Effect of Aluminum on the Solubility of Calcium in Liquid Iron at Low Calcium and Aluminum Contents

Abstract

The solubility of calcium in liquid iron as a function of aluminum content and calcium potential, at compositions relevant to production of aluminum killed steels, was studied experimentally at 1873 K. The measurements were made using a closed molybdenum chamber in which iron-aluminum alloys were held. The calcium potential was fixed using pure liquid calcium held at different temperatures. The calcium contents in the iron varied between 6 and 22 ppm by weight and the aluminum contents varied between 70 and 1900 ppm by weight. The results indicate that the effect of aluminum on the solubility of calcium in iron is very low in the composition ranges studied.

Study

Calcium additions during ladle treatment of aluminum-killed steels, referred to as calcium treatment, is a common practice with the aim of modifying solid alumina-containing inclusions, into liquid calcium aluminate inclusions. Despite calcium treatment being widely used there is still a lack of understanding of the basis of the process and research efforts are made continuously to provide more information on the topic. In general, there is little experimental data available on the thermodynamics of calcium dissolved in liquid iron at steelmaking temperatures. Studies on the effect of other alloying elements on calcium in iron based alloys are even more scarce. Sponseller and Flinn¹⁾ and Köhler et al.²⁾ measured the solubility of calcium in liquid iron alloys containing a number of different alloying elements, including aluminum. The measurements were based on equilibrating liquid calcium alloys with liquid iron alloys under high argon pressure to prevent excessive calcium vaporization. The focus of the mentioned studies were high contents of calcium in liquid iron as well as relatively high contents of the alloying elements. Song and Han³⁾ measured the solubility of calcium in liquid iron alloys at lower calcium contents. However, the effect of aluminum was not studied. The aim of the present study is the provide experimental data on the aluminum-calcium-iron system at temperatures and composition ranges applicable to steelmaking, specifically the production of low alloyed, aluminum-killed steels.

The effect of aluminum on the solubility of calcium in liquid iron-aluminum alloys at low calcium and aluminum contents was studied experimentally at 1873 K. Different iron-aluminum alloys were equilibrated with calcium vapor in a closed container. The vapor pressure of calcium was controlled by liquid calcium being held at lower temperature at the bottom of the container. By using a closed container and a surplus of calcium the vapor pressure of calcium can be held constant throughout the experiment.

The experimental setup used has been described in detail in a previous publication by the authors⁴⁾ and only a description of the experimental procedure is given below.

Before each run, 7 g of iron powder (99.9 pct) was held for 60 minutes at 773 K in flowing H₂ gas (99.995 pct) in order to reduce any surface oxides. After cooling down, the iron was mixed with an appropriate amount of aluminum granules (99.9 pct). The mixture was placed in a calcium oxide crucible (99.9 pct) and inserted into a cylinder shaped molybdenum container. The container was closed by inserting and securing a conical plug holding 4 g of calcium (99.5 pct). The sample holder was connected to a steel rod using a molybdenum extension and lowered into the hot zone of the furnace. The furnace was sealed tightly and the positions of the two sample thermocouples were adjusted carefully to ensure that they were level with the iron-aluminum alloy and the calcium respectively. An overview of the sample holder assembly is presented in Fig. 1. The furnace was evacuated for 3 hours using a vacuum pump. After evacuation the furnace was refilled with argon gas (99.999 pct). An argon flow of 0.1 l/min through the furnace was maintained during the experiments to prevent oxidation of the molybdenum parts. The furnace was heated to 1873 K at a rate of 1.6 K/min. After stabilization at the target temperature the sample was held for 48 hours before being withdrawn from the hot zone into a water cooled aluminum chamber. To increase the cooling rate, a high flow of argon gas, approximately 10 l/min, was injected directly in the aluminum chamber. After cooling down, the sample holder was removed from the furnace and opened.

Fig. 1.

Schematic illustration of the sample holder.

The iron-aluminum alloys were cut and carefully ground before being analyzed. Approximately 1 g of alloy was dissolved and analyzed using ICP-SFMS to determine the calcium and aluminum contents. The relative error of the chemical analysis is estimated as less than 20%. Samples were also examined in scanning electron microscope.

