Solubility of some transition metal oxides in cryolite-alumina melts: Part I. Solubility of FeO,FeAl2O4, NiO,and NiAl2O4

The solubilities of FeO, FeAl₂O₄, NiO, and NiAl₂O₄ were measured in cryolite-alumina melts at 1020 °C. FeO was found to be the stable solid phase at alumina concentrations below 5.0 wt pct, while FeAl₂O₄ was stable above that. The corresponding figure for the nickel system was 3.0 wt pct Al₂O₃. These values correspond to Gibbs energies of formation of the aluminates (from the constituent oxides) of −17.6 and −29 kJ/mol, respectively. In alumina-saturated melts in the range 980 to 1050 °C, the solubilities of both aluminates increased with increasing temperature, the apparent enthalpies of solution being 65 kJ/mol for FeAl₂O₄, and 249 kJ/mol for NiAl₂O₄. Investigation of the solubilities of the aluminates as a function of the NaF/AlF₃ ratio in alumina-saturated melts at 1020 °C showed maxima at a molar ratio of around 5. The results are discussed in terms of the species apparently existing in the solution, and are consistent with the solute species being fluorides, not oxyfluorides. The activity coefficients of FeF₂ (liquid) and NiF₂ (solid) in dilute solution in cryolite are found to be 0.22 and 1.2, respectively.     

The solubility of aluminum in cryolite-alumina melts and the mechanism of metal loss

The solubility of aluminum in cryolite-alumina melts has been determined in laboratory experiments by analyzing rapidly-quenched samples of the melt after equilibration with metal at temperatures between 960 and 1060°C. The solubility in pure cryolite increases from 0.085 at 1020°C to 0.12 wt pct Al at 1060°C. The addition of alumina decreases the solubility at 1020°C to 0.081 with 5 pct A1₂O₃ and to 0.073 wt pct Al in melts saturated with alumina. Quenched samples have been taken from operating 130 kA prebake cells at different heights above the metal pad, both in the center channel and beneath the anodes. Within about 10 mm of the cathode the metal content is close to the equilibrium value obtained in the laboratory but above this level it decreases rapidly. It is suggested that oxidation occurs in a central zone of the electrolyte. Mechanisms of metal loss and implications for current efficiency are discussed.     

Alumina Solubility and Diffusion Coefficient of the Dissolved Alumina Species in Low-Temperature Fluoride Electrolytes

The solubility of alumina was measured by rotating an alumina cylinder (~500 rpm) in a high-purity melt for ~3 to 6 hours, crushing and sampling the frozen melt, and determining the oxygen content in a Leco analyzer. The alumina solubilities determined were as follows: (1) 3.2 ± 0.3 wt pct in NaF-AlF₃ eutectic at 1023 K (750 °C); (2) 3.0 ± 0.3 wt pct in NaF-AlF₃-CaF₂ (5 wt pct) at 1023 K (750 °C); and (3) 5.2 ± 0.5 wt pct in a KF-AlF₃ eutectic at 1003 K (730 °C). The alumina solubility in the KF-AlF₃ eutectic was 2 wt pct more than in the sodium analogue, offering the possibility of operating a low-temperature aluminum smelting cell without the need for an alumina slurry. The diffusion coefficient of the dissolved alumina species was determined in the NaF-AlF₃ eutectic at 1023 K (750 °C) using the rotating disc method and applying the Levich equation. Through a limited range of rotation rates, the system seemed to be mass-transfer controlled, and the diffusion coefficient was estimated to be in the range 1.8 to 2.2 × 10⁻⁶ cm² s⁻¹. This value is about five times lower than the values encountered at traditional aluminum smelting temperatures (~1233 K (960 °C)) and would result in relatively low mass transfer coefficients.     

Solubility of hydrogen in liquid Fe−Co−Ni alloys

The solubility of hydrogen in the Fe−Co−Ni ternary has been determined by the Sieverts' method over the temperature range 1500° to 1700°C. The solubility of hydrogen at 1600°C and 1 atm hydrogen pressure is 0.00264 wt pct in iron, 0.00224 wt pct in cobalt, and 0.00448 wt pct in nickel. Hydrogen follows Sieverts' law for all alloy compositions. The solubility surface rises smoothly from the Fe−Co binary to the nickel corner of the ternary, and when expressed as the free energy of hydrogen solution the surface is planar. The enthalpy of hydrogen solution is 8.0 kcal per g-atom H in iron, 8.5 kcal per g-atom H in cobalt, and 5.2 kcal per g-atom H in nickel and is planar for the entire ternary. Interaction parameters with hydrogen for Al, Cu, and Mn were established: ɛ _H ^Al =2.0, ɛ _H ^Cu , and ɛ _H ^Mn and are constant for the entire Fe−Co−Ni ternary. This paper is based on a portion of a thesis submitted by R. G. BLOSSEY in partial fulfillment of the requirements for the degree of Doctor of Philosophy at The University of Michigan.     

On the interfacial rate of reaction of CO2 with a calcium ferrite melt

Measurements of the rate of dissociation of CO₂ have been made by the¹⁴CO₂-CO isotope exchange technique on calcium ferrite melts with Ca/Fe = 0.30 at 1300 °C. Studies have also been made of the interfacial rates of oxidation of calcium ferrite melts with an average CaO content of 19.45 wt pct (CaJFe ≃0.33) in CO₂-CO atmospheres at 1362 °C. It is shown that the rates of oxidation are consistent with the rates of isotope exchange, indicating a common rate determining step. Measurements of the equilibrium Fe³⁺/Fe²⁺ ratio as a function of the CO₂/CO ratio for 19.3 wt pct CaO melts at 1360 °C and for 28.7 and 18.6 wt pct CaO melts at 1300 °C are found to be in close agreement with the deductions of Takeda, Nakazawa, and Yazawa. Combination of the equilibrium data with the results of the isotope exchange studies indicate that the apparent first order rate constant for the dissociation of CO₂ is inversely proportional to the square of the Fe³⁺JFe²⁺ ratio of the melt, as has been previously found for liquid iron oxide, lime-saturated calcium ferrites, silicasaturated iron silicates, and an equimolar “FeO”-CaO-SiO₂ melt.     

Solubility of water in lime-alumina-silica melts

The water solubility in fused silicates of the CaO-SiO₂ and CaO-SiO₂-Al₂O₃ systems has been measured using a vacuum fusion technique. The melts were equilibrated with nitrogen as “carrier gas” containing an accurately known water content. The solubility of water increased on the addition of lime to the melts of both systems. The effect of alumina is initially, to decrease the solubility. Additions above about 20 wt pct at a constant basicity of 0.6 however raised the water solubility of lime-silica melts. From the plot of oxygen density against the hydrogen concentration of the silicate melts of the system lime-silica-alumina studied in this investigation, it may be concluded that free hydroxyls take up interstitial positions in the silicate network. Wustite additions to lime-silica melts of basicity equal to 1 do not appear to influence the water solubility, although the rate of gas absorption from the furnace atmosphere is accelerated significantly. P. L. SACHDEV, formerly Graduate Student, Institut für Eisenhüttenwesen R. W. Technische Hochschule Aachen, West Germany A. MAJDIČ, formerly Senior Engineer, Institut für Eisenhüttenwesen R. W. Technische Hochschule Aachen     

The apparent solubility of aluminum in cryolite melts

The apparent solubility of aluminum in cryolite melts saturated with A1₂O₃ has been determined by titration with electrolytically generated O₂. The results may be expressed by wt pct Al = − 0.2877 + 0.0268 (NaF/AlF₃ wt ratio) + 2.992 × 10⁻⁴ (temp °C) − 0.00192 (% CaF₂) −0.00174 (% Li₃AlF₆) −0.00288 (% NaCl) with a standard deviation of ±0.017. Ranges covered were ratio 0.8 to 2.3, temperatures 969° to 1054°C, CaF₂ ≤ 14 pct, Li₃AlF₆ ≤ 20 pct, and NaCl ≤ 10 pct. There was no significant effect of adding 0 to 38. pct K₃A1F₆ or 0 to 10 pct MgF₂. It was found that solubility was approximately proportional to activity of aluminum when Al-Cu alloys were used. Possible mechanisms of solution are discussed. Monovalent aluminum is ruled out on the basis of the variation of solubility with NaF/AlF₃ ratio and a_Al. The favored, but not proven, mechanism involves formation of both sodium atoms and a colloidal dispersion of aluminum.     

