Wetting of Al₂O₃ by Molten Aluminum: The Influence of BaSO₄ Additions

Abstract

The effects of BaSO₄ additions on the wetting of alumina by molten aluminum were studied by the sessile drop technique. To study the effect of BaSO₄ decomposition (1100-1150^∘C), the additions were treated at two temperatures 700^∘C (973 K) and 1450^∘C (1723 K), respectively. BaSO₄ additions at low and high temperatures did not improve the nonwetting character of these compositions. However, at higher firing temperature, the formation of BA₆(BaO•6Al₂O₃) has a nonwetting trend with increasing its content. To address the BA₆ specifically a pure BaO•6Al₂O₃ was produced and tested. It was more nonwetting than the pure alumina. After the analysis of the contact angles for the BaSO₄ and the BA₆(BaO•6Al₂O₃), it was concluded that these additions to alumina do not inhibit wetting by molten aluminum. In fact, at the addition levels common for refractories, the wetting tendency of molten aluminum is enhanced. Alternative explanations for the effectiveness of BaSO₄ additions to alumina refractories are discussed.

1. Introduction

The wetting of molten metals on solid ceramics such as described by the liquid/solid contact angle (𝜙) directly affects whether it is possible to use a specific ceramic for many different engineering applications. Potential uses include composites, brazed components, the bonding of ceramics with metals, and the use of ceramics as refractories to contain molten metals [1].

Molten aluminum containment during its production process has used certain chemical additives to alumina refractories starting from late 1970s. These have been suggested to decrease the wetting of the refractory by molten aluminum and thereby reduce the aluminum penetration into the monolithic refractory products. These chemical additives include zircon (ZrSiO₄), barium sulfate (BaSO₄), calcium fluoride (CaF₂), and some glass frits as well. The levels of these additions to the refractories are usually only from 0.5 to about 6 wt% [2–13].

The Handbook of Chemistry and Physics [14] reports that BaSO₄ melts at 1580^°C, however prior to melting, it decomposes. The decomposition reaction occurs from about 1100^°C to 1150^°C: BaSO₄⟶BaO+SO₂+12O₂. (1) This reaction defines two temperature regimes for the effects of BaSO₄ additions to Al₂O₃. One is below 1100^°C and the other is above 1150^°C, the latter after the sulfate has decomposed to the oxide, BaO.

Brosnan [2], in his review of refractory corrosion, has suggested that the BaSO₄ additives to aluminum refractories may act to provide a mechanism for the oxidation of the molten aluminum. This is important if the metal penetrates the exposed surfaces of the refractory. Oxidation could occur through reactions of the type: 3BaSO₄+8Al(l)⟶3BaAl₂O₄+Al₂S₃,Δ𝐺_933K=−3215kJ/mole3Δ𝑉=10%,BaSO₄+8Al(l)⟶3BaO+Al₂S₃+3Al₂O₃,Δ𝐺_933K=−2870kJ/moleΔ𝑉=−2%,BaSO₄+2Al(l)⟶BaAl₂O₄+S(g),Δ𝐺_933K=−596kJ/moleΔ𝑉=15%. (2) These reactions suggest that BaSO₄ can react with molten Al above its melting point 660^°C (~933 K) and form stable phases like BaAl₂O₄ (spinel), BaO, Al₂S₃, and Al₂O₃.

At temperatures above that of the Al melting point, barium aluminates can be formed after the BaSO₄ decomposes. Although BaSO₄ is reported to melt at 1580^°C, the study by L’vov and Ugolkov [3] suggests that the decomposition reaction is kinetically favorable from 1124^°C (1397 K) to 1154^°C (1450 K). Mohazzabi and Searcy [4] applied mass spectrometry to determine if SO₃ was present in the products of (1). They concluded that the decomposition process is correctly described by (1). Table 1 summarizes some of the kinetic results from these studies [3, 4].

Table 1 

Activation energies (Q) reported for (2).

The data in Table 1 suggests that the BaSO₄ decomposition occurs over the temperature range from 1124^°C (1397 K) to 1267^°C (1540 K). After the BaSO₄ decomposition, the barium oxide may react with alumina to form stable oxide phases. These may include the barium hexaluminate (BaO•6Al₂O₃) or the monoaluminate spinel (BaO•Al₂O₃), as illustrated in the two following reactions: BaO+6Al₂O₃⟶BaO•6Al₂O₃,ΔG_1373K=−1133.3kJ/moleΔ𝑉=5%,BaO+Al₂O₃⟶BaO•Al₂O₃,ΔG_1373K=−132.88kJ/moleΔ𝑉=15.5%. (3) As these reactions are highly exothermic, extreme care must be used during experiments to observe or to utilize these reactions.

