W. O. Philbrook

Non-Topochemical reduction of iron oxides

Non-topochemical behavior was studied during reduction of porous spheres of hematite by stages through the intermediate oxides and also continuously to iron by CO/CO2 mixtures at temperatures of 600 to 900°C (873 to 1173 K). The behavior became more nearly topochemical as temperature increased. Shrinking occurred during the reduction of hematite to magnetite and of magnetite to wüstite, whereas swelling was observed during the reduction of wiistite to iron. Shrinking was greater, and swelling less, at higher temperatures. The total surface area of the solid decreased with increasing extent of reduction during each of the three stages. A non-topochemical model was developed which satisfies, better than previously proposed models, the reduction data for the single reactions and the three reactions occurring simultaneously. The model provides for variation in particle size and local changes in porosity and effective diffusivity. An empirical “sintering exponent” was introduced to describe changes in reacting surface area.

Thermodynamic study of Na2O-SiO2 melts at 1300° and 1400 °C

The vapor pressures of Na above stirred Na2O-SiO2 melts in equilibrium with graphite and CO were determined at 1300° and 1400 °C using the transpiration technique. Compositions studied ranged from about 60 mole pct SiO2 to close to SiO2 saturation. Activities of components Na2O and SiO2 were calculated from the data. Log aNa2O (pure liquid as standard state) varies from about −8.7 and −8.5 at silica saturation to −6.3 and −6.1 at 40 mole pct Na2O at 1300° and 1400 °C, and the molar Gibbs energy of mixing, ΔGm, at the disilicate composition (XNa2O = 0.33) at each of these temperatures is −83.0 and −85.4 kJ, respectively. The Toop and Samis, Yokokawa and Niwa, and Lin and Pelton solution models for binary silicates were applied to the ΔGm data at 1350 °C and parameters for the models were estimated to give best fits. All three models show good correspondence with the measured ΔGm curve. The capabilities of the models in predicting activity data in this system have been compared.

A mathematical model to predict the growth and elimination of inclusions in liquid steel stirred by natural convection

A model is developed to predict the rate of removal and the change in the size distribution of inclusions in a melt stirred by natural convection. Difficulties in obtaining an exact solution to the problem due to lack of adequate knowledge for the velocity fields in the melts are discussed. The model is based on Smoluchowski’s Theory of Gradient Collision to obtain the probability of collision between two inclusions under an arbitrarily chosen velocity gradient. Initial size distributions obtained in experimental heats are used as the input to the model. Various conditions are proposed by which inclusions are removed from the melt. The rates of removal are compared with the experimentally obtained rate of removal of oxides. It is observed that a boundary layer effect and the presence of a thin liquid metal film prevent rapid removal of inclusions from the stirred melts. Inclusion size distribution predicted by the model agrees qualitatively with the experimentally observed size distribution. It is postulated that the surface forces play a significant role in coalescence and assimilation of inclusions. Finally, the application of similar models to understand the removal of inclusions in such processes as argon sparging, solidification, degassing and electroslag remelting are advocated.

Oxidation kinetic studies of zinc sulfide in a fluidized bed reactor

Kinetic studies of the oxidation of zinc sulfide were carried out in a fluidized bed reactor over a temperature range of 740° to 1000°C with O-N gas mixtures of 20 to 40 pct O2. A mathematical model was developed to describe the overall conversion of the solids. Application of the model to the experimental data indicated that the chemical reaction at the outer boundary of the unreacted sulfide core was the rate-limiting step for the process. The temperature dependence of the kinetic constant corresponded to an activation energy of 40,250 cal per mole. Oxygen starvation in the bed was not limiting in any of the experimental runs, but an increase in the inlet-oxygen mole fraction resulted in a substantial increase in reaction rate.

