Cooling curves of the purest iron show at least two well-defined discontinuities at temperatures more than 1,000F., below its freezing-point. It seems that the soft, magnetic metal so familiar as wrought iron, and called "alpha iron" or "ferrite" by the metallurgist, becomes unstable at about 1,400F. and changes into the so-called "beta" modification, becoming suddenly harder, and losing its magnetism. This state in turn persists no higher than 1,706C., when a softer, non-magnetic "gamma" iron is the stable modification up to the actual melting-point of the metal. These various changes occur in electrolytic iron, and therefore cannot be attributed to any chemical reaction or solution; they are entirely due to the existence of "allotropic modifications" of the iron in its solid state.

[Ill.u.s.tration: FIG. 45.--Inverse Rate Cooling Curve of 0.38 C Steel.]

Steels, or iron containing a certain amount of carbon, develop somewhat different cooling curves from those produced by pure iron.

Figure 45 shows, for instance, some data observed on a cooling piece of 0.38 per cent carbon steel, and the curve constructed therefrom. It will be noted that the time was noted when the needle on the pyrometer pa.s.sed each dial marking. If the metal were not changing in its physical condition, the time between each reading would be nearly constant; in fact for a time it required about 50 sec. to cool each unit. When the dial read about 32.5 (corresponding in this instrument to a temperature of 775C. or 1,427F.) the cooling rate shortened materially, 55 sec. then 65, then 100, then 100; showing that some change inside the metal was furnishing some of the steadily radiating heat. This temperature is the so-called "upper critical" for this steel. Further down, the "lower critical"

is shown by a large heat evolution at 695C. or 1,283F.

Just the reverse effects take place upon heating, except that the temperatures shown are somewhat higher--there seems to be a lag in the reactions taking place in the steel. This is an important point to remember, because if it was desired to anneal a piece of 0.38 carbon steel, it is necessary to heat it up to and beyond 1,476 F. (1,427F. _plus_ this lag, which may be as much as 50).

It may be said immediately that above the upper critical the carbon exists in the iron as a "solid solution," called "austenite" by metallographers. That is to say, it is uniformly distributed as atoms throughout the iron; the atoms of carbon are not present in any fixed combination, in fact any amount of carbon from zero to 1.7 per cent can enter into solid solution above the upper critical. However, upon cooling this steel, the carbon again enters into combination with a definite proportion of iron (the carbide "cement.i.te," Fe3C), and acc.u.mulates into small crystals which can be seen under a good microscope. Formation of all the cement.i.te has been completed by the time the temperature has fallen to the lower critical, and below that temperature the steel exists as a complex substance of pure iron and the iron carbide.

It is important to note that the critical points or critical range of a plain steel varies with its carbon content. The following table gives some average figures:

Carbon Content. Upper Critical. Lower Critical.

0.00 1,706F. 1,330F.

0.20 1,600F. 1,330F.

0.40 1,480F. 1,330F.

0.60 1,400F. 1,330F.

0.80 1,350F. 1,330F.

0.90 1,330F. 1,330F.

1.00 1,470F. 1,330F.

1.20 1,650F. 1,330F.

1.40 1,830F. 1,330F.

1.60 2,000F. 1,330F.

It is immediately noted that the critical range narrows with increasing carbon content until all the heat seems to be liberated at one temperature in a steel of 0.90 per cent carbon. Beyond that composition the critical range widens rapidly. Note also that the lower critical is constant in plain carbon steels containing no alloying elements.

[Ill.u.s.tration: FIG. 46.--Microphotograph of steel used in S. K.

F. bearings, polished and etched with nitric acid and magnified 1,000 times. Made by H. O. Walp.]

This steel of 0.90 carbon content is an important one. It is called "eutectoid" steel. Under the microscope a properly polished and etched sample shows the structure to consist of thin sheets of two different substances (Fig. 46). One of these is pure iron, and the other is pure cement.i.te. This structure of thin sheets has received the name "pearlite," because of its pearly appearance under sunlight. Pearlite is a const.i.tuent found in all annealed carbon steels. Pure iron, having no carbon, naturally would show no pearlite when examined under a microscope; only ab.u.t.ting granules of iron are delicately traced. The metallographist calls this pure iron "ferrite." As soon as a little carbon enters the alloy and a soft steel is formed, small angular areas of pearlite appear at the boundaries of the ferrite crystals (Fig. 47). With increasing carbon in the steel the volume of iron crystals becomes less and less, and the relative amount of pearlite increases, until arriving at 0.90 per cent carbon, the large ferrite crystals have been suppressed and the structure is all pearlite. Higher carbon steels show films of cement.i.te outlining grains of pearlite (Fig. 48).