Teiken and Krietz [5] have reported that the addition of 6%-BaSO₄ reduces the penetration of molten aluminum in conventional castables and in low-cement castables. A comparison of these refractories gives the advantage in penetration resistance to the conventional castable. Teiken and Krietz also have suggested that alumina or aluminosilicate refractories are mostly improved for the lining of aluminum-melting furnaces and other refractory containers just from 980^°C to 1090^°C.

Afshar and Allaire [6] have also studied the additions of BaSO₄ and those of CaF₂ to aluminosilicate castable refractories to reduce their corrosion. They reported that an addition of 7%-BaSO₄ improved the corrosion resistance of refractory samples for firing temperatures at <1000^°C (1273 K). For additions of CaF₂, they reported less efficient corrosion protection after firing below 1000^°C. The use of 3.5% CaF₂ actually provoked the formation of large cracks in the refractory, facilitating metal-penetration and enhancing corrosion. These studies suggest that these additives may be limited to temperatures <1000^°C (1273 K).

The addition of BaSO₄ to phosphate bonded, or phosphate-free plastics, ramming mixes, brick, mortar, or castables has been confirmed to reduce molten aluminum penetration and significantly reduces dross and molten aluminum adherence [5–9]. The principal additives effects studied by these researchers were their reactions with molten aluminum and its alloys. Investigators have also suggested that the additives may fill the voids in the refractories and inhibit penetration by reducing porosity at the interfacial contact region between molten Al and additives.

The above empirical corrosion tests were all completed at <1090^°C (1363 K) in air or by ASTM guidelines with the cup test. The barium sulfate also can be affected by the partial pressure of oxygen in the presence of molten aluminum as follows: BaSO₄(s)+Al(l)+O₂(g)⟶Ba_aluminates(s)+Al₂(SO₄)₃(s). (4) This reaction is even more exothermic than the previous ones as they are analyzed without any oxygen (2). The Gibbs free energy changes at 1000^°C (1363 K) for the formation of barium monoaluminate (BaAl₂O₄) and hexaaluminate (BaAl₁₂O₁₉) are −4866 kJ/mol and −6620 kJ/mol, respectively. These energies are high and are likely possible to occur in the dross zone of a refractory lining, which contains the reactants for the completion of (4).

The above observations suggest that additions of BaSO₄ to alumina refractories may have two temperature regimes of protection for these refractories. One is below the temperature of the BaSO₄ decomposition reaction at ~1100^°C and the second is above the decomposition reaction: >1250^°C. From previous studies, it was confirmed that additions of compounds, such as BaSO₄ to refractories for molten aluminum containment, reduced their corrosion [5–9]. However, the mechanism of this corrosion reduction is not clear. It could be either decreased molten Al wetting of the refractory, or a pore channel blocking effect by Ba-compound formation. To address the wetting aspect of this controversy, this molten aluminum wetting study of the effects of BaSO₄ additions to an alumina was undertaken.

2. Experimental Procedures

Consider a flat, undeformable, perfectly smooth, and chemically homogeneous solid surface in contact with a nonreactive liquid in the presence of a vapor phase. If the liquid does not completely cover the solid, the liquid surface will intersect the solid surface at a contact angle, 𝜙 or 𝜃, depending on the particular angular definition. The wetting relationship of the surface tensions of the solid-vapor interface (𝛾_𝑠𝑣), the solid-liquid interface (𝛾_𝑠𝑙), and the liquid-vapor interface or surface tension (𝛾_𝑙𝑣) in a sessile drop configuration is usually expressed by Young equation [15]. It can be written either as 𝛾cos𝜙=^𝑠𝑣−𝛾_𝑠𝑙𝛾_𝑙𝑣 (5) or 𝛾cos𝜃=^𝑠𝑙−𝛾_𝑠𝑣𝛾_𝑙𝑣. (6) Referring to Figure 1, the 𝜙 definition of the wetting contact angle is applied throughout this thesis. Because 𝜙+𝜃=𝜋, it naturally allows the interchangeable application of both definitions for the wetting angle. Some researchers use 𝜃.