The experimental conditions, including temperatures of the iron-aluminum alloys and the calcium, and the analysis results for each run are presented in Table 1. The calculated pressure of calcium in the container are also presented. The pressures of calcium were calculated based on the temperatures of the liquid calcium. According to Knacke et al.⁵⁾ the saturated vapor pressure of pure liquid calcium as a function of temperature is given as   

$$log{p}_{Ca(l)}^{\circ }=\frac{-9\mathrm{\hspace{0.17em} }103}{T}-1.71\times log(T)+10.688,$$(1)

where $p_{Ca(l)}^{\circ }$ is given in bar and T is given in K. Note that the calculated values of the vapor pressure of calcium, and consequently any value derived from them, depend on the reference data used. The results from the analyses show that the calcium content varied between 6 and 22 ppm by weight while the aluminum content varied between 67 and 1900 ppm by weight. Examination of the samples in scanning electron microscope revealed no reaction between metal and calcium oxide crucible. More specifically, no calcium aluminate layers or particles were seen at the interface. Furthermore, very few calcium containing particles were observed.

Table 1. Experimental conditions and analysis results. The analysis results are given as ppm by weight.

Sample id.	TFe [K]	TCa [K]	pCa [kPa]	ppm Ca	ppm Al
1	1873	1356	4.16	9.48	67.2
2	1873	1425	8.07	22.4	77.2
3	1873	1319	2.82	8.99	114
4	1873	1424	8.00	17.9	146
5	1873	1340	3.53	10.5	182
6	1873	1342	3.60	12.2	288
7	1873	1343	3.64	7.46	300
8	1873	1426	8.14	17.0	344
9	1873	1384	5.49	11.1	395
10	1873	1325	3.01	7.73	408
11	1873	1319	2.82	6.97	415
12	1873	1307	2.48	5.87	1020
13	1873	1383	5.43	10.6	1800
14	1873	1371	4.83	8.53	1900

The calcium contents obtained in the samples are in the lower range of what has previously been studied.^1,2,3,4) It should, however, be noted that the expected contents of dissolved calcium in iron during ironmaking and steelmaking are significantly lower. The aluminum contents cover what is normally seen in aluminum killed steels, which is the focus of the present study.

Under the assumption that the calcium vapor behaves as an ideal gas, the activity of calcium, relative to pure liquid calcium, can be expressed as   

$$a_{Ca(l)}=\frac{p_{Ca(l)}}{p_{Ca(l)}^{\circ }}=X_{Ca}\times {\gamma }_{Ca(l)},$$(2)

where $p_{Ca(l)}$ is the pressure of calcium at a given temperature, $p_{Ca(l)}^{\circ }$ is the saturated vapor pressure of pure liquid calcium at the same temperature, X_Ca is the mole fraction of calcium in a solution and ${\gamma }_{Ca(l)}$ is the activity coefficient of calcium relative to pure liquid calcium in the solution. It should be noted that the temperature of the iron-aluminum alloys in the present study is above the boiling point of pure liquid calcium. This means that the values of $p_{Ca(l)}^{\circ }$ are extrapolated to 1873 K based on Eq. (1). Table 2 presents the calculated values of the mole fractions of calcium and aluminum in the iron alloys and the activity of calcium relative to pure liquid calcium at 1873 K. The activities are calculated based on the assumption that the liquid calcium held in the lower part of the molybdenum container is pure and has an activity equal to unity.

Table 2. Calculated values of X_Al, X_Ca and a_Ca(l) at 1873 K.

Sample id.	XCa × 105	XAl × 104	aCa(l)
1	1.32	1.39	0.0244
2	3.12	1.60	0.0473
3	1.25	2.36	0.0166
4	2.49	3.02	0.0469
5	1.46	3.77	0.0207
6	1.70	5.96	0.0211
7	1.04	6.21	0.0213
8	2.37	7.12	0.0478
9	1.55	8.17	0.0322
10	1.08	8.44	0.0177
11	0.97	8.59	0.0166
12	0.82	21.1	0.0145
13	1.47	37.2	0.0319
14	1.19	39.2	0.0283

The activities of calcium relative to pure liquid calcium as a function of the mole fraction of calcium in the iron alloys are presented in Fig. 2(a). Even though there is some scatter in the data points, a clear trend can be seen with a relatively linear relationship between activity of calcium and calcium contents in the iron alloys in the composition ranges studied. This is in good agreement with previous studies by Song and Han³⁾ and the present authors.⁴⁾ To further illustrate this, Fig. 2(b) presents a comparison between the present data and data from the studies mentioned. Based on the data presented in Fig. 2(b), the relative standard deviation of the calculated activity coefficient of calcium is 19 pct. All points presented are based on measurements at 1873 K. The data presented indicates that there is no significant effect of aluminum on the activity coefficient of calcium under the present conditions.

Fig. 2.

