The purpose of this series is to make a thermodynamic analysis of the Fe-C system with a minimum of assumptions and to revise the equilibrium diagram of the same system in the light of the results of this investigation. In this paper, the first of the series, activities in liquid Fe-C alloys have been evaluated up to saturation, using Richardson and Dennis' data123 on dilute solutions of carbon in liquid iron and Darken and Smith's model~([1]) for carbon dissolved in austenite with certain modifications.A parameter α_c defined as logγc/N_Fe~2 with reference to graphite as the standard state is plotted against N_c for both austenite and Fe-C melt in order to facilitate the evaluation of α_(Fe) by graphical integration. Smith's data~([1]) on equilibrium between austenite and gaseous mixtures (CO_2/CO, CH_4/H_2) are re-treated to yield α_c~γ-N_c~γ curves for 800° and 1000℃ as shown in Fig. 1. On the assumption that L_c~γ the relative partial molal enthalpy of carbon in austenite, does not chan preciably with temperature, the α_c~γ-V_c~γ curve for 1153℃, the iron-graphite eutectic temperature, is obtained by extrapolation and found to lie above the graphite saturation point. This fact seems to indicate that the limit of application of Darken and Smith's model is reached somewhere around N_c~γ=0.0661 (1.50%) and a point of inflection should occur at this concentration. The above-mentioned assumption has been semi-quantitatively proved in this paper and will be discussed further in another paper of this series.In a similar manner, α_c~l-N_c~l curves for liquid Fe-C alloys are drawn through the experimental points of Richardson and Dennis on equilibrium between CO_2/CO mixed gases and dilute solutions of carbon in liquid iron at 1560° and 1660℃ as shown in Fig. 1. The curves are extended up to N_c~1=0.15 on the basis of Darken and Smith's model using 3600 cals. as the energy of interaction at 1560℃ between carbon atoms in the neighbouring interstitial sites as recommended by Richardson and Dens. Then, a suitable curve is drawn between N_c~l=0.15 and the graphite saturation point for 1560℃ to meet certain requirements, and a corresponding curve for 1660℃ is obtained by extrapolation, assuming that L_c~l, the relative partial molal enthalpy of carbon in liquid iron, does not change appreciably with temperature. Thus, α_c~l-N_c~l curves for 1560° and 1660℃ are completed from low carbon concentrations up to saturation. The activities of carbon in Fe-C melts at 1600℃ with reference to graphite as the standard state are readily obtained at different carbon concentrations by interpolation, from which the reversible electromotive forces of a concentration cell of the type Fe,C

Under the conditions of basic open hearth smelting at Anshan steel works which employ scrap-ore-liquid iron charge, a statistical analysis of the production data shows that within the limit investigated the manganese content of the molten bath has no appreciable influence upon the oxygen and sulphur content of the steel. Provided that the final chemical analysis of the steel was the same, no difference in mechanical properties was observed between the products derived from the higher (>0.2% Mn) and the lower(<0.1% Mn) manganese specifications during smelting. It is concluded that for the two steels (rail steel and a low-carbon steel for seamless tube) investigated at least, there is no necessity for adhering to the specification of manganese content above a certain level(>0.15-0.2% Mn) during smelting.On the contrary, with a view to maintaining a higher manganese-containing bath, it was usually found necessary to add iron-manganese or manganese ore at the end of melting down; in doing so, the rate of carbon-removal was to a certain extent retarded, thus lengthening the period of smelting and decreasing the rate of production.The use of low manganese-containing pig-iron for the smelting of certain qualitysteels is not only feasible but also advisable from the economic point of view, provided, of course, that the sulphur content of such pig-iron can be made low to meet the necessary requirement. A review of the iron production data in certain blast furnaces at Anshan shows that this is possible, although the best conditions for producing low manganese and low sulphur pig-iron in blast furnaces deserve further a more detailed investigation.

The study of the distribution of hydrogen in steel ingots, despite its practical importance, has not received due attention from previous workers. Available experimental results are mainly fragmentary and non-systematic, and therefore many disputable opinions exist. Desirous of investigating this problem in greater details, the authors employed several annealed ingots of high chromium steels which were considered to be particularly suitable because they evolved little gas at room temperature and consequently the inherent difficulty to avoid the loss of hydrogen during sampling was, to a very large extent, overcome. For this purpose also, suitable apparatus capable of determining relatively small amount of hydrogen was constructed. The results obtained show that the hydrogen distribution in the annealed ingots follows a significant and regular pattern, thus dismissing certain misgiving conclusions based on contradictory results given by previous workers. Although the average hydrogen content of the anealed ingots amounted to not more than half that of the liquid stael, yet in certain parts of their interior the local hydrogen content was found to be higher than that of the liquid steal. This affirms the existence of hydrogen segregation in steel ingots. Moreover, from maps of hydrogen contour lines drawn for the ingots it can be seen that the regions of the highest hydrogen content roughly coincide with the last solification. Indeed, the effect due to certain external irregularities encountered in the course of solification is detectable rather from the hydrogen maps than by the usual method of macro-etching.In the longitudinal or the transverse direction of the annealed ingots, the general trend of hydrogen variation based on average hydrogen content is shown to be governed by the law of hydrogen diffusion. Further examinations reveal that the ingot structure and its internal porosity exert considerable influence upon the distribution. It is likely that hydrogen diffusion may be faster in columnar crystals than in equi-axed crystal regions. The presence of porosities in ingots seems to retard the removal of hydrogen. Such implications have not been sufficiently realized in the past.Based on the discussion of the experimental results, certain immunizing treatment suitable for preventing hairline cracks in certain types of steel is explained.

Different methods of electrolytic separation of non-metallic inclusions in low carbon steels have been investigated, and certain suggestions for improvement in this connection are proposed.It is shown that for the separation of stable inclusions the more satisfactory procedure is to separate the inclusions and the carbides by the ferrous-sulfate method followed by the dissolution of the carbides either by ammonium-persulfide method or by nitric-acid method, whereas for the separation of unstable inclusions, the sodium-sulfite method gives better results.The modified method supplemented by microscopic examinations has been applied to the study of inclusions in a certain kind of constructional steel with favourable results.