T. I. Tikhonov.

CHANGE

IN THE STRUCTURE OF IRON

DURING

MECHANICAL WORKING.

Tomsk.

Type-Lithography of the Siberian Association of Printing Business. Corner of Dvoryanskaya St. and Yamskoy Lane, own building

1911.

Separate offprint from the Proceedings of the Tomsk Technological Institute of Emperor Nicholas II.

Introduction

The close connection between the mechanical properties of iron and its structure undoubtedly exists, and has been proved by a whole series of mechanical and metallographic investigations of a more recent period in the field of the study of the properties of iron. To elucidate the picture of those changes in the structural constitution of iron which occur in it during a whole series of mechanical processes of iron working constitutes the chief task of the present essay; but it must be noted that the principal task of this work has by no means been exhausted in full, often not through the fault of the author, but owing to the fact that local conditions were far from favorable to the completeness of such investigations. Partial preliminary investigations on the indicated question were begun long ago and were sometimes offered to student engineers of the Institute in the form of diploma works; but the subsequent works were always carried out under the close direction and with the participation of the author of the present essay. The entire exposition of the question touched upon is divided into three chapters: in the first of them the question of the crystalline structure of carbonaceous iron is considered; in the second—the influence of heating on the structure of carbonaceous iron; and in the third—the changes in the structure of iron during rolling, drawing, stamping, forging, cutting, and pressing.

The Crystalline Structure of Carbonaceous Iron.

Observing the various chemical compounds of iron encountered on the earth’s surface, we not infrequently find clearly crystalline forms of them. Thus, magnetic iron ore can often be found in the form of beautiful octahedral crystals. Figure 1 of Plate I shows a photograph of such an iron ore from our find in the Urals. The same regular octahedra are observed in chromic iron ore¹. According to the general law of isomorphism, the fundamental crystalline forms characteristic of a pure metal must be repeated, as such, also in the compounds of that metal; but one may also conclude, with a high degree of probability, conversely, that the crystalline forms of the octahedron must be repeated in pure iron as well. Beginning in the middle of the XVIII century, owing to the considerable interest in steel, very many experiments were made with the aim of obtaining distinct crystals of iron, and such attempts often ended successfully. Thus, in 1900 Osmond published the works of the XVIII-century metallurgist Grignon², devoted to the crystallization of cast iron, steel, and slags. Describing the phenomena of crystallization observed in cast iron, Grignon says³ that the first element of crystallization is a rhombus, upon which other constituent rhombi rise, forming, while gradually diminishing, a pyramid with a rhombic base. On the four sides of this principal and central pyramid there are arranged, at right angles from the base to the apex, other pyramids of the same form, diminishing in size according to the place occupied by them, in such a way that the section of the composite crystal is a quadrangular star, while in profile it forms little trees resembling small fir-trees. An entirely analogous crystallization in carbonaceous iron was observed by Martens and Knott⁴ and by our world-renowned professor Dmitrii Konstantinovich Chernov. Fig. 5 of Plate II shows Professor Chernov’s crystals, obtained by him from the ingot of a 100-t bloom of mild steel. The casting of this bloom was carried out in a metal mold, which entirely up to the upper feeder

was protected from rapid cooling by the earth. In the void of the ingot itself, owing to the shrinkage of the metal and owing to the very slow solidification of the ingot itself, the above-mentioned steel crystals were found. (The length of the large crystal is 39 cm and its weight 8½ pounds. At present this crystal has been placed at the disposal of the Naval Artillery Academy.) The crystal under consideration is double: one crystal is larger than the other; for one crystal the general conditions of its formation favored its growth more than for the other, neighboring one. In examining this crystal, we clearly notice that the fundamental form of the octahedron manifestly predominates in its structure, but the general arrangement of these crystals fully recalls the little fir-tree of which Grignon spoke earlier. Completely analogous octahedral crystals, but of considerably smaller dimensions, have been observed during quiet but very slow cooling of cast iron and steel by the investigators Bergman⁵, Berzelius⁶, Faraday⁶, and others. Iron crystals in the form of an octahedron under very slow and quiet cooling were observed by Bardeleben⁷ and Percy⁶.

Thus, on the basis of all that precedes, one may assert that the form of the octahedral crystal is quite characteristic of carbonaceous iron under very quiet and slow cooling, when the formation of this crystal is free; but it is interesting to trace the change in this fundamental crystallization of iron when the crystal, during its formation, encounters obstacles in the form of pressure upon its surface from the medium at the expense of which it is formed. The first condition for the beginning of the formation of a crystal from a solution, as every alloy is now regarded, is the condition of supersaturation of the solution during constrained crystallization. From the moment of the beginning of the formation of the crystal, the latter experiences the pressure of the medium at the expense of which it is formed. This pressure is proportional to the surface of the crystal and, obviously, it will be least at the more pointed cones of the crystal, i.e. at the ends of its crystallographic axes. Consequently, from the moment of the beginning of crystallization, the latter will begin to spread most intensively along the lines of least resistance and will give a series of axes of crystallization. As the crystals are separated out, the heat of crystallization is liberated, which raises the temperature of that medium through which the principal axis of crystallization passes. If the metal has weak thermal conductivity, or is artificially so arranged that it loses heat very slowly, and if around the principal axes of crystallization convection currents occur in the very substance at the expense of which the crystals are formed, then under favor—

...in favorable places for crystallization, secondary axes of crystallization begin to appear in directions perpendicular to the first principal axes. Then, under conditions entirely analogous and wholly favorable to the formation of second axes of crystallization, axes of a third crystallization may also form, perpendicular to the second, and so on. In this way we obtain a very complex system of axes of crystallization. Carbonaceous iron is not homogeneous in composition, and this inhomogeneity of composition is revealed still more at the moment of solidification of the metal, i.e., at the moment of the formation of crystallization. In places, the indicated axes of crystallization pass through foci (points) of liquid metal, which abundantly supply material for the formation of crystals; growth and thickening of the axes of crystallization occur, as a result of which the form of crystallization is not uniform, but has sharp changes in cross-sections—there results the so-called interrupted crystallization. The heat of crystallization of metals in general, and in particular also of iron, according to Wüstoff’s data,⁸) is very considerable; but carbonaceous iron also possesses considerable thermal conductivity; consequently, this heat can easily be lost and can contribute little to crystallization. But owing to the fact that carbonaceous iron is not homogeneous, the heat during cooling is radiated (lost) unevenly; in the mass of the cooling metal foci with an elevated temperature (relative to the average) appear, and in these places the passing axis of crystallization, provided that there are substances present at the expense of which the formation of crystals takes place, must acquire growth, a thickening in cross-section; i.e., the conditions again exist for obtaining interrupted crystallization. With respect to carbonaceous iron, a whole series of observations on the development and growth of crystals under the condition of the existence of pressure upon the crystals from the side of the medium from which these crystals are formed led Professor Chernov to conclusions of the following kind⁹).

“Examining individual crystals under the microscope, we see that they belong to torn crystals with the greatest development of shoots in the direction of the octahedral axes; moreover, one of them, in the direction of the principal growth of the crystal, always proves to be longer than the other two, so that each torn crystal appears in the form of the skeleton of an elongated square octahedron. Besides shoots in the direction of the octahedral axes, or shoots of the first order, as the distance from the apex to its base increases, we encounter, first in rudiments, and then more and more developed, shoots of the second, third, etc. orders, which sometimes literally form stepped rafters of the octahedral crystal.”

⁸) Theory of Alloys. p. 332.
⁹) Chernov’s articles in the Notes of the I.R.T.S. for 1868–69.

From the preceding it is not difficult to grasp that the principal condition for obtaining crystalline forms of carbonaceous iron is the slowness and uniformity of the cooling of the metal, and only under such conditions did the above-mentioned persons succeed in clearly observing crystalline forms in carbonaceous iron. With the aim of obtaining clearly crystalline forms in carbonaceous iron, experiments were recently carried out at the Putilov Works⁸ according to the directions of Prof. Chernov. For the experiments, a charge of soft iron and graphite was taken, placed in crucibles, melted, and then heated for more than two hours longer. The crucibles were then closed, the furnace was sealed, and the alloy was slowly cooled in the crucible together with the furnace. (The cooling time of the alloy reached 60 hours.)

For the indicated purposes, iron of the following chemical composition was taken:

	C.	Si	Mn	Ph	S
Demidov iron . .	0,10	0,14	0,22	0,043	0,019
Open-hearth . . . . .	0,12	—	0,39	0,007	0,012

From the indicated iron, mixed with graphite, after melting the following varied material of carbonaceous iron was obtained:

Nos. of ingots	C.	Si	Mn	Ph	S	Kind of melting.
1	1,80	0,30	0,15	0,026	0,022	Fireclay crucibles.
2	2,30	0,40	0,25	Traces.	0,024	Fireclay crucibles.
3	3,88	0,70	0,17	0,028	0,095	Graphite crucibles.
4	3,89	0,64	0,31	0,27	0,050	Graphite crucibles.
5	0,62	0,51	0,14	0,024	0,019	Fireclay crucibles.
6	0,60	0,48	0,10	0,026	0,026	Fireclay crucibles.

Thus, from the above table of analyses of the metal it is clearly evident that the carbonaceous iron in the experiments was very varied in chemical composition—from the composition of cast irons to medium steel in hardness inclusively. In fireclay crucibles during melting there occurred an enrichment of the metal with silicon at the expense of the walls of the crucible. Since the conditions of carrying out the latter experiments very favorably promoted the formation of crystallization of carbonaceous iron, as a result, in the fractures of the given alloys it was possible to notice clear traces of octahedral crystallization of iron. Fig. 2, Table I, represents the outer crust of alloy No. 2. In this fracture we see a clear [[unclear: continuation cut off on page]]

...a picture of primary crystallization; there is a series of crystalline axes gradually growing from the periphery toward the center; these principal axes have axes of the 2nd and 3rd order and, taken together, give a clear picture of discontinuous crystals. It was indicated earlier that on the axes of crystallization the growth of crystallization is most intense, as a result of which, on the outer crust, at the places where the axes of crystallization end, we observe distinct protuberances. These protuberances were especially characteristic on the outer crust of alloy No. 1; part of the upper crust of this alloy is shown in Fig. 3, Table I, with clear traces of the bases of the crystals themselves.

On the basis of all the foregoing, we involuntarily come to the conclusion that carbonaceous iron, under free crystallization, but under conditions of very slow cooling after melting, freely assumes a distinctly crystalline form in the shape of an octahedron. Under constrained crystallization, but under conditions of very slow cooling after melting, carbonaceous iron clearly shows an inclination toward the formation of discontinuous crystals, which in their form plainly resemble an octahedron (see Figs. 2 and 3), and from this we may conclude that the octahedron is the fundamental form of both of the just-considered types of crystallization of carbonaceous iron. The size of these crystals increases not only with the slowness of cooling, but also depends on the carbon content; namely: the character of crystallization, all other conditions being equal, will be the more relief-like the closer the carbon content is to 2%⁹, i.e., to point E of the fusibility diagram of carbonaceous iron according to Rosenbaum¹⁰.

In addition to the most frequently encountered form of crystallization, the octahedron, it has been found that carbonaceous iron is capable of crystallizing in the cubic system¹¹. In the latter form of crystallization, iron was found in one of the investigations in the metallographic laboratory of the Tomsk Technological Institute. (See Metallic Alloys. G. Tikhonov. 1910, p. 65, Fig. 95). Such variations in the crystallization of iron are entirely possible and find a complete analogy in variations in the character of crystallization of various salts depending on the conditions under which this crystallization is observed. Thus, according to the observations of Osmond and Carta¹², chrome alums, with rapid crystallization, give dendrites of the cubic system, and with slow crystallization—octahedra. Sodium chloride, which usually crystallizes in the form of cubes, gives an octahedron from solution—with an admixture of urea or Ca Cl₂ and Mg SO₄.

The reason for such a change in the principal form of crystallization, which is encountered rather often in many salts but is very rarely observed in carbonaceous iron, Curie¹³ explains in the following way: in order to form, a crystal expends a minimum of energy in overcoming external obstacles, and the magnitude of this energy for a given substance is a constant, independently of the conditions of crystallization. If the medium of crystallization of a given substance changes, then, while the previously indicated energy of crystallization remains constant, a change in the total external surface area of the crystal is possible, in the sense of its decrease or increase depending on the conditions of the medium in which crystallization occurs; with a change in the external surface area of the crystal, the acquisition of a new form of crystallization is possible. Only in this way can one explain the double form of crystallization of iron,—the very rare cube and the most widespread form of octahedra, formed in the first period of solidification of carbonaceous iron after melting, but before the appearance of allotropy in iron; and such crystallization, in distinction from allotropic crystallization, is called the primary crystallization of iron.

In the preceding, we considered the conditions for the formation of crystals in carbonaceous iron at high temperatures after melting, under slow cooling and before the onset of allotropic changes in iron; but if the cooling conditions fully favored the formation of crystals in iron before the onset of allotropic changes, then it is obvious that all further movement of the metal particles during the period of its allotropy must take place in the voids between the crystals of such metal and along the cleavage planes of these crystals. In fact, the later observations made in this area fully confirm what has just been stated. Thus, one of Prof. Chernov’s crystals¹⁴ from the shrinkage cavity of a 100 t ingot, when polished (Fig. 4, Plate I), has clear signs of crystalline structure within the crystal itself, with an evident grouping of the axes of crystallization; but in the cleavage planes of the crystals themselves (octahedra), at high magnification, there is clearly visible an accumulation of ferrite (Fig. 6, Plate III), i.e. of a substance that separated in these voids already after the formation of the primary crystals, at the moment of the allotropy of iron. In the present case the ferrite contains a very considerable quantity of slag embedded in it, which is explained by the fact that the given crystal formed in a shrinkage cavity, where there was a considerable quantity of slag squeezed out from the ingot itself during its solidification. A similar picture is dispo-

The ratio of the elements of secondary crystallization to the elements of primary crystallization can, evidently, be observed by us only in those cases when the cooling conditions of the carboniferous iron itself are wholly favorable to the appearance of the principal primary crystallization, and depending on the chemical composition of the carboniferous iron, the cementing substance between the principal primary crystals will in one case be ferrite, if the carbon content is less than 0.85%, and in the other case cementite, with a carbon content from 0.85 to 2%. (See Roozboom’s curve. Metallography, Part I, G. Tikhonov, pp. 27–29). A completely clear and distinct separation of ferrite grains in the cleavage planes of the primary crystals was distinctly observed in the above-mentioned experiments of Mr. Belyaev¹⁵ at the Putilov Works, and one of such sections from alloy No. 5 is shown in Fig. 7, Table III, with an entirely clear inclusion of ferrite between the primary crystals of the alloy itself.

During the period of secondary allotropic crystallization, the free mass of carboniferous iron—ferrite or cementite, depending on the chemical composition of the metal under consideration—fills the voids between the principal large crystals of primary crystallization; but the same elements of ferrite or cementite, under conditions of very slow cooling, are capable of penetrating into the cleavage planes of small crystals (of the 2nd, 3rd, etc. order) and producing independent branches. Under the indicated conditions, on the section we can clearly observe a picture of the division of the principal, large crystal of primary crystallization into crystals of the 2nd, 3rd, etc. orders, which was discussed earlier. Fig. 8, Table III, presents a clear picture of the distinct division of the principal crystal of carboniferous iron into the elements composing it. The alloy contains 0.55 C, 0.30% Mn, 0.020% Ph and 0.021% S, and was obtained by Mr. Belyaev¹⁶ at the Izhevsk Works. If, in parallel with the section just considered by us (Fig. 8), we carefully begin to examine the following (Figs. 9 and 10, Table III), depicting photographs from a section of one of the meteorites, then we shall find a very great similarity in the sense of the combination of lines in both drawings. In both drawings we can distinctly find triangles, parallelograms, and in some cases rectangles formed by the common network of ferrite on the sections under consideration. The figures under consideration on meteorites were first discovered in 1808 by the director of the Imperial Vienna porcelain factory, Alois von Widmanstätten, and in his honor were named “Widmanstätten figures.” The presence of Widmanstätten figures is an obvious sign of the crystalline structure of the body under consideration in the form of an octahedron. Depending on the position of the cutting plane, in the present case the plane of sec—

of the section under consideration, but with respect to the axes of the octahedron one obtains figures from the arrangement of the ferrite elements in the form of a triangle, a parallelogram, or a rectangle¹⁷. The basic elements of the Widmanstätten structure are clearly visible in a section of magnetic iron (Fig. 11, Table III); they are encountered in sections of steel (Fig. 12, Table IV) after slow cooling from high temperatures or after its overheating (Fig. 28, Table VI), and in general this structure appears in sections of carbon iron when the metal has been heated considerably above the initial moment of the allotropy of iron and then has been cooled very slowly. But the tendency of carbon iron to retain its crystalline structure is revealed in it even when the external conditions of cooling are very little favorable to such crystallization. Thus, under high heating (above the critical points) of carbon iron with a carbon content of about 2% and under very rapid and intense cooling, the principal structural element of such iron consists of austenite (Fig. 13, Table IV); but the structure of austenite, as is evident from the section under consideration, has clear signs of Widmanstätten figures. Finally, martensite, formed during the quenching of steel, under conditions in which the elements of the steel, owing to rapid cooling and to the change in volume of the steel during quenching, experience very strong lateral stresses (shears), has clear signs of Widmanstätten figures (characteristic triangles). (Fig. 14, Table IV). Thus, on the basis of all the preceding, we are involuntarily led to the following conclusion: carbon iron is capable of crystallizing; the most widespread form of such crystallization is the octahedron. With free, calm crystallization the development of octahedral crystals in carbon iron is clearly noticeable. Under less favorable conditions, with constrained crystallization, a distinct separation of the crystals is not possible; an interrupted crystallization is obtained with all the characteristic signs of complete octahedral crystallization. Finally, under very constrained conditions of crystallization, even under conditions of very rapid cooling, carbon iron manages to assume the rudiments of the crystalline structure peculiar to it, which is clearly revealed by the character of the arrangement of the particles capable of changing their position during the period of allotropy of carbon iron.

The influence of heating on the structure of carbonaceous iron.

From the preceding chapter we have ascertained that carbonaceous iron, when cooled very slowly after melting, is capable of assuming a distinctly crystalline form. Under ordinary conditions of cooling, however, and in the presence of mechanical treatment, we do not observe distinct phenomena of crystallization in iron. Thus, under the microscope, soft iron reveals a series of oval ferrite grains surrounded by a network of eutectic (pearlite). (Fig. 15, Plate IV). In harder grades of iron, close to cast, mild steel, the ferrite grains have the same form as in soft iron, but the pearlite network becomes more massive and thick (Fig. 16, Plate IV). Finally, in mild steel, instead of the thin pearlitic network of soft iron, a similar network of ferrite appears, and this network surrounds the pearlite grains (Fig. 17, Plate IV). The structure of cast irons in this respect is less definite; as they approach steels of medium hardness, the ferrite network is replaced by a dense network of cementite, while the pearlite grains acquire a distinctly fibrous structure (Fig. 18, Plate V). In white cast irons, when very slow cooling is observed, or after prolonged annealing, the fibrous structure of the pearlite grains is preserved (Fig. 19, Plate V), but the cementite network becomes discontinuous, with the formation of massive cementite inclusions standing out in relief within the pearlite field *).

The structure of carbonaceous iron under the influence of gradual heating from zero to the highest temperatures changes, and the first noticeable change in the structure of soft iron, as our observations in the Metallographic Laboratory of the Tomsk Technological Institute have shown, occurs at a temperature of about 300° C, at the temperature of the so-called “blue heat.”

The properties of blue heat manifest themselves at a definite temperature and in the presence of mechanical treatment, especially dynamic treatment.

Two factors play a role in blue heat: temperature and working. Therefore the entire work consists of a series of tests on bars of boiler iron subjected to thermal, mechanical, and combined treatment.

The temperature of blue heat has not been established; and it is most probable that it is a function both of the chemical composition of the iron and of its initial treatment. This is indicated already by the fact that all investigators who have studied blue heat have been unable to agree on a single definite temperature.

*) We have examined the structure of carbonaceous iron in greater detail in our brochure “Metallography,” Part I.

Before proceeding to the account of the course of the work and the material obtained, it is necessary to say a few words about the structure of the metal under blue heat. A comparison, made even with the naked eye, of the fractures of two bars—one raw and one treated at the temperature of blue heat—clearly indicates a very important fact, which partly clarifies the essence of the phenomenon of blue heat.

The fracture of iron under blue heat is striking because of the dissimilarity of the structure of its fracture to the structure of the fracture of a raw specimen. In the treated specimen under blue heat the grains become larger. The very appearance of the fracture, in form and character, little resembles the fracture of iron. The sharp crystalline, protruding ribs rather resemble the fracture of steel.

The considerable brittleness of iron under blue heat, as the work has shown, is due to a change in the structure of the material. Thus Fig. 20, Table V, depicts the structure of boiler iron before heating, and Fig. 21, Table V, after treatment under blue heat.

