No abstract

The purpose of this work has been to establish activity data on sodium in liquid aluminum-sodium alloys at temperatures applied by the industry in liquid metal refining processes. A coulometric titration technique using a galvanic cell employing CaF 2 as a solid electrolyte has enabled measurements to be done under very clean and well-defined conditions over the entire range of compositions from highly diluted up to nearly sodium-saturated solutions. Sodium in liquid aluminum of 99.9999 pct purity is found to exhibit strong negative deviation from Henry's law, corresponding to a large negative self-interaction coefficient Na Na as expressed by the equation Na Na ϭ 16,318 Ϫ (191.1 и 10 5 K) и T Ϫ1 . This behavior is normal for elements, which exhibit strong positive deviation from Raoult's law and is explained by formation of Na clusters. The activity coefficient at infinite dilution, ␥ o Na , is expressed by the equation: RT ln ␥ o Na ϭ 86,729 Ϫ 26.237T. The magnitude of ␥ o Na from this equation agrees with the value predicted from the Miedema's semiempirical model. Sodium in liquid Al-Si5 pct alloy of 99.9999 pct purity exhibits strong positive deviation from Henry's law, which is in agreement with earlier investigations of the activity of sodium in liquid Al-Si alloys. The activity coefficient of sodium in pure liquid aluminum at saturation, ␥ sat Na , is expressed by RT ln ␥ sat Na ϭ Ϫ67,476 ϩ 102.33T, which gives for the sodium concentration at saturation x sat Na ϭ exp (8115.5/T Ϫ 12.307). This implies that the solubility of sodium in liquid aluminum at temperatures around the melting point of aluminum is about 10 times higher than previously reported and decreases rapidly with increasing temperature, possibly due to a decreasing stability of Na clusters. Analysis of the experimental conditions used by previous investigators supports these findings.

alumina) is inert toward sodium at elevated temperatures. This may well be true; Reaction [3] demands that aluminum is produced, and in the absence of any aluminum at the start, it may be kinetically hindered. With aluminum present, on the other hand, it should be expected. The preceding argument may not be the final word regarding the results of Hansen et al., but it represents a viable explanation.The following are some words about the quenching experiments of Hansen et al., [1] clearly intended to corroborate the high solubilities found from the electrochemical measurements. In short, Hansen et al. postulate that previous workers have not allowed sufficient time for equilibrium to be established, assuming that diffusion of Na into liquid Al is a slow process. At the same time, they assume that previous authors have also failed to take account of the sodium lost as a result of segregation during the quenching, but this presupposes a very fast rate of diffusion. Their arguments are clearly contradictory, and their results indicate that substantial amounts of elemental Na have been included in their analysis of sodium in the aluminum. In fact, it is likely that the established solubilities, as reviewed by Murray [5] and shown by the dotted line in Figure 1, are essentially correct. This is also supported by the recent measurements of Fellner et al., [6] using an electrochemical method to determine the solubility of sodium in liquid aluminum.

The liquidus lines of the binary system Mn – C were studied, mainly by solubility determinations (at high carbon) and thermal analysis (at low carbon). The low carbon part of the liquidus was mapped in unprecedented detail. The poorly defined phase Mn4C1 – X is probably the highest melting manganese carbide (i. e. not Mn7C3). No (stable) peritectics exist below 1326 °C, but three eutectics were identified: one meta-stable and two stable. Also, transistion temperatures were identified for some Mn carbides, as well as γ-Mn (austenitic manganese), and δ-Mn (ferritic). Literature data were co-opted to quantify the enthalpy of melting for δ-Mn (8.3 ± 1.5 kJ · mol−1) which is 3 to 5 kJ lower than given by current tabulations and databases. Several transition temperatures within Mn – C (including the melting of pure δ-Mn) are markedly affected by even minor amounts of oxygen and nitrogen.

No abstract