Figure 1 

Surface energy balance and the contact angle equilibrium definition of the Young equation. Note the two different definitions of the contact angles, 𝜙 and 𝜃.

Figure 2 illustrates the definition of (5) using 𝜙 as the contact angle. When 𝜙 is equal to 180^°, no wetting occurs, and the droplet sits over the solid in point contact. When 𝜙 is >90^°, a partial nonwetting condition prevails. Partial wetting occurs when 𝜙 is <90^°. However, when the contact angle is equal to 0^°, it is understood that complete wetting and droplet spreading occurs over the solid surface.

(a)

(b)

(c)

(d)

Figure 2 

Wetting stages resulting from the interaction of a solid and a liquid.

2.1. Specimen Preparation

The effect of BaSO₄ on the wettability of alumina by molten aluminum was studied for several Al₂O₃-BaSO₄ compositions. One was pure alumina and other three had BaSO₄ additions at the 1, 3, and 5% levels. The specimen preparation consisted of mixing a 99.96% 0.5𝜇m alumina powder with 1, 3, and 5% of a 99.9% BaSO₄ (1𝜇m). Individual 100 g batches were prepared by ball mill mixing for 2 hours using a cylindrical container and spherical alumina balls with a ratio of balls to mix of 3 : 1. The ball-mixing fluid medium was isopropanol. A mix speed of 100 rpm was used. After mixing, the powder was dried at 70^°C (343 K) for 48 hours.

The dried powder was then passed through a −325 mesh sieve (45𝜇m) to break up the agglomerates. The resultant powder was weighed for 5 g specimens for axial pressing of ~1 inch diameter discs at 150 MPa. Sintering schedules were using a heating rate of 5^°C/min to 700^°C (973 K) and to 1450^°C (1723 K) for one hour hold each in air.

In addition, the pure BaO•6Al₂O₃ phase was produced using same experimental set up and used the theoretical composition of BaO•6Al₂O₃ by mixing 27.6 wt% BaSO₄ with 72.4 wt% Al₂O₃. To produce the pure BA₆ (BaO•6Al₂O₃) phase, the Al₂O₃-BaSO₄ specimens were sintered using a heating rate of 5^°C/min to 1700^°C (1973 K) for one hour hold in air.

The BaSO₄-Al₂O₃ specimens at different firing temperature were subsequently analyzed by density using the Archimedes method and by X-ray diffraction measurements (35 kV, 25 mA, and Cu_𝜆=0.15406 nm). The surfaces of all ceramic specimens prepared for the wetting studies were semipolished with an SiC # 325 mesh sandpaper to achieve a constant roughness effect.

Aluminum metal specimens were fabricated from a commercial 1020 aluminum ingot provided by Alcoa. The chemical analysis of the Al is shown in Table 2. Impurities were Si, Fe, and Mg, each at the <0.20% level. The Al test specimens were 0.5 cm diameter cylinders with a height of 0.5 cm. To minimize oxidation during the experiments, the Al specimens were first semipolished with an SiC # 400 mesh sandpaper then ultrasonically cleaned in isopropanol.

Table 2 

Chemical analysis of the 1020 aluminum (wt%).

2.2. Sessile Drop Experiments

Sessile drop tests of the molten 1020 aluminum on the three different types of BaSO₄-Al₂O₃ specimens were completed using a commercial sessile drop apparatus manufactured by Centorr Associates, Inc. (Suncook, NH, USA). It is schematically illustrated in Figure 3. The experiment consisted of placing the aluminum metal cylinder on the refractory substrate in the center of the furnace. After closing, the furnace was evacuated to less than 50 millitorrs. An ultrahigh purity argon (~99.999%) was introduced to a pressure of 1000 millitorrs. These vacuum/argon steps were repeated four times to reduce the experimental chamber oxygen level. The sessile drop experiments were then completed in the argon atmosphere to avoid any of the oxide skin effects that have been observed on Al droplets and reported by other investigators [16–25]. Measurements in the sessile drop furnace were under argon gas at ~1520 Torr. The partial pressures of several gas species were P_H2=10⁻¹⁶,P_H2O=10⁻¹⁹,P_N2/CO=10⁻¹⁹, and P_O2=10⁻¹⁷ Torr as determined by a gas analyzer (MKS Vision 100-B).

Figure 3 

Schematic of the sessile drop apparatus.