For the purpose of elucidating the change in the structure of the material under blue heat, for various kinds of thermal and mechanical tests there was simultaneously conducted an investigation of the structure of boiler iron under the microscope.

The first question of the work is the question of the temperature of blue heat. As the starting point in determining the temperature, the following proposition was adopted: under blue heat the structure of the metal changes in the direction of the formation of a larger grain than in the raw metal. This proposition was subsequently verified by a series of observations. For the purpose of determining the temperature of blue heat, the following work was carried out: 4 bars were prepared with dimensions \(12 \times 19 \times 19\) mm. One of them, designated No. 1, was heated in an electric furnace to a temperature of \(305^\circ\) C, and at this temperature was held for 20 min. The temperature fluctuations in this case did not exceed \(5^\circ\) in either direction. The bar was then rapidly cooled by immersion in water.

The polished section prepared from it had an average grain size \(=\) \(0.00163\) sq. mm.

The second piece, marked No. 2, was heated to a temperature of \(330^\circ\) C, was also held at a constant temperature for 20 minutes, and rapidly cooled. The polished section prepared from it had an average grain area \(=\) \(0.00180\) sq. mm.

The third specimen, marked No. 3, was heated to \(350^\circ\) C, under conditions analogous to those for bars 1 and 2; it was then also rapidly cooled. The polished section prepared from it and the photograph showed that the average area of its grain \(=0.002419\) sq. mm.

The fourth piece remained unheated. From it a polished section was also prepared, and the average grain area of the latter polished section \(=\) \(0.00180\) sq. mm.

Thus, examination under the microscope gave the following:

Nos. of sections	TYPE OF TREATMENT.	Heating temperature.	Heating time, min.	Grain size.
1	Heated in an electric furnace.	305	20	0.00163
2	” ” ” ”	330	20	0.00180
3	” ” ” ”	350	20	0.002419
4	Raw . . . . . . .	—	—	0.00180

On the basis of the material obtained, it follows that the temperature of blue heat, within the limits investigated, is equal to \(350^\circ\) C. As regards the practical indication—the blue color of the bars—it should be noted that all 3 specimens, heated and rapidly cooled, had colors of varying intensity, and in any case the coloration could not serve as a reliable means of recognizing blue heat.

The second series of experiments had the aim of proving the proposition that the metal at blue heat does not withstand mechanical working.

If a temperature of \(350^\circ\) is the temperature of blue heat of boiler iron, then, during the working and testing of specimens heated to \(350^\circ\), will the properties inherent in the metal at blue heat appear? To clarify this question the following was done: specimens were prepared for a bending test.

The first, marked No. 5 and No. 6 at the ends, was heated in a forge on charcoal to blue heat, which was judged by eye from the color; then it was forged with a sledgehammer, the bar being turned through \(90^\circ\) the whole time. The piece treated in this way was subjected to a bending test. Bending was carried out by blows of a hammer on one end (and not at the place of bending), while the other end was fastened in an anvil. During bending care was taken that there should be no crushing and damage to the fibers on the side adjacent to the supporting point.

The result obtained was the following: the bar, bent through \(90^\circ\), gave clearly visible cracks on the outer fibers; when bent back to its initial position the bar broke at the very first blow.

In parallel, a specimen of raw boiler iron was bent under exactly the same conditions. The result obtained was the following: the bar withstood 4 complete bends to one side and the other. The first signs of mechanical damage to the fiber appeared at \(2\frac{1}{2}\) turns; later the damage increased, but the material showed considerable toughness.

The result obtained may be presented in the following form:

Nos.	TYPE OF TREATMENT.	Number of bends through 90°	Number of bends in %.
5 and 6	Heated in the fire to a blue color and forged with a sledgehammer . . . . . Raw . . . . . . . . . . . .	1 1/4 8.0	15.6% 100%

The second specimen was also heated in the forge to a blue heat, then forged under the hammer with constant turning through 90° and bent into the form of the letter V, and then flattened under the hammer already completely.

The result of the inspection was as follows: inside the bend of the bar there appeared a clearly visible crack through the whole bar.

The same test was performed with a raw bar. No cracks were observed either inside or outside. The material showed great ductility. Comparing the results obtained, we shall have the following:

Nos.	TYPE OF TREATMENT.	Results of the experiments.
7 and 8	Heated in the forge to a blue color, forged under a sledgehammer and bent into the form of V. Raw piece was also bent into the form of the letter V.	A crack appeared inside. No cracks appeared.

The final conclusion of the tests carried out may be formulated as follows: mechanical treatment at a blue heat affects the metal in the direction of diminishing its viscosity and ductility, making it at the same time brittle and poorly resistant to bending forces.

The tested specimens were left to cool slowly. Then pieces were taken from parts close to the place of fracture and bend for making polished sections and for study under the microscope.

Work under the microscope gave the following:

Nos.	TYPE OF TREATMENT.	Average area of grains in sq. mm.
5 6	Treated under the sledgehammer and tested in bending . . . . . .	0.004195
7 8	Treated under the hammer and bent into the form of the letter V . . . . . .	0.00427

Change in the Structure of Iron during Mechanical Working

The subsequent work had the aim of verifying the results obtained by testing boiler iron in tension.

From the raw iron, 4 specimens were made with dimensions \(12 \times 19 \times 300\) mm. The first bar, numbered No. 1, was left without any treatment.

The second piece (No. 2) was heated in a forge, by eye, to a blue heat and was subjected to working under the hammer, with turning through \(90^\circ\).

The third bar (No. 3) was heated in an electric furnace to \(350^\circ\), held at this temperature for 20 minutes, and then slowly cooled.

The fourth was likewise heated in an electric furnace to \(350^\circ\), held at it for 20 minutes, and then rapidly cooled by being thrown into water (No. 4). Thereafter all four bars were used to make tensile specimens.

The specimens are marked as follows:

Specimen from raw material, marked ................. 1
Specimen from heated and worked material, marked .... 2
Specimen from heated and slowly cooled material, marked .... 3
Specimen from heated and rapidly cooled material, marked .... 4

From each piece, 5 tensile specimens were obtained.

Work on Gagarin’s tensile-testing machine gave the following result:

No.	Relative elongation, %	Relative contraction, %	Stress at elastic limit	Maximum stress	Stress at rupture	Treatment of the bar from which the specimen was made
1 I	28,8	70	18,95	30	—	none.
1 II	39,75	77	17,80	30,2	19,2	none.
1 III	37,80	74	17,3	30,25	16,4	none.
1 IV	37,50	76	17,8	30,75	18,4	none.
1 V	39,20	77,5	18,4	30,5	19,0	none.
Average	36,6	74,9%	18,06	30,5	18.25	none.
2 I	19,40	78%	36,3	41	19,12	Heated to \(350^\circ\) and worked under the hammer.
2 II	12,60	68	44,5	42,0	26,5	Heated to \(350^\circ\) and worked under the hammer.
2 III	17,60	63,5	38,5	42,5	24,75	Heated to \(350^\circ\) and worked under the hammer.
2 IV	13,50	1	39,8	42,6	25,5	Heated to \(350^\circ\) and worked under the hammer.
2 V	9,4	64	50,2	50,4	29,2	Heated to \(350^\circ\) and worked under the hammer.
Average	14,5	68,9	41,86	43,6	25,0	Heated to \(350^\circ\) and worked under the hammer.

No.	Relative elongation, %	Relative contraction, %	Stress at the elastic limit	Maximum stress	Stress at rupture	Treatment of the bar from which the specimens were made
3 I	—	—	15,85	29,6	—	Heated to 350° and slowly cooled.
3 II	33,2	77%	16,3	28,7	20,0	Heated to 350° and slowly cooled.
3 III	44,0	78	16,7	29,2	18,0	Heated to 350° and slowly cooled.
3 IV	38,5	80	15,2	29,3	17,9	Heated to 350° and slowly cooled.
3 V	37,5	74	18,8	30,0	18,6	Heated to 350° and slowly cooled.
Average	38,3	77,25	16,75	29,36	18,5	Heated to 350° and slowly cooled.
4 I	36,5	73%	18,0	30,0	18,6	Heated to 350° and rapidly cooled.
4 II	39,0	76	17,2	30,5	18,7	Heated to 350° and rapidly cooled.
4 III	30,5	75	18,1	30,7	17,1	Heated to 350° and rapidly cooled.
4 IV	36,5	77	18,1	29,5	19,6	Heated to 350° and rapidly cooled.
4 V	38,0	69	18,8	30,2	24,2	Heated to 350° and rapidly cooled.
Average	36,1	74	18,04	30,2	19,64	Heated to 350° and rapidly cooled.

On the basis of these numerical data, one may draw the following positive conclusion: 1) blue heating alone, neither with slow nor with rapid cooling, has any harmful influence on the properties of the metal; 2) blue heating together with mechanical working fundamentally changes the properties of the metal, raising the stress of rupture to the elastic limit,—from 18,06 in the raw state to 41,86 in the material worked under blue heating; this, in percentages, taking the stress of the raw material as 100, gives for the heated and worked material a value of 210,5%, or, in other words, the stress of rupture in the heated and worked material is 2,10 times greater than in the raw material.

Conversely, the elongation of a specimen worked under blue heating will be considerably less than the elongation of the raw material.

The average elongation obtained for the raw material . . . . . . 36,6
” ” ” worked under blue heat . . 14,5

That is, the elongation of a specimen worked under blue heating is 2,5 times less than the elongation of the raw specimen.

The results of the experiments gave on average:

No.	TYPE OF TREATMENT	Heating temperature	Stress to the elastic limit	% elongation to the elastic limit
1 (5)	Raw . . . . . .	—	18,06	36,6
2 (5)	Heated and worked under the hammer.	350°	41,86	14,5
3 (5)	Heated (cooled slowly).	350°	16,75	38,3
3 (5)	Heated (rapidly cooled).	350°	18,04	36,1

When examining the fracture diagrams of the bars, the following features of each series of diagrams became clear.

The diagrams for the raw iron proceed smoothly up to the elastic limit; beyond the elastic limit the curve runs for some time almost horizontally, which indicates a considerable viscosity of the metal, and then the curve again rises upward and, finally, descends quite smoothly.

In the diagrams obtained in rupturing 5 specimens of the 2nd piece, heated and treated, the curves up to the elastic limit run considerably more steeply and rise higher, which indicates a greater resistance to the rupturing force and a small capacity of the metal to elongate. Indeed, calculation showed that the stress up to the elastic limit in the heated and treated bar is almost 2.1 times greater than in the raw one, while the elongation, on the contrary, is 2.5 times greater in the raw one than in the heated and treated one. Beyond the elastic limit the curve rises considerably and then quickly begins to fall. Consequently, the viscosity is considerably reduced.

The diagrams obtained from rupturing specimens of the 3rd piece, heated and slowly cooled, have the following features.

The curve up to the elastic limit resembles the curve of the raw material, but beyond the elastic limit there is no longer a horizontal or slightly rising curve, as in the first bar. The diagram at this point presents a series of more or less considerable jumps, indicating a lower viscosity of the material than initially. Such a phenomenon can be attributed only to a change in structure. Earlier microscopic investigations showed that the structure of the metal heated to blue heat changes toward the formation of a coarser grain. In this fact one should seek the explanation of the phenomenon that the curve beyond the elastic limit makes jumps upward and downward. It follows that blue heating, having caused a change in structure, as it were somewhat diminishes the bond between the ferrite grains and the pearlite network.

The diagrams obtained from specimens of the 4ths, heated and rapidly cooled, proceed up to the elastic limit smoothly, but more gently than in the raw material. Beyond the elastic limit the curves also proceed smoothly and in character resemble the diagrams of raw specimens. On this basis, one may assume that the material heated to blue heat and rapidly cooled, as it were, restores its original capacity to resist rupturing forces.

In conclusion, it remains to mention also the tests carried out for the purpose of clarifying the question of whether, within the limits of the blue-heat temperature, there are characteristic critical points.

For this purpose three observations were made:

1) A bar of raw boiler iron was heated to 450° and slowly cooled in the furnace. With the aid of Kurnakov’s apparatus a curve was obtained which showed that, within the temperature range up to 450°, no such points are noticeable.

2) The experiment was repeated with a second bar, which had previously been heated to 350° and worked under the sledgehammer. The second curve of the diagram gave the same result.

3) For verification a third observation was made. A specimen, heated and worked, was heated to 1000° and slowly cooled. The critical point was found at about 700°.

The result of the latest work on determining the effect of blue heat can be formulated as follows:

For our boiler iron the temperature of blue heat is 350° Celsius; when heated to this temperature the iron becomes coarse-grained; iron heated to 350° and worked, becoming still more coarse-grained, becomes brittle and poorly resistant to bending, although the tensile stress rises considerably, with a significant decrease in ductility.

Iron heated to 350° and subjected to tensile tests resists almost in the same way as raw iron, irrespective of the rate of cooling.

From the foregoing it is evident that soft boiler iron, under the influence of heating at a temperature above 300° C, begins to enlarge the ferrite grains; a similar picture of grain enlargement continues further when the iron is heated. The duration and high temperature of heating only promote grain enlargement. The latter is quite vividly illustrated by the following example: soft boiler iron was placed in a burning forge hearth, and after 4, 5, and 7 days, sections were prepared from the iron specimens taken (Figs. 22, 23, Pl. V, and 24, Pl. VI). The hearth operated only during the day, for 8 hours. From the photographs of the sections taken, the enlargement and gradually increasing disintegration of the ferrite grains are clearly noticeable. At first, as the duration of heating of the iron increased, black points appeared on the ferrite grains (Fig. 22), which the specimen had not had before; then the number of these black points grows on the ferrite grains together with the growth of the latter (Fig. 23). Thereafter the size of the ferrite grains reaches considerable dimensions, together with the formation of torn cracks arranged along straight lines, evidently in the direction of the rolling of the iron, along the traces of elongated slags. We can observe a similar picture of the gradual change in grain shape in steel and cast iron. Thus, in steel of medium hardness (carbon 0.95) (Fig. 25, Pl. VI), when heated to about 700°, the ferrite network gradually begins to disappear, and the disintegration of the pearlite grains is noticeable (Fig.

at 26 (Plate VI). At temperatures above 700° C the ferritic network disappeared completely and, together with the melted grains of pearlite, formed a new structural form—martensite (Fig. 27, Plate VI). Upon further heating at temperatures close to 1000° C, when the steel begins to sparkle strongly, the same steel assumes a form close to primary crystallization, with an evident sign of Widmanstätten figures; the steel, as they say, is in the stage of overheating (Fig. 28, Plate VI). Cast irons change their structure somewhat differently under slow heating: in gray cast irons, under the influence of slow heating, graphite begins to separate abundantly and thereby causes considerable brittleness of the cast iron. Thus Fig. 29, Plate VI, depicts gray cast iron before heating, and Fig. 30, Plate VII, the same cast iron after prolonged heating; the sharp and abundant separation of graphite is clearly noticeable on the latter polished section. In white cast irons, under the influence of prolonged heating at high temperatures, decomposition of the carbon of hardening and of cementite occurs, with separation of temper carbon and grains of fibrous pearlite (Fig. 19, Plate V), but in the presence of oxidizers (red iron ores)—the temper carbon begins to separate in the form of CO, free carbon forms small inclusions (in the form of small circles on the polished section), and the carbonaceous eutectic of the polished section begins noticeably to divide into oval grains of ferrite; the result is malleable cast iron (Fig. 31, Plate VII). In the last two cases the cast iron, sharply changing its structure, at the same time acquires considerable softness, even malleability; but it is also possible, conversely, by rapid cooling to transform gray, fairly soft, cast iron into chilled, considerably hard cast iron; in the latter case, the cast iron is said to have become chilled (Fig. 32, Plate VII).

From the preceding we have seen that the three principal types of carbonaceous iron—wrought iron, steel, and cast iron—are capable, upon heating, of changing the structure of their constituent particles, and the change of this structure is observed most sharply and vividly at higher temperatures. Often such an allotropic change in the structure of the given body is accompanied by considerable absorption of heat during heating and, conversely, by evolution of heat during cooling. In the latter case the metal, at the indicated moment, noticeably changes its color to a higher color according to temperature and changes in volume. It is quite natural that such characteristic, or, as they say, critical points in various grades of carbonaceous iron were long ago noticed by experimenters. Thus, as early as 1868, our famous metallurgist D. K. Chernov¹, in his report to the Imperial Russian Techni—

... indicated to the [[unclear: -ical]] Society the existence of two critical points, “a” and “b,” in various grades of carbon steel. The significance of these two critical points D. K. Chernov characterized in the following way:

Each grade of steel has its own critical temperature, above which it must be heated in order to obtain hardening. Steel heated below this temperature does not take hardening. This first critical point D. K. Chernov calls the critical point “a.”

For each grade of steel there exists yet a second critical point, called by D. K. Chernov point “b”; when heated below it, steel after cooling does not change its structure. But if the steel is heated to the limit “b,” then it passes from a crystalline structure into an amorphous (wax-like) one, which it retains up to a certain limit very close to the melting temperature of the steel itself. Upon slow cooling of steel heated to a temperature above point “b,” the reverse phenomenon occurs with the steel—the steel gradually passes from an amorphous, wax-like structure into a distinctly crystalline one; the phenomenon of crystallization of the steel will be the more intense, the more slowly the cooling of the steel proceeds, and the closer the temperature is to point “b,” below which no change of crystallization is observed in the given steel. To retain a fine-grained structure of the steel it is necessary to cool the latter rapidly to a temperature below point “b.”

Thus in 1868 there was indicated the boundary of temperatures within which a distinct crystallization of iron occurs. In 1886 Pionchon and Le Châtelier, heating chemically soft iron and steel to a certain temperature, noticed that such heating is always accompanied by a certain absorption of heat at a definite temperature, and conversely by an evolution of heat upon cooling. Then in 1888 Professor Osmond², wishing to study the position of the critical points in various grades of iron, applied Le Châtelier’s thermo-electric pyrometer for determining such points. For his observations Osmond took samples of the following chemical composition:

	C.	Si.	Mn.
Soft iron obtained electrolytically	0.08	—	—
Very soft steel	0.16	0.01	0.14
Soft steel	1.29	0.06	0.27
Semi-hard crucible steel	0.57	0.08	0.23
Hard tool steel	1.25	0.19	0.1

The results of the above-mentioned observations of Osmond came down to the following: upon cooling soft electrolytic iron, heated—

...up to \(1000^\circ\mathrm{C}\), the pyrometer stopped at \(855^\circ\) and remained at this temperature for a certain period of time. This first arrest Osman called the point “\(Ar_3\).” A second, slight arrest of the pyrometer, called by him the point “\(Ar_2\),” was observed at a temperature of \(750^\circ\), reached a maximum at \(730^\circ\), and ended at \(690^\circ\), with weak oscillations at \(705^\circ\mathrm{C}\).

For the tested grades of steel the position of these same points varied in the following manner:

For very mild steel:

	Beginning	Maximum	End
\(Ar_3\)	\(845^\circ\)	\(825^\circ\)—\(819^\circ\)	\(800^\circ\)
\(Ar_2\)	\(755^\circ\)	\(735^\circ\)—\(725^\circ\)	\(710^\circ\)
\(Ar_1\)	\(680^\circ\)	\(662^\circ\)—\(650^\circ\)	\(645^\circ\)

For specimens of mild steel:

	Beginning	Maximum	End
\(Ar_{3-2}\)	\(780^\circ\)	\(721^\circ\)—\(715^\circ\)	\(690^\circ\)
\(Ar_1\)	\(680^\circ\)	—	—

For semi-hard steel:

	Beginning	Maximum	End
\(Ar_{3-2}\)	\(750^\circ\)	\(700^\circ\)—\(690^\circ\)	—

and ends with an arrest \(Ar_1\) at \(661^\circ\mathrm{C}\).

Specimens of hard steel had only one prolonged arrest \(Ar_{3-2-1}\) at a temperature of \(674^\circ\mathrm{C}\).

The position of the indicated critical points characterizes the following allotropic changes in iron: in very mild grades of iron \(Ar_3\) and \(Ar_2\) correspond to the first transition of iron from the “\(\gamma\)” form into the stage \(fe\beta\), and the second—from \(fe\beta\) into \(fe\alpha\), while the point \(Ar_1\) is the end of such transformations³. From the preceding table it is clearly evident that the duration of the position of the point \(Ar_1\) depends on the carbon content. For very mild grades of iron the point \(Ar_1\) is almost entirely imperceptible; for harder ones it attains a considerable magnitude. Since the ability of steel to take a hardening increases with the carbon content, it follows that the point \(Ar_1\) corresponds to the beginning of the allotropic changes of iron that are proper to the hardening, in consequence.

The temperature of the critical point \(Ar_3\) corresponds to the complete end of the allotropic transformations of the particles of carbonaceous iron during the transition from an evidently crystalline form into the fine-grained, almost amorphous structure characteristic of hardened steel. Thus, comparing the value of the critical...

