Surface and interfacial properties of Fe-C-O-Cr alloys in contact with alumina

In this paper, temperature and concentration dependencies on density and surface tension of liquid Fe-C-O-Cr alloys (1.93 - 4.80 wt.% Cr) were investigated in high-temperature resistance observation furnace by a sessile drop method during heating from liquidus temperature to 1600?C. The interfacial characteristics (interfacial tension, wetting angle, work of adhesion, and spreading coefficient) of liquid alloy/alumina system were also determined depending on temperature. The effect of temperature and chromium content on surface and interfacial properties was proven in case of all examined alloys. Based on the fact that the content of surface-active elements such as oxygen (up to 195 ppm) and sulfur (up to 545 ppm) was higher, the influence of activities of both mentioned elements on surface tension of alloy samples was assessed. Particular attention was paid to the dependence of the surface tension temperature coefficient on oxygen and sulfur activity.


Introduction
During numerous metallurgical processes, particularly casting and refining, the metallic melt is in contact with a solid ceramic refractory material. At the interface between these systems, the interactions influencing physicomechanical properties of the final product occur at high temperatures. Characterization of the ceramic material surface by determining the contact angle of metal drop resting on solid substrate contributes to the understanding of physicochemical processes at the phase interface and thereby optimizing the metal product. It can be assumed that non-wettable surfaces are more resistant to the melt exposure, i.e. during production, these surfaces will be more inert to chemical and mechanical action. The experimental study of the surface and interfacial properties of liquid multicomponent metals in contact with a ceramic material is demanding, especially, due to the possible high reactivity of metallic melts, chemical heterogeneity, roughness and porosity of the substrate surface and sensitivity of surface tension to impurities. Besides, the determination of surface tension and the characterization of the phase interface (metallic melt/ceramic) is also intricate due to discrepancies in methodology and inconsistency in experimental procedures [1].
In principle, the surface tension or surface energy arises from a phenomenon that atoms near a free surface have partially empty coordination shells and therefore they are at higher energy states than the atoms in the bulk of the solution. In other words, it is a surface physical force representing that the atoms in the liquid bulk pull their neighbours in all directions in zero net force, contrary to the surface atoms experiencing a net inward force from atoms below. Nevertheless, at the surface. In a multicomponent alloy, atoms whose energy state is affected least by the surface are segregated to the liquid surface region. Chromium is among metals segregating onto an alloy surface. Its energy changes concerning segregation are relatively negligible compared to strongly surface-active elements like sulfur and oxygen. Therefore, it influences the surface tension to a significantly smaller extent [2][3][4][5][6]. The effect of chromium has already been investigated by several researchers. Treťyakova et al. investigated the dependence of surface tension on chromium content (up to 1 wt.% Cr) in Fe-Cr-O systems. In this case, the minimum surface tension occurred when the melt was the most microheterogeneous. The presence of chromium increased surface tension due to its ability to penetrate between the clusters, which made the melt more uniform [7]. The influence of chromium on surface tension was also investigated by Mukai and Li [8,9]. They found that chromium slightly increases surface tension. Further, the chromium possesses a strong affinity to oxygen and facilitates its adsorption. Therefore, at higher concentrations (above 10 wt.%) oxygen lowers the surface tension of Fe-Cr-O systems [10]. It is assumed that for most liquid metals, alloys and steels, the temperature coefficient of surface tension is negative [11]. Nevertheless, especially for systems having considerable positive free energies, it can be positive [12]. Usually, the positive surface coefficient of surface tension is associated with surface segregation and higher values of the excess Gibbs energy [13]. Li noted that the coefficient of Fe-16 mass%Cr-S alloys increased with increasing sulfur content and was positive when sulfur content reached over 20 mass ppm [8]. Further, the effect of steel composition on temperature coefficient was investigated on several commercial 4-series ferritic stainless steels. It was found to change from negative to positive at a soluble sulfur content above 30 mass ppm in the steel [14]. Similar conclusions concerning sulfur influence were documented in [15,16]. The extensive research on this issue was provided by Brooks, who examined more than 40 austenitic and ferritic steels and observed that for those with high sulfur content, the coefficient was positive [17]. Measurement of surface tension of low alloyed AISI 4142, Fe-Cr-Ni stainless steel AISI and high-manganese Fe-Cr-Mn-Ni steel containing sulfur (> 50 ppm) revealed that temperature coefficient of surface tension was positive for all investigated alloys. It was suggested that the free surface of a liquid droplet is covered by a monolayer of surface-active elements, e.g. sulfur causing a reduction of surface tension. Whereas, when the temperature is increasing, sulfur is desorbed into the bulk of the liquid metal, and the surface tension is rising [3,18,19]. This work is focused on the experimental study of surface and interfacial quantities regarding Fe-C-O-Cr alloys in contact with alumina substrate depending on temperature and change in chemical composition. Moreover, the influence of chromium and surface-active elements (oxygen and sulfur) on surface tension has been determined.