Heating of the test assembly was manually controlled by applying current (mA) to the heating element until the desired temperature was achieved. The actual heating profile is illustrated in Figure 4. It had isothermal temperature holds from 800^°C (1073 K) to 1200^°C (1473 K) at intervals of 200^°C (473 K). Once the solid aluminum cylinder melted during heating, it immediately formed a droplet. A camera lens and a video monitor (black and white) connected with a VCR recorded the experiments. The droplet shape was monitored for 20 minutes at each of the holding temperatures. This length of time, ~20 minutes, was the same as that reported by Li [18].

Figure 4 

The heating schedule followed in the sessile drop measurements.

A principal difficulty when studying the wetting of the molten Al and a solid Al₂O₃ system couple is the presence of a high partial pressure of oxygen. Aluminum metal may react with oxygen at high temperatures [16–25] to form gas by the reaction: Al₂O₃(film)+4Al(l)=3Al₂O(g). (7) Although this equation is not thermodynamically favorable (Δ𝐺_1173K=1,040.7 kJ), it has been reported to occur during experimental observations and by spectrographic analysis [18, 21].

Droplet shape analysis is required to accurately determine the contact angle (𝜙). Videotapes of the sessile drops were transferred to a PC by using the DFG/LC1 frame grabber with Grab and View 2.1 software. This allowed the direct observation of individual frames at any time during the duration of the experiment. It has an efficiency reading of 0.01 second. Figure 5 illustrates a droplet and its representation. Analysis determined the droplet profiles and fitted them to theoretical shapes obtained from the solution of the Laplace equation of capillarity [22].

(a)

(b)

(c)

Figure 5 

Typical droplet acquisition data to obtain the contact angle (𝜙) by image analysis and optimization of the Laplace equation of capillarity.

3. Results

3.1. Wetting of the Al₂O₃-BaSO₄ System Fired at 700^°C (973 K)

After firing in air at 700^°C (973 K), the BaSO₄ content in the alumina was confirmed by microchemical analysis in a scanning electron microscope (SEM) using the EDS technique (35 kV and 25 mA) getting a 99.85% of reading accuracy. It was also checked by X-ray diffraction (XRD) as shown in Figure 6 (35 kV and 25 mA) using Rietveld measurements. The EDS results were for an average of five specimens. The BaSO₄ contents were found to be 0.98%, 2.97%, and 4.98%. Firing at a temperature of 700^°C does not affect the BaSO₄. The densities of the Al₂O₃-BaSO₄ specimens fired at 700^°C (973 K) are illustrated in Figure 7. Open porosity increased as the amount of BaSO₄ increased. The open porosity changed from 34.0 to 40.0%, and the density decreased from 2.4 to 1.9 g/cm³ for 1.0 and 5.0% of BaSO₄, respectively. These Al₂O₃-BaSO₄ specimens were quite porous as well as the pure aluminas.

Figure 6 

X-ray diffraction patterns of the Al2O3-BaSO4 specimens fired at 700°C (973 K) for one hour. Observe that BaSO4 is present in each sample and its intensity increases proportionally to the amount of BaSO4 added.

Figure 7 

Density measurements by Archimedes method for Al2O3-BaSO4 specimens at different additive amounts of BaSO4, sintered at 700°C (973 K) for one hour.

Figure 8 illustrates the Al wetting contact angles as a function of time for the Al₂O₃-BaSO₄ specimens at 800^°C (1073 K), 900^°C (1173 K), and 1000^°C (1273 K). These measurements were at temperatures above the sintering temperature of the mixes. The wetting of the BaSO₄-free alumina is slightly temperature dependent. The wetting contact angles, the 𝜙_𝑜-values at zero time, are 134^° at 800^°C (1073 K), 132^° at 900^°C (1173 K), and 128^° at 1000^°C (1273 K). The contact angle is >90^° at all temperatures, a nonwetting condition as previously illustrated in Figure 2.

(a)

(b)

(c)

Figure 8 

Wetting contact angle (𝜙) variation with time at different temperatures for the Al2O3-BaSO4 systems fired at 700°C (973 K) for one hour. The regression equations are listed on the figures.

Additions of BaSO₄ affect the wetting contact angles (𝜙) differently at the three different temperatures. At the lowest temperature, 800^°C (1073 K), the 1% addition of BaSO₄ actually decreases the contact angle to 105º, nearly 30^° less than that for pure alumina. At 800^°C, the 3% addition of BaSO₄ also increases the wetting tendency of the alumina by the molten aluminum. The 5% BaSO₄ addition has nearly the same contact angle profile as that for the pure alumina, but the 3% addition decreases the contact angle (𝜙) from 134^° to 126^°.