...of the points “a” and “b” with the points obtained by Osmond, we arrive at the conclusion that point “a” of D. K. Chernov corresponds to Osmond’s point “Ar₁,” and point “b” to point “Ar₃.” The hardening of steel is possible only under the condition of heating the latter to the temperature of the critical point Ar₃ and of the corresponding rapid cooling. At this same temperature, carbon iron, under the condition of slow heating to this temperature and rapid cooling from it, loses its magnetic properties, and in addition the electrical conductivity resistance increases considerably. Thus, the temperature of the point Ar₃ is the critical temperature for the allotropic change of iron. It is necessary to say a few words about the fact that the positions of the critical points in one and the same grade of carbon iron under conditions of slow heating and cooling do not coincide; namely, the positions of the critical points on heating are always somewhat higher than on cooling. Such an anomalous displacement of the critical points, in Osmond’s opinion, is explained by the existence in iron of a special factor, namely the initial temperature⁴; thus, according to Osmond’s observations, hard steel heated to 745° C showed Ar₁ at 680° C, and after heating to 1410° C (melting), at 655°; in this way, an increase of the initial temperature by approximately 670° lowered the position of Ar₁ by 25°.

From the preceding table of the positions of the critical temperatures obtained by Osmond, it is clearly apparent that the interval of the positions of the critical points narrows with an increase in the carbon content in iron, and, moreover, the position of the critical points is lowered in temperature. In addition, according to the experimental investigations of Carpent, Keeling, and A. L. Baboshin,⁵ manganese also affects the position of the critical points in the same way as carbon; the latter circumstance has especially great significance in deciding the question of the choice of rolling temperature for rails with a high manganese content.

Let us now consider the question of changes in the grain size of carbon iron upon heating. The clear moment of the beginning of such changes coincides with the position of the point Ar₄ in the iron being tested. It does not prevent us from noting that an increase in grain size in soft grades of iron, as we saw earlier (blue heating) and shall see subsequently, was also observed at a temperature below Ar₁. The general law, formulated by Howe with respect to the dependence of grain size on temperature, consists in the fact that the grain size of iron is strictly proportional to the temperature. On the basis of his observations, Sauveur⁶ for the steel specimens tested by him formulated this dependence of the grain—

...as follows from the temperature: T max = 680 + 281,250A, where T max is equal to the temperature attained, and A is the actual area of the grain in square millimeters at a linear magnification of 250 times.

Further, in Howe’s opinion, to each given temperature there corresponds a quite definite grain size. The growth of the iron grain does not occur instantaneously up to the limiting size upon the iron’s reaching a definite temperature, but gradually; at first the growth of the grain proceeds very rapidly, and then becomes slower and slower as the grains approach the normal size.

Finally, in Howe’s opinion, if there exists a grain larger than the normal one for the given temperature, then it does not return again to the normal size.

The general law formulated by Howe concerning the size of the iron grain as a function of temperature is shown rather clearly in graphic form in Fig. 1, Table XXIII. The line jDG expresses the grain size as a function of temperature. Suppose that the given iron has been heated to temperature C, and that the grain of this iron has reached size H; but the grains of the given iron at the same temperature tend to increase in size until they reach size D, corresponding to this temperature C. After the grain size reaches D, let us cool the iron to temperature E, but in doing so the grain size remains equal to D, and will not be equal to K, as this would correspond according to the diagram. In constructing the curve of the dependence of grain size on temperature, Howe noted that direct proportionality of grain size to temperature exists only up to a certain limit M, after which the dependence indicated by him is violated: the grain size grows appreciably faster than the corresponding change in temperature. In observations made by us in the Metallographic Laboratory of the Tomsk Technological Institute, the opinion just expressed by Professor Howe was fully confirmed.

Direct proportionality between the size of the iron grain and the corresponding heating temperature exists only up to a known limit of heating—within the limits up to 1000° C for soft and medium grades of iron in hardness. A similar dependence of grain size on heating temperature is also indicated by A. L. Baboshin⁷.

If we now represent the material available to us, concerning the dependence of grain size on heating temperature, in the form of curves (Figs. 2, 3, Table XXIII, and 4, Table XXIV), then on the indicated curves there will be clearly noticeable a point of inflection at a temperature of about 1000°, corresponding to the cessation of direct proportionality of grain size to temperature.

The position and significance of this point of cessation of the direct dependence of grain size on temperature are of profound importance for engineers engaged in the hot working of iron. At this temperature, and especially above it, iron passes into the stage of coarse crystallization, and it is known that, in the presence of coarse crystallization, iron cannot possess good mechanical properties.

It is quite possible to suppose that D. K. Chernov, in his report of 1868, had in mind precisely the just-mentioned point of inflection of the direct dependence of grain size on temperature, when he formulated his proposition as follows: ⁸) “If we wish to increase the density of steel, to bring its grains closer together and thereby attain a tighter bond between them, then forging must be carried out at a temperature below point ‘b’.” This latter assumption of ours is all the more probable and close to the truth if one takes into account that at the time indicated there were as yet no simple, accurate methods for determining temperature, and the structure of the observed piece of iron was determined by fracture, without a microscope.

Thus, from the foregoing it is evident that carbonaceous iron, within the limits of considerable heating, is highly prone to change its structure. The original supposition that a change in the crystallization of iron is possible only above the position of the critical point $Ar_1$ is refuted by the preceding case with boiler iron at a “blue heat” temperature, and was disproved by Stead, who observed an increase in the grain of mild steel with $0.10\%$ C at temperatures from $500^\circ$, the grain growth increasing with the duration of heating.

A similar picture was observed also by von A. Joisten ⁹) in his experimental investigations of iron wire, a soft and very homogeneous material. In the experiments, a whole series of precautions was employed to prevent oxidation of the metal during heating in an electric furnace. The heating time of the specimens ranged from one to 20 hours at one and the same temperature. The results of these experiments are presented in the form of curves in Fig. 5 of Table XXIV. From an examination of the just-mentioned curves of the growth of iron grain as a function of temperature and duration of heating, it is clearly evident that growth of the iron grain is possible beginning at $500^\circ$ C and beyond, and that the grain size increases with increasing duration of heating.

Similar observations were made by us in the metallographic laboratory of the Tomsk Technological Institute on soft wrought iron and rail steel. The experiments were carried out in a Heraeus electric furnace in a nitrogen atmosphere in order to reduce oxidation of the metal—

⁸) G. Tat’ev. Homogeneity of Rails, p. 86.
⁹) Mitteilungen aus dem Eisenhüttenmännischen Institut des Königl. Techn. Hochschule Aachen. 1910.

steel; the specimens were heated in the aforementioned furnace for 1; 1.5; 2; 2.5 and 3 h, and from the middle of the observed specimens sections were prepared for measuring the grains over an area equal to 1 sq. cm, with the total magnification when taking the photograph equal to 100. The results of the observations obtained are presented in the following table, in which the number of grains per 1 sq. cm is indicated.

Heating temperature	Material	1 hour	1.5 hours	2 hours	2.5 hours	3 hours
500° C.	Soft wrought iron	10	10	9,5	9	9
550° C.	Soft wrought iron	9,5	9,5	9,5	9	9
600° C.	Soft wrought iron	10	9	9	8,5	8,5
650° C.	Soft wrought iron	10	8,5	9	8,5	8
700° C.	Soft wrought iron	8	8	8	7,5	7,2
750° C.	Soft wrought iron	8	8	7	7	6
800° C.	Soft wrought iron	8	7,5	7,2	7	7
850° C.	Soft wrought iron	5	7	7	7,5	8
500° C.	Rail steel	3,5	3,5	3,5	3,5	3,4
550° C.	Rail steel	4	3,5	2,8	2,8	2,7
600° C.	Rail steel	4	3,5	3,3	2,9	2,6
650° C.	Rail steel	3,5	3,3	3,1	3	2,6
700° C.	Rail steel	3	3	—	—	2,5
750° C.	Rail steel	2,8	—	—	—	2,4
800° C.	Rail steel	2,6	—	—	2,2	2,3
850° C.	Rail steel	2,5	2,5	2,4	2,3	2,2

From the last table and the preceding Fig. 5 it is quite clearly seen that the grain size of iron increases noticeably with an increase in the heating time. This latter proposition concerning the increase in the grain size of soft iron with the duration of heating is rather vividly illustrated by the photographs examined earlier (Figs. 22, 23, Plate V, and 24, Plate VI).

The causes producing a change in the grain during prolonged heating may be explained as follows: every carbon steel is a conglomerate of underdeveloped crystals soldered together by eutectic. Under the influence of heating, diffusion of substances occurs between neighboring crystals through the film of eutectic. Diffusion is the result of the different solubility of two adjacent planes of crystals in the eutectic, as a result of which, owing to the appearance of pressure, there occurs a gradual transfer of particles of one crystal toward another, whereby po-

the latter begins to grow at the expense of the former. The reason for the different etchability of two adjacent planes of crystals lies in the nonuniform orientation of the axes of the crystals, which is very clearly revealed when a polished section of soft iron is etched; namely, ferrite grains under the microscope have different shades of coloration depending on the different disposition of the crystal axes. The eutectic, whose melting temperature is lower than that of the crystals themselves, on heating is closer to the liquid state in comparison with the crystals themselves, and thereby only promotes the growth of the latter. But if the eutectic binding the crystals of a given iron is in the solid state, then at a temperature from \(690^\circ\) to \(670^\circ\) there occurs a mutual displacement and attraction of like particles. This phenomenon was called by Stead coalescence. Only in steel with a carbon content corresponding to the eutectic is the structure in equilibrium, and the movement of the particles of the structure is barely noticeable. Mutual attraction of like particles takes place in the case when there is no equilibrium in the structure, but there is an excess of one of the constituents of the eutectic: ferrite or cementite. The slower the cooling proceeds, the more complete is the separation of the excess elements, and the less eutectic remains. With a very small pearlite content, the latter, after slow cooling, turns into the thinnest bands separated by ferrite, as we observed earlier in Fig. 8, Plate III for steel with content \(C = 0.55\%\) (the ferrite on the polished section is white). In hard grades of steel, after prolonged annealing within the temperature limits indicated by Stead, cementite, being the connecting link between separate pearlite grains, attracts to itself cementite grains from the eutectic, as a result of which the pearlite acquires a fibrous structure (Fig. 18, Plate V). Under more rapid cooling conditions, less favorable to coalescence of the observed piece of steel, the fibrous structure of pearlite is not visible, but the beginnings of enrichment of the cementite network are clearly noticeable (Fig. 12, Plate IV). Under slow-cooling conditions of eutectic steel, the usual form of fibrous pearlite is formed (Fig. 33, Plate VII), but if we carry out cooling of the eutectic steel with some slowing in the critical interval, then after rapid cooling at the temperature of this interval we can obtain pearlite that is not fibrous but granular; the reason for this phenomenon lies in the property of coalescence of metal particles even in the eutectic itself (Fig. 34, Plate VII). Fibrous pearlite does not at once pass into granular pearlite, but gradually, according to the duration of heating: at first the cementite fibers acquire thickened constrictions, and in this form pearlite was called by Benedicks—reticular; and then reticular pearlite, under analogous conditions, passes into granular. In the latter case we have complete separation of the constitu-

...forming pearlite, i.e. cementite, and ferrite, i.e. we obtain, according to Osmond’s expression⁸, the so-called structural equilibrium.

Speaking of granular pearlite, we have already passed somewhat into the interval of critical temperatures, in which the allotropic change of the particles of carbonaceous iron takes place, accompanied by the transition of grains of ferrite and pearlite, or pearlite and cementite, or pearlite alone (in eutectic steel) into the higher allotropic modifications with respect to temperature—martensite and austenite. Such a transition takes place smoothly as the temperature rises, with the formation of certain intermediate forms bearing special names. Thus the transitional form of fibrous pearlite, with the formation of cementite grains in the ferrite field, is not infrequently called sorbite, after the scholar Sorby, who discovered this element. With a further rise in temperature the structural elements indicated earlier form a new allotropic modification, called troostite. The latter is formed in carbonaceous iron when heated within the critical interval at the moment when the shortening of the length of the specimen under test reaches its greatest magnitude. On a polished section, troostite appears as dense masses with a faint disposition of granules, rudiments or needles of martensite (Fig. 35, Table VII). The homogeneous structure of troostite permits one to suppose that it represents an intermediate equilibrium between the phases of martensite and cementite + ferrite.

With further heating of carbonaceous iron, at the moment of the critical temperature \(Ar_3\), there is formed the next, more stable allotropic structure, named, after the person who discovered it, Prof. Martens—martensite. Martensite has a characteristic needle-like structure (Fig. 14, Table IV). With an increase in the carbon content in iron the general character of the structure does not change, but the hardness increases and the coloration on etching changes somewhat. The homogeneity of the structure of martensite permits it to be regarded as a solid solution of carbide in iron. Pure martensite is obtained when quenching steel with a carbon content up to \(1.1\%\); with a further increase in the carbon content and with a rise in temperature to \(1000^\circ\mathrm{C}\), provided there is rapid cooling after the indicated temperature, a new structural element appears in steel, named by Osmond austenite (Fig. 13, Table IV). Austenite in pure form is not obtained on polished sections of steel; it always represents a dense mass having the structure of Widmanstätten figures, the dark bands of these figures being filled with martensite. The ratio of the areas occupied by austenite and martensite is in great dependence on the carbon content in the steel and on the rate of cooling. Thus, with a carbon content in steel of \(1.1\%\), after the above-mentioned thermal operation, there appear—

only traces of austenite remain; at 1.57% C in steel the amounts of martensite and austenite are equal, and at 1.8% C the amount of austenite constitutes 70% of the area of the polished section, the rest being martensite (according to Osmond). Austenite should be regarded as a less stable allotropic phase, corresponding to the hardening of carbon iron (iron $\gamma$ from Roseboom’s diagram⁹), while martensite is only a modification of this phase with incipient decomposition into ferrite and cementite.

It is necessary to say a few more words about the burning of carbon iron—after cooling from temperatures of about $1000^\circ$ C; but under conditions of prolonged heating at this temperature—in soft iron and steel there occurs a disintegration of the ferrite grains, with the disappearance of the outlines of the grains themselves (Fig. 24, Table VI), while in harder steel there is a considerable thickening of the very network of ferrite surrounding a significant area of pearlite (Fig. 28, Table VI). During high heating, atmospheric oxygen penetrates into the loose crevices between the grains of ferrite or pearlite, oxidizes the surface of these grains, and thereby only contributes to the disruption of the close bond between the grains, causing considerable brittleness of the metal. The phenomena of primary crystallization, which immediately follow the phase of allotropic changes (called by us secondary crystallization), were considered by us in the first part of this work.

Change in the structure of iron during rolling, drawing, stamping, forging, cutting, and pressing.

The temperature at which the rolling of iron is completed at the best modern works is considerably below $1000^\circ$ C—usually the finishing temperature of rolling is very close to $850^\circ$ C. At one of the southern Russian works (namely the Dnieper works), Engineer I. I. Efron carried out prolonged observations of the rolling of hard, manganese rails with contents of $C = 0.48\%—0.55\%$ and $Mn = 0.92—1.11\%$; by means of an optical pyrometer, Mr. Efron determined the temperature $T$ at the beginning of rolling, at the exit of the bloom from the first pass of the blooming mill, when the bloom was freed from scale, and the temperature $t$—at the end of rolling, at the exit from the last pass. For seven rails of the indicated composition these temperatures fluctuate¹:

1) $T = 1240^\circ$ . . . . . . . . . . . . . . . . . . $t = 738^\circ$ C.
2) „ $\ \ 1195^\circ$ . . . . . . . . . . . . . . . . . . „ $\ \ 919^\circ$ C.

3) \(T = 1220^\circ\) . . . . . . . . . . . . . . \(t = 1029^\circ\) C.
4) „ \(1200^\circ\) . . . . . . . . . . . . . . „ \(1000^\circ\) „
5) „ \(1164^\circ\) . . . . . . . . . . . . . . „ \(964^\circ\) „
6) „ \(1283^\circ\) . . . . . . . . . . . . . . „ \(1055^\circ\) „
7) „ \(1308^\circ\) . . . . . . . . . . . . . . „ \(1035^\circ\) „

From this table it is evident that the temperature at the end of rolling for the rails observed ranged from \(738^\circ\) to \(1055^\circ\). Only one rail was finished by rolling at \(738^\circ\), but according to Mr. Efron’s testimony, to finish rolling a hard rail at so low a temperature is extremely difficult for the mill. The normal temperature for the end of rolling should be taken as \(850\)—\(900^\circ\) C. The indicated temperature at the beginning and end of rolling is such that a change in the grains of iron at it is quite possible, but this is hindered by the mechanical working of the metal. To illustrate the latter proposition let us turn to the analysis of Fig. 6 of Plate XXIII. In this drawing there is graphically represented the general law, considered by us earlier, of the dependence of the magnitude of the grain on the heating temperature, with the magnitudes of the grains plotted on the axis of abscissas, and the corresponding temperatures on the axis of ordinates. Let us suppose that, having begun rolling at \(1400^\circ\) C, by mechanical working we have brought the grain size to \(OP\), corresponding to temperature \(B\). If after this the iron under observation begins to cool quietly, then the grain size, in accordance with the propositions formulated earlier, will reach the normal value \(OE\) along the cooling curve \(BCN\). But in reality we do not allow the metal to cool quietly to temperature \(NE\), but continue the rolling of the metal further, from some average temperature \(C\); in this case, under the influence of the mechanical forces of rolling, the grain size of the metal again decreases to a certain limit equal to the abscissa of point \(D\). Then, in the interval of time of rolling between two successive passes, the grain size of the metal under observation, under conditions of slow cooling, will grow to the abscissa of point \(K\). After this the continuation of rolling will begin again, and, finally, rolling will be finished at some temperature \(G\). After this temperature the metal will begin to cool slowly, whereby its grains will begin to grow to the limit of normal ones along \(GT\), to a grain size equal to \(OP\). Thus, under the condition of the existence of mechanical rolling, the previously considered law of the change in grain size with temperature for the case under consideration will be expressed by the broken line \(BCDKMOPqrSGT\). If the observed specimen after the last pass is immediately cooled rapidly from temperature \(G\), then its grain size will remain equal to \(OH\) and will not reach the value \(OP\). If, however, upon leaving the last pass, the grain size of the given iron is greater than the normal one corresponding to the temperature of this pass, then with further quiet cooling, the mag-

the grain size will not diminish to the normal one, but will retain its value. The ideal case of rolling will be that in which the grain size of the metal at the moment rolling is completed is equal to the normal size corresponding to the temperature at which rolling is completed (point G), but in actual rolling this grain size can be obtained only after annealing to temperature G, whereas under ordinary rolling conditions the final grain size for the indicated finishing temperature of rolling will be equal to OP.

In practice there are cases in which steel after rolling has a grain size greater than normal. Thus Howe²) cites an example where steel (0.5% C), heated to 1394°, then slowly cooled to 963° and rolled at this temperature, had a coarser grain in comparison with the same steel heated to 1339° and annealed at 966° C; moreover, the cooling conditions following the rolling and annealing were identical.

For most cases of hot working of iron, the well-known proposition of D. K. Chernov and Sauveur that “the grain size does not depend on the degree of mechanical working” is entirely applicable. This proposition is indisputable in the case when the degree of mechanical working is such that the grains of iron upon completion of the mechanical working are smaller than normal; then, under this condition, the grains, in accordance with the dependencies indicated earlier, will attain normal dimensions. But if the grain, upon completion of working, is larger than normal, it will remain so in the metal in the cold state and thereby impair its mechanical properties.

For a more detailed clarification of the question of changes in the structure of iron during rolling, with the very kind assistance of the administration of the Alexandrovsk Plant of the Bryansk Company in Ekaterinoslav, we took several specimens from the rolling of wire through passes and of various grades of strip iron; the specimens were taken from one and the same piece successively as it passed through the passes. A detailed study of the changes in structure during the successive passage of rolling through the passes led to the conclusion that the picture of such structural changes in the different specimens is almost identical for all the cases observed, and therefore, to illustrate this question, we shall examine the picture of changes in the structure of soft iron during the successive rolling of wire through 17 passes. The successive change in the profile of the wire rolling under consideration is shown schematically in drawing 7 of Plate XXV. From a specimen of each caliber, two polished sections were made, one transverse and the other longitudinal to the rolling. Photographs of the prepared specimens were successively made—

²) Iron Steel and other Alloys. Pp. 255 and 265.
A. L. Baboshin. Z. I. R. T. O. 1907. Issues IX–X.

are shown at a magnification of 100, in Figs. 36, Plate VIII, through 69, Plate XIII, inclusive; the even numbers depict microsections taken across the rolling direction, and the odd numbers—along the rolling direction.

