Extractive Metallurgy Division - The Absolute Seebeck Coefficient of the Molten Ni-S System (TN)

In an earlier study1 in this laboratory, it was found that there is a minimum in the specific conductance of Ni-S melts in the region of the stoichiometric composition, Ni3S2. This and similar observations on other sulfide1 and telluride2 systems suggest that short-range order exists in the molten phase for several hundred degrees above the melting point, and a qualitative energy-band model has been proposed, with intrinsic semiconduction at the stoichiometric composition and degenerate electron and hole conduction at metal- and tellurium-rich compositions. This suggestion has been substantiated for the telluride systems by Seebeck coefficient measurements which showed a change in sign from negative on the metal-rich side to positive on the tellurium-rich side. In the present investigation, in order to substantiate still further the proposed model, Seebeck coefficient determinations have been extended to the Ni-S system over a composition range in the vicinity of Ni3S2. The experimental method was essentially the same as that used for the telluride systems, with minor modifications on cells for high sulfur melts to permit the use of tungsten wires and W— W-Re th~rmocouples instead of the chrome1 or alumel wires and chromel-alumel thermocouples originally used. Chrome1 and alumel both deteriorate very rapidly in atmospheres containing sulfur. The Ni-S alloys were prepared as described in the conductivity paper. Initially, a portion of the solidified melt was analyzed at the end of a run, but after this was found to be inaccurate because of segregation during cooling, the whole sample was analyzed. The nickel content, only, was determined, the sulfur content being obtained by difference. The results are summarized graphically in Fig. 1, and clearly show a change in the sign of the Seebeck coefficient in the region of the stoichiometric composition. This change of sign near Ni3S2, coupled with the minimum in the conductivity, is suggestive of an intrinsic semiconductor at the stoichiometric composition, with degenerate electron and hole conduction on either side of this composition, in keeping with the proposed model. The negative temperature coefficient of conductivity, normally indicative of metallic conduction, is not inconsistent with this model. In the completely degenerate region it would be expected, while in the vicinity of Ni3S2 the suggestion originally made by Argyriades, Derge, and pound3 for the Fe-S system can probably be applied. They proposed that the forbidden gap between the valence band and the conduction band is small, and that the increase in the number of the conducting species with increase in temperature is not large enough to compensate for the loss of mobility through thermal scattering. The research was sponsored by the Office of Naval Research under Contract NONR-760(22) Project NR 039-081. 'E. A. Dancy and G. J. Derge: Trans. Met. Soc. Aime, 1963, vol. 227, pp. 1034-38. 'E. A. Dancy: Trans. Vet. soc. AIME. 1964, vol. 233, pp. 270-77. 'D. Argyriades, G. Derge, and G. M. Pound: Trans. Met. Soc. . 1959, vol. 215, pp. 909-12.