The effects of higher temperatures on the wetting of the Al₂O₃-BaSO₄ specimens are always that of increasing the wetting tendency. At 800^°C (1073 K), only the 5% BaSO₄ addition appeared to decrease the wetting tendency of pure alumina, all the wetting contact angles decreased, but those for the BaSO₄ additions decreased more rapidly than those for pure alumina. The result is that from 900^°C (1173 K) to 1000^°C (1273 K), the pure alumina is more difficult to wet than the alumina with the BaSO₄ additions.

Figure 9 illustrates the contact angle (𝜙_o) as a function of temperature. These results show that by adding BaSO₄ to alumina, the wetting contact angles are less than those for pure alumina. Although the contact angles remain >90º, their decreases indicate an increased wetting tendency with the BaSO₄ additions. The 5%-BaSO₄ specimen has a slightly larger contact angle than alumina at 800^°C (1073 K). However, at higher temperatures the contact angle becomes less than that for the pure alumina.

Figure 9 

Variation of the wetting contact angle (𝜙𝑜) at time zero with temperature for the pure alumina and BaSO4 additions of 1%, 3%, and 5% after firing at 700°C (973 K).

Variation of the 𝜙_𝑜-values with temperature can be expressed as 𝜙_𝑜=𝜙^′_𝑜−(𝜕𝜙_𝑜)𝜕𝑇^𝑡𝑇, (8) where 𝜙^′_𝑜 is the value at zero Kelvin and the slope (𝜕𝜙_𝑜/𝜕𝑇)_𝑡 represents the decrease of the wetting contact angle with increasing temperature. These two parameters increase with the BaSO₄ content as illustrated in Figure 9. The BaSO₄ addition effects in alumina at temperatures from 800^°C (1073 K) to 1000^°C (1273 K) are generally to increase the wetting tendency by molten aluminum. However, at no times the contact angles were reduced to less than 90º. As 𝜙<90^∘ is the criteria for partial wetting, it must be concluded that no wetting process was observed, but the additions of BaSO₄ to Al₂O₃ increased the wetting tendency.

3.2. Wetting of the Al₂O₃-BaSO₄ System after Firing at 1450^°C (1723 K)

After firing the Al₂O₃-BaSO₄ specimens at 1450^°C, the BaO content in the alumina specimens was analyzed by microchemical analysis in an SEM (EDS) (35 kV and 25 mA) getting a 99.85% of reading accuracy and by X-ray diffraction (XRD). The EDS results represent an average value for five specimens. Only Ba, Al, and O were detected. Sulfur was absent. The amounts of BaO increased in the order of 0.64, 1.92, and 3.25% for the initial introductions of 1.0, 3.0, and 5.0% of BaSO₄. From (1), the reaction for BaSO₄ decomposition, the theoretical amounts of BaO are 0.66, 1.97, and 3.28% for additions of 1.0, 3.0, and 5.0% BaSO₄. These values are nearly the same as those obtained by EDS microanalysis. This confirms that the BaSO₄, which was introduced to the alumina and then fired at 1450^°C (1723 K), decomposed as described by (1). The BaO obtained then reacted with the Al₂O₃ to form BaO•6Al₂O₃ (3). This is confirmed by the XRD in Figure 10, where the two phases coexist. The amounts of BaO•6Al₂O₃, obtained by X-ray diffraction analysis, are of 3.10, 9.50, and 16.20%, respectively. These values are in agreement with the theoretical ones of 3.28, 9.85, and 16.42%, respectively.

Figure 10 

X-ray diffraction patterns of Al2O3-BaSO4 specimens fired at 1450°C (1723 K) for one hour. Observe that the BaSO4 has decomposed and reacted with the Al2O3, forming the stable BaO•6Al2O3 phase.

The equilibrium binary phase diagram of Al₂O₃ and BaO illustrated in Figure 11 can be used to explain the above results. It is shown that when adding BaO to Al₂O₃ the first stable compound formation is BaO•6Al₂O₃. This compound has about 20 wt% BaO and 80 wt% Al₂O₃. It is congruently melting at ~1915^°C (2188 K) and it is a highly refractory material.

Figure 11 

Binary phase diagram of the BaO-Al2O3 system [23].

