Regular Article
Effect of Fluoride Ions in Slag on the Dynamic Change of the Interfacial Tension between Liquid Iron and Molten Slag
2024 Volume 64 Issue 15 Pages 2217-2225
Details
2024 Volume 64 Issue 15 Pages 2217-2225
The interfacial tension between the liquid steel and molten slag is one of the key properties to control the entrapment of mold flux in molten steel in the continuous casting process. A dynamic change of the interfacial tension is observed when deoxidized iron and silicate slag are in contact, which can be explained by the oxygen absorption and desorption at the iron/slag interface. However, the dynamic change of the interfacial tension is influenced by other surfactant components of the molten iron and slag. Fluoride ions are fundamental component of mold flux, and recognized as the surface active component of molten slag. The effect of fluoride ions in slag on the interfacial tension has not been critically evaluated. Here, the effect of fluoride ions in slag on the interfacial tension between molten iron and molten silicate slag was evaluated at 1823 K, where the fluoride-containing slag compositions were designed to exhibit the same SiO2 activity and slag viscosity as those of the fluoride-free slag. Compared with the case of molten iron and fluoride-free slag, the interfacial tension between the molten iron and fluoride-containing slag was initially lower. Except the effect of oxygen adsorption, fluoride ion was considered to directly decrease the interfacial tension. However, as the fluoride content in slag was higher, the interfacial tension tended to show the higher value at the final state. This behavior was attributed mainly to fluoride vaporization as SiF4, which reduce the SiO2 activity in slag and thus equivalent oxygen content at the iron/slag interface.
For many metallurgical processes, the interfacial tension between the liquid metal and molten slag is considered to be the key property to control the stability of the metal/slag interface. In the continuous casting process in steelmaking, a metal/slag interface forms at the top surface of the liquid steel, where the provided mold flux becomes molten oxide–fluoride slag and acts as a lubricant between the steel and mold. For clean steel production, the slag inclusion in the liquid steel should be minimized. To prevent slag inclusion, the interfacial tension between the liquid steel and molten slag should be kept high so that the steel/slag interface remains flat. Therefore, the interfacial tension between the liquid steel and molten oxide–fluoride slag should be taken into account when designing the mold flux composition.
The equilibrium interfacial tension has been experimentally evaluated for various types of iron-based alloy and slag compositions.1,2,3,4,5,6,7,8,9) It has been concluded that the equilibrium interfacial tension can be mainly determined by the oxygen content in the alloy.3,8) Because oxygen is considered to be a surface active element for liquid metal, its adsorption at the metal/slag interface can significantly reduce the interfacial tension.
A dynamic change of the interfacial tension has been observed when the redox reaction occurs at the metal/slag interface.5,6,7,10,11,12,13) Tanaka et al.14) reported that this behavior is observed even in the case where liquid iron with a small amount of a deoxidizing element (e.g., aluminum) is in contact with molten silicate slag. To explain the dynamic change of the interfacial tension, they proposed a model of over-adsorption and desorption of oxygen, which is provided to the metal/slag interface by the decomposition of the reducing slag component (e.g., SiO2).14,15,16,17)
Our group has investigated the effects of other surface active elements (e.g., sulfur) and reducing oxide components (e.g., B2O3) in slag on the dynamic change of the interfacial tension between liquid iron and molten slag.18,19) We found that sulfur addition to silicate slag promotes the reduction of the interfacial tension in the early stage. The co-adsorption of sulfur and oxygen at the metal/slag interface is considered to be the main reason for the reduction of the interfacial tension in the early stage. The addition of B2O3 to slag also promotes the reduction of the interfacial tension in the early stage, while it increases the time for the interfacial tension to increase from the minimum to the equilibrium. This behavior is well explained by the decomposition of B2O3 and SiO2 to provide an abundance of oxygen to the metal/slag interface.
The above cases all focused on the adsorption of surface-active elements at the metal/slag interface to reduce the interfacial tension for liquid metal. However, the effects of the surface-active components of slag, such as fluoride ions, on the interfacial tension have not been clarified.
Fluoride species is a fundamental component of mold flux, because it optimizes many physicochemical properties of the molten slag as a lubricant, as well as a heat transfer material.20,21,22,23,24) In addition, oxide-fluoride slag is used for secondary refining flux because of its effects to decrease melting temperature and viscosity of slag. Due to the environmental impacts by vaporized fluoride species, many researches have been recently proceeded on the development of fluoride-free fluxes.24,25,26,27,28,29,30) However, to design the optimum condition of flux without fluoride species for sustainable steel production, it is necessary to first understand the roles of fluoride ions on the various kinds of physicochemical properties of molten slag.