Thermodynamic activity of Na2O in Na2O-CaO-SiO2, Na2O-MgO-SiO2 and Na2O-CaO-SiO2-Al2O3 melts at 1400°C

A gas phase equilibration technique was used to generate Na2O isoactivity data at high-silica compositions in the systems Na2O-CaO-SiO2, Na2O-MgO-SiO2, and Na2O-CaO-SiO2-(10 and 20 wt pct) A12O3 at 1400 °. LogaNa2O values referenced to pure liquid Na2O at 1400 ° as standard state ranged from about −8.0 to −7.0. Silica activities were calculated in the Na2O-CaO-SiO2 system using the Gibbs-Duhem equation. Richardson’s model for ideal mixing of basic oxides in silica was applied to the Na2O-CaO-SiO2 system. The model shows best fit when cationic mixing is assumed to occur on divalent sites. The alkali retention in slags has been described using a defined “sodium capacity”. The temperature variation of alkali retention was estimated, and the resulting sodium capacity was used to evaluate alkali stability in blast furnace slags.

The rate of CO bubble nucleation at oxide metal interfaces within liquid iron alloys

Baker, Warner, and Jenkins found that levitated droplets of Fe-0.8 pet C alloys exploded when decarburized at 1660°C, whereas during the present investigation, the drops remained intact during decarburization at temperatures above 1850°C. Therefore, the object of this work was to determine whether heterogeneous nucleation of CO bubbles at an iron-iron oxide interface could occur at 1900°K but could not occur at 2200°K. An equation was developed to calculate the nucleation rate of CO bubbles at an iron-iron oxide interface in iron at 1900°K containing 0.8 pct C and in iron at 2200°K containing 0.1 pct C. The results of the calculation showed that an iron-iron oxide interface could not serve as a site for CO bubble nucleation. Therefore, a new mechanism is postulated in which cavities swept into the levitated droplet from the surface serve as nuclei for CO bubble formation instead of nuclei formed at the iron-iron oxide interface.

Application of Phase Diagrams to Predict Phases Formed During Deoxidation of Steel

Phase diagrams for simple ternary systems such as Fe-Al-O, Fe-Si-O and more complex systems Fe-Mn-Al-O, Fe-Mn-Si-O have been used to rationalize some of the observations made during deoxidation. The formation of less stable oxides such as liquid aluminates and spinel is shown to be a consequence of the reaction path followed by the metal-inclusion system. The significance of univariant conditions in which three phases are in equilibrium is brought out. During cooling and solidification inclusions may change from a solid oxide to a liquid oxide phase and vice versa. This causes a change in solidification from eutectic type to a monotectic type. The existence of a saddle point at which either a eutectic or monotectic solidification begins is established and its significance in semi-killed steel is brought out.

Thermal conductivities of some Ni-S, Ag-S, and Ag-Te melts

Thermal conductivity data were obtained on a total of five melts in the Ag-Te, Ag-S, and Ni-S systems. Analysis of the data showed that electronic conduction was present in all of the melts, and one solution, Ni-34.2 at. pct S, was a dominant electronic conductor. The thermal conductivities of the four Ag-Te and Ag-S melts were lower, and phonon conduction probably was important in these materials. The phonon conduction tended to mask any enhanced electronic heat transport, and ambipolar effects could not be conclusively identified in any of the melts studied.

A Simple Continuously Weighed, Rotating Disk Reactor

A disk is rotated by a drive consisting of an electric motor, batteries, and variable resistor speed control. The entire drive is suspended from an analytical balance which weighs both the disk and the drive. This apparatus is used in fluid‐solid kinetics to achieve uniform mass transfer to a weighed solid.

Kinetics of reactions between magnetite and carbon monoxide between 723 and 823 K

The kinetics of some reactions between carbon monoxide and magnetite were investigated between 723 and 823 K. Gas-phase mass-transfer effects within and near the porous samples and surface-reaction kinetics were considered. The experiments were done on sintered magnetite disks whose void fraction was 15 to 20 pct. A continuously weighed rotating disk reactor provided uniform mass transfer to the sample faces. The reacted disks were analyzed by Mössbauer spectroscopy and other methods. The only product species found were cementite and carbon, which suggested the existence of two reactions: the formation of cementite and carbon dioxide directly from magnetite and carbon monoxide, and the disproportionation of carbon monoxide to carbon dioxide and free carbon. These reactions were modeled mathematically. The surface kinetic data for both reactions were found to be compatible with rate expressions which were directly proportional to the square of the partial pressure of carbon monoxide and approximately inversely proportional to the partial pressure of product carbon dioxide.