If carried out at a high temperature or for a long time this may result in the formation of a coarse and brittle structure. For tunately this occurs only rarely in the non-ferrous metals, but in very mild steel it is frequently encountered. In this material, however, it is possible to produce a refining of the structure by "normalising." This consists in raising the steel to a temperature just above the critical range, maintaining it there only long enough to ensure that the whole of the piece has reached the desired temperature and then removing the steel from the furnace and allowing it to cool rapidly in still air. The result is a marked refining of the crystal structure with a corresponding improvement in the mechanical properties. If, however, the steel is of a very soft variety and is required to possess the greatest ductility with out much regard to strength, the final heat-treatment may be simply annealing at a much lower temperature, such as 68o° C, which is just below the critical range of the steel. This causes the carbide of the steel to become "balled up" into little spherical masses making the material very soft and relatively weak.
The most important forms of heat-treatment applicable to steel, whether a plain carbon steel or an "alloy steel" containing nickel, chromium, manganese, etc., is a double treatment known as "hardening and tempering." In some cases the first stage of this treatment is a true "hardening," but in a great many cases it can only be called "hardening" by analogy. A full discussion of these transformations could only be given after an explanation of the metallography of steel and especially of the equilibrium diagrams of carbon steels and of alloy steels. Since space will not permit of such treatment, the subject can only be treated in the broadest outline. The general nature of the changes involved in heat-treat ment of steel can, however, be explained in an approximate manner if it is realized that the metal iron, which constitutes the great bulk of all steels, even of the more complex alloy steels, can exist in at least two allotropic states, known as the a and 7 states respectively. In pure iron, the y state exists at temperatures between goo° C and 1,450° C. It is a soft non-magnetic substance which crystallizes with the face-centred cubic lattice. Below goo° C, pure iron assumes the a form, which is also relatively soft and ductile, but is—below 75o° C—strongly magnetic. The two kinds of iron differ most widely, however, in their power of absorbing carbon in the condition which is known as "solid solution." Gamma
iron has a considerable power of holding carbon in solid solution— up to about 1•2 per cent—while in alpha iron carbon is only soluble in very minute amounts. "Soluble" as here used simply means that the metal—in this case iron—can take up a certain proportion of the other element into its own crystal structure without the f orma tion of a second constituent or kind of crystal. Thus a carbon steel at a high temperature consists of 7 iron crystals holding carbon in solid solution, while on cooling through the transformation range— if the cooling is slow—the carbon is separated. As the iron itself undergoes transformation into the a form, crystals of iron carbide are separated. This is the condition of slowly cooled or an nealed steel. By very rapid cooling—chilling or quenching—the transformation can be more or less suppressed or delayed. If cer tain alloying elements are present, particularly chromium and nickel, the retardation of the transformation is much facilitated and if enough nickel or manganese is present the transformation may be entirely suppressed. In that case the iron retains the 7 condition, the steel is non-magnetic and soft. It is in the inter mediate condition, where the transformation has been retarded rather than entirely suppressed that the steel becomes hardened. The capacity of hardening was at one time believed to be a unique property of steel but it is now known in the alloys of other metals also. It was long regarded as mysterious and controversy still turns on its theoretical explanation. Here, however, we are mainly concerned with the fact that by suitable quenching, from a temperature above the "critical range"—i.e., above the tempera ture at which the y a transformation occurs—all but the very softest steels can be more or less strongly hardened. In the high carbon tool steels hardening is utilized to provide cutting tools, cutlery, etc. Even for these special purposes, however, where the greatest hardness is desirable, it is not possible to use the steel in the fully hardened state produced by direct quenching if that quenching has been drastic, as in water. The steel in that state is too brittle. The steel is therefore slightly heated and thereby be comes "tempered." When quench-hardened steel is gently heated, as the temperature rises the suppressed transformation gradually takes place to a slight but progressive extent. The steel is thereby rendered tougher but at the same time slightly softer. The degree of tempering applied is adjusted to suit the purpose for which the implement is required.