The rate of reaction of solid iron with oxidized “FeO”-CaO-SiO2-Al2O3 slags at 1360 °C—the chemical diffusivity of iron oxide

Measurements have been made of the rate of reduction of oxidized iron oxide-containing 41CaO-38SiO₂-21Al₂O₃ (wt pct) slags at 1360 °C by a rotating disc of solid iron. For initial total iron concentrations of between 1.8 and 13.4 wt pct and rotation speeds up to 1000 rpm, the rate is shown to be determined by mass transfer in the liquid phase. The chemical diffusivity of iron oxide (in cm² s⁻¹) is found to be given by the empirical expression log D = −6.11 + 0.08 (wt pct Fe). It is concluded that the values of the diffusivity are for melts at close to iron saturation. It is shown that the available measurements of the diffusivity of iron oxide in liquid slags are consistent with increasing diffusivity with increasing state of oxidation, with about a tenfold increase between melts in equilibrium with iron and those in equilibrium with oxygen at 1 atm.     

The solubility and solution behavior of antimony and tin in α-lron and the effects of nickel and chromium additions

The solid solubilities of Sn and Sb in α-Fe have been determined by means of lattice parameter measurements. The Sb solubility ranges from a maximum of 11 wt pct (5.4 at. pct) at 1000°C down to 5.3 wt pct (2.5 at. pet) at 600°C; the Sn solubility ranges from a maximum of 17.7 wt pct (9.2 at. pet) at 900°C to 6.5 wt pct (3.2 at. pet) at 600°C. These solubilities are remarkably large in view of the large sizes of the Sb and Sn atoms in relation to the Fe atom. It was not possible to rationalize the variation of the α-phase lattice parameter with Sb or Sn content from the point of view of atomic diameter or atomic volume. The addition of 1 wt pct Ni lowers the Sb solubility at 600°C from 5.3 to 3.5 wt pct (2.5 to 1.6 at. pet); the effect of Cr on the Sb solubility appears to be small. The addition of 1 wt pct Ni or 1 wt pct Cr lowers the Sn solubility from 6.5 to 5.2 wt pct (3.2 to 2.5 at. pet). It was found that a substantial amount of Ni substitutes for Fe in both the FeSb phase and the Fe₅Sn₃ phase. Formerly Research Fellow, Department of Metallurgy and Materials Science and LRSM, University of Pennsylvania     

The effect of TiO2 additions and oxygen potential on liquidus temperatures of some CaO-Al2O3 melts

The effect of TiO₂ on the liquidus temperature of CaO-Al₂O₃ melts was determined in an argon atmosphere which had oxygen potentials of 0.21, 10⁻¹¹, 10⁻¹⁶, and 10⁻²¹ atm, using hot-wire microscopy. Compositions ranging from 44 to 52 wt pct A1₂O₃ and CaO and between zero and 10 wt pct TiO₂ were investigated. Experiments done in air show that TiO₂ lowers the liquidus temperature at low concentrations and raises the liquidus temperature at high concentrations. At low oxygen potentials, however, there is a continuous increase of the liquidus temperature with increasing TiO₂ concentrations. Liquidus temperature diagrams for the system at oxidizing (Po₂ = 0.21) and reducing (Po₂ = 10⁻¹⁶) conditions are presented. formerly a Research Assistant at the University of Kentucky     

Activities in liquid Fe-V-O and Fe-B-O alloys

The activities in liquid Fe-V-0 and Fe-B-O alloys have been determined using the following galvanic cells Cr-Cr₂O₃(s) | ZrO₂(CaO) | Fe-V-O (l, saturated with oxide) Cr-Cr₂O₃(s) | ThO₂(Y₂O₃) | Fe-V-O (l, saturated with oxide) Cr-Cr₂O₃(s) | ZrO₂(CaO) | Fe-B-O (l, B₂O₃ saturated with Al₂O₃) The solubility of oxygen in Fe-V alloys at 1600°C decreases with increasing vanadium content to a minimum of about 180 ppm at 3 wt pct V, and then increases to over 4000 ppm at 36.3 wt pct V. Vanadium was found to decrease the activity coefficient of oxygen and the value of the interaction coefficient e_o ^V at infinite dilution of vanadium is -0.14. The activity of vanadium was calculated from the measured electromotive force, and log γ_v was found to be represented well by the quadratic formalism for N_v < 0.4: log γ_V = -0.70N ² _Fe -0.30 At 1550°C boron decreases the solubility of oxygen down to about 80 ppm at 0.67 wt pct B in Fe-B melts in equilibrium with B₂O₃ saturated with A1₂O₃ (N_{Al 2} O₃ = 0.087). The boron deoxidation product, ’K′ = (wt pct B)²(wt pct 0)³ at infinite dilution of boron is 4.4 × 10_-9 and 1.5 × 10^-8 at 1550° and 1600°C, respectively. Boron decreases the activity coefficient of oxygen in liquid iron, and the value of the interaction coefficient e_o ^B is -2.6 at infinite dilution of boron. The activity coefficient of boron at infinite dilution (γ^° _B) is 0.083 at 1550°C relative to solid boron.     