A consecutive examination of the microsections leads us to the following conclusions: in the first four passes (Figs. 36 to 43) the metal is in a state of particular ductility, and the outline of the grains, owing to the high rolling temperature, is not always distinct in the microsections. In the subsequent passes up to and including the 17th, the iron has a clearly granular structure; the rolls of the rolling mills act strongly upon these grains, drawing them out into regular rows in the direction of rolling. Especially characteristic microsections for determining the correction of rolling are Figs. 47, 61, 59, and 65.

In the first four passes, during the rolling of the wire observed by us, the grains of the metal itself were deformed very slightly, while at first the voids and slags among the metal grains were compressed. The weak deformation of the metal grains may explain that general disorder of the grains which we observe in the microsections of the first four passes, and at the same time this fully confirms the words of D. K. Chernov that the working of iron at welding heat does not affect its structural constitution. On the basis of the microsections obtained and the photographs from them, a count was made of the number of grains in the transverse sections of the microsections by passes, and the resulting diagram of such a count is shown in Drawing 8, Plate XXIV.

From an examination of this diagram one sees a very sharp change in the area of the grain on passing from the 9th to the 10th pass. It must be supposed that at the temperature of these passes the interval of critical temperatures for the given metal had ended, as a result of which, under the influence of mechanical working, the grains of the metal began vigorously to reduce their area.

In passing from the 3rd to the 9th pass, the area of the grains changes very slightly, and thereby the previously stated proposition of D. K. Chernov is indirectly confirmed, concerning the very slight change of grains at high temperatures. Conversely, at low temperatures, during rolling in the last passes, a strong change of the grains is observed, with a strong stratification of the latter.

On examining microsections of longitudinal and transverse sections, a considerable nonuniformity of the grain was discovered. Almost everywhere, the grain area at the edge of the microsection for the first passes is less than the same area in the middle; and for the subsequent passes it is the reverse—the grain area at the edge of the microsection is greater than in the middle. The cause of such a phenomenon for the first passes may be explained by the influence of the cooling of the iron by the rolls. At the beginning of rolling, under the influence of rapid cooling, grains are formed on the surface of the wire being rolled with that grain size which they acquired after mechanical working, not having had time to grow to the normal size owing to rapid

...of cooling. Between these the grains in the middle of the rolled pass, under the influence of the high temperature in the intervals of time between the rolling passes, have time to increase in size, following the law stated earlier concerning the normal grain size for each temperature. In the last rolling passes, however, where the rolling temperature will be low, the opposite phenomenon is observed: the size of the grains on the surface of the section (at the edge) is somewhat greater than that in the middle of the section. The latter can be explained by the fact that in the last passes, owing to the low temperature, the effects of mechanical working show themselves most distinctly in the middle of the pass, namely: the grains on the surface of the metal, because of their cooling by the rolls, had formed earlier; then in the subsequent pass they undergo mechanical working of such a character as increases their ability to resist mechanical stresses (equivalent to cold working), as a result of which they preserve the size previously acquired by them. At the same time the grains of the middle of the pass, under the influence of the mechanical drawing by the rolls, are strongly broken up and reduced in size. For a more detailed elucidation of the question of the degree of nonuniformity of the grain over the area of the pass and from pass to pass, polished sections were prepared from other wire-rolling samples of the same diameter. The polished-section specimens were taken across and along the rolling direction, and photographic images were made from the middle and from the edge of each polished section. Thus 68 images were made, with a total magnification equal to 100. To determine the size, the number of grains in an area of 300 mm² was counted. On the basis of such a count the diagram, drawing 9, table XXVII, was constructed, on which the broken line Ac refers to the middle of the observed polished sections, and Ak—to the edge of the transverse sections of the passes; the broken line Ac^x refers to the middle, and Ak^x—to the edge of the longitudinal sections of the same passes.

Examining the broken lines Ac and Ak, we note that, with the exception of the first pass, the line Ak for all the remaining passes lies above Ac, i.e. the grains at the edges of the polished sections of the transverse sections are larger than the grains in the middle. From the character of the lines Ac^x and Ak^x for the longitudinal sections we note that in the first pass the grain size in the middle is somewhat greater than at the edges of the polished section; in the second pass this inequality is smoothed out, and then, up to the 7th pass, the grain size in the middle is smaller than at the edge. In the seventh pass the grain size in the middle increases in comparison with that at the edge; in the ninth both sizes almost coincide; then, up to the 16th pass, the grain size at the edges is greater than in the middle, while in the 17th pass it is the reverse. The influence of high temperatures and rapid cooling on the structure of iron in the case of wire rolling under consideration is, with a small exception, rather distinctly confirmed by the character of the examined lines relating to rolling in the first passes; the influence of mechanical

of drawing on the structure of the wire is distinctly confirmed in the last passes (beginning with the 10th). The deviation of the last, 17th pass, in the sense that the size of the grains in the middle proved greater than that at the edges, must be explained by the considerable crushing of the grain in the last pass in the middle, with the formation of voids between the grains, which, when counted, only increase the area of the observed grains.

In the following diagram (Fig. 10, Plate XXVII), in the form of lines analogous to the preceding diagram, are shown the mean values of the grain size for transverse sections from the middle and the edge of the pass (solid line), and for longitudinal sections—likewise from the middle and edge of each pass (dashed line); and then, on the basis of the two preceding lines, a resultant line has been constructed (dashed line with dots), expressing the general character of the changes in grain size during rolling of the wire, depending on the sequence of alternating passes. On the basis of the last line of the diagram under consideration, we come to the conclusion that the change in the grains during rolling, in the sense of a decrease in their size, is only slightly noticeable in the first passes, owing to the considerable rolling temperature. Further: in the present case, beginning with the 7th pass and up to the end of rolling, there is observed a general decrease in the grains of the metal in passing from one pass to the next; moreover, in the last passes, owing to the strong influence of the mechanical working itself on the grain size of the metal being worked—which promotes preservation of the size of the surface grains and, on the contrary, strongly diminishes the size of the grains at the middle of the section—no markedly strong decreases in grain size are noticeable. It must also be noted that the grain size in the longitudinal sections does not increase according to the passes of rolling; consequently, it must be assumed that at the beginning of rolling the grains increase slightly in length, and then, under the influence of mechanical forces and strong lateral pressure, they are divided into parts, with a preliminary transitional formation of necks (constrictions) in the grains, of which Prof. Neup speaks³. In the ingots from the rolling of strip iron that were observed, the character of the change in grain structure was entirely analogous to the preceding one, and the character of the lines expressing this dependence was close to the last passes of wire rolling. In thin sections of strip iron (\(1/8'' \times 1 1/2''\)) longitudinal cracks in the direction of rolling were very noticeable. The latter can be explained by the unequal amount of drawing of the upper and lower parts of the section of the profile under consideration, owing to improper operation of the rolls.

From the preceding we have noted that the grain size of iron does not remain the same even over the entire section of one and the same bar, but, depending on the temperature and the mechanical conditions of rolling, changes greatly in magnitude. Evidently, the more pronounced this difference in grain size is, the more complex the profile of the specimen being made. The general picture of changes in the size of the metal grains along the bars remains also in the finished specimen, provided only that the latter is not subjected to annealing after rolling. Such a difference in grain size, depending on the temperature and rolling conditions over the sectional area of the specimen, exists in all grades of iron that have come from rolling without annealing; and, to illustrate the latter proposition, in the Metallographic Laboratory of the Tomsk Technological Institute the profiles of ordinary rails were investigated. For this purpose, from the rail being tested there was first prepared a general macrostructure for judging the general structure of the rail, and then the rail was divided into belts; the belts were divided into parts, from which sections were prepared transverse to the rolling. The sections were photographed at a magnification of 120, and on a circle of such photographs with a diameter of 7.2 cm the number of grains was determined. Such a division into parts of one of these rail sections is shown in Drawing 11, Table XXVIII, and the results of the overall count of the number of grains over the rail sections are represented by broken lines 1 to VI inclusive in Table XXIX. From examination of the first broken line (the number of grains is indicated on the ordinates, and the figures on the abscissae designate the numbers of the sections along the rail section) it is clearly seen that the grain size in the middle of the rail head became considerably larger (almost twice as large) than that at the rail edges. In the transition from the rail head to the neck, owing to the considerable decrease in the rail section and the rapid cooling of the metal associated with this, the difference in grain size between the middle and the edges is smoothed out. In the section of the neck there is a noticeable increase of the grains to one (right) side, which is explained entirely by the nonuniformity of cooling of the rail after rolling. This last proposition is fully confirmed by the control sections made from specimens taken from the left and right sides at the extreme points, the number of grains in the photographs of which is shown by the figures in Fig. 11, Table XXVIII.

From comparison of the latter figures of grains on the sections it is clearly seen that the right side of the rail after rolling cooled more rapidly than the left, as a result of which the grain size at the edge is smaller than on the left side. Diagrams IV and V clearly illustrate the proposition that the grain size of iron in the larger sections of the observed specimen is smaller at the edges than in the middle. Then diagram VI shows a general decrease in grain size from the head to the base of the rail, which was quite natural to expect, since the process of cooling of the metal and the action of mechanical forces during rolling most strongly affect the grain size in the smaller sections of the rail, at

of its base. Thus, briefly formulating the investigation of the structure of rails by section just presented, we arrive at the following conclusions: the head of the rail possesses a rather coarse crystallization, while in its middle it is fairly homogeneous. The neck of the rail has a finer structure in comparison with the head, but the decrease in grain size toward the right side of the section is noticeable, which is fully confirmed by the investigation of the outer lateral points and supports our supposition regarding the nonuniform cooling of the rail after rolling on the mill tables. The base of the rail has a finer structure in comparison with the head, which is explained by the influence of rapid cooling and by the greater influence of mechanical working during rolling. The mechanical working during rolling was most sharply reflected at the outer points of the rail, causing fine fragmentation of the grain.

In the preceding section we dwelt at considerable length on the consideration of the question of the change in the grains of iron during rolling, taking, for a detailed elucidation of this question, specimens from a single piece of rolled wire rod. We shall now consider further how the structure of the wire rod changes during subsequent working—drawing. The specimen for drawing was taken from one coil of wire according to the measure of its passage during drawing through the die, but, owing to practical inconveniences, only one annealing of the wire in the middle of the drawing was omitted. The chemical composition of the last wire is approximately the same as the chemical composition of the wire that was examined by us during rolling⁴. Photographs of the specimens studied are shown, at a magnification of 150, in Figs. 70 of Plate XIII through 103 of Plate XIX inclusive, the odd numbers referring to the longitudinal sections of the specimens investigated, and the even numbers—to the transverse sections.

In order better to understand the picture given by the microscope, we must consider the methods of mechanical working to which our specimens were subjected.

The material for drawing, the structure of which we see in the microsections of Figs. 70 and 71, is supplied to the plant and is a product of rolling, which is also noticeable from the remaining voids along the axis. At the plant, before being put into the drawing plate, the material is annealed. Drawing plates are metal plates provided with holes of various sizes. A piece of wire is threaded into these holes, then grasped on the other side of the plate with special tongs, by which the wire is pulled. But since the cross-section of the hole in the drawing plate is smaller than the section of the wire being worked, the latter is compressed and thinned, assuming a section in form and size

approximately coinciding with the hole of the drawplate. We say “approximately,” since upon emerging from the eye of the drawplate, owing to the cessation of the compressive force on one side and to the elasticity of the material on the other, the wire thickens somewhat. This is proved by the fact that if a wire that has just been drawn is pushed again into the same eye and drawing is begun again, the force will be greater than that which would be required for friction against the walls of the eye. It has been found experimentally that for iron, for a second drawing, \(^{10}/_{22}\) of the initial force is required. In drawing, a large role is played by the shape of the eye itself and the condition of its surface. The simplest form of an eye is a cylindrical opening, but at the same time it is also the most unsuitable, since all the work of compressing and compacting the wire is performed only by the section of the eye in the plane of the drawplate. Such an instantaneous change in the cross-section of the incoming wire may serve as the cause of shearing and biting through. Owing to the material being, as it were, scraped by the outer edge of such an eye and to the absence of participation in drawing by the rest of the section, this edge becomes chipped and the drawplate becomes completely unfit for use. The most perfect form is an eye consisting of two truncated cones (Fig. 12a, Plate XXVIII), joined by their smaller bases, the middle edge being rounded. Here there is no abrupt change in the cross-section and no scraping of the material. The manufacture of such an eye and the necessary careful polishing of it present great difficulties, which is why it is used only for such varieties of wire made of precious metals. The usual form of the eye is the cylindro-conical form with a rounding of the middle edge (Fig. 12b, Plate XXVIII). Thus, the part of the wire between the tongs and the drawplate is stretched, while the part passing through the drawing opening is compressed, the material at the surface being brought into a plastic state and experiencing a stress corresponding to the flow pressure. Let \(d\) and \(d_1\) denote the diameter of the wire before and after drawing, \(K_z\) the stress of the stretched part, \(\delta\) the pressure on the surface of the cone of the drawing opening; and \(f\) the coefficient of friction between the wire and the drawplate (see drawing 13, Plate XXVIII). From the condition of equilibrium we may write the following equality:

\[ \frac{\pi d_1^2}{4}\cdot K_z = Z\,dF\cdot \delta \cdot \operatorname{Sin}\alpha + Z\,dF\cdot \delta \cdot f\cdot \operatorname{Cos}\alpha; \]

but \(Z\,dF\) is the lateral surface of the truncated cone; consequently:

\[ Z\,dF = \frac{d+d_1}{2}\,\pi\cdot a\,b = \frac{d+d_1}{4}\,\pi\cdot \frac{d-d_1}{2\operatorname{Sin}\alpha} = \frac{\pi}{4}\cdot \frac{d^2-d_1^2}{\operatorname{Sin}\alpha}. \]

(where \(ab\) is the generatrix of the cone).

Substituting this value, we obtain:

\[ \frac{\pi d_1^2}{4}\cdot K_z=\frac{\pi}{4}(d^2-d_1^2)\cdot \delta \cdot (1+f\cdot \operatorname{ctg}\alpha), \]
whence

\[ K_z=\left[\left(\frac{d}{d_1}\right)^2-1\right]\cdot \delta\cdot (1+f\cdot \operatorname{ctg}\alpha); \]

the stress \(\delta\) may be taken as \(40\ \mathrm{klgr.\ m/m}\); the angle \(\alpha=15^\circ\), the coefficient of friction \(f=0.1\).

To reduce the friction of the metal against the walls of the die opening, the wire is lubricated with tallow, oil, or soapy water.

The change in the diameters of the wire under investigation is as follows:

2) \(7.12\ \mathrm{m/m}\).	7) \(2.62\ \mathrm{m/m}\).	12) \(1.33\ \mathrm{m/m}\).
3) \(6.06\)	8) \(2.15\)	13) \(1.14\)
4) \(5.02\)	9) \(1.98\)	14) \(1.00\)
5) \(4.08\)	10) \(1.76\)	15) \(0.82\)
6) \(3.26\)	11) \(1.58\)	16) \(0.75\)

The change in the ratio of the diameters between the 2nd and 3rd dies:

2 and 3 \(=1.18\)	6 and 7 \(=1.27\)	10 and 11 \(=1.11\)
3 and 4 \(=1.2\)	7 and 8 \(=1.22\)	11 and 12 \(=1.18\)
4 and 5 \(=1.23\)	8 and 9 \(=1.08\)	12 and 13 \(=1.17\)
5 and 6 \(=1.23\)	9 and 10 \(=1.13\)	13 and 14 \(=1.14\)
14 and 15 \(=1.22\)	15 and 16 \(=1.09\)

The tensile stress experienced by the material of the wire in passing from \(d\) to \(d_1\) is obtained approximately as:

2 and 3 \(K_z=(1.18^2-1)\cdot 40\cdot(1+0.37)=(1.18^2-1)\cdot 54.8=21.4\ \mathrm{klgr.}\)
3 and 4 „ \((1.2^2-1)\cdot 54.8=24.1\)	9 and 10 \(K_z=(1.13^2-1)\cdot 54.8=15\)
4 and 5 „ \((1.23^2-1)\cdot 54.8=28\)	10 and 11 „ \((1.11^2-1)\cdot 54.8=12.6\)
5 and 6 „ \((1.25^2-1)\cdot 54.8=30.4\)	11 and 12 „ \((1.18^2-1)\cdot 54.8=21.4\)
6 and 7 „ \((1.24^2-1)\cdot 54.8=28.4\)	12 and 13 „ \((1.17^2-1)\cdot 54.8=20.2\)
7 and 8 „ \((1.22^2-1)\cdot 54.8=26.8\)	13 and 14 „ \((1.14^2-1)\cdot 54.8=16.4\)
8 and 9 „ \((1.08^2-1)\cdot 54.8=8.8\)	14 and 15 „ \((1.22^2-1)\cdot 54.8=26.3\)
	15 and 16 \(K_z=(1.09^2-1)\cdot 54.8=10.1\ \mathrm{klgr.}\)

Such an approximate calculation shows that the stress experienced by the material of the wire exceeds, in the majority of cases, the elastic limit. Let us turn to the experiments carried out by Prof. Hein (Z. d. V. J. 1906, No. 14). For the experiment he took wire rolled from soft cast iron, consisting exclusively—

but from ferrite grains, the size of which was the same both in the longitudinal and in the transverse section. The specimen was torn apart in a testing machine. It turned out that near the place of fracture the grains had become elongated, while the transverse section of the grains had decreased and the length of the grains had increased in comparison with their original dimensions. As a result the volume of the grains decreased. Wire drawn through a drawplate to 3.7 mm in diameter had elongated grains, with the transverse section decreased, while the length remained unchanged (slightly decreased). Wishing to give an idea of the dimensions of the grains in the different directions, Prof. Hein represents the grains in the form of a prism with a square base, whose area is equal to the area of the transverse section of the grain, and the length of the prism is taken such that the volume of the prism is equal to the volume of the grain. The dimensions of the grains, measured in thousandths of a millimeter, were as follows:

	Side of the square base in H	Length of the prism	Volume in H³
1)	41.5	41.5	68890
2)	14.4	77.1	16194
3)	19.0	34.5	12178

In both cases the grains decreased in their dimensions. Their division occurred, but for different reasons. During rupture the grains were stretched out; when a certain limit was passed, they broke into parts. The length of the grains increased almost twofold. During drawing the grains were subjected chiefly to transverse compression. The division of the grains occurred because the material, being compressed unevenly in different places, first formed necks, and then one part of the grain was finally separated from another. Turning to the microsections prepared by us, we observe the following picture. Microsection, Figs. 70 and 71, Plate XIII. Wire obtained from the factory, but drawn through a die in order to give the section a round form; preliminarily annealed. The shape of the ferrite grains is almost the same both in the longitudinal and in the transverse section. The increase in the size of the ferrite grains in comparison with the microsection, Figs. 72 and 73, Plate XIV, indicates a rather high annealing temperature. (The latter microsections refer to wire obtained from the factory.) The voids remaining after rolling along the longitudinal axis became somewhat smaller in size. Some densification of the material occurred. Microsection, Figs. 74 and 75, Plate XIV. The elongation of the ferrite grains, already having a smaller size in comparison with the microsection, Fig. 71, is still not so sharply expressed. The presence of voids becomes less noticeable. Microsection, Figs. 76 and 77, Plate XIV. Here the elongation of the grains and the reduction of their size in the transverse section already come sharply to the fore; the form of the latter deviates strongly from the outline of the grains of the microsection, Fig. 74. The ferrite grains

CHANGE OF THE STRUCTURE OF IRON DURING MECHANICAL WORKING.