In relation to interfacial properties, it should be noted that fluoride species is recognized as a surface-active component of slag, because the addition of a small amount of fluoride ions to slag significantly reduces the surface tension.31,32,33) The segregation of the fluoride ions to the slag surface is a possible mechanism for reducing the surface tension. However, the effect of fluoride ions as a surface-active component of slag on the interfacial tension between molten iron and slag is unclear, which is important for designing the mold flux composition.
In this study, we directly evaluated the effect of fluoride ion addition to slag on the dynamic change of the interfacial tension between liquid iron and molten slag at 1823 K as a fundamental study to investigate the chemical reaction between liquid steel and mold flux, by using molten fluoride-containing slag designed to have fixed values of the physicochemical properties, except for the interfacial properties.
The dynamic change of the interfacial tension between molten iron and silicate slag is influenced by the SiO2 activity and viscosity of the slag.14) The SiO2 activity of the slag determines the driving force of the decomposition reaction of SiO2 to provide oxygen to the interface. The slag viscosity controls the mass transfer of the slag components, such as SiO2, from the bulk to the metal/slag interface. Therefore, to directly evaluate the effect of fluoride ions in slag on the dynamic change of the interfacial tension, these physicochemical properties should be fixed by the fluoride addition.
To design appropriate fluoride-containing slag compositions, we performed estimations of the SiO2 activity and viscosity of molten slag at 1823 K with FactSage thermodynamic computation software using the latest thermodynamic database for oxide–fluoride systems.34) The slag viscosity of the silicate systems was originally calculated by Pelton’s model as functions of the concentrations of the oxide components and the polymerized network species,35,36) which were obtained by thermodynamic calculations. We confirmed that the above viscosity model is applicable to fluoride-containing slag systems.
In this study, 40 mass% SiO2–40 mass% CaO–20 mass% Al2O3 was selected as the mother slag composition, and fluoride ions were introduced by substituting part of CaO with CaF2. The calculated SiO2 activity as a function of the mass ratio of CaF2 to (CaO + CaF2) is shown in Fig. 1(a). The results showed that the SiO2 activity can be simply increased by increasing the amount of fluoride addition. When the Al2O3 content was increased with a fixed composition ratio, the calculated SiO2 activity of the fluoride-containing slag decreased to the value of the fluoride-free mother slag. The calculated SiO2 activity as a function of the Al2O3 content is also shown in Fig. 1(a). The slag viscosity linearly decreased with increasing addition of fluoride ions, whereas it increased with increasing Al2O3 addition and achieved the value of the fluoride-free slag, as shown in Fig. 1(b). Thus, around the mother slag composition, the changes in the SiO2 activity and slag viscosity by the addition of fluoride ions can be adjusted by increasing the Al2O3 content.
The designed compositions of the fluoride-containing slag are summarized in Table 1. Slags B and C contained 2.0 and 4.5 mass% fluorine ions, respectively. It was verified that the calculated SiO2 activity and slag viscosity were comparable with those of the fluoride-free mother slag A. Using these slag compositions, we evaluated the effect of fluoride ions in the slag on the dynamic change of the interfacial tension between liquid iron and the molten slag.
| Name | Chemical composition in mass% | SiO2 activity at 1823 K | Viscosity at 1823 K | |||
|---|---|---|---|---|---|---|
| CaO | SiO2 | Al2O3 | CaF2 | |||
| Slag A | 40 | 40 | 20 | 0 | 0.0907 | 0.513 |
| Slag B | 33.7 | 37.5 | 24.0 | 4.8 | 0.0926 | 0.509 |
| Slag C | 27.0 | 34.2 | 29.7 | 9.0 | 0.0977 | 0.516 |
We used a low-carbon steel sample with 0.020 mass% Si (iron X), whose detailed composition is provided in our previous studies.18,19) The Al content in iron X was too low for the redox reaction between SiO2 in the slag and Al in the liquid iron to occur.
