The relationships of the compounds in the Mg–Ti–Al–O system in equilibrium with molten iron are investigated at temperatures ranging from 1873 to 1973 K, and the thermodynamic calculations are conducted in avoiding Al2O3 or spinel MgAl2O4 formation and for inclusion control. The equilibrium relations between the compounds (Mg2TiO4, MgTi2O4, or MgAl2O4) and the composition of solutes in steel are clarified. The conditions are shown that the transformation of the stable compound from Mg–Ti spinel to Mg–Al spinel occurs at Mg contents ranging from 1 to 10 ppm by mass. It is also found, on the Mg–Ti spinel, the stable compound is transformed from inverse-spinel Mg2TiO4 to normal-spinel MgTi2O4 at Ti = about 60 ppm by mass. In addition, the stable compound is transformed from MgTi2O4 to Ti2O3 at Ti = about 1000 ppm by mass.
Nonmetallic inclusions in steel are mainly composed of deoxidation products, and they adversely affect the toughness, fatigability, and ductility of steel because of their high melting points and hardness. For example, it is very important to avoid the formation of Al2O3 and spinel MgAl2O4 during the deoxidation of molten iron for high-grade wire and bearing steels. Several thermodynamic studies on the deoxidation equilibria of the Fe–Mg–Al–O and Fe–Ca–Mg–Al–O systems have been reported, and the phase equilibrium relationships among each oxide (or compound) and the composition of the solutes in molten iron has been clarified in the MgO–Al2O3 and CaO–MgO–Al2O3 systems.[1-3]
“Oxide metallurgy” is the control of inclusions and is used to improve the properties of steel. The toughness of steel can be improved by fine dispersed nonmetallic inclusions in the steel and by causing the fine inclusions to act as nuclei for ferrite formation in grains. Fine dispersed Ti oxides through complex deoxidation involving Ti are considered to be effective for the above-mentioned purpose. Determination of the equilibrium relationships between the oxide compounds in the MgO–Al2O3–Ti2O3 system and molten iron is important in inclusion control for avoiding Al2O3 or MgAl2O4 formation. However, the study on this system is very limited, and the phase equilibrium relations have not been clarified yet.
In this work, the equilibrium relationships between the deoxidation products (MgTi2O4, Mg2TiO4, or MgAl2O4) and the composition of solutes in steel are investigated. The conditions that MgTi2O4 or Mg2TiO4 forms instead of MgAl2O4 nor Al2O3 are cleared up in the Fe–Ti–Mg–Al–O system at temperatures ranging from 1873 to 1973 K.
2.1 Equilibrium Experiments at 1973 K
Following equilibrium experiments at 1973 K were done for the condition of low Ti content in steel: Ti = 22–209 ppm by mass, in order to know the equilibrium relations between the deoxidation products (MgTi2O4, Mg2TiO4, or MgAl2O4) and steel and in order to obtain the phase stability diagram of those deoxidation products.
An induction furnace was used for the deoxidation of high-purity electrolytic iron. The electrolytic iron in an alumina crucible was preliminary deoxidized in H2 gas for over an hour at 1873 K and was used in the following experiments. The oxide mixtures were made by mixing the reagent grade of TiO2 (purity > 99%), MgO (purity > 99%), and Al2O3 (purity 99.6%). The initial composition of oxide mixtures is shown in Table 1.
|No.||Crucible||Initial composition of oxide mixtures (% by mass)|
Sixteen grams of the deoxidized iron and the reagent grade of TiO2, MgO, and Al2O3 oxides totally weighing 8 g were put in a magnesia or an alumina crucible (OD 25 mm, ID 20 mm, height 35 mm), and the crucible was inserted in a carbon crucible with a carbon lid, as shown in Figure 1a. The experiments were carried out by using a multipurpose electric resistance furnace, of which the heating element is carbon. After the sample was set in the furnace, the inner space of the furnace was decompressed to 8 Pa and was switched to an inert atmosphere by introducing Ar gas. By carrying out this operation repeatedly two times, the inside of the furnace was kept to be Ar gas atmosphere (0.011–0.013 MPa). Then, the sample was heated to 1823 K at a heating rate of 9.5 K min−1 and then to 1973 K at a heating rate of 3.75 K min−1. The sample was equilibrated in the furnace by being held for 3-h in Ar gas atmosphere at 1973 K. The oxide mixtures were completely melted at 1973 K, so that the metal sample contacted with the magnesia or the alumina crucible directly. The sample was quickly cooled in the furnace by switching off the button of the power supply. After each experiment, metal and oxide mixtures in the crucible were divided, and the Ti, Mg, and Al contents of metal and oxide mixtures and the O content of metal were analyzed by an inductively coupled plasma (ICP) emission spectrometry and an inert gas impulse infrared absorption spectroscopy, respectively.