This represents the structure of annealed, slowly cooled steels.

It is possible to change the relative sizes of the ferrite and cement.i.te crystals by heat treatment. Large grains are a.s.sociated with brittleness. Consequently one must avoid heat treatments which produce coa.r.s.e grains.

[Ill.u.s.tration: FIG. 47.--Structure of low carbon steel, polished, etched and viewed under 100 magnifications. Tiny white granules of pure iron (ferrite) have small acc.u.mulations of dark-etching pearlite interspersed between them. Photograph by H. S. Rawdon.]

[Ill.u.s.tration: FIG. 48.--Slowly cooled high-carbon steel, polished, etched and viewed at 100 magnifications. The dark grains are pearlite, separated by white films of iron carbide (cement.i.te). Photograph by H. S. Rawdon.]

In general it may be said that the previous crystalline structure of a steel is entirely obliterated when it pa.s.ses just through the critical range. At that moment, in fact, the ferrite, cement.i.te or pearlite which previously existed has lost its ident.i.ty by everything going into the solid solution called austenite. If sufficient time is given, the chemical elements comprising a good steel distribute themselves uniformly through the ma.s.s. If the steel be then cooled, the austenite breaks up into new crystals of ferrite, cement.i.te and pearlite; and in general if the temperature has not gone far above the critical, and cooling is not excessively slow, a very fine texture will result. This is called "refining" the grain; or in shop parlance "closing" the grain. However, if the heating has gone above the critical very far, the austenite crystals start to grow; a very short time at an extreme temperature will cause a large grain growth. Subsequent cooling gives a coa.r.s.e texture, or an arrangement of ferrite, cement.i.te and pearlite grains which is greatly coa.r.s.ened, reflecting the condition of the austenite crystals from which they were born.

It maybe noted in pa.s.sing that the coa.r.s.e crystals of cast metal cannot generally be refined by heat treatment unless some forging or rolling has been done in the meantime. Heat treatment alone does not seem to be able to break up the crystals of an ingot structure.

HARDENING

Steel is hardened by quenching from above the upper critical. Apparently the quick cooling prevents the normal change back to definite and sizeable crystals of ferrite and cement.i.te. Hardness is a.s.sociated with this suppressed change. If the change is allowed to continue by a moderate reheating, like a tempering, the hardness decreases.

If a piece of steel could be cooled instantly, doubtless austenite could be preserved and examined. In the ordinary practice of hardening steels, the quenching is not so drastic, and the transformation of austenite back to ferrite and cement.i.te is more or less completely effected, giving rise to certain transitory forms which are known as "martensite," "troost.i.te," "sorbite," and finally, pearlite.

Austenite has been defined as a solid solution of cement.i.te (Fe3C) in gamma iron. It is stable at various temperatures dependent upon its carbon content, which may be any amount up to the saturated solution containing 1.7 per cent. Austenite is not nearly as hard as martensite, owing to its content of the soft gamma iron. Fig.

49 shows austenite to possess the typical appearance of any pure, crystallized substance.

In the most quickly quenched high carbon steels, austenite commonly forms the ground ma.s.s which is interspersed with martensite, a large field of which is ill.u.s.trated in Fig. 50. Martensite is usually considered to be a solid solution of cement.i.te in beta iron. It represents an unstable condition in which the metal is caught during rapid cooling. It is very hard, and is the chief const.i.tuent of hardened high-carbon steels, and of medium-carbon nickel-steel and manganese-steel.

Troost.i.te is of doubtful composition, but possibly is an unstable mixture of untransformed martensite with sorbite. It contains more or less untransformed material, as it is too hard to be composed entirely of the soft alpha modification, and it can also be tempered more or less without changing in appearance. Its normal appearance as rounded grains is given in Fig. 51; larger patches show practically no relief in their structure, and a photograph merely shows a dark, structureless area.

[Ill.u.s.tration: FIG. 49.--Coa.r.s.e-grained martensite, polished and etched with nitric acid and magnified 50 times. Made by Prof. Chas.