The slag samples A–C were prepared in the following way. First, special-grade calcium carbonate (Fujifilm Wako Chemicals, Osaka, Japan) was calcined at 1223 K in air to remove carbon dioxide and synthesize solid CaO. Second, special-grade silicon dioxide and aluminum oxide (Fujifilm Wako Chemicals) were mixed with the CaO powder, and the mixture was melted at 1823 K in a platinum crucible under an air atmosphere. The melt was quenched by pressing it between two thick copper plates. To obtain the fluoride-containing slags B and C, the slag was crushed into a powder and mixed with high-grade calcium fluoride (Fujifilm Wako Chemicals), and the mixture was pressed into a pellet with diameter of 10 mm and thickness of 5 mm. The pellet was then placed in a molybdenum plate and heated at 1823 K in a furnace for measurement of the interfacial tension under a Ar gas atmosphere, which was aforehand dehydrated by passing through silica gel and magnesium perchloride and then deoxidized by metallic magnesium flakes heated at 773 K.
The interfacial tension between the liquid iron and molten slag was measured by the floating lens method, where a small slag droplet was placed on a large surface of liquid iron and the apparent contact angle at the gas/slag/iron triple point was observed. This method enables the accurate observation of the dynamic change of the interfacial tension in an optical way. The interfacial tension is determined by the measured contact angle and the surface tension of liquid iron and molten slag:
| (1) |
where σFe/Slag, σFe, and σSlag are the interfacial tension between liquid iron and molten slag, surface tension of liquid iron, and surface tension of molten slag, respectively. The surface tension of liquid iron was estimated using the empirical model reported by Ogino et al.4) as a function of the oxygen and sulfur contents, where the oxygen content in the iron was measured by an oxygen sensor. The surface tension of the molten slag was estimated by the semi-empirical model reported by Nakamoto et al.33)
The apparatus used to observe the apparent contact angle between the liquid iron and molten slag is described in our previous studies.18,19) The heating material was made of graphite, so a reducing atmosphere was prepared in the furnace when it was heated. The experiment was conducted as follows. First, 80 g of the iron sample was placed on a flat alumina crucible, and then the crucible was placed in the center of the furnace. A 0.8 g slag pellet was placed on a molybdenum plate with a small hole, and the plate was set above the iron sample in the furnace. Here, the density of each slag sample was recognized as comparable in the molten state by additive rule of those of pure substances.37) The furnace was sealed and evacuated to vacuum, and it was then filled with purified Ar gas dehydrated and deoxidized aforehand as mentioned above. The furnace was heated and held at 1823 K to separately melt the iron and slag samples. The slag droplet was placed on the flat surface of the liquid iron, and the shape of the droplet was captured as an optical image through a quartz window. The apparent contact angle at the gas/slag/iron interface was evaluated from the image by ImageJ software, where the measurement was performed three times for both right and left hand sides and the average value was taken. Particularly, the functions of image binarization and drawing outline were used to minimize the reading error of the apparent contact angle. As a result, the error in measured contact angle for each experiment was determined as less than 5%.
To evaluate the composition changes of the liquid iron and molten slag during contact, the furnace was rapidly cooled at the predetermined holding time to solidify the iron, and it was then gently cooled to room temperature. The Si content in the iron and the fluorine content in the slag were analyzed by an induction-coupled plasma method, whereas the contents of other slag components were analyzed by X-ray fluorescence spectroscopy.
First, the apparent contact angles between molten iron X and fluoride-containing slags A–C (Table 1) and their dynamic changes were evaluated. Typical images of the slag droplets in contact with the molten iron at different holding times are shown in Fig. 2. In each case, the slag droplet initially spread over the surface of the molten iron, and it then gradually narrowed to reach the steady state. With increasing fluoride content in the slag, the degrees of both the slag spreading in the initial stage and the slag narrowing in the later stage became more significant.
The apparent contact angle between the molten iron and slag as a function of the holding time is shown in Fig. 3, where the moment for the iron/slag contact is defined as zero. For the case of iron X/fluoride-free slag A contact, the contact angle initially decreased, exhibited the minimum contact angle at 10 min, and then gradually increased to achieve a certain constant value at 60 min. For the case of iron X/fluoride-containing slag B contact, the contact angle rapidly showed the minimum value at 2 min, and the minimum value was lower than that for the case of iron X/slag A contact. The contact angle then remained at the minimum value for a while. Finally, the apparent contact angle increased for a long time and achieved a higher constant value than the case of iron X/slag A contact. For the case of iron X/slag C contact, the contact angle rapidly exhibited the minimum value at an even lower value than for iron X/slag B contact, and it then increased for a long time to reach an even higher value at the steady state than the other cases. The measurement of contact angle was performed 2–3 times for the above each case, where certain degrees of scattering were observed for absolute values. This may be attributed to the scattering in the size of slag droplet which was held on the molten iron surface. When the size of the slag droplet is small, the apparent contact angle could be measured as a high value because the slag droplet is partially suspended by Mo rod as well.