Solubility of Chromium(III) Oxide in the Molten Cryolite System

The solubility of Cr(III) species originating from dissolution of Cr₂O₃ in cryolite-based melts was studied in the temperature range 1173 K to 1293 K (900 °C to 1020 °C). The molar ratio n(NaF)/n(AlF₃) was in the range of 1.4 to 2.6. It was found that the solubility depends markedly on the molar ratio n(NaF)/n(AlF₃), high ratios resulting in higher solubility. A semi-empirical model describing the solubility of Cr₂O₃ was developed. The standard deviation between calculated and experimental data is 10 pm (ca 2.4 pct).     

Evolution and coarsening of Al2O3 dispersoids at 500 °C to 600 °C in a mechanically alloyed Al-Ti-O based material

The formation and coarsening of Al₂O₃ dispersoids have been investigated at 500 °C, 550 °C, and 600 °C in a mechanically alloyed (MA) extrusion of composition Al-0.35wt pct Li-1wt pct Mg-0.25wt pct C-10vol pct TiO₂ for times up to 1500 hours. In the as-extruded condition, the dispersed phases included Al₃Ti, Al₄C₃, MgO, cubic TiO (C-TiO), monoclinic TiO (M-TiO), TiO₂, and a small amount of Al₂O₃. However, numerous Al₂O₃ dispersoids (various polymorphs: η, γ, α, and δ) with “block-shaped” morphology were formed after heat treatment due to reduction of C-TiO, M-TiO, and TiO₂. Transmission electron microscopy (TEM) and X-ray diffraction (XRD) showed conclusively the transformation of these phases to additional Al₂O₃ and Al₃Ti. High resolution TEM showed that the α-Al₂O₃ dispersoids exhibited some lattice matching with the α-Al matrix. Coalescence of the block-shaped Al₂O₃ dispersoids occurred after heat treatment, and Al₄C₃ also became attached to them. The length and width of the block-shaped Al₂O₃ dispersoids increased by a factor of ∼1.55 between 340 and 1500 hours at 600 °C.     

Effects of transition metals on the kinetics of slag-refractory reactions

The dissolution kinetics of dense alumina discs in calcium aluminosilicate based melts was determined with a rotating disc technique at 1560 °C to 1590 °C, under a controlled atmosphere of Ar-CO-CO₂. The effects of rotation speed and the concentration of iron and manganese oxides on the dissolution rate of alumina into slags were measured by monitoring the concentration of species in the slag. Analysis of the results obtained indicated that at low concentrations of these transition metal oxides in 53 pct CaO-5 pct MgO-12 pct SiO₂-30 pct Al₂O₃ slags, the dissolution rate is most likely controlled by mass transfer in the slag phase. The rate data obtained also showed that the addition of iron oxide or manganese oxide results in considerable increase in the mass transfer by increasing the apparent diffusivities of species in the slag. Comparison of these results with published data on the diffusivities of species in similar slags are made and practical implications of the findings are briefly discussed.     

The solubility of nitrogen in liquid pure nickel

A constant-volume Sieverts’ method was used to determine the solubility of nitrogen in liquid nickel. This method has been used for the first time on this type of material. It was found that the solubility of nitrogen in pure nickel at 1600 °C and 1 atm is equal to 0.0020 wt pct (with an experimental error of ± 0.0002 pct). The solubility increases with increasing temperature. The temperature coefficient of nitrogen solubility within the temperature range from 1500 °C to 1750 °C is equal to 1.6 × 10⁻⁶ wt pct/°C. The solution of nitrogen in pure nickel at 1600 °C and pressures up to 1 atm was found to obey Sieverts’ law.     