can no longer retain a more or less regular shape, owing to the strong lateral pressure experienced as they pass through the aperture of the drawplate. Such a strong change in the shape of the ferrite grains is explained by the rather considerable tensile stress in the wire as it passes through this opening. According to Osmond, iron may exist in three allotropic states, \(\gamma\), \(\beta\), and \(\alpha\). The modifications of iron \(\gamma\) and \(\beta\) have no magnetic properties, whereas the modification of iron \(\alpha\) is magnetic. The transition of iron \(\gamma\) into \(\beta\) takes place at about \(900^\circ\). The modification \(\beta\), containing a larger quantity of carbon, is harder and more brittle in comparison with \(\alpha\). The difference in hardness of the ferrite grains becomes noticeable beginning with the polished sections of Figs. 78 and 79, Plate XV. Whereas the grains of ferrite \(\alpha\) are more elongated and thinner, the grains of \(\beta\) tend to preserve their size. The length of the \(\beta\) grains is smaller, and they are much thicker, and they stand out rather sharply against the general background owing to the roughness of their surface. Etching was carried out with Kurbatov’s reagent (a 4% solution of nitric acid in amyl alcohol). Nitric acid acts the more strongly the harder and less pure the metal is. Therefore the \(\beta\) grains also turned out to be rougher. If we turn to the transverse section, we shall see the sinuous form of the section of the grains, and here too the difference between the \(\alpha\) and \(\beta\) grains is quite noticeable. The strong transverse compression of the material exerts its influence, and it becomes difficult to separate one grain from another. The difference between \(\alpha\) and \(\beta\) is observed still more sharply in the subsequent polished sections, Figs. 80 and 81, Plate XV, and further. The size of the grains decreases in the transverse section, and their length greatly increases, but the \(\beta\) grains have a larger transverse-section size in relation to their length than \(\alpha\). The \(\beta\) grains, while being preserved during stretching and compression, begin to suffer, owing to their brittleness, during sliding. When the wire is drawn through the aperture, strong friction is produced between the lateral surface of the wire and the aperture. The force of friction causes sliding of the material lying closer to the surface relative to the middle part. If attention is paid to the end of the drawn wire, an elongated surface is always seen, which clearly shows the displacement of the material in the central part relative to the surface. Considering transverse sections, we encounter the following circumstance. The size of the grains changes noticeably from the periphery to the center. For greater clarity in the difference in the size of the grains lying near the periphery and the center, photographs of polished sections were made at a magnification of \(800 : 1\). See Figs. 88 and 89, Plate XVII. One photograph shows the sizes of the grains near the surface, the other in the middle part. Let us try to clarify somewhat the cause of such a phenomenon. Since the grains turned out to be smaller in the middle, it means that the metal in the middle part was more elongated in comparison with the outer part and, experiencing lateral pressure, proved to be more compressed. As we have already said, the wire during pass—

when passing through the drawplate hole experiences strong friction. The work of the frictional force is spent on heating. This phenomenon is very noticeable; workers are sometimes even forced to use gloves. Let us approximately calculate the amount of heat that will be released due to the frictional force. In doing so we shall make the assumption that all the released heat goes into heating the wire. The eye angle $\alpha$ is about $15^\circ$. The compressive pressure is $40\ \mathrm{klg./mm.^2}$. Taking into account the lubrication of the wire before the drawplate hole, let us take the coefficient of friction to be $0.1$. The speed at which the wire passes, according to the data of Knabe, “Manufacture of Wire,” shall be taken as follows. The lateral surface is

\[ = \frac{\pi}{4}\cdot \frac{d^2-d_1^2}{\operatorname{Sin}\alpha};\quad \operatorname{Sin}\alpha = 0{,}259;\quad \frac{\pi}{4}\operatorname{Sin}\alpha = 3{,}04; \]

	Wire diam. in m/m.	Lateral surface, in m/m.$^2$	Friction force in klgr.	Speed in mtr. in 1″	Work in klgr. mtr.
2)	7,12	$(50{,}7-36{,}7)\cdot 3{,}04=426$	170,4	0,255	43,5
3)	6,06	$(36{,}7-25{,}2)\cdot 3{,}04=38$	152	0,3	45,6
4)	5,02	$(25{,}2-16{,}7)\cdot 3{,}04=25{,}8$	123,2	0,3	36,9
5)	4,08	$(16{,}7-10{,}6)\cdot 3{,}04=18{,}5$	74	0,45	33,3
6)	3,26	$(10{,}6-6{,}9)\cdot 3{,}04=11{,}3$	45,2	0,6	27,1
7)	2,62	$(6{,}9-4{,}6)\cdot 3{,}04=7{,}0$	28	0,65	16,8
8)	2,15	$(4{,}6-3{,}9)\cdot 3{,}04=2{,}1$	8,4	0,7	5,8
9)	1,98	$(3{,}9-3{,}1)\cdot 3{,}04=2{,}4$	9,6	0,75	6,7
10)	1,76	$(3{,}1-2{,}4)\cdot 3{,}04=2{,}1$	8,4	0,8	6,7
11)	1,58	$(2{,}4-1{,}76)\cdot 3{,}04=1{,}94$	7,7	0,9	6,9
12)	1,33	$(1{,}76-1{,}3)\cdot 3{,}04=1{,}4$	5,6	1,05	5,6
13)	1,14	$(1{,}3-1)\cdot 3{,}04=0{,}9$	3,6	1,2	4,3
14)	1	$(1-0{,}67)\cdot 3{,}04=1{,}01$	4	1,3	5,2
15)	0,82	$(0{,}67-0{,}56)\cdot 3{,}04=0{,}33$	1,3	1,4	1,8

If we make the assumption that all the heat goes into uniform heating of the wire over its cross-section, then we obtain approximately the following rise in temperature $r$ for each case:

	Volume of wire in cm.$^3$	Weight in klgr.	Amount of heat.	Rise in temperature.
3)	$0{,}288\cdot 25=7{,}2$	0,0554	0,1	$0{,}0554\cdot 11\cdot r=0{,}1;\ r=16{,}5$
4)	$0{,}198\cdot 30=4{,}94$	0,038	0,106	$0{,}038\cdot 0{,}11\cdot r=0{,}106;\ r=25{,}3$
5)	$0{,}13\cdot 30=3{,}9$	0,03	0,086	$0{,}03\cdot 0{,}11\cdot r=0{,}086;\ r=26$
6)	$0{,}0835\cdot 45=3{,}76$	0,0289	0,078	$0{,}0289\cdot 0{,}11\cdot r=0{,}078;\ r=24{,}8$
7)	$0{,}0539\cdot 60=3{,}23$	0,0279	0,063	$0{,}0279\cdot 0{,}11\cdot r=0{,}063;\ r=23$
8)	$0{,}0363\cdot 60=2{,}36$	0,0182	0,039	$0{,}0182\cdot 0{,}11\cdot r=0{,}039;\ r=19{,}5$
9)	$0{,}0307\cdot 70=2{,}15$	0,0166	0,013	$0{,}0166\cdot 0{,}11\cdot r=0{,}013;\ r=7{,}1$

Wire volume in cm³	Weight in klgr.	Quantity of heat.	Temperature increase.
10) 0.0243 · 75 = 1.84	0.0142	0.015	0.0142 · 0.11 · r = 0.015; r = 9.6
11) 0.0196 · 80 = 1.56	0.012	0.015	0.012 · 0.11 · r = 0.015; r = 11.3
12) 0.0139 · 90 = 1.25	0.0096	0.016	0.0096 · 0.11 · r = 0.016; r = 15
13) 0.0102 · 105 = 1.07	0.0082	0.013	0.0082 · 0.11 · r = 0.013; r = 12.7
14) 0.0078 · 120 = 0.94	0.0074	0.010	0.0074 · 0.11 · r = 0.010; r = 12.3
15) 0.0052 · 130 = 0.68	0.0052	0.012	0.0052 · 0.11 · r = 0.012; r = 21
16) 0.0044 · 140 = 0.62	0.0048	0.005	0.0048 · 0.11 · r = 0.005; r = 25

Such an approximate calculation of the temperature rise has been made on the assumption of uniform heating of the wire, and consequently of instantaneous propagation of heat. Such an assumption in reality cannot hold at such a rather considerable speed of drawing of the wire. If it is even assumed that the heat, on passing through the draw-plate opening, will be distributed according to a straight-line law from a maximum at the periphery to 0 near the center, then the greatest rise of temperature near the surface is obtained from the following relation:

\[ {}^{2}/_{3}\, R \cdot \frac{\pi D^{2}}{4} = r \frac{\pi D^{2}}{4}; \quad \text{whence} \quad R = {}^{3}/_{2} r \]

or, according to the calculation, the maximum rise is \(26 \cdot {}^{3}/_{2} = 39^\circ\). Let us now see how the properties of iron change with respect to resistance to tensile stresses upon heating. For this we shall use the diagram, given by Bach, of the influence of temperature on the tensile strength of iron. See Fig. 14, Table XXVIII. The most illustrative for the present case is the curve of decrease of cross-section \(\psi\) and the curve of elongation \(\varphi\).

The curve \(\psi\) of the change in cross-section under tension shows that the viscosity of iron falls sharply with increasing temperature. The maximum change in cross-section corresponds to about \(300^\circ\), which gives the blue-temper temperature at which iron becomes brittle. The minimum \(\varphi\) corresponds to a temperature of \(160^\circ\). The curves correspond to experiments performed on bars of cast iron, which had these properties at room temperature.

Tensile strength.	Tensile strength.	Elongation per 100 mm.	Elongation per 100 mm.	Transverse contraction.	Transverse contraction.
In the original form.	In annealed form.	In the original form.	In annealed form.	In the original form.	In annealed form.
4100 klgr.	3840	26.9%	30.4%	48.3%	58.6%

If the initial temperature of the wire is taken as \(20^\circ\), and the increase at the surface layer as \(40^\circ\), then we obtain a decrease in elongation

for the outer ones by as much as 25.7%, and a reduction of viscosity by 6.5%. It may be supposed that the change in the resistance of iron as a function of heating also serves as the cause of the unequal change in the size of the grains. Since the material in the middle part is more viscous and is capable of stretching more strongly, undoubtedly the middle fibers of the wire, when passing through the drawing-hole, are deformed more than the outer ones, which are held back by the force of friction. The slip that arises as a result of this also affects the grains β in such a way that a stratification of the material appears. If we turn our attention to the photographs, we shall easily notice that the observed stratification of the material is always located around the grains β. At first a weak lamination is observed along the longitudinal axis of the grain. With further drawing, the destruction of the grain itself becomes noticeable. A weak cracking of the grain β passes into a large longitudinal lamination. The transverse dimension of this lamination is somewhat reduced in the subsequent passes as a result of the considerable transverse compression of the material when it passes through the die. Thus, the pictures of our polished sections give us the right to make the following conclusions concerning the change in the structure of iron during drawing.

1) In passing through the drawing-hole, the magnitude of the transverse section of the grain decreases. The presence of high tensile stress influences the elongation of the grains. Strong transverse compression of the material causes the division of the grains along their length, first forming necks.

2) The sizes of the grains are not the same throughout the entire transverse section. Their magnitude increases from the middle part toward the periphery, owing to the difference in resistance.

3) The different resistance of the grains and the force of friction against the die cause the inner fibers to slip relative to the outer ones, serving as the cause of stratification of the material. After several draws in the manufacture of wire, which becomes hard and begins to tear, it is annealed in order to change the fibrous structure to a granular one and thereby improve the quality of the material. To show how the structure of the material changes during annealing, we have presented specimens of annealed wire and of wire etched to remove scale. Figs. 104, 105, Table XX; the first is annealed. The size of the grains depends on the annealing temperature. At first the grains divide along their length, and then, only at a higher temperature of about 900°, larger grains may be formed. In this connection it should be noted that, if the annealing temperature was not sufficiently high, then the difference in the size of the central grains and those close to the periphery is also observed after annealing, as, for example, this is especially sharply noticeable at the end of the wire, polished section fig. 104. We did not succeed in tracing the influence of etching on the formation of grains.

The desire to reduce costs in mass production and the need to have identical dimensions for the articles being manufactured led to stamping. Stamping may be defined as giving a material the required form by means of a die, by pressure. Hence it is assumed that the material must possess plasticity under compression, but quite often, in order to reduce the pressure, the work is carried out at a high heating temperature.

As to how the resistance of iron to compression changes, one may cite Daelen’s experiments. To immerse a die into a billet to one and the same definite depth, the following pressure was required:

Condition	Pressure
At yellow heat	5.14 kg on 1 m/m. of billet
At orange heat	5.99 kg on 1 m/m. of billet
At red heat	6.31 kg on 1 m/m. of billet
At dark-red heat	14.90 kg on 1 m/m. of billet
At brown heat	16.40 kg on 1 m/m. of billet
At black heat	20.80 kg on 1 m/m. of billet

Stamping of heated material is analogous to forging, only the corresponding form is obtained by means of a recess in the die, which is filled by the material owing to impact loading or smooth pressure. In stamping by impact, the live force of the striking device is utilized, and it is possible to use steam, air, or a simple drive to set the hammer in motion; the latter is used quite often in small shops, where power hammers or even lever hammers are installed. Thus small articles are produced in large quantities, for example bolts, braces, rivets, and so forth. The magnitude and form of the ferrite grains are strongly influenced by the temperature at which the working is completed. If forging is finished below the point \(Ar_1\), or close to it, then during the entire time of their formation the ferrite grains are subjected to working, and owing to the blows and shaking the structure becomes fine-grained. This explains why smiths continue to strike the article being forged until it has cooled below the point \(Ar_1\). If the work is finished at a fairly high temperature, then the ferrite grains can assume their normal form quite independently. We compare stamping in the hot state with forging, only the number of blows here is reduced to a minimum, for example to one, rarely to two. Here we, of course, do not include stamping in the hot or cold state by smooth pressure, such as, for example, stamping of boiler bottoms, coining, drawing, hydraulic stamping. At our disposal were only specimens of stamping in the hot state—heads of rivets, and in the cold state—nails. For illustration on this question we shall take rivets \(1/4''\) and \(5/8''\) and nails \(1/16''\) and \(1/8''\).

Photographs of parts of rivet heads in longitudinal section (see Figs. 106 and 108, Plate XX) were made along a line serving as a continuation of the extreme generatrix of the cylindrical part of the body. Approximately in the same place, specimens of transverse sections were taken (see Figs. 107 and 109, Plate XX). How the material was deformed during the manufacture of rivet heads can be seen from the form of the interlayers remaining in the body of the rivets. The longitudinal fibers, while remaining parallel in the shank, changed direction in the head. They diverged strongly to the sides at the middle of the head, retaining their direction somewhat only at the very top of the head. The size of the grains changes greatly in different places of the head. At the end of the shank there still remains an untouched spherical part, the grains of which, in comparison with the shank, have changed hardly at all. From this bulge a fibrous transition to the side parts of the head is observed. The outer, superficial part of the head, both the upper part that came into contact with the die and the bearing plane below, have a finer-grained structure. The part between the spherical bulge and the surface layer appears, as it were, flattened. Thus we find that stamping produces crushing and refinement of the material in the parts lying closer to the surface in contact with the die, and compression in the middle part. The outer part of the article, having a fine-grained structure, is denser; this also explains the appearance of a kind of crust on stamped objects. In stamping, in the cold state, the heads of machine nails, which are formed by the impact of a special punch, the picture changes somewhat. Here, too, a distinction is noticeable between the untouched and the deformed part along the spherical surface. The structure of the shank is fibrous, as a result of drawing, especially in thin nails. The photographs in Figs. 110 and 111, Plate XXI, show longitudinal and transverse sections of a nail \(1/8''\) in diameter near the head; similar sections of a nail \(1/16''\) are shown in Figs. 112 (longitudinal) and 113 (transverse) of Plate XXI. The fibers are visible at the beginning of the transition to the side part of the head and then become crushed. This transition is sharply visible on the polished section of the \(1/16''\) nail. In the photograph of the transverse section of the head of this nail, the grains are not uniform. On one side we see elongated grains, on the other small crushed ones. We see the same picture on the polished section of the \(1/8''\) nail. A mass of voids is visible between the grains. The full head, from the outside, does not have a dense structure inside.

For a more detailed clarification of the question of the change in the structure of iron during stamping, let us also consider the picture of similar deformations during hydraulic and hand riveting. For this purpose, specimens of hand and hydraulic riveting were made by experienced workers at one of the large plants in the south of Russia (the Aleksandrovsk Works of the Bryansk Society); Fig. 114, Plate XXI, refers to a 22 mm rivet.

in diameter, with the thickness of the sheets being riveted \(m\) mm. The filling of the seam is very dense; the character of the deformations is very fine, with weak formation of small cracks along the middle of the rivet shank. Fig. 115 of Table XXI relates to hand riveting with a rivet diameter of \(18\,m/m\), and the same thickness of the sheets being riveted. In the latter case the deformation of the rivet takes place under the blow of a hammer, and riveting ends at a lower temperature in comparison with the preceding one, as a result of which the body of the rivet inside has acquired considerable voids (cracks). The last two photographs illustrate very fully the general picture of deformation in hot and cold stamping. The character of the deformations at the head of the rivet, the crutch, the nail, etc., remains as before without substantial changes; only the size of the voids (cracks) changes, which in hot working are smaller, and in cold working considerably wider and deeper. The shank of the stamped object usually preserves the structure which it received during the preceding working (rolling or drawing).

In ordinary smith-forgings the oval form of the grain of soft iron, under the influence of the blows of the hammer, gradually changes its form into a wedge-shaped one, sometimes with strong splitting of the grain itself. No significant difference in the deformation of the grain of iron during smith-forgings was observed; on the contrary, the character of these deformations was very uniform and is shown quite clearly in Fig. 116 of Table XXII. The previously indicated wedge-shaped form of the grain and the crushing of the grain itself under the influence of forging are clearly visible in the latter microsection.

Finally, in order to investigate the question of changes in the structure of iron during cutting and punching of holes, in the mechanical workshops of the Tomsk Technological Institute a considerable number of observations were made on the cutting of iron \(1.5\,m/m\), \(8.5\,m/m\), \(10\,m/m\), \(15\,m/m\), and \(19\,m/m\) thick, and on punching holes in it with a punch of diameter \(7/16''\), \(5/8''\), \(3/4''\), and \(7/8''\). When sheets of iron are cut on ordinary frame shears, strong compaction of the metal occurs at the surface of separation. The strongest deformation of the metal is found at the place where the upper movable blade of the shears enters, and the deformation of the metal gradually diminishes to zero as it approaches the lower surface of the sheet being separated. The greatest magnitude of the indicated deformations when cutting sheets on frame shears is, on average, equal to \(1\,m/m\).

For a detailed elucidation of the question of deformations during the punching of holes, sections of the punched sheets were made, photographs were taken of the general picture of the deformations found, and then, at high magnification, the character of the deformations at the entrance of the punch, its middle position, and its exit was photographed. Studying, from the photographs obtained, the change in the structure of the material, we observe the following picture: at the entrance of the punch into the sheet, the material of the latter is at first compressed, [[unclear: text continues]]

there occurs shearing of the metal along the edge of the punch. The material situated in front of the punch is also strongly compressed, spreading out to the sides. With the further travel of the punch, the material of the sheet seems to flow. On the opposite side of the sheet being pierced, the edge of the die exerts on the sheet the same pressure as the punch, as a result of which the piece being sheared is displaced. The surface of separation is not located exactly in the direction of motion of the punch, but somewhat inclined to this direction, which evidently occurs as a result of the lateral displacement of the compressed particles. As the punch moves, the surface of internal cohesion of the particles diminishes, and the material, no longer being able to resist the acting force, breaks away.

The general character of the deformation of iron grains during punching is shown by the photographs in Figs. 117, 118, 119, and 120 of Table XXII, the first giving the general view of such deformations, and the following three (at a magnification of 100) the character of such deformations at the entrance, the middle, and the exit of the punch. The magnitude of such deformations, as is seen from an examination of the just-cited polished sections, increases strongly toward the exit of the punch, and it was then observed that the magnitude of the deformation increases with the thickness of the sheet. The mean values of the observed deformations as a function of the thickness of the sheet are given in the following table:

Thickness of iron in m/m.	Diameter of the punch in inches	Magnitude of deformation in m/m.
1.5 m/m	7/16″	0.25 m/m.
1.5 m/m.	5/8″	0.25 m/m.
8.5 m/m.	7/16″	0.9 m/m.
8.5 m/m.	3/4″	0.9 m/m.
10 m/m.	5/8″.	1 m/m
15 m/m.	3/4″	1.4 m/m.
15 m/m.	7/8″	1.4 m/m.
19 m/m.	7/8″	2 m/m.