The Al₂O₃-BaO•6Al₂O₃ specimen densities, after firing at 1450^°C/1hour (1723 K), are presented in Figure 12. The densities are much higher and the porosities are much lower than the previous series of Al₂O₃-BaSO₄ specimens. The results show an increase of porosity as the amount of BaO•6Al₂O₃ increases. The total porosity changes from 12.0% to 17.0% for the 1.0% and 5.0% additions of BaSO₄, respectively.

Figure 12 

Density measurements by the Archimedes method in Al2O3-BaO•6Al2O3 specimens after firing 1450°C (1723 K) for one hour.

The wetting contact angle (𝜙) variations with BaO•6Al₂O₃ content and their changes with temperature are illustrated in Figure 13. Large contact angle differences are obtained from the BaSO₄ additives after firing at 1450^°C. The temperature effect is evident. It normally decreases the contact angle. Figure 13 illustrates the aluminum contact angles of the Al₂O₃-BaO•6Al₂O₃ specimens at the three temperatures, 800^°C (1073 K), 1000^°C (1273 K), and 1200^°C (1473 K). The wetting of the alumina without BaSO₄ additions after firing at 1450^°C is only slightly temperature dependent. Its contact angles (𝜙_𝑜) at 𝑡=0 are 125^° at 800^°C, 121^° at 1000^°C, and 118^° at 1200^°C.

(a)

(b)

(c)

Figure 13 

Wetting contact angle (𝜙) variation with time at different temperatures for the Al2O3-BaO•6Al2O3 system fired at 1450°C (1723 K) for one hour.

The BaO•6Al₂O₃ effects on the contact angles (𝜙) are similar for the three temperatures. At the lowest temperature, 800^°C (1073 K), the 3.1% BaO•6Al₂O₃ decreases the wetting contact angle to 95º, about 30^° less than that for the pure alumina. The addition of BaO•6Al₂O₃ has the effect of increasing the wetting tendency of the alumina by molten aluminum. At 800^°C (1073 K) the 9.5% and 16.2% of BaO•6Al₂O₃ the wetting contact angles decrease to 105^° and 107^°, respectively. This suggests that by increasing the BaO•6Al₂O₃ phase in the microstructure the contact angles will increase toward less wetting. One can speculate that further increases of BaO•6Al₂O₃ phase content may increase the wetting contact angles beyond that obtained for the pure alumina.

The effect of temperature decreases the wetting contact angles (𝜙) for all of the Al₂O₃-BaO•6Al₂O₃ specimens. From 800^°C (1073 K) to 1200^°C (1473 K) contact angles are reduced for all levels of BaO•6Al₂O₃ additions. The result is that at all three temperatures, the pure alumina is the most difficult to wet. The presence of BaO•6Al₂O₃ consistently decreases the tendency toward wetting. It must be concluded that the amounts of BaO•6Al₂O₃ additions to alumina tested at temperatures from 1000^°C (1273 K) to 1200^°C (1473 K) do not reduce or inhibit the wetting of pure alumina by molten aluminum. Figure 14 summarizes the effects of temperature on the wetting contact angles for the addition of BaO•6Al₂O₃ additions.

Figure 14 

Variation of the wetting contact angle (𝜙o) with temperature for the BaO•6Al2O3 additions to alumina and after firing at 1450°C (1723 K) for one hour.

Figure 14 shows a linear tendency of the decrease of the wetting contact angle (𝜙_𝑜) with temperature. This describes the effects of temperature on the contact angle variation. With BaO•6Al₂O₃ additions to alumina, it is observed that the contact angles are increased from those obtained with pure alumina. However, the resultant slopes (𝜕𝜙_𝑜/𝜕𝑇)_𝑡 are similar to those for the pure alumina.

The decrease of 𝜙_𝑜 with increasing temperature is described by (8). It is probable that the BaO•6Al₂O₃ phase may increase the resultant contact angle of molten aluminum beyond that for pure alumina. The trend in Figure 14 certainly suggests that larger additions of AB₆ (BaO•6Al₂O₃) may achieve contact angles above from those of the pure alumina and actually decrease the wetting by molten Al.

3.3. Wetting of the Pure BaO•6Al₂O₃ Phase

Davis et al. [26, 27] suggest different processing routes for the synthesis of pure BaO•6Al₂O₃ and establish that this barium compound is of interest as a refractory. Barium monoaluminate BA (BaAl₂O₄) hydrates, m.p. ~1800^°C (2073 K), and could be the basis of cement or grouting material. The hexaluminate BA₆ (BaO•6Al₂O₃), however, is the more likely engineering ceramic, m.p. ~1925^°C (2198 K) [26, 27].