Y. Clayton.]

Sorbite is believed to be an early stage in the formation of pearlite, when the iron and iron carbide originally const.i.tuting the solid solution (austenite) have had an opportunity to separate from each other, and the iron has entirely pa.s.sed into the alpha modification, but the particles are yet too small to be distinguishable under the microscope. It also, possibly, contains some incompletely transformed matter. Sorbite is softer and tougher than troost.i.te, and is habitually a.s.sociated with pearlite. Its components are tending to coagulate into pearlite, and will do so in a fairly short time at temperatures near the lower critical, which heat will furnish the necessary molecular freedom. The normal appearance, however, is the cloudy ma.s.s shown in Fig. 52.

Pearlite is a definite conglomerate of ferrite and cement.i.te containing about six parts of the former to one of the latter. When pure, it has a carbon content of about 0.95 per cent. It represents the complete transformation of the eutectoid austenite accomplished by slow-cooling of an iron-carbon alloy through the transformation range. (See Fig. 46.)

[Ill.u.s.tration: FIG. 50.--Quenched high-carbon steel, polished, etched and viewed at 100 magnifications. This structure is called martensite and is desired when maximum hardness is essential. Photograph by H. S. Rawdon.]

[Ill.u.s.tration: FIG. 51.--Martensite (light needles) pa.s.sing into troosite (dark patches). 130 X. From a piece of eutectoid steel electrically welded.]

[Ill.u.s.tration: FIG. 52.--Sorbite (dark patches) pa.s.sing into pearlite (wavy striations). Light Areas are Patches of Ferrite. 220 X. From a piece of hypo-eutectoid steel electrically welded.]

These observations are competent to explain annealing and toughening practice. A quickly quenched carbon steel is mostly martensitic which, as noted, is a solid solution of beta iron and cement.i.te, hard and brittle. Moderate reheating or annealing changes this structure largely into troost.i.te, which is a partly transformed martensite, possessing much of the hardness of martensite, but with a largely increased toughness and shock resistance. This toughness is the chief characteristic of the next material in the transformation series, sorbite, which is merely martensite wholly transformed into a mixture of ultramicroscopic crystals of ferrite (alpha iron) and cement.i.te (Fe3C).

"Tempering" or "drawing" should be restricted to mean moderate reheating, up to about 350 C., forming troost.i.tic steel. "Toughening"

represents the practice of reheating hardened carbon steels from 350 C. up to just below the lower critical, and forms sorbitic steel; while "annealing" refers to a heating for grain size at or above the transformation ranges, followed by a slow cooling.

Any of these operations not only allows the transformations from austenite to pearlite to proceed, but also relieves internal stresses in the steel.

Normalizing is a heating like annealing, followed by a moderately rapid quench.

JUDGING THE HEAT OF STEEL

While the use of a pyrometer is of course the only way to have accurate knowledge as to the heat being used in either forging or hardening steels, a color chart will be of considerable a.s.sistance if carefully studied. These have been prepared by several of the steel companies as a guide, but it must be remembered that the colors and temperatures given are only approximate, and can be nothing else.

[Ill.u.s.tration: FIG. 53.--Finding hardening heats with a magnet.]

_The Magnet Test_.--The critical point can also be determined by an ordinary horse-shoe magnet. Touch the steel with a magnet during the heating and when it reaches the temperature at which steel fails to attract the magnet, or in other words, loses its magnetism, the critical point has been reached.

Figures 53 and 54 show how these are used in practice.

The first (Fig. 53) shows the use of a permanent horse-shoe magnet and the second (Fig. 54) an electro-magnet consisting of an iron rod with a coil or spool magnet at the outer end. In either case the magnet should not be allowed to become heated but should be applied quickly.

[Ill.u.s.tration: FIG. 54.--Using electro-magnet to determine heat.]

The work is heated up slowly in the furnace and the magnet applied from time to time. The steel being heated will attract the magnet until the heat reaches the critical point. The magnet is applied frequently and when the magnet is no longer attracted, the piece is at the lowest temperature at which it can be hardened properly.

Quenching slightly above this point will give a tool of satisfactory hardness. The method applies only to carbon steels and will not work for modern high-speed steels.

HEAT TREATMENT OF GEAR BLANKS

This section is based on a paper read before the American Gear Manufacturers' a.s.sociation at White Sulphur Springs, W. Va., Apr.

18, 1918.