From the table given it is evident that the magnitude of the deformation grows with the thickness of the sheet and on the average fluctuates around \(1/10\) of the thickness of the latter. There is also no doubt that the magnitude of the deformation is influenced by the magnitude of the clearance between the punch and the diameter of the die. To investigate the latter question, a series of experiments was carried out with iron 9 mm thick, in which holes were successively punched with a \(3/4\)″ punch at clearances in the die of \(1/2\), 1, \(1\frac{1}{2}\), and 2 mm; for this purpose the diameter of the punch remained constant, while the diameter of the die was successively increased. From the experiments it was found that the magnitude of the deformation increases with the clearance: thus, for a clearance of

Change in the Structure of Iron During Mechanical Working.

between the punch and the diameter of the die of \(1/2\) and \(1\ \mathrm{m/m}\), the amount of deformation is, on average, equal to \(0.9\ \mathrm{m/m}\); with a clearance of \(1\frac{1}{2}\ \mathrm{m/m}\), the amount of deformation is \(1.1\ \mathrm{m/m}\); with \(2\ \mathrm{m/m}\), \(1.5\ \mathrm{m/m}\). With a clearance of \(1/2\) and \(1\ \mathrm{m/m}\) the edges of the punched sheets are smooth, whereas with a further increase of the clearance the lower edge of the punched hole has large tears (Fig. 120, Table XXII). It is evident that the most advantageous clearance between the diameters of the die and the punch lies within the limits of \(1/2\) and \(1\ \mathrm{m/m}\). To eliminate the deformations obtained, the holes are usually drilled out by the thickness of the deformations, or the sheets are subjected to annealing before riveting. Annealing, as our observations have shown, restores the grains of the disturbed structure, and, most importantly, apparently promotes the separation of the deformed part from the sound part. Thus Fig. 120, Table XXII, quite clearly illustrates this action of annealing on the deformed part after annealing. The shape of the punch itself also has a great influence on the magnitude of the deformation of the metal when holes are punched. To investigate this question, punches of \(1/2''\) diameter were made (Fig. 121, Table XXII), the punch first being made without a central guiding cone, and the angle at the cutting edge of the punch being \(75^\circ\). Such a punch, when pressed through, produced torn holes, passed through the opening with great difficulty, and, from excessive axial pressure, burst diametrically (see Fig. 121). Next a punch was made with a central guiding cone, with a small cutting edge in the form of a ring around the circumference of the punch (the angle of sharpening at this annular edge being \(75^\circ\), and the width of the annular edge at its base \(2.5\ \mathrm{m/m}\)); but the middle part of the punch, on which the guiding cone was located, was flat (see Fig. 121). This punch worked somewhat better than the preceding one; the tears at the edges of the hole were smaller, but the cutting annular edge, owing to the strong expansive force, quickly broke off. The third punch was made with an angle of sharpening at the cutting edges of \(65^\circ\) and with considerable taper toward the guiding center, and this punch proved poor in operation, produced torn holes, and finally, owing to the strong central pressure, itself broke down (see Fig. 121). The best result in all respects was obtained for a punch of the same diameter with a central guiding cone, but with the angle at the cutting edges equal to \(87^\circ\). Thus, on the basis of the last experimental observations, it may be concluded that the presence of a guiding cone on the punch, for reducing the deformations of the metal being punched and for preserving the punch itself, is necessary; such a guiding cone on the punch promotes a more uniform distribution of the pressure of the metal on the shank of the punch, for the metal, during its flow in the period of punching, under the influence of the guiding cone and of the general conicity of the end surface of the latter, is displaced most strongly at first from the center

the punch to the periphery under the action of the guiding cone, and then, under the influence of the conicity of the cutting edges of the punch, from the periphery toward the center; as a result, the greatest compaction of the pressed-out disk of metal occurs along a certain middle annular surface, owing to which the cutting edges of the punch are in more favorable working conditions, and the deformations of the metal being forced through at the edges of the hole have the smallest magnitude. The clearance between the punch and the die in the latest experimental observations was about \(1/2\) mm.

In the preceding, we considered the question of the change in the structure of iron under various processes of mechanical working. We shall now try to elucidate the connection that actually exists between the structure of a given metal and its mechanical properties. The close connection between the mechanical properties and the structure of a metal was already pointed out in 1868 by D. K. Chernov, who said that “in the manufacture of steel products one always strives to obtain the grain as fine as possible, especially if the objects being made must have great toughness and great resistance, since numerous experiments have shown that the more crystalline the structure of the specimen, the larger and more regular the crystals, the less resistance to rupture the steel offers, and the less its toughness.”

Since a coarsely crystalline structure of various grades of iron is observed when heated within the limits of the critical points and above them, under conditions of slow and calm cooling after such heating, it follows from the foregoing that iron after such heat treatment cannot have good mechanical properties. If iron with a considerable carbon content is heated to the interval of the critical points and then rapidly cooled, it plainly changes its mechanical properties; and this change in mechanical properties will be the more intense the higher the heating temperature and the greater the carbon content in it. Such a process of heat treatment of carbonaceous iron is known in technology under the name of hardening. The intensity of the degree of hardening depends not only on the chemical composition of the given steel, but also on the conditions under which the hardening process itself is carried out. Thus, the intensity of hardening is strongly influenced by the heating temperature of the body being tested, the temperature of the bath, the ratio of the volumes occupied by the specimen and by the bath, the properties of the hardening liquid, and a whole series of minor causes, sometimes very difficult to take into account. During hardening, steel undergoes an internal structural change, and together with this the mechanical properties of the steel also change greatly. With complete hardening in a medium rapidly removing heat from the steel, the latter retains the higher structural form of its allotropic transformation, austenite (if the carbon content is more than 1.1%) or martensite; but if the harden—

Change in the Structure of Iron during Mechanical Working

...the medium weakly absorbs heat (oil), or is itself considerably heated (molten lead), then an intermediate structure is fixed in the specimen being tested: troostite or sorbite. The first case corresponds to quenching in water or mercury and bears the name full quenching, while the second case (in oil or lead) is sometimes called negative quenching. The process of quenching steel is in general always accompanied by the appearance of internal stresses between the particles of the given specimen. The cause of such internal stresses may be explained as follows: during quenching the specimen cools unevenly. Its inner part still retains heat when the outer shell has already cooled. Therefore the inner part is compressed by the outer part, which tends to expand. Such a picture is easy to verify experimentally: for this it is enough to cut a sufficiently long quenched rod lengthwise in half, and immediately a noticeable curvature of its halves will be observed as a consequence of the above-mentioned internal stresses. Similar experiments were performed by Howe⁵ and Barba⁶. During quenching the internal stress sometimes reaches such considerable magnitudes that the specimen after quenching cracks or breaks easily. The magnitude of the internal stresses can be considerably reduced by subjecting the specimen after quenching to slow heating (tempering) at temperatures between 200 and 600°, whereby the brittleness and hardness of the steel, under the influence of such tempering, decrease, but at the same time its toughness increases. Such tempering, it must be supposed, restores the disturbed equilibrium between the particles of the metal and thereby weakens the magnitude of the internal stresses between the particles of the given specimen.

The hardness of quenched steel increases chiefly with the carbon content and rises with the heating temperature. Thus, if for an experiment one takes a long steel rod and heats it for quenching from one end, then after quenching this specimen will possess the greatest hardness at the place of highest heating; moreover, the fracture of the given specimen successively changes its structure from coarsely crystalline to fine-grained, found at the place of highest temperature.

On the question of the change in the mechanical properties of various grades of carbon iron during various processes of heat treatment, including quenching, very thorough experimental investigations were carried out by Brinell⁷, undertaken by him at the Fagersta works and presented by this works at the Paris Exhibition of 1900.

All the tested specimens, 403 in number, are divided into 4 groups. Each group includes several series of specimens with different carbon contents.

Group	Carbon content
I	\(C = 0.84\%\).
II	\(C = 0.39\%\).
III	\(C = 0.20\%\).
IV	\(C = 0.09\%\).

Specimens of 18 mm diameter were turned from round iron 32 mm in diameter, rolled in one heat from bars \(5 \frac{1}{2}'' \times 5 \frac{1}{2}''\), which in turn had been rolled from ingots \(8'' \times 8''\) and \(10'' \times 10''\).

Then each grade of steel, comprising 13 specimens, was subjected to 13 different methods of treatment:

I. Without further treatment (rolling in the hot state).

II. Annealing at temperatures: \(350^\circ\), \(750^\circ\), \(850^\circ\), \(1000^\circ\), \(1100^\circ\), and \(1200^\circ\).

III. Hardening in water \(+20^\circ\),

Treatment	Temperature
hardening temperature	\(750^\circ\)
a) without tempering
b) with tempering to	\(350^\circ\)
c) with tempering to	\(550^\circ\)
d) with tempering to	\(650^\circ\)

The same hardening in water \(+20^\circ\).

Treatment	Temperature
hardening temperature	\(850^\circ\)

The same hardening in water \(+20^\circ\)

Treatment	Temperature
hardening temperature	\(1000^\circ\)

Hardening in oil \(+80^\circ\)

hardening temperatures \(850^\circ\) and \(1000^\circ\), a) without tempering and with tempering to temp., b) \(350^\circ\) and c) \(550^\circ\).

IV. Hardening in lead \(+550^\circ\)

hardening temperatures \(750^\circ\), \(850^\circ\), and \(1000^\circ\). (See Table XXX).

First of all, let us note the circumstance that hardening produces a real effect only from \(850^\circ\); the temperature \(750^\circ\) is still insufficient.

In this case the greatest effect was obtained in the specimen with \(0.09\%\ C\). Its resistance increased considerably: by \(60\%\), while elongation decreased by \(40\%\).

The resistance of the other groups changed barely at all, increasing somewhat. The elongation, however:

Group	Change in elongation
III	decreased by \(33\%\)
II	decreased by \(14\%\)
I	increased by \(15\%\).

If these changes depended on the transformation of carbon and iron, they would proceed in the reverse order—being greatest for the hardest steel itself, i.e. group I, and smallest for the specimen \(C=0.09\%\). The changes for group I are rather analogous to the effect of moderate annealing, since they give the greatest increase both in resistance and in elongation. All this indicates that at \(750^\circ\) hardening does not act, and the changes obtained should be ascribed to purely physical causes—arising during hardening from internal stresses, the action of which, moreover, is completely destroyed by reheating to \(350^\circ\).

I believe that hardening takes full effect at temperatures of \(850^\circ\) and \(1000^\circ\), in some cases at the first, in others at the second.

The tensile strength increases in all cases of hardening. Tempering, as an action opposite to hardening, restores to some degree the properties of annealed steel and precipitates \(\alpha\)-iron from martensite in the structure. The resistance raised by hardening usually decreases, and the dependence curves have the form of descending branches \(abc\) (Fig. 14a, Table XXVIII). These curves have, at first glance, anomalous branches \(ad\).

The presence of the branches \(ad\) cannot be explained by the influence of the elements on which the hardening effect expressed by the branch \(abc\) depends, i.e. by the transformation of carbon and iron and, correspondingly, of the structure. Under the influence of these transformations, the curve \(abc\) should be continued further in the direction \(ad'\), and the curves \(b_1, c_1\) and \(b_2, c_2\)—in the direction toward \(d_0\) and \(d_1\). Apparently some unknown cause, manifesting itself during hardening in water, obstructs the natural course of the curves \(abc\) and lowers the resistance to a greater degree the harder the steel is. For specimens of group I this resistance falls to \(o\), for group II to the value \(d\), and for group III to \(d'''\).

Specimens of group I made of hard steel, when hardened in water at \(850^\circ\) and \(1000^\circ\), became so brittle that they did not withstand straightening to correct the warping produced during hardening, and cracked. Only for the softest grade \(C=0.09\%\) does the branch \(a'''d^{iv}\) rise upward. Thus, the same cause which in hard steel lowers the resistance as compared with the resistance of simply annealed steel (group I), while in soft steel (group III) and semihard steel (group II) prevents an increase—in the softest specimen \(C=0.09\%\)—raises the resistance.

This phenomenon can be explained by the appearance, as a result of rapid cooling, of internal stresses, which in some cases reduce the resistance and in some cases may increase it.

For a more detailed elucidation of the results of Brinell’s experimental investigations, engineer Tagiev gives a series of diagrams, presented—

...are in Table XXX, which clearly illustrates the results of experimental investigations; but in general, on the basis only of the investigations cited and of Howe’s data⁸), one may arrive at the conclusion that quenching generally increases tensile strength. But in the case when the carbon content and the rate of cooling reach higher limits, the tensile strength under quenching rapidly falls⁹). Thus, according to Howe’s data, steel with a content of \(C = 0.50\%\), when quenched in water, increases its strength by 47%. Steel with a content of \(0.75\%\ C\), after quenching from any temperature, always gives a decrease in tensile strength, which after very rapid and severe quenching may even be reduced to zero, i.e. the specimen is destroyed during the very process of immersion in the quenching medium. The medium in which the quenching process is carried out has a great influence on the change in the mechanical properties of steel during quenching. Thus, the greatest intensity of quenching is attained by using mercury for the quenching bath; then come water, oil, tallow, tar, and, finally, molten metals, such as, for example, lead.

Quenching in oil, at any carbon content, always gives an increase in tensile strength and increases the elastic limit. Moreover, with respect to the change in the elastic limit of steel during quenching, it is necessary to note the following: the elastic limit is closely connected with tensile strength and, during the quenching of steel, undergoes the same changes as the tensile strength.

The percentage elongation of steel during quenching always decreases. Thus, out of 81 examples from Howe’s data, elongation during quenching in water decreased by 63%, and in oil by 45%.

The action of annealing on carbon iron is generally the reverse of the action of quenching. Brinell’s experimental investigations prove that the influence of annealing begins to be manifested even at a temperature of \(350^\circ\), at which there is already a noticeable increase in tensile strength and in elongation at the elastic limit. When annealed by heating to \(750^\circ\), there is a noticeable fall in the initial elastic limit and in elongation, which nevertheless remain above the initial values characteristic of quenched steel. Such a fall in the mechanical properties of steel upon annealing up to \(750^\circ\) must be explained by the appearance of an allotropic change in the structure of the steel. Annealing up to \(1000^\circ\) increases still more the mechanical properties of the metal, and especially the elastic limit. Further annealing, causing coarse crystallization of the metal, contributes to a decrease in the elastic limit and elongation, but the tensile strength decreases very little; moreover, the total values of these mechanical coefficients will be greater than the initial ones characteristic of unannealed steel. Thus, from experi-

⁸) Howe. The Metallurgy of Steel. S. 33.
⁹) Howe. The Metallurgy of Steel. S. 33.

... from Brinell’s investigations it follows that uniform annealing up to 1000° improves the quality of steel, in %, in the following manner⁸:

	Elastic limit.	Beginning of local extension.	Tensile strength.	Elongation.
Group I	44	19	8	11
Group II	44	20	2	12
Group III	40	28	3	15
Group IV	48	26	2	39

From consideration of the various cases of hardening and annealing of steel, a close connection between the structure of the metal being tested and its mechanical properties is clearly noticeable. But, on the other hand, conversely, from the fracture curve of a given bar, with known experience one may obtain an idea of the structure of the bar being tested.

It is beyond doubt that tensile strength depends on the force of cohesion between the grains of the metal, or on the cohesion of the masses of the grains themselves, if the fracture occurs within the latter. The force of cohesion is proportional to the magnitude of the fracture surface, which will be the greater the finer the grains. Thus, the finer the grain, the higher the elastic limit and the greater the breaking load.

The breaking load at the beginning of the test, it must be supposed, is uniformly distributed among all elements of the structure of the specimen being tested, and all elements of this structure take part to a relatively equal degree in accepting the elongation and together give the total elongation of the bar. Up to a certain limit of loading of the tested bar, proportionality is observed between the elongation and the load of the bar; consequently, within this limit of elastic changes the elongations of the principal grains of the given metal and of the substances connecting them with one another are equal. In soft grades of iron and steel the main mass of the structure consists of ferrite grains, and the cementing layer is pearlite. In tensile testing of bars of soft grades of carbon iron, complete equality of the elongations obtained between the principal grains of the structure and their cementing substance is observed only up to a certain limit; thereafter the ferrite grains begin to undergo a noticeably more rapid elongation, without change in the load of the bar, as compared with the elongation of the cementing layer, and on the tensile diagram of the given bar a characteristic horizontal yield step is obtained (see Fig. 15 A, Table XXV). With an increase in the carbon content in the bar being tested, i.e., with a decrease in

of a ferrite structural element existing independently, the horizontal ledge in the rupture diagrams gradually disappears. According to Charpy’s experimental investigations, this horizontal ledge in the rupture diagram of the bar disappears completely in steel with a carbon content of 0.09%, and the rupture curve of the tested bar becomes steeper (see Fig. 15 B, Table XXV). The more carbon the tested specimen contains, the more independently separated pearlite it will have, and the steeper will be the rupture curve of the bar. Hardening of steel promotes the destruction of ferrite and pearlite grains with the formation of martensite, and the rupture curve for such steel is very close to a straight line with a break. (See Fig. 15 C, Table XXV). Annealing of such steel promotes the appearance either of transitional elements, troostite and sorbite, or of the original structure with a predominance of pearlite. As a result, the diagram of breaking stresses for an annealed bar of steel will not have a break, but will have a steep turn (shown by a dotted line) (Fig. 15 C, Table XXV).

Thus, from the character of the diagram of breaking stresses of the specimen tested, one can determine with a high degree of probability its principal structural elements—ferrite, pearlite, and martensite.

Earlier we considered the question of the change in the structure of iron during rolling, forging, and drawing, and noted the strong change in the structure of the metal under such mechanical treatment; but since a close connection between the structure of a metal and its mechanical properties undoubtedly exists, it is natural also to assume a strong influence of mechanical treatment of the indicated type on the mechanical properties of the metal. To determine this dependence, experimental investigations were repeatedly carried out. Thus, Ferbert¹¹) cites tests made by Clay at the Mersey Steel Works iron-rolling plant, which consisted in the following: one and the same iron was welded and rolled successively several times. Several mill-bars No. 1 were welded together and rolled; from the resulting mill-bar No. 2 a packet was made, welded, and again rolled, etc.

From tests of each of the rolled Nos. the following strength coefficients were obtained:

Mill-bar No.	1—30.8	Mill-bar No.	5—40.3	Mill-bar No.	9—40.3
”	2—37.1	”	6—43.4 max.	”	10—38.0
”	3—41.8	”	7—41.8	”	11—36.5
”	4—41.8	”	8—40.3	”	12—30.8

¹¹) Ferbert. Iron, p. 251.
A. Gavrilenko. Mechanical Technology. P. 80.

The greatest coefficient of strength was obtained after the 6th rolling, and then it began to decrease.

Thus, successive welding heats and rolling of iron increase the coefficient of strength up to a certain limit, and then it begins to decrease.

Forging has the same effect on the strength of iron. It densifies the metal and increases its tensile resistance. As a result, small parts, which can be forged better, have greater strength than large ones. In large parts, however, the metal is forged better at the surface, and therefore it is always stronger there than inside. Thus, for example, on the American steamship Dolphin a shaft broke during a trial voyage. Specimens cut from the shaft gave the following results in testing:

	Elastic limit, kg per □ mm	Coefficient of strength per □ mm	Elongation %
From the center of the shaft	23.9	37.9	2
From the surface of the shaft	22.5	56.2	18

In steel, during forging, the coefficient of strength changes less markedly. According to Wickenden’s experiments at the Baldwin Locomotive Works, in Philadelphia, the following results were obtained on cast steel:

	Coefficient of strength, kg per □ mm	Elongation %	Fracture
Bloom with square section, 125 mm on a side	35.7	13	Crystalline.
Bloom forged to a thickness of 180 mm	35.6	16	Partly crystalline.
Same to 145 mm	34.7	18
Same to 125 mm	35.0	20	Silky.
Same to 100 mm	35.1	22

From the table it is evident that the coefficient of strength changes little, but the ductility of the metal, and consequently also its tensile resistance, increases greatly.

Experimental investigations have also repeatedly been carried out to determine the changes in the mechanical properties of metal under cold working. Thus, Considère¹⁰ gives the following data, characterizing, to a certain extent, the influence of cold working:

	Elastic limit in kg per □ mm.	Coefficient of strength in kg per □ mm.	Elongation.
Very soft steel	25.3	42.5	26.5
The same, pressed with a hydraulic press, 6 tons per □ sant.	35.7	44.6	17.0
Shipbuilding steel	29.6	52.5	18.0
The same, rolled cold from 16 to 9.45 mm.	42.3	54.5	11.5
Iron sheet	22.8	37.4	15.0
The same, rolled cold from 8 to 7.1 mm.	41.6	46.9	7.0

As is evident from this table, the coefficient of strength changes not very significantly, while the elastic limit rises very greatly, but at the same time the viscosity decreases.

Cold rolling is used on a large scale in the United States of North America for the manufacture of drive shafts. Mainly the purest, soft, puddled iron is subjected to such treatment.

The numerous investigations of Prof. Thorston¹¹ on iron treated in this manner led him to the following conclusions.

The tensile strength increases from 25% to 40%, and the bending strength by 50–80%.

The elastic limit rises by 80–125%.

The live elastic resistance increases considerably.

The total live resistance, however, decreases by as much as 50%; thus the increase in elasticity and strength is obtained at the expense of viscosity; therefore iron rolled in the cold state is applicable only in those cases where the stresses in the material cannot go beyond the elastic limit.

In addition to the indicated changes, cold rolling imparts to iron a smooth, shiny surface, entirely free from scale. Iron can be rolled with great accuracy in dimensions (at the Johnes and Laughlins Pittsburgh works); the rolls are rolled with an accuracy of up to 1/1000″, requiring no further processing.

From examination of the tables presented above, the influence of the general conditions of mechanical treatment on the mechanical properties of iron is clearly noticeable; but, on the other hand, the mechanical properties of iron vary very strongly with changes in its chemical composition, and therefore

Change in the Structure of Iron during Mechanical Working

Jüptner, for expressing the coefficient of rupture, gives the following formula:

\[ R = A + {}^{2}/_{3} C + {}^{2}/_{7} L + {}^{1}/_{7} Mr, \]

where \(A\) is a quantity dependent on the working—the coefficient of working. It must be noted that this formula gives only approximate results, since it does not take into account the percentage content of other elements, such as phosphorus, sulfur, chromium, etc., which noticeably alter the mechanical properties of iron¹².

Although Jüptner’s formula is approximate in character, in its application in practice it fully confirms many observations of engineers. Thus, according to the determination of engineer A. Baboshin, from a considerable number of rails examined by him that had served on the track, it turns out that the best qualities were possessed by those rails for which the coefficient of mechanical working was the greatest. A. Baboshin¹³ divides the rails examined by him into three groups:

Group I—rails removed from the track, without any damage after long service.

Group II—rails with crushed heads and worn out.

Group III—rails that burst in the track and had cracks.

The results of Engineer Baboshin’s observations are brought together in the following table:

Rail Nos.	Resistance to rupture and elastic limit, top of head	Resistance to rupture and elastic limit, middle of head	Ratio of elastic limit to rupture resistance, %, top of head	Ratio of elastic limit to rupture resistance, %, middle of head	Coefficient of working \(A\), top of head	Coefficient of working \(A\), middle of head	Group
14—2	83—31	83—33	37	40	24	24	From group I
15—3	70—34	70—32	49	46	22	22	From group I
28—1	70—29	56—33	42	59	25	30	From group I
36—3	80—27	82—36	33	44	28	11	From group I
Average	76—30	73—34	40	47	25	22	From group I
10—2	73—26	74—28	36	38	13	18	From group II
20—2	74—27	74—34	37	46	18	18	From group II
22—4	70—29	69—31	42	44	17	17	From group II
25—5	70—23	73—19	33	26	16	18	From group II
Average	72—26	73—28	37	38	17	18	From group II
20—1	81—36	82—38	44	47	22	22	From group III
21—2	74—29	71—32	39	45	28	27	From group III
21—4	66—38	68—34	58	50	12	13	From group III
26—4	72—33	50—45	46	89	11	(−11)	From group III
29—1	83—31	69—35	37	50	28	14	From group III
Average	75—35	68—37	45	56	20	?	From group III

Thus, the rails in which the coefficient of mechanical working reached the greatest value were distinguished by the best qualities: on average for the first group of rails—25.

In the rails of the second group (crushed), the working coefficient decreases considerably, on average to 17.

All the rails whose tests are given in the table are fairly close in chemical composition (C about 0.5 and Mn about 1%) and have the same coarse pearlite grain in their structure. The average ratio of the elastic limit to the tensile strength for the middle of the head in rails of the second group (crushed) is 38%, while the same ratio for the good rails of the first group is 47%, and a similar dependence of the elastic limit on the coefficient of rupture is in close connection with the values of the coefficient of mechanical working given above. Only in the rails of the third group, brittle ones which broke prematurely, do we not see such a dependence of the coefficient of mechanical working, the elastic limit, and the coefficient of rupture; moreover, for some of them the ratio of the elastic limit to the tensile strength exceeds the same ratio for the first group, and hence it must be assumed that the main cause of the inferior quality of the rails of the third group lies in accidental causes of a secondary nature that accompanied either the production of the rails (local hardening, improper cooling on the racks, nonuniform heating, and so forth) or in causes that manifested themselves during the rail’s service in the track and abnormally changed the temporary tensile strength in the middle of the rail head.

Of considerable interest, according to the results of the investigation by the rail commission of the Ministry of Ways of Communication ^16) on the question of the relation between the structure of rails and their service in the track, are the general conclusions of the commission on this matter, which are as follows:

a) Rails that have served well in the track (having carried much freight and removed without any damage) have, under the microscope, a fine-grained structure.

b) Poor rails, removed from the track because of crushing, have a coarse-grained structure.

c) Brittle rails, which broke in the track after the passage of a small amount of freight, also very often have a coarse-grained structure.

According to the calculations of Engineer Baboshin, the average grain size for sound rails that had served a long time and were removed from the track without noticeable damage was, on average, equal to 0.0019 mm²; the same average value for poor, coarse-grained rails was 0.0195 mm² ^17).

^16 and ^17) Investigation of rail steel. Proceedings of the Mechanical Laboratory of the Institute of Engineers of Ways of Communication. 1906, pp. 239–246.

The conclusion of the rail commission is likewise not new in technology, since analogous conclusions had repeatedly been drawn earlier¹⁸) and clearly follow from the general proposition of D. K. Chernov cited by us above concerning the dependence of the qualities of worked steel on the structure of the latter.

In its turn, the same commission observed that coarse crystallization of steel is most characteristic of crushable rails with a low elastic limit, and that such a joint combination of these two factors is an obvious sign of poor mechanical working.

On the basis of the general picture considered above of the change in the structure of iron during rolling, we inevitably arrive at the conclusion that, in order to obtain a fine grain in the metal being worked, it is necessary to finish its rolling at as low a temperature as possible, but within certain limits of the latter, when no noticeable disunion between the grains of the metal has yet occurred; on the other hand, the temperature at which rolling is finished must not be higher than, or close to, that critical point (Chernov’s point “b”) at which the direct proportionality of grain size to heating temperature is interrupted. But since the position of the critical points at which a distinctly noticeable growth of grain size with heating temperature is observed depends very strongly on the chemical composition of the iron, and, specifically, the purer the iron the higher the position of the critical points, it follows from this how important it is in each particular case to know the position of the critical points and, in accordance with this, to choose the temperature at which rolling is finished. Thus it was indicated earlier that manganese greatly lowers the position of the critical points and, together with carbon, narrows the interval of the critical points; hence it clearly follows that, in order to obtain fine crystallization, manganese rails must be rolled at a considerably lower temperature than rails containing little manganese.

Without entering in particular detail into determining the position of the limiting points for finishing rolling in order to obtain a fine-grained product, very good in its mechanical properties, I shall allow myself to cite two excerpts from letters to the above-named Rail Commission by the well-known practical engineers Mr. Arthur Yuz and Mr. Yasyukovich, Managing Director of the South Russian Metallurgical Society¹⁹).

In Mr. Yuz’s letter he points out that the high quality of the steel rails of an earlier make is explained by the low temperature at which rolling was then finished; this circumstance was the reason—

¹⁸) R. Sob. “Steel rails: relations between structure and durability” The Metallographist. 1902. No. 3.

¹⁹) “Investigation of rail steel.” Proceedings of the Mechanical Laboratory of the Institute of Railway Engineers. 1906. pp. 304 and 308.

a natural consequence of the fact that, in the transition from iron rails to the rolling of steel rails, the works were compelled to make use of the former weak machines. In view of this, rolling proceeded slowly, and the temperature at which the rail left the rolls was considerably lower than that observed when rolling on modern powerful machines. Rolling at a low temperature is accompanied by the formation of a fine grain. Rolling at a high temperature causes the formation of a coarse grain; and the size of the grains has an essential influence on the resistance of the rail to wear.

It is further indicated that the Keppe process, practiced in America for rerolling rails of heavy types, withdrawn from service owing to wear, into a lighter type, fully confirms this view; when rerolling at a low temperature the grain always turns out finer. For the same purpose, the Kennedy-Morrison method is practiced in America, consisting in the fact that before passing through the last rolls the rails are held on specially arranged traverses until they reach a low temperature.

G. Yasyukovich, in his letter, expresses the following views: As became evident from the results of investigations carried out by the commission, as well as from the work of engineer Miller on the Bavarian roads, there is no clear and direct connection between the duration of service of rails in the track, their chemical composition, and the mechanical testing of rail samples. Despite the constant changes in the technical conditions for the acceptance of rails, both in our country and abroad, it has proved impossible to prevent the acceptance and laying in the track of rails unsatisfactory in track service. While acknowledging unconditionally that mechanical control must exist so that obviously unsuitable rails cannot enter service, engineer Yasyukovich believes that, owing to the constancy of the materials used in the conversion—chemical composition of rail steel should not be subjected to control. But since, on the other hand, experiments have shown that fine-grained steel with a high elastic limit and with the smallest possible ratio of tensile strength to elastic limit better resists wear and in general possesses better and more reliable mechanical qualities than coarse-grained steel with a low elastic limit and with a high ratio of tensile strength to elastic limit, all attention and control must be directed to these circumstances. This aim is achieved by proper mechanical and, chiefly, thermal treatment of steel ingots that are as dense as possible (without bubbles and shrinkage cavities). The production of dense ingots, in the opinion of engineer Yasyukovich, is not a difficult task. Likewise, the more fully the mechanical treatment affects the various parts of the ingots—the more uniform—

formity and perfection the structure of the metal will have. But in order to obtain fine-grained steel, an appropriate heat treatment is also necessary. It turns out that heated steel has three critical points; when these are passed through, by cooling or heating, the composition of the steel is rearranged with the liberation or absorption of heat. For mild steel with a carbon content below 0.2%, these critical points are 860°, 740°, 650° Celsius; for rail steel containing 0.45–0.55% carbon, the three critical points merge into one and lie at about 700–725° C. Consequently, if rolling is finished above the critical temperature, then the rail, cooling to the temperature of the air, must pass through its critical point and there undergo a rearrangement of its composition. This rearrangement is accompanied by an increase in the grain size of the steel. Therefore rolling must be finished below a certain temperature; in that case the structure of the steel no longer changes on cooling, and the grain remains as it was obtained with the metal-working at our disposal. Thus, decisive importance attaches to finishing rolling at a temperature of about 700° C.

Next, Mr. Yasyukovich explains how production should be organized so that the rails leave the rolls at a low temperature, and how control should be arranged so that the rails leave the rolls at the proper temperature.

As for the first question, at the present time there are two methods by which the rail is made to leave the last grooves sufficiently cooled. The first method, the American one, of Kennedy–Morrison, consists in holding the rail bar before the last groove so that it cools to the proper temperature. This method, used as early as 1887 at the Putilov Works, requires no special devices and can be applied at any works, provided that the proper temperature of the rail before passing through the last groove is checked by some pyrometer.

However, Mr. Yasyukovich finds that this method is not so expedient that it can be recommended to our works, for the following reasons.

The change in structure is closely dependent on the degree of mechanical working. In the last groove the metal is reduced very slightly, by no more than 5%; such a slight reduction is insufficient to affect the change to any considerable depth in the structure of the steel that was produced during cooling from a temperature above the critical one; reduction in the last grooves affects the change in the structure of the head to a depth of no more than 3–6 mm. This is best confirmed by St. Martin’s experiments in England. In one case the rail was held for 1¾ minutes before passing through the last groove, and in the sec—

…in the second case it was delayed for two minutes before entering the first pass and then, with a pass, continuously through all the passes. In the first case, micrographs of the structure of the steel, taken from the middle of the head and at a distance of 3 mm from the surface, showed that the center of the head has a coarse-grained structure, while near the surface of the head it is fine-grained. In the second case, the fine-grained structure was found both in the middle of the head and near the surface. Thus, the American Kennedy-Morrison method does not ensure the necessary fine-grained structure of the head. Therefore the rail bloom should be passed into rolling at such a temperature that it emerges from the last pass as a rail with a temperature below the critical one; consequently, control over the rational heat treatment must be transferred from the end to the beginning of rolling. The temperature of the blooms entering the rolling mill must be measured by some pyrometer, and the most practical for this purpose prove to be light (optical) pyrometers by Wanne, Hase, Mesuré and Nowl, Lejeune, and others.

Overheating of the blooms can be avoided by dispensing with their preheating, which is achieved by passing the blooms through Gjers soaking pits.

As for the sorbite method of treating materials, consisting in local cooling (hardening) of rail heads, according to Mr. Yasyukovich it should not be recommended to works, since Stead, who proposed this method, stated that good results were obtained only in the treatment of small specimens; in the treatment of long specimens, however, satisfactory results were not achieved.

At the end of his letter Mr. Yasyukovich briefly criticizes the sorbite method of manufacturing rails, which consists in the fact that the rail, having left the rolling mill, is immersed head-first in water, as a result of which the head is first hardened, and then the steel is tempered by the heat present in the head, whereby the rail head receives a sorbite structure. A similar method of obtaining rails with a very hard, but also tough, head was proposed by S. Stead and W. Richards¹⁸, and in its results the hardening of the steel head quite resembles the hardening of steel in oil; but in view of purely practical difficulties, namely the impossibility of obtaining uniform hardening of the rail head, this manufacture of rails has not become widespread. Yet one cannot deny the usefulness of the sorbite method as applied to forged articles, the treatment of which Stead proposes, for the stated purpose, to carry out in the following manner: after heating the article to a temperature somewhat above Ar₃, to the fullest possible transformation of the carbon, it is necessary to cool the article in air or by—

... be quenched, and then tempered to a temperature below \(Ar_3\) and cooled in air¹⁹.

The mechanical advantages of a sorbitic structure are apparent from the following comparison of two specimens:

	Elastic limit, kg	Tensile strength, kg	Elongation, %
Heating to \(1150^\circ\) and cooling in the furnace (pearlitic structure)	39,901	81,162	17
Heating to \(1150^\circ\) and cooling in air (sorbitic structure)	53,260	99,979	12

Such is the change in the mechanical properties of steel of medium hardness when it is treated by the sorbitic method.

Thus our preceding arguments concerning the change in the structure of the metal during rolling and concerning the dependence of the mechanical properties of the metal on the structure of the latter lead us to the general conclusion that, in order to impart good mechanical properties to a given metal, it is necessary that the rolling and the hot working of it be completed at a definite temperature. The beneficial influence of this temperature on the mechanical qualities of iron has been proved by a number of laboratory investigations and has been acknowledged by well-known persons standing at the head of large industrial enterprises. At works manufacturing critical articles by rolling (rails), careful control is organized over the heating of the billets being rolled; in most cases the heating of the billet for rolling is carried out in a single operation, and the rolling of the rail is carried out in a single heat. The temperature not only at the end of rolling is strictly observed, but also that of the initial heating of the billet, since a strongly overheated billet, owing to its coarse crystallization, cannot give a good rail and, during the rolling itself, will not have time to cool sufficiently. To obtain uniform heating of the billets, they are heated in soaking pits, most often with additional firing.

The temperature at the end of rolling must lie within the critical interval and in no case should be close to the temperature called by D. K. Chernov the point “\(b\),” when iron noticeably and readily assumes a coarse-crystalline structure with the cessation of the direct proportionality between the heating temperature and the grain size.

At the present time, not infrequently, when receiving highly responsible orders, plant management receives fairly broad instructions regarding the chemical composition (in the sense of deviations from one specified standard of chemical composition), but, on the other hand, strict, definite conditions for mechanical tests, which are reduced chiefly to determining the elastic limit as the principal indicator of the results of the mechanical working of iron.

To control the temperature of rail rolling, in addition to the usual thermoelements and optical pyrometers, a special magnetic instrument has recently begun to be used, based on the property that steel above the critical point is not magnetic and acquires this property only below that critical point. Such an instrument makes it possible automatically to sort rails into those rolled above the critical point and below it, i.e., into rails with coarse- and fine-grained structure. This method may in time prove very useful in the manufacture of rails, but from the constructive side it is still little developed. Thus, without denying the great influence of the chemical composition of iron on its mechanical properties ²²), it must be acknowledged that the conditions of mechanical working have an enormous influence on the change in the mechanical properties of iron.

Under the influence of mechanical working, elements of one and the same profile acquire different structure and different mechanical properties. Thus, after rolling, iron emerges from the roll groove structurally non-uniform across the section, and to illustrate a similar picture we previously examined the section of a rail. The non-uniformity of structure is directly dependent on the complexity of the profile being rolled, but, on the other hand, it must be supposed that the non-uniformity of the metal also depends strongly on the intermediate pass profiles, to which heightened attention must be paid when calibrating the rolls.

Thus, depending on the complexity of the profile being made, after rolling through one and the same section the iron proves to be structurally non-uniform; moreover, it also enters the profiles of the pass rolls and passes through them chemically non-uniform, under the influence of segregation of the metal in the ingots, as is clearly indicated by the chemical analysis of rails made by engineer Baboshin ²³) and presented in the following table:

²²) The influence of chemical elements on the mechanical qualities of iron is discussed in greater detail in Chapter II of Metallic Alloys, T. Tikhonov. 1910.

²³) Investigation of rail steel. Proceedings of the Mechanical Laboratory of the Institute of Engineers of Ways of Communication. P. 6.

Rail No.	Total carbon C total	Total carbon C total	Carbon soluble in river [[unclear: “p.”]] Ck	Carbon soluble in river [[unclear: “p.”]] Ck	Quenching carbon C quen.	Quenching carbon C quen.	Phosphorus P	Phosphorus P	Silicon Si	Silicon Si
	On the head surface	In the middle of the head	On the head surface	In the middle of the head	On the head surface	In the middle of the head	On the head surface	In the middle of the head	On the head surface	In the middle of the head
14—2	0.605	0.736	0.264	0.236	0.341	0.500	—	—	—	—
22—6	0.604	0.718	0.295	0.310	0.309	0.408	—	—	—	—
13—1	0.400	0.418	0.086	0.082	0.314	0.336	0.094	0.112	—	—
9—4	0.304	0.386	0.209	0.327	0.095	0.059	0.107	0.144	—	—
9—3	0.373	0.582	0.195	0.373	0.178	0.209	0.101	0.209	—	—
14—7	0.210	0.310	—	—	—	—	—	—	—	—
22—1	0.255	0.268	—	—	—	—	0.247	0.230	—	—
22—7	—	0.290	—	—	—	—	0.185	0.165	0.224	0.201
19—4	—	0.160	—	—	—	—	—	—	0.485	0.489
22—5	0.350	0.410	—	—	—	—	—	—	—	—
14—5	0.355	0.370	—	—	—	—	—	—	—	—
20—4	0.273	0.360	—	—	—	—	—	—	—	—
37—1	0.250	0.380	—	—	—	—	—	—	—	—

From it one can see that the amount of total carbon is always smaller on the surface of the head than in its middle, and the difference is sometimes very considerable.

To eliminate the noticeable influences of iron segregation, a number of precautions are taken in the manufacture of the metal and of the blooms themselves from it, and in finished blooms the upper third is often cut off and not used for the manufacture of critical sections. The chemical composition of iron undoubtedly affects the structure of its elements; consequently, the principal aim of the modern practical engineer—to give chemically heterogeneous iron entering rolling, by changing the conditions of mechanical working, an identical structural constitution (fine-grained) over the cross-section of the section (sometimes of very complex shape)—is a very difficult task. To attain this aim it is necessary to take not only a whole series of the above-mentioned precautions in the sense of precisely determining the temperature at the beginning and end of rolling, but it is also necessary subsequently, after the completion of rolling of the metal, to place it under such conditions that the metal would not have the possibility of changing its structure. The metal leaves the last pass of the finishing rolls with a significantly high heating temperature, as a result of which it is capable of undergoing local hardening, provided the local conditions favor this; and therefore it is necessary to give the hot finished sections (strips) the opportunity to cool slowly and uniformly, protecting them for this purpose from accidental drafts and local cooling. The arrangement of special

premises intended for the uniform cooling of hot strips after rolling, is especially necessary to recommend for places with sharp fluctuations of temperature.

A hot finished section (strip) after rolling, when cooling even under very favorable conditions as regards the uniformity of its cooling, owing to the nonuniform distribution of the mass of metal over the section and the nonuniform cooling of the parts of the section in the straightening rolls, is always, after cooling to the cold state, warped along its length. To eliminate such distortions one has to impart to the section (strip) residual deformations in the cold state by blows with sledgehammers or by special pressing mechanisms. The production of such residual deformations is always accompanied by a disturbance of the previously obtained structure of the metal and causes the appearance of additional internal stresses, at times having a very harmful effect on the mechanical qualities of the metal. In order to eliminate the just-mentioned harmful effects of cold straightening of finished strips, it is necessary to carry out such straightening in the hot state, when the particles of the metal are still sufficiently elastic; therefore, in order to eliminate the amount of warping of the strip, it is necessary for the latter, in the hot state, immediately after the completion of rolling, with the aid of special devices (templates prepared by experiment or movable mechanisms), to be given a curvature opposite to that which the cold strip acquires as a result of the above-mentioned causes; in the latter case the extent of cold straightening and its harmful consequences can be reduced to a minimum.

In concluding the survey of the change in the structure of iron during hot working, it is necessary to note one more interesting fact, observed by Charpy ^24) and verified by Le Châtelier ^25), namely that steel, when work-hardened at a temperature of 600—700° C, shows a tendency toward coarse crystallization, as a result of which, being annealed at this temperature, it becomes very brittle. Le Châtelier illustrates the tendency of work-hardened metal toward coarse crystallization on specimens of stamped aluminum ware that had begun to undergo rebirth. In the latter specimens the author observed a granular structure, especially clearly noticeable in proximity to the place where the rebirth had begun. Observations of a similar character, illustrating the phenomenon that work-hardened metals in general, and among them iron, are in a structurally unstable equilibrium, so that a quite insignificant impulse (of temperature or chemical action) is sufficient to cause a transition to a stable state with the appearance of distinct crystallization, have in recent times been—

^24) R. de Métall. 1910. 655.
^25) R. de Métall. 1911. No. 4. 364.

given sufficiently fully²⁰, but it is still impossible to regard the question raised as in any way exhausted.

