Mass Transfer Model for the De-oxidation of Molten Copper LAMBERTO DI ´ AZ-DAMACILLO, 1 FIDEL REYES, 1 ALBERTO INGALLS, 1 CLAUDIO ME ´ NDEZ, 2 and GABRIEL PLASCENCIA 2,3 1.—Ingenierı ´a Quı ´mica Metalu ´ rgica, Facultad de Quı ´mica, UNAM, Circuito de la Investigacio ´n Cientı ´fica s/n, 04510 Cd. de Me ´xico, Mexico. 2.—CIITEC – Instituto Polite ´cnico Nacional, Cerrada Cecati s/n, 02250 Cd. de Me ´xico, Mexico. 3.—e-mail: g.plascencia@utoronto.ca In this paper, we present a mass transfer model that predicts two different mechanisms that control copper de-oxidation: (1) the transport of the reducing gas from the gas bubbles towards the melt/bubble interface, and (2) the transport of dissolved oxygen from the melt towards the melt/bubble interface. The model accounts for gas fluid flow and other process parameters such as lance submergence and nozzle diameter. The model was validated with pub- lished data and predictions from our model are in good agreement with the values reported. The key parameters to determine are the mass transfer coefficients of the reducing gas and that of the dissolved oxygen in the melt. INTRODUCTION Fire refining is typically conducted in rotary fur- naces with the capacity to process up to 270 tonnes of molten copper, just before anode casting. It consists of two stages. Initially, air is injected to oxidize trace impurities and residual sulfur from the copper. After this, de-oxidation with reducing gases takes place. This occurs while probes are used to register the sulfur and oxygen contents in the copper. In this paper, we are testing a mass transfer model created to describe the mass transfer during the de-oxidation stage. To do so, the model predic- tions will be compared with data already published. LITERATURE SURVEY When blister copper contains around 1 wt.% oxygen, 1 the injection of air stops and the excess oxygen is then lowered by a reducing agent. It was common to throw lumber (poling) into the copper melt to reduce its oxygen content; 2 however, strin- gent environmental laws and better process control has shifted to the use of reducing gases. Brantley and Shack 3 tested several gases to de- oxidize liquid copper between 1398 K and 1598 K. They lowered the oxygen content to 0.1 wt.%. Of the gases tested, butane yielded the best results. Inde- pendently, 4 ammonia gas was blown into copper melts, reducing the oxygen content in the copper below 300 ppm. Andreini et al. 5 reported results on copper de- oxidation with CO gas between 1386 K and 1446 K. They found that oxygen is the rate-limiting stage. These findings are in good agreement with a previous report. 6 Themelis and Smith 7 found that the rate of reaction of a CO jet is closely related to process parameters such as nozzle diameter and submergence into the melt, as well as to the gas flow rate. A different research 8,9 on de-oxidizing copper with CO 2 -CO mixtures resulted in finding an overall rate constant. Such constant is comprised by differ- ent mass transfer contributions, like the one in the gas phase, that in the liquid copper and also including the chemical reaction at the gas/melt interface. The transport of oxygen in the liquid was determined as the rate-limiting step. Kang et al. 10 reduced the oxygen concentration in pure copper from 190 to less than 6 ppm at 1523 K after bubbling CO for up to 15 min using gas flow rates of 1.66 9 10 6 m 3 /s and 3.3 9 10 6 m 3 /s. Soltanieh and Karimi 11 compared the efficiency of copper reduction when using liquid and gaseous reductants. Gaseous de-oxidation offers better results. Diffusion of oxygen in the melt controls the rate of copper reduction. Additionally, liquid copper de-oxidation has been conducted with solid graphite. 12,13 In both cases, it was found that the chemical reaction between CO gas and liquid copper is the rate-limiting step. JOM DOI: 10.1007/s11837-017-2356-0 Ó 2017 The Minerals, Metals & Materials Society