The compound in equilibrium with metal and oxide mixtures was identified for each sample as follows: The oxide compound layer newly formed at the interface of metal and magnesia (or alumina) crucible was mainly identified. The vicinity of the metal interface was analyzed by using FE-SEM/EDS. In the FE-SEM/EDS analysis, the point analysis was carried out from the interface of metal/crucible at the bottom part of sample to the inside of the crucible at the interval of a few µm, and the contents of Mg, Ti, Al, and Fe were analyzed. The oxide compound layer, which adhered to the metal surface was also analyzed by X-ray diffraction (XRD).
2.2 Equilibrium Experiments at 1873 K
Following equilibrium experiments at 1873 K were done for the condition of high Ti content in steel: Ti = 420–1960 ppm by mass, in order to obtain the equilibrium relations between the oxides and steel under the condition of Ti2O3 saturation.
At first, Ti2O3 powder was made using reagent grade TiO2 (purity > 99%) and Ti (purity 98%) by mixing them in a molar ratio of 3:1. The mixture was put in an alumina crucible and the crucible was put in a carbon crucible with a lid and inserted in an electrical resistance furnace. The sample was held in the furnace for 18-h at 1023 K under Ar gas atmosphere. After removal from the furnace, the compound was cooled under Ar gas, and the formed Ti2O3 was then crushed to a powder before being used for the following experiments.
The Ti2O3 powder, reagent grade MgO (purity > 99%) and reagent grade Al2O3 (purity 99.6%) were mixed to obtain a total weight of 12 g. The initial composition of the oxide mixtures is shown in Table 2. The oxide mixture was at first put in an alumina crucible (OD 25 mm, ID 20 mm, height 35 mm), as shown in Figure 1b. High-purity electrolytic iron weighing about 6 g, which was deoxidized with H2 gas at 1873 K using an induction furnace, was then added to the oxide mixture. The oxide mixtures were not melted at 1873 K, so that the metal sample did not contact with the alumina crucible directly, as shown in Figure 1b.
|No.||Crucible||Initial composition of oxide mixtures (% by mass)||Composition of metal [ppm by mass]||Equilibrium phase|
The alumina crucible was put in a carbon crucible with a lid and was inserted in an electrical resistance furnace. The sample was heated to 1273 K at a heating rate of 6.7 K min−1 under Ar gas and then the temperature was maintained for 1 h. The sample was then heated to 1873 K at a heating rate of 6.7 K min−1 under Ar gas and kept at that temperature for 13 h. The sample was then removed from the furnace and cooled in an Ar gas stream. After each experiment, the oxide compound layer newly formed at the oxide/metal interface were identified by XRD.
The Ti and Mg contents of the metal were analyzed by ICP. The O content of the metal was analyzed by inert gas impulse infrared absorption spectroscopy. In order to suppress the effects of inclusion entrainment, the samples were analyzed more than twice for each element, and the content of each element was determined from the average of two or more analyzed values.
3.1 Results of Equilibrium Experiments at 1973 K
The composition of oxide mixtures after the experiments is shown in Table 3 and is plotted on the TiOX–MgO–Al2O3 diagram at 1973 K in Figure 2, which is described by regarding the value for X of TiOX as to be 2. From Figure 2, it is clarified the liquidus curve of the TiOX–MgO–Al2O3 system in equilibrium with molten iron at 1973 K. The liquidus curve described from the experimental results is much different from the calculated on from Fact Sage 5.4.1, which is shown by the dotted curve in Figure 2. One of the causes of this discrepancy is considered to be the solid solution of MgAl2O4. It is known that MgAl2O4 has a relatively wide solid solution range at high temperature.[2, 3] Taking the solubility of Al2O3 or MgO in MgAl2O4 into consideration, the phase diagram is described as shown in Figure 2 at 1973 K. The composition of metal after the experiments is also shown in Table 3. In the present experiments, the Ti content of molten iron is from 22 to 209 [ppm by mass]. In this Ti content range, the kind of Ti oxides is reported to be Ti2O3, Ti3O5, or Ti4O7 in Fe–Ti–O system at 1873 K. Accordingly, although the coexistence of Ti2O3(Ti3+) and TiO2(Ti4+) in the oxide mixture phase is considered in this work, the experimental results are plotted by regarding all the Ti oxides as to be TiO2 in Figure 2, as a matter of convenience.
|No.||Composition of oxide mixtures (% by mass)||Composition of metal [ppm by mass]||Equilibrium phase|
On the identification of the equilibrium oxide phase, as examples, the results of the FE-SEM/EDS analysis for No. A11, A22, and A30 are shown in Figure 3a–c, respectively. The analysis points are also shown in Figure 3a–c. From Figure 3a and b, the equilibrium phases for Nos. A11 and A22 are considered to be normal-spinel Mg2TiO4 and inverse-spinel MgTi2O4, respectively. From Figure 3c, the sample of No. A30 is expected to be in equilibrium with both MgAl2O4 and Al2O3 (or Al2TiO5).
From the results of XRD, each equilibrated phase is also confirmed. The XRD patterns for both normal-spinel MgTi2O4 and inverse-spinel Mg2TiO4 are almost identical, and it is difficult to distinguish MgTi2O4 from Mg2TiO4 from the XRD. With regard to the experiments of No. A1–A14, which are conducted by using magnesia crucible, the formation of Mg2TiO4 is confirmed from the results of FE-SEM/EDS, in which the content of Mg is about twice as much as Ti, as shown in Figure 3a (No. A11). On the other hand, with regard to No. A15–A23, the formation of MgTi2O4 is confirmed, as shown in Figure 3b (No. A22). Through these analyses, it is judged that the samples of No. A1–A14 and No. A15–A23 are in equilibrium with Mg2TiO4 and MgTi2O4, respectively, in this work. In addition, we found that the Ti content for samples A1–A14, with equilibrium phase of Mg2TiO4, is generally higher than that for samples A15–A23, with equilibrium phase of MgTi2O4.
With regard to the experiments, which are conducted by using alumina crucible, the samples of No. A25, A31, and A32 are identified to be in equilibrium with MgAl2O4. On the other hand, the coexistence of Al2O3 or Al2TiO5 with MgAl2O4 is identified for the other samples of No. A24–A32.
3.2 Results of Equilibrium Experiments at 1873 K
We initially investigated the equilibrium compounds at higher Ti2O3 contents in the MgO–Al2O3–Ti2O3 system at 1873 K by XRD. From the results, the equilibrium experiments with molten iron were done at 1873 K at the two compositions: one is the Ti2O3 and MgTi2O4 doubly saturated composition, and the other is the Ti2O3 and MgA12O4 doubly saturated composition. The equilibrium relationships between the oxide compounds with molten iron was investigated. After the experiments, the Ti, Al, Mg, and O contents in the molten iron for each sample are shown in Table 2. The equilibrium phases are investigated by XRD and are also shown in Table 2. We found that the Al content for samples B1–B6, with equilibrium phases of Ti2O3 and MgAl2O4, is higher than that for samples B7–B11 that have equilibrium phases of Ti2O3 and normal-spinel MgTi2O4. Inverse-spinel Mg2TiO4 is not found as an equilibrated phase. In this work, Mg2TiO4 is considered to be unstable because of the higher Ti contents of 420–1960 ppm by mass. Overall, the Mg content is low and ranges from 1.3 to 8.4 [ppm by mass].
4.1 Determination of Thermodynamic Data on the Equilibrium of Normal-Spinel MgTi2O4 and Molten Iron at 1973 K
where and ai (i = Mg, Ti, or O) denote the activity of MgTi2O4 relative to the pure solid and the activity of i relative to the dilute solution of 1% by mass, respectively. The activity of MgTi2O4 is set to be unity by assuming the existence of the pure solid. The values for the equilibrium constant of Equation (1) and the interaction parameter between Mg and Ti in molten iron are determined from the present experimental results at 1973 K. The activity of Mg, Ti, and O in Equation (2) can be rewritten as Equation (3), using the activity coefficient and the concentration in molten iron. The activity coefficient can be written by Equation (4) from Wagner's equation, using the interaction parameters and the solute contents of molten iron
where [% i, j, or k] denotes the content of component i, j, or k in molten iron [% by mass], fi the Henrian activity coefficient of component i relative to the dilute solution, the first and the second order interaction parameters of i on j in molten iron, respectively. Equation (5) is derived from Equation (2), using Equation (3) and (4)
where the values for the interaction parameters has been already known except and . Equation (5) is rearranged by transposing log K(1) and the terms including and to the right-hand side, and by using Equation (6), which is a conversion equation of the interaction parameter, Equation (7) is derived
The values for the left-hand side of Equation (7) can be determined by substituting each interaction parameter and the experimental results of each solute content of molten iron. The values for the left-hand side of Equation (7) are plotted against ([%Ti] + 2 × (47.9/24.3)[%Mg]) in Figure 4 in accordance with Equation (7). From the slope and the intercept in Figure 6, the values for log K(1), , and are determined, respectively, as follows:
The thermodynamic data used in this work are summarized in Table 4. The data are usually given in the literatures as the values at 1873 K. Accordingly, the values are converted to those at 1973 K by using Equation (9) 
where T and denote the temperature (K) and the interaction parameter of i on j in molten iron at T (K), respectively.
