Results of cooling steel slowly from different high temperatures are given in Tables 18, 20, 23, 24, and 25, groups VII. and VIII.
In some cases (group VII.) this slow cooling was complete; in others it was interrupted by again raising the temperature, and after this the slow cooling was resumed.
Bars marked T closed down without checking. When hammered close, 45 and 50 cracked half through. The rest did not crack.
Let us, as before, designate as T max the temperature from which slow cooling begins, i.e., the temperature which is reached in the heating which precedes our slow cooling.
In case of the hard tool-steel of Series 10 and of the soft tool- steel of Series 9, we get the greatest ductility when T max is but little above W.
As we raise T max, we progressively and rapidly lessen the ductility of the slowly-cooled metal; but, at least in case of Series 9, we appear to increase its tensile strength. The results are too few in number to permit strong inferences.
The loss of ductility thus caused by raising T max from just above W to higher temperatures is, in case of both these steels, removed, and probably completely, by again heating to W and cooling slowly thence (group VIII., Tables 18 and 20).
It is not necessary that the metal be cooled completely before this re-toughening heating to W begins. If metal has lost its toughness by heating far above W, we restore its toughness, probably completely (bars 117, 118, 120, of Table 20), by cooling it to below V, then reheating it to W, then cooling it slowly.
Thus heating to W has in some way a wonderful toughening effect, which is lessened if the temperature rise even very slightly above W.
General Note to Tables 17 to 23 (inclusive), and to Table 26.
Except when expressly asserted to the contrary, all quenchings were in cold water, at or slightly below the temperature of the room.
The bending tests were made in my private laboratory, between dies under a hand press on bars 5/16 in. square and 4 in. long, not machined, and as they left the rolling-mill.
The tensile tests were made in the mechanical laboratory of the Massachusetts Institute of Technology, by Mr. Wm. Haskins, and through the kind permission of Professors Lanza and Miller.
A double heating to W probably further intensifies this effect. Thus bar 118 first had its ductility lessened by heating to 956°. It was then cooled below V, reheated above W, again cooled below V, heated above W a second time, and cooled slowly. It bent 170°, or even more than any of the bars the ductility of which had
NOTE.—a. Bent double without cracking.
b. Bent double, but cracked very deeply.
not been lessened by too high heating. Yet, if it had not been thus reheated, it should have bent only about 34°, judging from bar II.
I looked for some of these effects in the extra hard steel of Series 17 and in the soft steels of Series 15 and 16, but I failed to find them. If, as is probable, the steel of Series 17 was injured by heating it to 880°, this injury was not removed by reheating to a
NOTE.—Bar 37 had previously been, heated to 880° C., and cooled slowly. Bar 41 had not been heated before the heating to 737° of this experiment.
little above W (bars 1 and 8, Table 25). The soft basic steel of Series 15 was as tough when cooled slowly from 1002° or 886° as when slowly cooled thrice from just above W (Table 23), whence I
NOTE.—During this time the temperature of bar 11 varied between 693° and 704°.
That of bar 52, between 697° and 701°.
That of bar 13, between 675° and 683°.
That of bar 27, between 697° and 701°.
That of bar 19, between 674° and 684°.
infer that it is not injured by heating even to 1002°. The soft steel of Series 16, too, was practically as tough when cooled slowly from 880° as when cooled from 733°, just above W (Table 24), and cooling it slowly thrice from just above W gave no additional toughness.
We see, from Series 9 and 10, that variations, and sometimes slight ones, in the maximum temperature reached, i.e., in that whence slow cooling starts, apparently have a wonderful effect on the properties of the metal. The cooling was in all cases so slow that we cannot refer these effects to stress, nor can we suppose that they are due to our not allowing sufficient time for the carbon to change from the hardening to the cement state, or for the iron to change from the β to the α state. Should further experiments corroborate this, we should clearly be in the presence of some cause of titanic power, a cause which neither the carbon theory nor the β—α theory of the hardening of steel takes into account. Should we hold either that hardening-carbon or that β-iron is the true cause of hardening, we should yet have to admit the existence of another influence, which, as an independent variable, powerfully affects the properties of the metal. Hence we could not expect to find a complete accord between those properties and the proportion of hardening carbon or of β-iron, as the case may be.