Samples preparation
Four alloys (samples 1-4) were selected to study the surface and interface properties by a sessile drop method, a steady and accurate method for surface tension measurement at high temperatures having experimental error lover than ± 2 %, as noted in [20]. These samples were prepared from pure metals (Fe and Cr, purity 99.99 %), carbon (purity 99.99 %) and Fe2O3 tablets (purity 99.999 %) by vacuum induction melting in furnace Leybold Heraeus at our working site. The melt was cast into the vertically oriented mould to obtain 3 kg ingots from which rods of a diameter 5 mm were made. Their chemical composition determined by GDA 750 HP optical emission spectrometer (GDOES) is listed in Table 1. Carbon, oxygen and sulfur contents were obtained by combustion analysers Eltra 2000 CS and Eltra 2000 ONH. Prior to the experiment, cylindrical alloy samples (5 mm diameter x 5 mm height) were thoroughly mechanically cleaned to remove surface oxides. For the purposes of the experiment, alumina plates (99.8 % Al2O3) were annealed at 1150 °C for 6 hours, and their surface was cleaned by acetone immediately preceding the measurements. The content of other elements (Ni, Si, Ti, Mo, P, Al, Cu and Zr) present in alloys was less than 0.005 wt.%, and the rest represents iron.

Experimental procedure
Experimental determination of density, surface tension and wetting angle was carried out with using the sessile drop method in the high-temperature observation furnace CLASIC ( Figure 1) in a temperature range from the melting point of the alloy sample to 1600 °C. The prepared sample (alloy/Al2O3 substrate) was placed in the furnace tube, which was then hermetically sealed, evacuated to approximately 0.1 Pa and purged with argon of high purity (> 99.9999 %). The latter two steps were repeated. The heating rate and maximum temperature were set to 5 °C min -1 , and 1600 °C. The temperature was measured with a Pt-13% Rh/Pt thermocouple close to the sample. To prevent subsequent oxidation of the sample, which substantially affects surface tension and temperature coefficient, as reported by Yuan [21], all measurements were carried out in an inert atmosphere of argon. Each sample was measured four times. The images of drop forming during a heating ramp were taken by a CANON EOS550D and then saved in a PC. Then, they were evaluated by the Laplace -Young equation, describing the equilibrium pressure at the interface. The evaluation was performed by in house software involving ADSA (Advanced axisymmetric drop shape analysis) method [22,23]. This method allows to determine parameters such as wetting angle, surface tension, density, etc. The interaction of alloy samples with alumina substrate after the experiment was investigated by scanning electron microscope (SEM) JEOL 6490LV. The X-ray microanalysis of microstructural particles and their chemical composition was carried out by EDS (energy dispersive X-ray spectroscopy) analyzer INCA in the mode of backscattered electrons (BSE). The device setup was: thermo-emission cathode LaB6, voltage 20 kV, resolution of 3.0 nm, vacuum 2.5⋅10 -6 Pa.