The niobium (columbium)-palladium constitution diagram

The Nb-Pd system was investigated over the entire composition range by metallography and X-ray diffraction analysis. The solubility limits of terminal and intermediate phases and solidus temperatures were determined. α-Nb dissolves ∼36 at. pct Pd at. 1520°C and ∼20 at. pct Pd at 800°C; α-Pd dissolves ∼31 at. pct Nb at 1610°C and ∼18 at. pct Nb at temperatures below 1500°C. The presence of three intermediate phases NbPd₂ (MoPt₂-type), α-NbPd₃ (TiAl₃-type), and β-NbPd₃ (β-NbPd₃-type) was confirmed; NbPd₂ melts at 1610°C and one of the NbPd₃ phases transforms at the same temperature into α-Pd solid solution which melts at 1625°C. In addition, an approximately equiatomic high-temperature phase α-NbPd with a homogeneity range of ∼11 at. pct was found which melts at 1520 to 1565°C and probably is an extension of and isomorphous with the α-Pd solid solution. Five three-phase reactions are described, and crystal chemical relationships are discussed. D. P. PARKER formerly with MIT . R. C. MANUSZEWSKI formerly with the ADAHF Research Unit at NBS.     

Discussion of “On the free energy of formation of TiC and Al4C3”

Several years ago, Banerji and Reif^[1] reported some very interesting studies on a process to react Ti and C in molten Al to form particles of TiC. The process was used to prepare a master alloy with a fine dispersion of TiC to inoculate Al for grain refinement. Approximately 2 wt pct of preheated graphite particles were stirred into the Al-5 to 10 pct Ti melts. The authors explained that the melts needed to be superheated above 1000 °C to avoid the undesirable formation of A1₄C₃ and Ti₃AlC at the TiC/melt interface. Their explanation for this phenomenon was based on thermodynamics. They observed that the standard free energy of formation curves for AI4C3 and TiC cross near 1175 °C, with A1₄C₃ having the lower free energy of formation below this temperature. There are several aspects of this work which merit further discussion.     

Determination of the Fe-Ni and Fe-Ni-P phase diagrams at low temperatures (700 to 300 °C)

The α + γ two-phase fields of the Fe-Ni and Fe-Ni (P saturated) phase diagrams have been determined in the composition range 0 to 60 wt pet Ni and in the temperature range 700 to 300 °C. The solubility of Ni in (FeNi)₃P was measured in the same temperature range. Homogeneous alloys were austenitized and quenched to form α₂, martensite, then heat treated to formα (ferrite) + γ (austenite). The compositions of the α and γ phases were determined with electron microprobe and scanning transmission electron microscope techniques. Retrograde solubility for the α/(α + γ) solvus line was demonstrated exper-imentally. P was shown to significantly decrease the size of the α + γ two-phase field. The maximum solubility of Ni in α is 6.1 ± 0.5 wt pct at 475 °C and 7.8± 0.5 wt pct at 450 °C in the Fe-Ni and Fe-Ni (P saturated) phase diagrams, respectively. The solubility of Ni in α is 4.2 ± 0.5 wt pct Ni and 4.9 ± 0.5 wt pct Ni at 300 °C in the Fe-Ni and Fe-Ni (P saturated) phase diagrams. Ternary Fe-Ni-P isothermal sections were constructed between 700 and 300 °C. Formerly Research Assistant in Department of Metallurgy & Materials Engineering, Lehigh University, Bethlehem, PA.     

The solution rate of alumina in CaF2-Al2 O3 slags

The rate of solution of A1₂O₃ in CaF₂ + 30 wt pct A1₂O₃ (at 1518° and 1509°C) and CaF₂ + 20 wt pct A1₂O₃ (at 1500°C) liquids has been determined. The operative process is diffusion-controlled, with an interdiffusion coefficient,D for the process varying between 8.5 and 8.1 x 10^-5 sq cms ^- 1 in the CaF₂ + 30 wt pct A1₂O₃ slags, and 4.0 × 10^-5 sq cms ^- 1 in the CaF₂ + 20 wt pct A1₂O₃ slag. Estimations of the rate at which alumina inclusions would react with these slag during the electroslag processing of steels, indicate that electrode inclusions approaching 100 μ in diam will be dissolved.