these ions easily substitute each other and, therefore, the solid solution is formed in the entire composition range of Ca1–xMnxO. In the FeO–MnO system, which has no direct relation to the lime phase in the present study, the formation of all proportional solid solutions12) is also explained as resulting from the small difference in ionic radii.
Crystallized lime contained in the slag can be treated as a solid solution formed by dissolutions of FeO and MnO into CaO, if low-level MgO is ignored. FeO and MnO are sufficiently similar to result in the formation of all the corresponding proportional solid solutions. The lattice parameter of the lime phase represented by Eqs. (1) and (2) are very similar. Therefore, in this study instead of considering each divalent metal oxide of FeO, MnO, and MgO in the lime phase, they were considered together as divalent metal oxide species, MO. Hereafter, the solubility x of divalent metal oxides in the lime phase will be expressed by that of MO estimated using Eq. (1) and characterization of the Ca1–xMxO solid solution will be performed.
3.2. Analysis of Lime Phase in Converter Slag 3.2.1. Quantitative Determination of Free Lime Using Ethylene Glycol Extraction ProcedureThe XRF measurement of the Ca Kα intensity of the standard solutions prepared by dissolving the CaO reagent in ethylene glycol resulted in a linear calibration curve for concentrations in the range of ~3000 ppm. The calibration curve was used to determine the Ca concentrations, which were extracted by the ethylene glycol from the aged and non-aged slag samples. Table 3 shows the free lime content of the slag, calculated from the Ca concentration and the weight of the slag powder, and ethylene glycol used. The Ca content is expressed in terms of oxide (CaO) weight. Ethylene glycol extraction combined with XRF analysis allows for quick and easy determination of the Ca content. However, the extracted quantity significantly increased (Table 3) when ball milling was used as the extraction procedure. As was reported for an expansion test of converter slag by water immersion,13) hydration of the lime phase becomes difficult if it is surrounded by stable phases such as wüstite and ferrite. Since CaO extraction from slag into ethylene glycol is a solid-liquid extraction, it is influenced by the delay of material transfer depending on the size and shape of the slag particles, and the texture of the slag constituent solid. Therefore, a longer time elapses before an equilibrium state is reached. The agitation in the conical flask needs sufficient time to reach extraction equilibrium. If extraction equilibrium is not reached, then the free lime content will be underestimated. On the other hand, wet milling by ball milling results in finer slag powders and migration from the solid phase is promoted by increasing the surface and denudation of the lime phase.
| Slag sample | CaO concentration in ethylene glycol (ppm) | Amount of CaO extracted from 0.50 g of slag (g) | Amount of free lime in slag (mass%) | |
|---|---|---|---|---|
| Non-aged | flask | 1808 | 0.040 | 8.0 |
| ball mill | 2788 | 0.062 | 12.4 | |
| Aged | flask | 1395 | 0.031 | 6.2 |
The free lime content determined by extraction in the flask was compared in the aged and non-aged samples. The free lime content did not decrease significantly with aging. In addition to the CaO and lime phase, more than 50% of Ca(OH)2 and up to 10% of CaCO3 were extracted from the slag by the EG method.6) Therefore, it is believed that hydration species, stabilized by the aging treatment, are included in the value of the free lime content. Conversely, if the hydration is incomplete then it is possible that part of the lime phase remained. Crystallized lime which contains significant amounts of divalent metal oxides has particularly low hydration reactivity and is, therefore, not converted by hydration. If the lime phase has sufficiently low hydration reactivity and is permanently inactive, then it may be used as base course material or other resources. In this regard, if ethylene glycol extracts lime phase in which divalent-metal oxides are densely-doped, then the hydration reactivity tends to be overestimated by the EG method. Consequently, in order to make ethylene glycol extraction useful for hydration reactivity assessment of steelmaking slags, then the extraction condition in which only hydration-reactive CaO and lime phase are selectively separated must be determined. Other analytical methods may be required in order to subtract the amounts of Ca(OH)2 and the solid solution from the ethylene glycol extracted value.
3.2.2. Determination of Solid Solubility in Lime Phase Using XRD MethodIn addition to the lime phase, several crystalline phases were identified from XRD measurements of the converter slag namely, calcium carbonate CaCO3, calcium hydroxide Ca(OH)2, wüstite, ferrite Ca2Fe2O5, silicate β–Ca2SiO4, Ca2SiO4·H2O, and magnetite Fe3O4. As previously mentioned, amounts of Ca(OH)2 and CaCO3 are contained in the quantitative value of the lime phase that was determined by the EG method. However, XRD can completely distinguish between the lime phase and the other crystalline phases. In the case of Ca(OH)2, hexagonal-disc-shaped crystals of the trigonal (rhombohedral) system are typically oriented with their c axes parallel and, therefore, the 0001 reflection appears strongly at 2θ= around 18˚ (Cu Kα X-rays). Since this is the only reflection that appears in the low angle region, XRD allows for easy identification of the Ca(OH)2 phase. XRD patterns also confirmed that the aging treatment resulted in weaker and much stronger intensities of the lime and Ca(OH)2 phases, respectively compared to those of the non-aged samples.
