INCONGRUENT EVAPORATION EXPERIMENTS ON IRON SULFIDE (FE1-DELTA-S) UNDER H-2-RICH (AT-1-ATM) AND EVACUATED CONDITIONS

Evaporation experiments using pyrrhotite single crystals (Fe0.886S) we re carried out at temperatures between 500 and 1300 degrees C at 1 arm in an H-2-CO2 gas flow (0.62-0.64 arm H-2), and at 500 and 900 degree s C under an evacuated condition. Under the H-2-rich condition, spongy metallic iron layer was formed on the sulfide crystal surface at temp eratures below the Fe-FeS eutectic point as a result of incongruent ev aporation, and developed inward almost conserving its original shape. The thickness of the iron layer increases linearly with time at consta nt temperatures (linear rate law) due to transportation of evaporated gas species through pores in the spongy iron layers. If incongruent ev aporation is controlled by diffusion of element(s) in an evaporation r esidue layer, a parabolic rate law is expected. The linear rate law sh ows that FeS evaporates more efficiently than expected based on a para bolic rate law. The linear rate constant obtained at various temperatu res obeys the Arrhenius relation: k(FeS)=(1.61 +/- 0.42) x 10(-3)exp(- 115 +/- 2 [kJ/mol]/RT) [m/sec]. A minor part of metallic iron in the s urface layer diffused into the inner sulfide to form stoichiometric Fe S (troilite) in the early evaporation stage. Thus, the experiments can be almost regarded as evaporation of troilite. Evaporation coefficien ts of FeS were obtained by comparing the experimental results with cal culated rates using the Hertz-Knudsen equation. They are small(1.4 x 1 0(-4)similar to 9.4 x 10(-6)) due to slow surface reaction and/or slow escape of S-bearing gas species into the gas flow. Mass-dependent iso topic fractionation of S by the evaporation was not detected within an error of +/-3 parts per thousand probably due to slow diffusivity of S in the sulfide crystal. In the evacuated experiments, evaporation oc curred very slowly due to the absence of H-2 gas, which acts as a redu cing agent. Iron residue layer was very thin or sometimes not detected probably because the evaporation rate of S from FeS became comparable to the evaporation rate of metallic iron, which can be neglected unde r the H-2-rich condition.