Activity of calcium as function of calcium content at 1873 K, a) present data and b) including data without aluminum from a previous study⁴⁾ and from Song and Han.³⁾

The concept of interaction parameter is sometimes used to quantify the effect of a solute on the activity coefficient of another solute. Due to earlier studies reporting the interaction parameter for the present system, the concept will be used for comparison here. The interaction parameter of solute i on solute j is expressed as   

$${\epsilon }_j^i={\frac{\partial ln{\gamma }_j}{\partial X_i}|}_{X_i\to 0},$$(3)

where ${\epsilon }_j^i$ is the interaction parameter, γ_j is the activity coefficient of solute j and X_i is the mole fraction of solute i. This gives the interaction parameter of aluminum on calcium in liquid iron as   

$${\epsilon }_{Ca}^{Al}={\frac{\partial ln{\gamma }_{Ca(l)}}{\partial X_{Al}}|}_{X_{Al}\to 0},$$(4)

corresponding to the slope of a plot of γ_Ca(l) as a function of X_Al as X_Al goes to zero at constant calcium content. Figure 3 presents the natural logarithm of the activity coefficient of calcium in liquid iron, relative to pure liquid calcium, as a function of aluminum content for the present data as well as for data without aluminum from Song and Han³⁾ and the present authors.⁴⁾ The relative standard deviation of the plotted values is 2 pct. It should be noted that the calcium contents for the different data points are not the same, but vary in the range 6–110 ppm. For the aluminum containing samples the range is 6–22 ppm as mentioned above.

Fig. 3.

Natural logarithm of activity coefficient of calcium as function of aluminum content including data from a previous study by the authors⁴⁾ and Song and Han.³⁾

Previous studies by Sponseller and Flinn¹⁾ and Köhler et al.²⁾ reported the interaction coefficient of aluminum on calcium in liquid iron as −7.5(±1) and −5.5(±0.5) respectively. These studies used very high calcium contents in iron as well as a wide range of aluminum contents. Apart from samples with no aluminum, no measurements at aluminum contents lower than approximately 1 pct by weight were made. The data plotted in Figs. 2 and 3 suggests that the activity coefficient of calcium in liquid iron is practically constant in the ranges studied. However, from a linear regression based on the data presented in Fig. 3, a positive slope of 87 is obtained. The authors believe that this depends largely on the experimental uncertainties combined with the relatively few experimental points at higher aluminum contents. To confirm this more experimental data is needed. An estimation of the effect of aluminum on the activity coefficient of calcium in the range of the present study can be made by using the more negative value reported by Sponseller and Flinn¹⁾ and the equation   

$$ln{\gamma }_{Ca(l)}=ln{\gamma }_{Ca(l)}^{\circ }+{\epsilon }_{Ca}^{Al}\times X_{Al},$$(5)

where ${\gamma }_{Ca(l)}^{\circ }$ is the activity coefficient of calcium liquid iron, relative to pure liquid calcium, at infinite dilution. Using the highest aluminum content in the range presently studied, meaning a mole fraction of 39 × 10⁻⁴, the activity coefficient of calcium is estimated to be approximately 3 pct smaller compared to in pure iron. This is well within the experimental uncertainty band of the present study. The present results in combination with the data reported by Sponseller and Flinn¹⁾ and Köhler et al.,²⁾ indicate that the relative effect of aluminum on the activity coefficient of calcium in liquid iron is very small in the composition range applicable to aluminum killed steels. It should further be noted that the estimated equilibrium contents of calcium in steels at 1873 K are extremely low, meaning that any reasonably large relative change of the activity coefficient of calcium would lead to very small absolute changes in the equilibrium calcium content. Instead, the effects of different alloying elements on the activity coefficient of calcium could be expected to become more important at higher contents of calcium, such as during calcium treatment. However, concerning aluminum, the available data indicate that the effect would be relatively small unless aluminum contents are higher than what is commonly found in aluminum killed steels.

References

• 1)   D. L.  Sponseller and  R. A.  Flinn: Trans. Metall. Soc. AIME, 230 (1964), 876.
• 2)   M.  Köhler,  H. J.  Engell and  D.  Janke: Steel Res., 56 (1985), 419.
• 3)   B.  Song and  Q.  Han: Metall. Mater. Trans. B, 29 (1998), 415.
• 4)   M.  Berg,  J.  Lee and  D.  Sichen: Metall. Mater. Trans. B, 48 (2017), 1715.
• 5)  Thermochemical Properties of Inorganic Substances, ed. by O. Knacke et al., Springer-Verlag, Berlin, Heidelberg, (1991), 326.