As mentioned above, converter slag is a mixture of various materials including amorphous constituents, and, hence, the overlapping of diffraction lines makes analysis difficult. The strongest 200 reflection from the lime phase is, however, relatively distinct and was, therefore, examined in detail. Figure. 5 shows the XRD pattern which focuses on the 200 line of the lime phase that occurs at 2θ=37–38°. The 200 peak of the lime phase of the aged sample is shifted relative to that of its non-aged counterpart. This implies that part of the lime phase, which had low levels of solute (mostly undissolved lime), was transferred by the conversion to Ca(OH)2 during aging. On the other hand, as previously discussed in 3.2.1, the other part of the lime phase, that had high levels of solute (crystallized lime), remained after aging. Although the non-aged slag contains both crystallized and undissolved lime, the diffraction lines overlapped and formed a broad profile. This broad profile was treated as a single peak, and was fitted to the pseudo-Voigt function. The diffraction angle 2θ was determined from the peak position and the lattice parameters were calculated for both aged and non-aged samples. In addition, the solid solubility x of MO in the CaO matrix of the lime phase was determined using Eq. (1) and was found to be 0.04 and 0.11 for non-aged and aged samples (Table 4). These values are, so to speak, average solid solubilities, and the value for the non-aged sample is small owing to the presence of undissolved lime and lime containing low levels of solute. The survival, after aging, of the component with high levels of solute confirms, as was previously reported, that the hydration reactivity depends on the solid solubility and that the lime phase becomes insensitive to water with increasing solute.2,3) The exact reason for the suppression of the hydration reaction of lime by Fe2+ or Mn2+ solutes is unclear. FeO and MnO usually do not become hydrated. Moreover, the calculated change in enthalpy and the Gibbs energy for the hydration reaction are both much smaller than that of CaO hydration. Therefore, it is believed that the dissolving of FeO and MnO into CaO lowers the overall reactivity of hydration. Iguchi et al. measured the enthalpy changes for hydration of CaO–FeO and CaO–MnO solid solutions.14) They showed that FeO or MnO added in excess of certain amounts result in a decrease of the absolute values of enthalpy per mole of CaO. Water molecules become associated into a substance during hydration. Since the entropy of the association reaction is negative, a decrease of the exothermic energy results in a depression of reactivity. The small exothermic energy indicates that the solid solutions are relatively inert and stable. DV–Xα15) calculations were made for the solid solutions in the current work. The results revealed that Fe and Mn make bonds with O, which are more covalent than that of Ca–O thereby weakening the covalency of CaO. Ca, therefore, becomes more ionic in the solid solution than in CaO. This, in turn, results in a decrease in the energy levels of the electron-occupied orbitals and the electrons that contribute to chemical reactions become inactive.
X-ray diffraction patterns of aged and non-aged converter slags.
| Slag sample | Non-aged | Aged |
|---|---|---|
| 2θ200 (degree) | 37.487 (1) | 37.701 (9) |
| Lattice parameter, a(nm) | 0.4798 (4) | 0.477 (3) |
| Solid solubility, x(Ca1-xFexO) | 0.04 | 0.11 |
| Peak area of 200 reflection (counts) | 8702 | 3011 |
The background was subtracted from each of the curves fitted to the diffraction profiles of the aged and non-aged samples (Fig. 5) and the peak area (integrated intensity) was obtained (Table 4). The portion of the lime phase that remained after aging, was determined from the ratio of the peak areas of the aged and non-aged samples and was found to be 35%. The remaining portion of lime was also calculated from the amount obtained using the EG method (Table 3) and a value of 78% was obtained. This rather large value includes, as discussed in 3.2.1, amounts of Ca(OH)2. Discrimination of the lime phase by XRD proved that the hydration, by means of aging, progressed to about two-thirds completion. X-ray scattering power from the lime phase increases with substitution of Fe and Mn for Ca, since the 3d-transition atoms have relatively larger atomic scattering factors. Therefore, a correction, based on the solid solubility, is needed for a proper quantification. For example, the 200 diffraction intensity from the Ca1–xFexO lime phase for a solid solubility of x=0.11 is about 4.6% larger than that for x=0.04. This small difference has almost no effect on the matter being discussed here. The absolute quantity of lime phase in the slag was not estimated in the current XRD analysis, although quantification is possible using the standard addition method. The quantity of lime in the slag will be determined and verified in such a way as to examine the consistency of the value obtained by the EG method.
In this study, the free lime contained in a converter slag was analyzed using XRD by focusing on the diffraction line from the lime phase, which causes hydration swelling of the slag. Doping of crystallized lime with divalent metal oxides results in changes in the diffraction angle. These changes allow crystallized lime to be distinguished from undissolved lime. The difference in diffraction line profiles before and after aging, confirmed that hydration reactivity varies with the solute content in the solid solution. Moreover, crystallized lime tends to remain unchanged by the aging treatment. The change in the lime content of the slag by aging was successfully estimated. The major findings of the current study can be summarized as follows:
(1) Lime phase solid solution that simulates crystallized lime was synthesized. Ca1–xFexO solid solution (x≈0−0.10 and 0.60–1.0) was obtained by mechanochemical method and high-temperature solid-state reaction. In the case of Ca1–xMnxO, all proportional solid solutions were obtained by high-temperature solid-state reaction. Equations expressing the relationships between the solid solubility x and the lattice parameter a were determined.
(2) The solid solubility x of divalent metal oxide species (MO) in the lime phase (Ca1–xMxO) of the converter slag was estimated. Although lime phases of different solid solubility are mixed in the slag, representative values of x=0.04 and x=0.11 were obtained for the slag before and after aging, respectively.
(3) XRD measurement of the lime phase in the converter slags confirmed that undissolved lime and lime containing low levels of solute are easily hydrated by the aging treatment. However, crystallized lime, with particularly high levels of solute tends to remain after aging.
(4) The change of diffraction intensities confirmed that aging of the converter slag resulted in about two-thirds of the lime phase being converted by hydration.
This work was supported by the ISIJ research group in the technical section of analysis (technical development B-type). The authors greatly acknowledge helpful discussions with the members of the research group.