The only study in the literature reporting properties of the barium hexaaluminate BA₆ (BaO•6Al₂O₃) is that by Zawrah and Khalil [28]. They prepared the compound from both low and high purity materials by sintering at 1700^°C (1973 K). The first was prepared from high purity BaSO₄ and Al₂O₃, while the second was prepared from natural minerals, Egyptian barite and Chinese bauxite. They found that by firing the pure materials at 1700^°C (1973 K), a mix of BA₆ (BaO•6Al₂O₃) and Al₂O₃ was present in the microstructure. However, when firing the natural minerals at 1700^°C (1973 K), a mix between BA₆ (BaO•6Al₂O₃), Al₂O₃, and BAS₂ (BaO•Al₂O₃•2SiO₂) was present in the microstructure. They identified these phases by X-ray diffraction. The X-ray diffraction results for their pure materials are in agreement with those presented in Figure 15 for the present study.

Figure 15 

X-ray diffraction patterns of the Al2O3-BaSO4 specimens fired from 1200°C (1473 K) to 1700°C (1973 K).

Figure 15 illustrates the reaction of the Al₂O₃-BaSO₄ specimens to form BA₆ (BaO•6Al₂O₃) as a function of the firing temperature. After firing at 1200^°C (1473 K) there remains a large amount of BaSO₄ that has not reacted with the Al₂O₃. However, on firing at 1400^°C (1673 K) is observed that the reaction process has initiated and as a result there is a formation of BA₆. After firing for one hour at 1700^°C (1973 K), the formation of the BA₆ phase is complete.

The densities of the BA₆ (BaO•6Al₂O₃) specimens are illustrated in Figure 16. There is a reduction of open porosity as the temperature increases. The open porosity is 6.5% after firing at 1700^°C (1973 K). The density is 3.97 [g /cm³]. The pure alumina fired at the same conditions had 5% open porosity and a density of 3.52 [g /cm³].

Figure 16 

Density and open porosity variation with temperature after firing BaO•6Al2O3 at 1700°C (1973 K) for one hour.

The Al wetting contact angle (𝜙) for the BA₆ phase is illustrated in Figure 17. The results for the pure alumina fired at 1700^°C (1973 K) are also presented. Distinct differences with respect to the pure alumina are obtained for all of the temperatures. This result was predicted by Figures 13 and 14 that show increases of the Al contact angles proportional to the BA₆ phase content. The BA₆ phase actually increases the wetting angle trend of molten aluminum beyond that for pure alumina. The BaO•6Al₂O₃ phase is consistently more nonwetting than pure alumina.

Figure 17 

Wetting contact angle (𝜙) with time at constant temperatures. Note the Al contact angles obtained with the pure BA6 (BaO•6Al2O3) are larger than those obtained with the pure Al2O3.

The Al contact angles (𝜙_o) at time zero for the BA₆ phase changed from 135.3^° for 800^°C (1073 K) to 127.4^° for 1200^°C (1473 K). These 𝜙_os are above those obtained for the pure alumina of 122.7^° and 115.3^° at the same temperatures. These represent a nearly 10% increase of the contact angles, respectively. Figure 18 illustrates this variation of contact angle with temperature. It is observed that the BA₆ (BaO•6Al₂O₃) has larger wetting contact angle than that for the pure alumina for all temperatures. Both have similar slopes, (𝜕𝜙_𝑜/𝜕𝑇)_𝑡. This clearly illustrates that the pure BA₆ (BaO•6Al₂O₃) phase is less wetting to the molten aluminum than pure alumina.

Figure 18 

Variation of 𝜙o with temperature for the couple systems of BaO•6Al2O3/Al and Al2O3/Al. Note that the BaO•6Al2O3 phase is more nonwetting than pure alumina.

3.4. The Pure Alumina Wetting Contact Angle

In the assessment of the effect of BaSO₄ and BaO•6Al₂O₃ additions, the fired refractory ceramics were compared with the pure alumina fired under the same conditions. The contact angles of the three aluminas were presented in Figures 8, 13, and 17 at the three different temperatures. The molten aluminum wetting contact angles are different for the three different alumina firing conditions. These three different firing schedules for the pure alumina produced three different levels of porosity, 34.5%—700^°C (973 K), 11%—1450^°C (1673 K), and 5%—1700^°C (1973 K), respectively. This shows the densification process of alumina with temperature and coupled with the microstructural differences, this naturally created different surfaces roughnesses. The reason for the different contact angles for these pure aluminas is their different structural surface roughnesses [28].