Results and discussion
Figure 2 depicts measured density data plotted against temperature for all investigated alloy samples. The temperature dependencies of density were investigated in the temperature interval from alloy melting point to 1600 °C during which a linear decline in density with increasing temperature was observed. The densities further depended on the chromium content. At 1600 °C, the highest value (7.04 g·cm -3 ) was obtained for Sample 1 containing 1.925 wt.% Cr and the lowest (6.79 g·cm -3 ) for Sample 4 holding 4.760 wt.% Cr. In summary, the density decreased with higher chromium content, which was consistent with the findings published in articles [14,19]. Parameters listed in Table 2 were obtained through the least square fitting of measured data to Equation (1) [19] where ρref is density (g cm -3 ) at reference temperature (liquidus temperature) Tref (°C), and dρ/dT (g cm -3 °C -1 ) is the temperature coefficient of the density. Experimental densities were compared with those calculated by Thermo-Calc software operating with the TCFE8 Steel/Fe-Alloy database. The database is applicable for different grades of Fe steels and alloys with the recommended content of specific alloy elements. Figure 3 presents a comparison of both densities, whereas a relative error was not larger than 3.5 %. The temperature dependence of surface tension followed an upward trend for all examined alloys, as can be seen from Figure 4. In all cases, the temperature coefficient of surface tension dσ/dT was positive and dependent on the activity of sulfur and oxygen ( Figure 5). This phenomenon can be explained by the Gibbs adsorption isotherm which assumes that the free surface of a metal drop is covered by a monolayer of surface-active elements like oxygen and sulfur tending to lower surface tension. During heating, these elements are desorbed into the bulk and, therefore, surface tension rises [19]. Such behavior was observed for samples 1-3 where the sulfur content was highest, unlike for sample 4 having the lowest content of surface-active elements (sulfur and oxygen). The linear Equation (2) was fitted to the measured data of surface tension according to [19] where σref is surface tension (mN m -1 ) at reference temperature (liquidus temperature) Tref (°C), and dσ/dT (mN m -1 °C -1 ) is the temperature coefficient of the surface tension. The values of the calculated parameters for individual samples are summarized in Table 3. The influence of chromium content on surface tension was assessed by the Kruskal-Wallis test [24], which revealed that the effect was statistically significant (p-value << 0.001). The differences in surface tension were statistically significant at all observed chromium levels. From Figure 4, it can be assumed that the surface tension of the examined samples increased with the increasing content of chromium. This is probably due to its strong affinity for oxygen, which results in reduced oxygen activity and increased surface tension. Comparable results were achieved in article [9]. Another significant quantity obtained by a sessile drop method is the wetting angle, i.e. the contact angle between the Fe-alloy and the alumina substrate. The temperature dependencies of wetting angles are shown in Figure 6. The rising temperature caused a slight increase in wetting angle values. Besides, they also grew with an increasing chromium content in the studied samples (Figure 7).

Figure 7. Images of alloy droplets taken during measurement at a temperature of 1550 °C.
The interfacial tension (σsl) between the molten Fe-alloy and the solid alumina substrate was calculated using Young's equation (3). This equation involves experimentally determined surface tension values (σlg) of the examined steel samples, the wetting angles (θ) between the steel melt and alumina and the surface tension of the alumina substrate (σsg).
Where σsl is interfacial tension (mN m -1 ), σsg surface tension of the alumina substrate (mN m -1 ), σlg is surface tension and θ (deg.) denotes wetting angle. Nogi and Ogino [25] reported that the surface tension of solid alumina is 750 mN m -1 at 1600 °C and temperature coefficient of alumina surface tension is -0.1 mN m -1 °C -1 . The latter quantity is acceptable in the measured temperature interval [26]. Figure 8 depicts the temperature dependence of the interfacial tension of all examined samples.   The surface tension of the investigated steels was also plotted against the activity of oxygen ( Figure 11) and sulfur ( Figure 12) at temperatures of 1520, 1550 and 1600 °C. It follows from these figures that surface tension decreases with increasing activity of surface-active elements. Further, at temperature 1520 °C, the surface active elements influence surface tension in a larger scale and drop in this quantity is more pronounced, in contrast to the situation at 1600 °C. Therefore, it can be assumed that the surface of molten steel contains more surface-active elements at lower temperatures. It also confirms that as the temperature grows, these elements are desorbed into the bulk, having less effect on surface tension. Additionally, desorption caused by chromium and oxygen evaporation should be taken into consideration. As a consequence, the surface tension rises with increasing temperature. Figure 11. Surface tension as a function of the logarithm of oxygen activity at selected temperatures.