|i||j||(1873 K)||(1973 K)||Refs.|
|Ti||Ti||0.048||0.046||Janke and Fischer|
|Al||0.0037||0.0035||Yanchang and Changzhen|
|O||−1.619||−1.486||Pak et al.|
|Al||Ti||0.0040||0.0038||Yanchang and Changzhen|
|Al||0.038||0.036||Satoh et al.|
|O||−1.867||−1.906||Satoh et al.|
|Mg||−0.047||0.016||Satoh et al.|
|0||−0.602||−0.203||Satoh et al.|
|O||Ti||−0.541||−0.496||Pak et al.|
|Al||−1.105||−1.128||Satoh et al.|
|Mg||−0.396||−0.133||Satoh et al.|
|O||−0.175||−0.128||Satoh et al.|
4.2 Stable Conditions of Normal-Spinel MgTi2O4, Inverse-Spinel Mg2TiO4 and MgAl2O4 at 1973 K
In order to investigate the stable conditions of normal-spinel MgTi2O4, inverse-spinel Mg2TiO4, and MgAl2O4, the calculations are conducted, using the thermodynamic data derived in Section 'Determination of Thermodynamic Data on the Equilibrium of Normal-Spinel MgTi2O4 and Molten Iron at 1973 K'.
The reactions of MgTi2O4, Mg2TiO4, and MgAl2O4 with molten iron are expressed by Equation (1), (10), and (12), respectively. The equilibrium constant for each reaction is given as Equation (8), (11),[9, 12-15] and (13), respectively
The doubly saturated curve with MgTi2O4 and MgAl2O4 can be described by solving Equation (5) and (15), simultaneously. Similarly, the doubly saturated curve for MgTi2O4 and Mg2TiO4 or Mg2TiO4 and MgAl2O4 can be described by solving Equation (5) and (14) or Equation (14) and (15), respectively. Here, we assumed that MgTi2O4, Mg2TiO4, and MgAl2O4 are immiscible each other. The relationship among the four solutes (Ti, Mg, Al, and O) in molten iron needs to be considered. By fixing the content of one component, the relationship among the other three components can be determined from the two equilibrium equations.
The relationships of Mg and Al contents of molten iron are described in Figure 5-7, which show the stable region of MgTi2O4, Mg2TiO4, or MgAl2O4 when Ti = 30, 60, and 150 ppm by mass, respectively. It is known that MgAl2O4 has a relatively wide solid solution range at high temperature. The solubility of Al2O3 in MgAl2O4 is rather large, and the composition of the spinel solid solution coexisted with Al2O3 is reported to be 64.2 mol% Al2O3 at 1900 K. The activity of MgAl2O4 coexisted with Al2O3 is reported to be 0.56 at 1948 K by assuming the ideal solid solution. It is also reported that stoichiometric MgAl2O4 exhibits a negative deviation from ideality in the spinel solid solution and the activity of MgAl2O4 at Al2O3 saturation is 0.47 at 1873 K. From these facts, the activity of MgAl2O4 is assumed to be 0.5 in this work, and the above calculations are conducted. The experimental results are also plotted in Figure 5-7, of which the Ti content ranges are 22–32, 42–88, and 92–209 ppm by mass, respectively. The numerical values in these figures show the experimental results of the O content of molten iron [ppm by mass]. The O contents of several samples are not much in agreement with the calculated iso-oxygen curves. One of the causes of this discrepancy is considered to be the sensitivity of the O content to the Ti content of molten iron, which is divided into three levels (Ti < 40 ppm by mass, 40 < Ti < 90 ppm by mass, 90 ppm by mass < Ti) in this work.
