Interaction of carbon-based refractories with liquid PGM-furnace melt

Abstract

The upper-sidewall of a typical PGM smelter is lined with synthetic graphite at the hot-face. Copper waffle coolers located behind the graphite extract the heat away from the refractory. The sufficient cooling provided by the cooled copper waffle coolers causes the process material to freeze at the hot-face of the graphite, forming a protective skull (freeze-lining). The protective skull retards the penetration of the process melt and gases through the graphite block. It is desired to extend the graphite block to the hot-face of the lower side-wall but the interaction between graphite and matte is not well understood. The principal aim of this work was to determine the prominent wear mechanism of graphite when in contact with a PGM furnace melt (slag and matte). Wettability tests and crucible tests were done at temperatures from 800 °C to 1550 °C to determine the effect temperature has on the interaction between graphite and primary PGM matte. The contact time was varied from 1 hour to 5 days, to determine the effect of exposure time on the wear of graphite by the primary PGM matte. Penetration and graphite dissolution were used as a measure of the compatibility of graphite with the liquid PGM matte. The melting behaviour of pure sulfides (Cu2S, FeS, and Ni3S2) was assessed; the observed melting temperatures were in agreement with the published figures. Wettability of graphite by pure sulfides (Cu2S, FeS, and Ni3S2), synthetic matte and industrial PGM-furnace matte was determined using a sessile drop method. The interfacial contact-angle between graphite and all samples was >90°; therefore graphite is poorly wetted by all the tested materials except FeS, reactive wetting was observed between FeS and graphite. Synthetic matte did not penetrate through the graphite under all the tested conditions. The formation of the Fe-Ni alloy through the metallization of (Fe, Ni)S increased with operating temperature and contact time. Fe-Ni alloy dissolved up to 0.3 mass percent carbon, whereas the sulfide phase dissolved up to 0.03 mass percent carbon. The industrial-matte samples had silicates and oxide impurities. During the exposure of industrial matte to graphite, the impurities interacted with the matte and graphite, this lead to high consumption of graphite. The silicates were extracted to a sulfide phase according the following reaction: [M]SiO3 + (Fe, Ni)S + 3C→ (Fe, Ni)Si + [M]S + 3CO This reaction caused severe wear of graphite owing to the formation of the CO gas; the forward reaction was favoured by high temperature. Sulfides in their pure state were not corrosive towards graphite. The erosion of graphite at the slag/gas interface increased with exposure time. Penetration of industrial matte was observed only during the initial stages of melting, the penetration ceased as the contact time increased. The penetration of industrial matte through graphite was driven by the foaming of matte and the excess gas pressure that forced the liquid matte through the pores of the graphite wall. The crucible was cooled from one end to determine if the frozen skull would form on the graphite-matte interface. The matte formed a frozen skull at temperatures below 900 °C. Penetration of matte was not observed where the frozen skull had formed but at temperatures above 900 °C matte penetration occurred and no skull was observed on the graphite surface. The industrial PGM-slag did not wet graphite, the contact angle between graphite and slag samples was greater than 90°. The industrial PGM-slag was exposed to graphite using crucible test method. Physical penetration of slag through the graphite wall was observed. Slag penetration was attributed to the slag foaming and gas generated during melting, excess gas pressure forced the slag into the graphite pores. Matte and slag were exposed in one graphite crucible to simulate the layout of the material in the industrial setup. Erosion of graphite was observed at the matte-slag interface. Graphite thickness of up to 1.5 mm was consumed after 12 hours of exposure. The prominent wear mechanism of graphite was the sulfidation of silicates, followed by the dissolution of graphite in a metal and finally the physical penetration of graphite by matte and slag. The silicates and oxides were more reactive and corrosive towards graphite than the matte (sulfides). Cooled-graphite is currently used against slag in industrial application; cooled-graphite has performed well at the slag-zone. Since the slag is more corrosive towards graphite than matte, it is envisaged that cooled-graphite can be used against a matte-layer if sufficient cooling is applied. Micropore-carbon was tested against matte for comparative reasons, to determine if micropore-carbon can out-perform graphite. Micropore-carbon reacted with both synthetic matte and industrial matte. A sulfide phase formed in the refractory matrix and the alumina and silicon compounds in the refractory dissolved into the matte. Micropore carbon cannot be used in contact with matte since it has poor resistance towards chemical attack by matte.