Figure 12. Surface tension as a function of the logarithm of sulfur activity at selected temperatures.
The interaction of the steel droplet with the ceramic substrate was studied by SEM / EDS analyses for all examined samples. However, it was most significant in sample 4 containing the most chromium (Figures 13 -16). Based on the EDS analysis of the metal droplet, it was found that there has been no dissolution of aluminum in the melt, and their free surface showed no signs of oxidation. Interaction at high temperatures resulted in the inhomogeneous surface of the ceramic plate as presented in Figures 13 and 14, where four concentric regions were identified. Unreacted area A corresponded to the non-affected surface where the following elements O, Al, Ca and Si, were detected by EDS analysis. Thus, it can be assumed that the dominant component was Al2O3. Also, CaO and SiO2, the binders used in sintering, were determined, but to a lesser degree (Figures 14-15A). In addition, particles of the melt, predominantly Fe and Cr, were present in areas B and C (Figures 14-15B and 14-15C). As for the chromium, it may be considered that the trivalent Cr(III) could substitute Al(III) in the alumina since it has a similar atomic radius. This is also indicated by the reddish coloring of these areas as chromium has two strong adsorption bands in the visible spectrum [28]. There was also an increased calcium content in area C. The increase in calcium was most pronounced in area D (Figures 14-15D), where the metal drop was located, and the interaction altered the substrate most substantially. As the calcium-rich particles have a hexagonal crystal structure, calcium was presumably involved in the formation of high-temperature mineral hibonite (CaAl12O19). Above mentioned conclusions are also confirmed by Figure 16.
Towards the metal droplet, there is a rise of calcium content and decline in aluminum content, which correspond to the occurrence of hibonite, especially in the area under the metal droplet.
The silicon content is more or less the same in all regions. The increased presence of chromium in the B and C regions correlates with the occurrence of the metallic melt.

Conclusions
The influence of temperature and chromium content on the surface and interface properties of four model alloys was assessed in this work. These dependencies were studied in a concentration range of 1.93 -4.76 wt.%. The following conclusions can be drawn from this experimental research:  The linear decrease in the density of the examined alloy samples with increasing temperature was observed. Density also declined with a rising chromium content in the samples. The comparison of experimentally determined density values with theoretically calculated by Thermo-Calc software showed a small relative error not exceeding 3.5 %.  The surface tension of the samples in contact with Al2O3 substrate showed a linear increase as a function of temperature. Positive temperature coefficient of surface tension can be explained by a higher content of surface-active elements (oxygen and sulfur). Besides, the chromium content increased the surface tension values since this element has a relatively strong affinity for oxygen, with the result that oxygen activity in the melt is reduced and consequently surface tension increases.  The contact angle (wetting angle) between the alloy samples and the alumina substrate depend on the temperature and chromium content, i.e. it raised with their increase.  Further, as the temperature and chromium content increase, the interfacial tension between the steel melt and the alumina substrate increases, unlike the work of adhesion and the spreading coefficient.  The surface tension depends on the activity of the surface-active elements, decreasing with the increase of oxygen and sulfur activity.  Upon high temperatures, the interaction between alumina substrate and alloy samples resulted in the formation of high-temperature mineral hibonite in the area under the droplet. The calcium content raised toward the metal droplet, reaching the highest values in the area below it. In addition, a higher amount of iron and chromium was detected in the area close to the metal droplet by EDS analysis.