It should be noted that the apparent contact angle at each time is determined by the balance between the surface tension of the liquid iron, surface tension of the molten slag, and interfacial tension between the iron and slag. The fluoride content, which determines the surface tension of the molten slag, could change owing to its vaporization. Therefore, to evaluate the dynamic interfacial tension, it is necessary to examine the change of the fluoride content in the slag with the holding time.
The analytical results of the fluorine content in the slag for the cases of iron X/slag B and iron X/slag C are shown in Fig. 4. Before contact with the iron, the fluorine contents were comparable with the desired values. However, in both cases, the fluorine content decreased with increasing holding time.
It was presumed that the above fluorine content decrease in the slag was mainly attributed to silicon fluoride vaporization. This is because the gas-slag-graphite (as heating material) equilibrium calculation by FactSage34) for the slags B and C indicated silicon fluoride species as the highest fugacity in the gas phase, whereas those of other fluoride species are relatively low, as shown in Table 2. In this case, the following chemical reaction was considered to generate the silicon fluoride:
| (2) |
| Gaseous species | Fugacity |
|---|---|
| SiF4 | 4.00×10−1 |
| CO | 1.72×10−1 |
| SiF3 | 1.64×10−1 |
| AlF3 | 1.51×10−1 |
| AlF | 3.98×10−2 |
| AlF2 | 3.45×10−2 |
| SiF2 | 2.41×10−2 |
| SiO | 9.25×10−3 |
| CaF2 | 3.61×10−3 |
| F | 2.14×10−9 |
| O2 | 8.71×10−18 |
As the chemical reaction (2) proceeds to the right-hand side, CaF2 and SiO2 in slag are consumed, while CaO content increases.
Taking account of the above chemical reaction, the changes in the surface tension and the SiO2 activity in slags B and C by fluoride loss were predicted. The surface tension was estimated by Nakamoto’s semi-empirical model, whereas the SiO2 activity was calculated by FactSage.34) The results are shown in Fig. 5, where these properties are exhibited as a function of fluorine mass content in slag. As decreasing fluorine content, surface tension monotonously increased, whereas SiO2 activity mildly decreased.
The above change of the slag surface tension by fluoride loss (Fig. 5(a)) was reflected to evaluate the dynamic interfacial tension by Eq. (1) using the observed contact angle and calculated surface tension values of liquid iron and molten slag.
The change of the interfacial tension between the liquid iron and molten slag as a function of the holding time is shown in Fig. 6. Similar to the change of the contact angle, the interfacial tension decreased in the early stage, exhibited a minimum at a certain time, and then increased in the later stage, finally reaching the steady state. Additionally, the interfacial tension for higher fluoride content in the slag was lower in the early stage, continuously increased with longer holding time in the later stage, and achieved higher interfacial tension in the final state.
As previously discussed,14,15,16,17,18) the decrease of the interfacial tension in the early stage can be attributed to the decomposition reaction of SiO2 in the slag to provide oxygen together with silicon to the interface by proceeding the following chemical reaction (3) to the right-hand side:
| (3) |
where the silicon would transfer to the liquid iron. Therefore, the Si content in the iron was analyzed at several holding times to verify the degree of the chemical reaction at the interface. The results for the cases of iron X and slags A–C are shown in Fig. 7. The increases of the Si content in the iron were comparable for the three cases. That is, it can be presumed that the degrees of oxygen supply at the interface, and thus the decreases of the interfacial tension in the early stage, were comparable for the three cases. Therefore, it can follow that the initial interfacial tension decreased as increasing the fluoride content in the slag. However, in this experiment, the effect of fluoride ions to the initial interfacial tension cannot be directly observed, because the SiO2 decomposition to supply oxygen atoms to the iron/slag interface would also occur.
The continuous increase of the interfacial tension in the later stage can be partially attributed to the oxygen transfer from the interface to the bulk side of the iron, and the other effect by the fluoride ions may also occur. Particularly, the decrease of the SiO2 activity by fluoride loss (Fig. 5(b)) should be taken into account, which decreases oxygen content in iron equilibrated with slag at the steady state.
