New York Paper - Direct Electrolysis of Black-copper Anodes of High Nickel-lead Content (with Discussion)

Some years ago, at the plant of the Baltimore Copper Smelting & Rolling Co., the receipt of large quantities of copper blister running high in lead, nickel, and arsenic resulted in the formation of a great amount of very foul slag, in the anode furnaces. A typical analysis of this slag was: Gold, oz. per ton......... 0.09 Arsenic, per cent.......... 0.60 Silver, oz. per ton......'.. 13.20 Antimony, per cent........ 0.50 Copper, per cent......... 19.50 Silica, per cent............ 28.30 Lead, per cent........... 11.90 FeO, per cent............. 19.20 Nickel, per cent.......... 4.60 CaO, per cent............. 1.60 The problem was to get out the copper in marketable form, to save the gold, silver, and particularly the nickel, and to get rid of as much as possible of the lead, arsenic, and other elements that were valueless to the Baltimore plant. In a copper refinery that does not take custom material for matte smelting, the usual method is to smelt the anode slag with coke in a reducing furnace, the black copper thus formed being sent back to the anode furnaces; this was the custom at Baltimore. An average anode slag will contain, perhaps, 40 per cent. copper and such small quantities of lead, nickel, etc., that smelting will get rid of most of these impurities in the black-copper furnace slag. To treat such an impure slag in this way, however, would result in a large nickel loss, and would send back to the anode furnaces all the lead and nickel not so lost, where it would be slagged off again and returned to the black-copper furnace as a slag still more concentrated in impurities. The result would probably be a constantly increasing amount of anode slag containing a likewise constantly increasing percentage of impurities. To avoid such a disastrous closed circuit some other process had to be devised. A small multiple-tank room, built for impure copper as an adjunct to the main series electrolytic refinery, was available, so the following standard process was suggested: