Two other methods of reducing metals from their oxides to the metallic state deserve mention. The first is applicable to metals which, while themselves reasonably fusible, are difficult to reduce and have a strong tendency to form carbides. Chro mium and manganese are examples of this type. In some cases the method of "alumino-thermic" reduction invented by Gold schmidt is applicable. This depends upon the powerful reducing action of metallic aluminium. The process consists in mixing the pure oxide of the metal with finely divided aluminium metal and igniting the mixture by the local application of intense heat. The reaction once started, sufficient heat is as a rule generated not only to melt the resulting metal but even to bring the aluminium oxide which is produced into fusion. The process, when iron oxide is used with aluminium, is in fact employed as a means of generat ing molten iron at a temperature high enough to serve for weld ing purposes. With chromium and manganese, also, the heat generated is sufficient to melt the resulting metal. On cooling after the reaction, a solid mass of the desired metal is found in the bottom of the crucible in which the operation has been carried out—or alternatively, the molten metal can be poured out into moulds. The resulting metals are free from carbon or carbides. but they contain an appreciable quantity of aluminium and the impurities derived from that metal, viz. : iron and silicon, while impurities derived from the containing vessel may also be pres ent. For many purposes, however, these aluminium-reduced met als are sufficiently pure and their use makes it possible to prepare alloys free from carbon.
In the case of a few very refractory metals—notably tungsten and molybdenum, another method of reduction has to be em ployed. Here the oxide is first reduced to the metallic state, generally by the action of hydrogen gas at a high temperature. The metal is thus obtained in the form of a finely-divided powder which is useless except for the preparation of alloys. To bring the metal into the massive form sintering and swaging processes are employed. The metal powder is strongly compressed mechani cally and the rather fragile rods thus produced are raised to a very high temperature in an atmosphere of hydrogen and are then submitted to prolonged and vigorous hammering in a swaging machine, the metal being kept surrounded by hydrogen through out. Finally, the swaged rods are strongly heated by the passage of an electric current, when the particles of the original powder coalesce to such an extent that the rod presents the external ap pearance and the internal micro-structure of a solid piece of metal. The large quantities of tungsten used in the manufacture of electric lamp filaments are produced by this method. The process, it may be remarked, as a resemblance, in principle, to that by which wrought iron is produced without the fusion of the purified iron, while certain modern processes of reducing iron from the ore by a "direct" process of gas reduction without fusion come into the same category.
iron and steel but their application differs in important respects. These differences arise mainly from the fact that the value per ton of iron and steel is much lower than that of the non-ferrous metals and very low cost of production is therefore essential. This factor, together with the immense quantities and large sizes in which ferrous products are required have led to the develop ment of processes on a scale much larger than that generally employed in connection with non-ferrous metals.
In the preliminary treatment of iron ores considerations of cost do not allow of elaborate concentration methods, while the relatively widespread and plentiful occurrence of suitable ores tends rather to the selection of those best fitted for a particular purpose than to special pre-treatment. The raw material of fer rous metallurgy may therefore be regarded as some form of iron oxide, contaminated in varying degrees by silicates and com pounds of sulphur and phosphorus. The beginnings of modern ferrous metallurgy were located at points where coal and iron ore were both available. The relation to coal-fields has been main tained, but modern facilities of transport now make it possible to use ores derived from distant sources—such as the Spanish and Swedish ores which are largely used in England. Further, while in the earlier days of steel production ores containing more than a small amount of phosphorus could not be used, the development of the "basic" processes, following the discoveries of Thomas and Gilchrist, has rendered available great deposits of phosphoric iron ore, such as the "minette" ores of Lorraine.
In modern practice the oxide of the iron ore is reduced to the metallic state by the action of carbon ; the crude and impure product is known as "pig iron." This corresponds to some extent with the matte or at most to "blister copper" in copper smelting and is also carried out in blast furnaces, but the iron blast fur naces are of very great size. The blast furnace is a vertical shaft or tower which is fed from the top with alternate layers of ore (iron oxide), coke (carbon) and limestone. A blast of pre-heated air is driven into the tower at the sides near the bottom by means of water-cooled nozzles or "tuyeres" and the coke in the charge is burnt, with the formation of a large amount of carbon monoxide. At a higher level in the furnace, part of this gas, and also to some extent the solid carbon of the charge, react with the iron oxide, metallic iron being liberated. This gradually falls to the bottom of the furnace and accumulates as liquid iron at the bottom. Meanwhile, most of the silica present in the ore combines with the lime of the limestone and with other oxides, to form a fairly fusible slag which also runs down to the bottom of the furnace and forms a layer above the molten iron. From time to time the furnace is "tapped"—i.e., molten iron is allowed to run from an opening made at the base of the furnace tower. The molten iron flows into a sand-bed.