Figure 8 shows the compositional changes of slag components during the iron and slag contact, where the elemental amounts detected by the X-ray fluorescence spectroscopy were translated to mass fraction of corresponding oxide species. However, the CaO content was determined by subtracting the equivalent Ca amount counted as CaF2 (evaluated from analyzed F content in slag) from the total Ca content in slag. The result clearly shows that SiO2 content decreased but CaO content increased while contacted with molten iron, whereas Al2O3 content was almost stable. This supports the occurrence of the chemical reaction (Eq. (2)) to vaporize silicon fluoride. However, because the reaction of SiO2 decomposition (Eq. (3)) also brings the decrease of SiO2 content, it was not clear whether the fluoride loss in the slag affected the increase of the interfacial tension in the later stage.
To clarify the effect of the fluoride ions in the slag on the increase of the interfacial tension, we conducted the following additional experiments. First, the iron Y sample was prepared by contacting iron X and fluoride-free slag A in the molten state for at least 2 h, where the Si transfer from the slag to the iron was regarded as saturated. After the sample solidified, the slag droplet was removed from the iron, crushed, and placed around the iron to maintain the equilibrium between the liquid iron and molten slag. Iron Y was then melted again in the furnace, and fluoride-containing slag B or C was placed on the surface of the liquid iron. Subsequently, the apparent contact angle between the iron and slag was observed to evaluate the dynamic change of the interfacial tension. By the above method, we could extract the effect of the fluoride ions in the slag on the dynamic change of the interfacial tension under suppression of the Si transfer from the slag to the iron.
The changes of the apparent contact angle and dynamic interfacial tension between liquid iron Y and molten slags A–C with the holding time are shown in Figs. 9(a) and 9(b), respectively. It was assumed that the surface tension of liquid iron Y was comparable with that of liquid iron X and the changes of the fluoride content in the slag for the above cases were identical to the cases for the contact of iron X with slags A–C. For iron Y/fluoride-free slag A contact, both the apparent contact angle and interfacial tension did not change with the holding time. In contrast, for iron Y/fluoride-containing slag B or C contact, the interfacial tension initially exhibited a lower value than for iron Y/slag A, gradually increased, and finally reached a higher value than that of iron Y/slag A. Thus, it was experimentally verified that the fluoride addition in the slag affected the interfacial tension in both the early and later stages.
The Si contents in iron Y and fluorine contents in the slag after the experiments are summarized in Table 3. In each case, the Si content in iron Y was slightly higher than that in iron X after contact with slag A, whereas the effect of the increase of the Si content on the interfacial tension was small. Conversely, the fluorine content in fluoride-containing slag B or C became lower than the designed content, similar to the cases where the slags were in contact with iron X in the molten state.
| Si content in iron Y in mass% | F content in slag in mass% | |
|---|---|---|
| Iron Y (Before): 0.060 | ||
| Iron Y/Slag A Run1 | 0.069 | – |
| Slag B (Designed): 2.0 | ||
| Iron Y/Slag B Run1 | 0.091 | 0.50 |
| Iron Y/Slag B Run2 | 0.072 | 0.39 |
| Iron Y/Slag B Run3 | 0.083 | 0.44 |
| Slag C (Designed): 4.5 | ||
| Iron Y/Slag C Run1 | 0.085 | 0.71 |
| Iron Y/Slag C Run2 | 0.079 | 0.68 |
The dynamic change of the interfacial tension between liquid iron Y and molten slags A–C showed the following two trends: (I) the initial value decreased, and (II) the final value became higher, with increasing fluoride content in the slag (Fig. 9(b)). It should be noted that the viscosity and SiO2 activity of the slags were initially fixed against the fluoride content, and that Si transfer from the slag to liquid iron Y was prevented beforehand. Thus, the above trends on the interfacial tension between the molten iron and slag should be directly attributed to the behaviors of fluoride ions in the slag.
In the following sections, possible mechanisms to explain the above trends in the early and later stages are proposed.
(I) Early Stage: The Initial Value of the Interfacial Tension was Lower for Higher Fluoride Content in the SlagA large amount of fluoride ion was remained in the slag at that stage, and the fluoride ions should directly affect the initial interfacial tension by the following reason.
As previously mentioned, the fluoride species has been originally known to significantly decrease the surface tension of molten slag.31,32,33) Because fluoride ion (F−) has a lower ionic potential than the oxygen ion (O2−) due to its larger ionic size and lower electric valence, they exhibit lower bonding energy when unsatisfied. Thus, the segregation of the fluoride ions to the slag/vapor interface (slag surface) is considered as possible mechanism of decreasing the surface excess energy.
In the case of the iron/slag interface, because they show totally different bonding manners, the adjustment of chemical bonds of each side should be separately made to minimize the total interfacial excess energy. If it is assumed that structural relaxation occurs in the slag side of the iron/slag interface similar to the slag/vapor interface (slag surface), the segregation of fluoride ions at the slag side of the iron/slag interface is possible to decrease the interfacial excess energy.
(II) Later Stage: Interfacial Tension Showed a Higher Value in the Final State for Higher Fluoride Content in the SlagAs most of fluoride was removed in the slag at the final state (Table 2), the silicon fluoride vaporization and the decrease of the SiO2 activity by the chemical reaction (Eq. (2)) would have proceeded. The SiO2 activity was calculated for the analyzed slag composition in the final state by FactSage,34) and it was revealed that the SiO2 activity in the final state became lower as the case where higher content of fluoride was initially included in the slag. When the chemical reaction (Eq. (3)) completed to the right-hand side, the oxygen content in the iron equilibrated with SiO2 in the slag should have been lower, and then the equilibrium interfacial tension should become higher than initial. Therefore, the above mechanism could well explain the trend (II).
The fluoride loss from the slag would increase the interfacial excess energy for both slag and iron side by the following reasons. From the slag side, the fluoride loss would decline the tendency of the fluoride segregation to the iron/slag interface. In addition, from the iron side, as the chemical reaction (Eq. (2)) proceeds to the right-hand side, the activity of SiO2 in the slag decreases, and then oxygen would be partially transferred from the iron side to the slag side of the interface to satisfy the local equilibrium of the chemical reaction (Eq. (3)) at the iron/slag interface. These phenomena both increase the interfacial tension. Because the reaction of the fluoride vaporization proceeded mildly, it took long time to increase the interfacial tension.
The proposed behavior of fluoride ions in the slag during the dynamic change of interfacial tension is schematically described in Fig. 10. The trend (I); the decrease of initial value, is directly caused by fluoride ions in the slag, possibly by their segregation to the slag side of the interface. In contrast, as fluoride vaporizes from the slag by the reaction (Eq. (2)), the SiO2 content and its activity decreases. Because the oxygen content in the iron side of the iron/slag interface decreases to satisfy the equilibrium of the chemical reaction (Eq. (3)), the interfacial tension at the final state achieves a high value than the initial (trend (II)).
In this study, the effect of fluoride addition to slag on the dynamic interfacial tension between liquid iron and molten silicate slag was evaluated under fixed slag viscosity and SiO2 activity against the fluoride content. The following behaviors were revealed:
(1) When molten deoxidized iron is in direct contact with molten silicate slag, the interfacial tension decreases in the early stage and shows a minimum at a certain holding time, and it then increases in the later stage to reach the equilibrium value. As the fluoride content in the slag increases, the initial value of the interfacial tension decreases, whereas the time for the interfacial tension to increase becomes longer and the final value tends to become higher. Chemical analysis of the metal indicated that the SiO2 decomposition reaction occurs at the iron/slag interface to supply oxygen and silicon, which partially controls the above dynamic change of the interfacial tension. In addition, the degree of the SiO2 decomposition reaction is comparable against fluoride content in the slag.
(2) When molten iron is in contact with fluoride-containing slag under the condition where Si transfer from the slag to the metal is suppressed, the interfacial tension simply increases and achieve the higher value than initial at the equilibrium state. As the fluoride content in the slag increases, the initial value of the interfacial tension decreases, whereas the equilibrium value tends to become higher. The former is directly recognized as the effect of fluoride ions in slag on the dynamic interfacial tension. The segregation of the fluoride ions to the slag side of the interface to decrease the interfacial excess energy was proposed as a possible mechanism. On the other hand, the latter was attributed to the fluoride loss from slag: as silicon fluoride is vaporized, SiO2 content and thus SiO2 activity in the slag decreases, which decreases the oxygen content in the molten iron. This leads to the increase of the interfacial tension at the equilibrium state.
We kindly appreciate Professor Toshihiro Tanaka (Osaka University, Japan) for many suggestions to the comprehension of interfacial phenomena at molten metal-slag interface.