From Figure 5-7, the equilibrium relations between the compounds (Mg2TiO4, MgTi2O4, or MgAl2O4) and the solute contents in steel are clarified. It is shown the transformation of the stable compound from Mg–Ti spinel to Mg–Al spinel at Mg contents ranging from 1 to 10 ppm by mass. It is also found, on the Mg–Ti spinel, the stable compound is transformed from MgTi2O4 to Mg2TiO4 at Ti = about 60 ppm by mass at 1973 K.
4.3 Stable Conditions of MgTi2O4, MgAl2O4, and Ti2O3 at 1873 K
Taking the temperature dependence of the equilibrium constant for the reaction of MgTi2O4 with molten iron expressed by Equation (10) into consideration from the literatures,[8, 9, 16] the equilibrium constant at 1873 K is given by Equation (16)
The reactions of MgAl2O4 and Ti2O3 with molten iron are expressed by Equation (12) and (18), respectively. The equilibrium constant for each reaction at 1873 K is given as Equation (17)  and (19), respectively
The relationship of Mg and Al contents of molten iron is described in Figure 8, which shows the stable region of MgTi2O4, MgAl2O4, or Ti2O3 when Ti = 1000 ppm by mass. Here, the stable compound is transformed from MgTi2O4 to Ti2O3 at Ti = about 1000 ppm by mass. It is not necessary to consider the formation of Mg2TiO4, because the Mg2TiO4 compound is unstable at Ti > 1000 ppm by mass. The experimental results are also plotted in Figure 8, of which the Ti content range is 890–1960 ppm by mass. The numerical values in the figure show the experimental results of the O content of molten iron [ppm by mass].
The relationship between the Al and Ti contents of molten iron was calculated at 1873 K. The phase stability region of MgTi2O4, MgAl2O4, or Ti2O3 when Mg = 1 or 10 ppm by mass is shown in Figure 9 together with our experimental results. The numerical values near the open and closed circles show the O content [ppm by mass]. The calculated iso-oxygen curves are also described in Figure 9. Overall, it was found that the obtained experimental results are in reasonable agreement with the calculated results. When the Ti content of molten iron is more than 0.1% by mass, Ti2O3 forms as a stable compound. MgAl2O4 forms at Al > 0.005% by mass when the Ti content is 0.1% by mass (1000 ppm by mass), and the region of MgAl2O4 formation widens as the Ti content decreases. It is necessary to lower the Al content to at least under Al = 0.005% by mass and to adjust the Ti content to within an appropriate concentration range, so that MgTi2O4 can form in the range of Mg = 1–10 ppm by mass. From the viewpoint of Mg/Ti ratio, it is necessary to keep the Mg/Ti ratio under from 0.001 to 0.01 with an increase of Mg content from 1 to 10 ppm by mass when Al = 0.005% by mass. When Al content decreases to 0.001% by mass, the upper limit of the Mg/Ti ratio is eased to from 0.006 to 0.05 with an increase of Mg content from 1 to 10 ppm by mass. Moreover, when Al content decreases to 0.0001% by mass, the upper limit of the Mg/Ti ratio is eased to from 0.06 to 0.5 with an increase of Mg content from 1 to 10 ppm by mass. That is to say, it is desirable to lower the Al content as much as possible in order to form not MgAl2O4 but MgTi2O4.
The relationships of the compounds in the Mg–Ti–Al–O system in equilibrium with molten iron are investigated at 1873 and 1973 K. The conclusions are as follows:
- The liquid phase region of TiOX–MgO–Al2O3 system in equilibrium with molten iron is clarified at 1973 K. The equilibrium compounds, which are coexisted with the oxide mixture melts on the liquidus curve in the system are identified to be Mg2TiO4, MgTi2O4, and MgAl2O4.
- The equilibrium constant of the reaction of MgTi2O4 with molten iron and the interaction parameter of Mg on Ti at 1973 K is determined as follows:
- The equilibrium relations between the compounds (Mg2TiO4, MgTi2O4, or MgAl2O4) and the composition of solutes in steel are clarified. It is shown the transformation of the stable compound from Mg–Ti spinel to Mg–Al spinel at Mg contents ranging from 1 to 10 ppm by mass. It is also found, on the Mg–Ti spinel, the stable compound is transformed from inverse-spinel Mg2TiO4 to normal-spinel MgTi2O4 at Ti = 60 ppm by mass. In addition, the stable compound is transformed from MgTi2O4 to Ti2O3 at Ti = about 1000 ppm by mass.
- It is necessary to lower the Al content to at least under Al = 0.005% by mass, so that MgTi2O4 can form instead of MgAl2O4 in the range of Mg = 1–10 ppm by mass.