4. Summary and Conclusions

The additions of BaSO₄ to alumina have two different Al wetting regimes as the firing temperature is increased to produce refractory ceramic bodies. The decomposition temperature of BaSO₄ divides these regimes. For 𝑇<1124^∘C (1397 K) (700^°C (973 K) firing was used in this study) the BaSO₄ does not decompose. It remains as the sulfate mixed with alumina. It is observed that the resultant molten Al contact angles are less for the Al₂O₃-BaSO₄ mixtures than those obtained with the pure alumina. Specimens do not show higher contact angles than the pure alumina. The BaSO₄ does not reduce the wetting of alumina to molten aluminum. In fact, the BaSO₄ reduces the wetting contact angle toward more wetting conditions.

The Al₂O₃-BaSO₄ specimens fired at 𝑇>1450^∘C (1723 K) cause the decomposition of the BaSO₄ (2). The reaction results in a stable compound BA₆ (BaO•6Al₂O₃). The amount of BaO•6Al₂O₃ is proportional to the content of BaO from the sulfate decomposition reaction in (1). The BA₆ (BaO•6Al₂O₃) amounts obtained by SEM (EDS) analysis from 1%, 3%, and 5% of BaSO₄ additions are 3.1%, 9.5%, and 16.2%. The Al wetting angles of the Al₂O₃-BaO•6Al₂O₃ mixtures are decreased below that of the pure alumina. They do increase proportionally with the amount of BA₆ (BaO•6Al₂O₃) presents in the system. This suggests that increments of BA₆ (BaO•6Al₂O₃) in the system would increase the Al wetting contact angle, possibly significant at higher BA₆ phase levels.

The investigation of the wetting of the pure BaO•6Al₂O₃ ceramic by molten aluminum yielded wetting contact angles greater than those for pure Al₂O₃. This indicates that BaO•6Al₂O₃ is less wetting toward molten aluminum than pure alumina. However, the contact angles for BaO•6Al₂O₃ ceramic do not approach the 𝜙=180^∘ mark indicative of an absolute nonwetting condition.

All of the contact angles decreased with increasing temperature. Higher temperatures promote wetting. They also might be expected to enhance refractory penetration and increased corrosion at higher temperatures.

The contact angles (𝜙) resulting from these refractory systems and 1020 Al decrease linearly with increasing temperature even in the presence of the additives. Measurements of the Al contact angles on the BaSO₄ additive containing alumina do not confirm that this additive inhibits wetting. In fact, in most cases the contact angles were reduced toward more wetting, suggesting that the BaSO₄ addition may actually promote wetting characteristics. However, BaSO₄ addition did not decrease the contact angles below 90º, the criterion for the transition from nonwetting to wetting. One must conclude that the nonwetting theory for the improvement of aluminum refractories by BaSO₄ additions does not appear to be the correct explanation.

There remain several other alternatives for the obvious corrosion improvements of aluminum containing refractories by BaSO₄ additions. One is the formation of compounds such as BaO•6Al₂O₃ or BaO•Al₂O₃ that fill the refractory pores and channels reducing porosity and the penetration of molten aluminum. Another is the suggestion by Brosnan [2] in his review, namely, that the sulfate participates in a reaction, or a series of reactions to promote the oxidation of molten aluminum if it does penetrate the refractory structure. Both of these are viable explanations.

Finally, it must be noted that this study was for the additions of BaSO₄ and its derivative compounds on the wetting of a pure alumina by a pure aluminum metal. These results may not directly translate to the wetting of other refractories such as the different aluminosilicates and cement-containing refractories by more complex highly alloyed aluminum metal.

Acknowledgments

The author is grateful to R. C. Bradt for his guidance and study of refractories materials. Special thanks are addressed to A. Saavedra and D. Stewart from Alcoa for the supply of the 1020 aluminum and to L. Krietz of Plibrico Co. (Chicago, Ill, USA) for the supply of the BaSO₄. The author would like to thank the Consejo Nacional de Ciencia y Tecnologia (CONACYT-Mexico) for its financial support of the graduate studies at the University of Alabama, Tuscaloosa, Ala, USA.