Of the change in the structure of iron under cold working, one may form a quite definite idea from the previously considered picture of the change in the structure of iron during drawing, for the general character of this picture does not change for the majority of cases of cold working; the change in the mechanical properties of iron during cold working is illustrated by the tables of experimental observations given earlier; but, for the sake of completing the general picture of the change in the structure of iron in the cold state, it is necessary to say a few words about the change in the structure of iron under the influence of repeated loads. Many experimenters have worked on this last question; thus, in 1858, Wöhler, by commission of the Prussian government, undertook a series of experiments on the action of repeated loads, and in 1871 published the results of his observations. According to Wöhler’s data, it follows that the number of repeated loads which a tested specimen can withstand increases as the magnitude of the loads decreases; but, besides this, the number of loads depends not only on their absolute magnitude alone, but also on the limits between which these loads vary.

Under the action of repeated loads, iron and steel may be destroyed even in the case where each of the individual loads is not capable of tearing the metal apart. But, on the other hand, there always exists a certain magnitude of stress whose repetition the metal can withstand an arbitrarily large number of times.

Further experiments, carried out by Bauschinger and Bacher, fully confirmed the general conclusions from Wöhler’s investigations. Of interest is the fact that, under the action of repeated loads, the metal does not lose its strength and toughness; thus, specimens destroyed by repeated loads, when tested on tensile machines, exhibit their former strength and toughness. It must be assumed that the destructive action of repeated loads extends to a very thin, weakest layer of the cross-section, along which the metal ruptures with an evident development of crystallization. Such a consideration is fully confirmed by the fact that, under repeated loads, iron and steel break without any elongations, exactly like brittle metals.

The same picture of fracture is not infrequently observed in the breaking of wagon axles, tires, and similar objects, and the cause of such fractures, in the majority of cases, must be supposed to lie in the influence on the metal of repeated loads.

The cause of such destruction of a metal under the action of repeated loads may be explained as follows: iron, as we saw earlier, under the influence of mechanical working always has a somewhat elevated elastic limit, the magnitude of which is determined on tensile-testing machines for the calculation of various machine parts. The action of repeated loads, as the experimental investigations of Bauschinger²¹ have shown, tends to lower this artificially raised elastic limit and to bring it to its natural value, at which the acting load produces, in the weakest places in the cross-section of the working machine part, an impulse toward the formation of distinct crystallization; as a consequence there is a local weakening of the cohesive force of the metal grains (for their size has increased) and fracture of the part itself. By a whole series of experiments Bauschinger endeavored to determine the natural value of the elastic limit, and obtained it for iron as equal to 12.6 klg. per sq. mm., and for mild steel 14.9 klg. per sq. mm.; these values are very close to those which were given earlier by Wöhler as safe stresses for producing a large number of repeated loads.

Thus, summarizing the principal propositions from all that has been said above, we arrive at the following conclusions:

1) Carbon iron, under conditions of very slow cooling after melting, is capable of assuming a distinctly crystalline form. Most often, carbon iron crystallizes in the form of an octahedron, and considerably more rarely in the form of a cube. In the presence of all the above-mentioned conditions, the most sharply expressed forms of crystallization in carbon iron are observed when the carbon content is about 2%. Carbon iron is capable in some cases of assuming a crystalline structure even with very rapid cooling from high temperatures and in the presence of external conditions that constrain the formation of crystals.

2) In the absence of very slow cooling and under the influence of the forces of mechanical working, the particles of carbon iron cannot separate out in the form of sharply expressed crystalline forms, but acquire oval, though in most cases distinctly expressed, contours.

3) With gradual heating, the size of the individual grains in carbon iron grows. The first noticeable change in the size of the grains in carbon iron appears at about 300°. Mild iron, heated to the latter temperature and worked at it, has not only a coarser crystallization in comparison with raw, unheated metal, but also possesses greater brittleness with a signif—

appreciable decrease in viscosity. The latter property of soft iron—to change the structure and mechanical properties of the metal itself—is known in technology under the name “blue heat.” The temperature of the blue heat of soft iron is close to 300° C, but the change in the mechanical properties of the metal depends not only on the temperature of the blue heat, but also on the mechanical working of the metal at this temperature, under the influence of which the characteristic change in the mechanical properties of the metal itself also occurs. The change in the growth of the grains of soft iron at blue heat is not accompanied by any noticeable evolution or absorption of heat, as a result of which it is not possible to register this characteristic point by using a pyrometer.

4) With further gradual heating, at a temperature above 600° carbon steel noticeably changes its internal structure; moreover, such a change in the structure of iron is accompanied by a noticeable change of heat and is easily registered by a pyrometer under the name of the critical points \(Ar_1\), \(Ar_2\), and \(Ar_3\). The first of the critical points, \(Ar_1\), corresponds to the beginning (during heating) of allotropic changes in the particles of iron, namely to the transition of \(Fe\alpha\) particles into \(Fe\beta\); the second, \(Ar_2\), to the transition of \(Fe\beta\) particles into \(Fe\gamma\); and the third point, \(Ar_3\), corresponds to the end of the allotropic changes of the indicated character. The position of the above-named critical points in carbon steel depends strongly on the carbon content in the latter, namely: \(Ar_1\) is almost entirely imperceptible for the softer grades of iron, while for the harder ones it reaches a considerable magnitude. As the carbon content in the iron increases, the position of the critical points \(Ar_1\) and \(Ar_2\) approaches the position of point \(Ar_3\), and in hard grades of steel all three points often coincide. The more carbon the iron contains, the lower in temperature is the position of the critical points. Along with carbon, manganese also affects the position of the critical points in the same way. The position of the critical points for one and the same grade of iron during heating will always be somewhat higher than during cooling, and this difference will be the more noticeable the higher the initial heating of the iron from which the observed cooling proceeds.

5) The size of the individual grains of carbon steel gradually increases during heating, but the most intensive growth of the grains of iron begins with the appearance of the critical point \(Ar_1\). The size of the individual grains of carbon steel is directly dependent on the heating temperature, and, in Howe’s opinion, to each given temperature there corresponds a definite grain size; if there exists a grain larger than normal for the given temperature, then it does not return again to the normal size. Direct proportionality between the size of the iron grain and the corresponding heating temperature exists only up to a certain limit of heating; within the limits—

to 1000° C for soft and medium grades of iron in hardness (in carbon content); with further heating, the iron grain grows faster than the corresponding increase in the heating temperature, as a result of which the iron rapidly passes into the stage of coarse crystallization, with the appearance of poor mechanical properties in the metal.

6) The cause of the growth of iron grains upon heating lies in the diffusion of substances between neighboring grains (crystals) through the film of eutectic binding them. This diffusion of the latter character itself appears as a result of the unequal orientation of the axes of the crystals (grains), owing to which the particles of two adjacent crystals dissolve unequally in the eutectic binding them; as a result there appears a one-sided pressure and a progressive displacement of the particles of one crystal (grain) toward the other, the latter beginning to grow at the expense of the former.

7) With prolonged heating at a temperature of about 640–670° C, in the structure of carbon iron the phenomenon of coalescence is observed, i.e. the mutual displacement and attraction of like particles. Under the influence of coalescence, with prolonged heating and cooling of eutectic steel in the critical interval, a complete separation of the structural elements occurs, and, in Osmond’s words, a picture of structural equilibrium is observed.

8) At a high welding heat, the mechanical forces of rolling change the size of the iron grains very little, this size depending mainly on the heating temperature. With a further lowering of the heating temperature and under the influence of the mechanical forces of rolling, a markedly strong change in the form and size of the iron grain is observed: at first the grains are elongated in the direction of rolling, and then they are divided into parts along the rolling direction under the action of lateral pressures, as a result of the action of the rolling rolls. The size of the grains of the iron being rolled does not remain the same over the cross-sectional area of the pass, namely: for the first passes of rolling, at a higher heating temperature, the grain area at the surface of the pass is smaller than that at the center of the pass section. In the last passes of rolling the reverse picture is observed—the iron grains at the surface of the pass are larger than at the middle of the latter. The cause of the first phenomenon is that, under the influence of the rapid cooling by the rolls of the hot metal, the metal grains at the surface retain the size that they acquired after mechanical working, not having time to grow to the normal size (for the temperature) owing to rapid cooling. Meanwhile, the grains in the middle of the first passes being rolled, under the influence of the high heating temperature, manage to increase their size, following the previously indicated law concerning the normal grain size for each temperature. In the last passes of the iron being rolled, at lower temperatures…

heating of the grain from the surface of the groove, owing to cooling by the rolls and to the mechanical forces of the latter, become more stable and retain their previously acquired form; whereas the grains from the middle of the latter grooves, having a higher heating temperature in comparison with the surface grains, under the influence of the mechanical drawing action of the rolls are strongly fragmented and change their size.

9) The size of the grains of iron is not constant for all points of the cross-section of the finished profile, but varies very considerably with the character of the change in relief of the profile itself. At the places of greatest section of the profile under consideration the grain size also has a larger magnitude, and, conversely, in the thin sections of the profile the grains have a smaller size, and the mechanical forces of rolling have a very noticeably strong influence on their form. Rapid and non-uniform cooling of the metal after rolling only increases the non-uniformity of the grain size throughout the profile, and conversely—calm and uniform cooling of the metal after rolling promotes the restoration of uniformity in grain size.

10) In drawing, the size of the grains of the iron being worked decreases in the transverse section of the profile. The presence of high tension during stretching affects the elongation of the grains. Strong transverse compression of the material causes division of the grains along their length, first forming necks. The dimensions of the grains are not the same throughout the transverse section. Their size increases from the middle part toward the periphery, owing to the difference in resistance. The different resistance of the grains and the force of friction at the die causes slipping of the inner fibers relative to the outer ones, serving as the cause of stratification of the material.

11) In stamping, a strong change in the form and arrangement of the grains of iron is observed in the worked end of the object, while the shank of the latter usually preserves the structure that it acquired during the preceding treatment (rolling or drawing). The longitudinal arrangement of the fibers, characteristic of the shank of the object being worked, in the middle of the head changes into a winding one, with splitting of the grains and with a noticeable formation of cracks. The grain size changes greatly in different places of the head, while at the place where the stem ends there remains an untouched spherical part, the grains of which, in comparison with the stem of the shank, have hardly changed, but from there a fibrous transition to the lateral parts of the head is already observed. The outer surface part of the head, both the upper part, which was in contact with the stamp, and the supporting plane below, have a more fine-grained structure. The part between the spherical convexity and the surface layer is, as it were, flattened. The magnitude of the grains and the general character of their deformation are enormously influenced by the temperature of treatment and by the character of the treatment itself. The more

the treatment temperature is higher, and the more smoothly the deformation of the body being treated is carried out, the more homogeneous is the structure of the treated end, and the fewer are the cracks between the individual grains in the head.

12) In hot forge forgings, the oval grains of the treated piece of iron acquire a wedge-shaped form with the direction of the longitudinal axis of the grain in accordance with the action of the external impact forces; at the same time, the division of individual grains into smaller parts is also observed.

13) In cutting and punching holes there is observed a characteristic deformation of the separated edges of the metal; in cutting, the magnitude of these deformations is about 1 m/m in thickness, while in punching holes it is about \(1/10\) of the thickness of the object being treated. The magnitude of the deformations in punching holes is noticeably influenced by the form of the punch and the size of the clearance between the latter and the die; namely: when the size of the indicated clearance is within the limits from \(1/2\) to 1 m/m, the deformations of the metal being treated have the smallest magnitude, but the punch, while maintaining the magnitude of this latter clearance, must have a central guiding cone and cutting edges with an angle close to \(87^\circ\). Additional annealing of cut edges or punched holes promotes the restoration of the grains of the disturbed structure, and chiefly, it apparently promotes the separation of the deformed part of the metal from the sound part.

14) The mechanical properties of the metal tested (iron) are in close connection with the chemical composition and with its structure, which in turn are in close connection with the preceding thermal and mechanical treatment of the metal. From the character of the fracture curve of the bar under test, one may, with a high degree of probability, judge the structural constitution of the bar.

15) Successive heating and rolling of iron increase the strength coefficient of the latter up to a certain limit, after which it begins to decrease. Forging has a similar influence on the strength of iron; for steel, however, the strength coefficient in forging changes less than for iron. In cold working the strength coefficient changes not very significantly; the elastic limit, however, rises very strongly, but at the same time the toughness of the metal decreases.

16) Those articles made of carbonaceous iron for which the coefficient of mechanical treatment has the greatest value (according to Jüptner’s formula) are distinguished by the greatest durability in service.

17) In order to obtain a fine-grained structure, and consequently also to obtain carbonaceous iron with good mechanical properties, it is necessary to finish the hot rolling of iron at a tempe-

temperatures below the critical interval (about 700° C), and in any case below the temperature at which the direct proportionality between the grain size and the corresponding heating temperature ceases. To eliminate rolling defects, it is necessary to straighten shapes (strips) in the hot state.

Brief Description of the Atlas of Phototypes

Table I.

Fig. 1. Magnetic iron ore from Mount Blagodat in the Urals. Clearly crystalline forms of magnetic iron ore were found in cavities of the latter.

Fig. 2. Part of the outer crust of a very slowly cooled alloy of carbon steel \((C = 2.27\%)\). The direction of the crystallization axes is noticeably apparent in relief in the middle of the piece, while toward the periphery of the latter the crystallization axes gradually disappear.

Fig. 3. Ends of the crystallization axes on the outer upper crust of carbon steel alloy No. 1 \((C = 1.8\%,\ Si = 0.30\%,\ Mn = 0.15\%,\ Ph = 0.026,\ S = 0.022\%\). See p. 6).

Fig. 4. A polished section from a cut of D. K. Chernov’s crystal, the direction of the section of the polished specimen being perpendicular to the principal axis of the crystal. The carbon content in the iron of the indicated crystal is equal to \(0.78\%\).

Table II.

Fig. 5. A crystal of carbon steel obtained by D. K. Chernov in the shrinkage cavity of a 100 t ingot of mild steel. The casting of this ingot was carried out in a metal mold, on which there was an earthen mold for the riser portion. Owing to the slight thermal conductivity of the earthen walls of the riser, the metal solidified in it very slowly; in the shrinkage cavity of the riser of this ingot, D. K. Chernov found distinct crystals of carbon steel, and the largest of them is shown in Fig. 5. The crystal is double, the length of the large crystal reaching 39 cm, and its weight—\(8\frac{1}{2}\) lb.

Table III.

Fig. 6. The polished section shows a cut of one of D. K. Chernov’s crystals in a direction perpendicular to the long axis of the crystal. In the field of the polished section, the separation of ferrite (white bands), surrounded by pearlite, is clearly noticeable. The ferrite segregations contain a fairly considerable quantity of slag mechanically enclosed in them (black spots on the white bands of ferrite).

Fig. 7. The polished section shows a section of alloy No. 5, prepared by N. Belyaev at the Putilov Works. The direction of the section of the polished specimen is perpendicular to the outer crust. On the polished specimen the ferrite grains (white bands) are surrounded by pearlite grains. The chemical composition of the indicated alloy was as follows: 0.62% C; 0.51% Si; 0.14% Mn; 0.024% Ph and 0.019% S. The alloy was prepared in a fireclay crucible and, after melting, cooled very slowly (approximately about 60 hours).

Fig. 8. The polished section shows a section of alloy No. 8, prepared at the Izhevsk Works under conditions of very slow cooling (carbon in the alloy 0.55%). The white bands on the polished specimen are ferrite, and the dark ones are pearlite.

Note. The photographs of Figs. 2, 3, 4, 5, 6, 7 and 8 are borrowed from the atlas in the book by N. T. Belyaev, “Crystallization, Structure and Properties of Steel during Slow Cooling.”

Fig. 9 and 10. Widmanstätten structure of meteorites. The parallelograms, rectangles and triangles of the named structure are clearly visible on the polished sections shown.

Fig. 11. The polished section shows the internal structure of one of the crystals of magnetic iron ore from the piece shown in Fig. 1.

Table IV.

Fig. 12. Cast iron with a content of 3.35% C; 0.28% Si and 0.41% Mn at an intermediate cooling rate. Cementite (white bands on the polished specimen) separated along the cleavage planes of the crystals.

Fig. 13. The polished section shows the structure of austenite, the white fields on the polished specimen being austenite, and the dark bands on them—martensite, the product of the incipient decomposition of austenite.

Fig. 14. Shows the structure of martensite at high magnification. The features of the Widmanstätten structure in Figs. 11, 12, 13 and 14 are clearly noticeable.

Fig. 15. Shows the structure of soft wrought iron. The ferrite grains on the polished specimen have an oval form and are cemented together by pearlite, while the black bands represent remnants of slags.

Fig. 16. The polished section shows the structure of soft cast steel, the grains having the preceding oval form, but the element binding them (pearlite—the black mass on the polished specimen), owing to the higher carbon content as compared with the preceding soft iron, occupies a larger area on the polished specimen.

Fig. 17. Tool steel of medium hardness. The pearlitic mass of the polished specimen (dark fields on the polished specimen) is uniformly penetrated by a network (white) of ferrite. Steel without heat treatment, contains about 0.46% carbon.

Table V

Fig. 18. The polished section shows the structure of a steel rich in carbon, a tomlyanka steel. The pearlite grains have a distinctly fibrous structure, and cementite has appeared as the element binding them; its hard, mirror-like (white) surface is clearly expressed in the phototype.

Fig. 19 shows the structure of Thomas iron after annealing for 108 hours in charcoal. The structure of the cementite (white fields on the polished section) stands out clearly among the fibrous pearlite grains. The dark inclusions on the polished section belong to the carbon from the annealing. The polished sections in Figs. 18 and 19 have relief polishing.

Figs. 20 and 21. Both polished sections present the structure of boiler iron, the second polished section showing the structure of this iron after treatment at “blue heat,” and the first—without treatment. On the second polished section, the increase in ferrite grains after treatment of the boiler iron at “blue heat” is noticeable in relief, as compared with the structure of the same iron in the untreated state.

Figs. 22 and 23 show the change in the structure of soft, forgeable iron after prolonged heating in a blacksmith’s hearth for 4 and 5 days, the hearth working daily for about 8 hours. As the duration of the indicated heating increases, the oval ferrite grains grow in size, and on their surface and between them (dark spots on the polished section) accumulations of iron oxides appear.

Table VI.

Fig. 24. The polished section shows the structure of the preceding soft, forgeable iron after heating it in a blacksmith’s hearth for 7 days, 8 hours daily. The ferrite grains (white oval spots on the polished section) have noticeably increased in size. The oxidation of the metal, as compared with the preceding case, has greatly increased, and the iron oxides are arranged in the form of powerful (black) continuous bands. The rectilinear arrangement of the iron oxides may be explained by the fact that the oxidation of the metal spread most intensively along its weakest places—along the traces of the pressure of small remnants of slags squeezed out of the metal and stretched out by rolling in its subsequent direction.

Figs. 25, 26, 27 and 28 show the structure of tool steel of medium hardness (carbon 0.95%) under successive thermal treatment. In raw steel (Fig. 25), the pearlite grains (dark spots on the polished section) are closely bound by cementite (white network on the polished section). When heated to 700° C the cementite network begins gradually to disappear (Fig. 26), and the pearlite grains change their structure somewhat;

upon further heating at the end of the critical interval, the grains of pearlite and cementite together form a new allotropic form—martensite, the needles of which are clearly visible in Fig. 27. Finally, Fig. 28 shows the structure of the same steel when overheated, the steel plainly acquiring a Widmanstätten structure.

Fig. 29 shows the structure of soft cast iron under ordinary cooling. The ferritic eutectic of the cast iron (the white mass on the polished section) abounds in segregations of graphite (black cracks on the polished section).

Table VII

Fig. 30. The phototype gives a clear picture of the change in the structure of soft cast iron of the preceding chemical composition under the influence of prolonged annealing. The segregations of graphite occupy a larger area on the polished section in comparison with Fig. 29.

Fig. 31 refers to the polished section of malleable iron, in which the ferritic eutectic (white fields on the polished section) has separated into independent grains (of oval form) of ferrite, while the carbon, in the form of annealing carbon, is clustered (black spots and circles on the polished section).

Fig. 32. The phototype shows the structure of cast iron during the process of its whitening. The direction of the rays of hardening is clearly noticeable on the polished section (straight white bands), with mirror-like segregations of cementite upon them; moreover these rays of hardening gradually disappear in the remaining, unwhitened mass of soft cast iron (the dark spot on the polished section). The indicated cast iron contains 4.03% C; 1.13% Si and 0.60% Mn.

Fig. 33 shows the structure of pearlite at high magnification. The white threads consist of cementite, and the intervals between them are filled with ferrite. The steel contains 0.45% C, was heated to 800° C and slowly cooled to 700° C. Relief polishing.

Fig. 34. Granular pearlite at high magnification; the oval, elevated grains consist of cementite, while the mass filling the intervals between them consists of ferrite. The steel contains 1.24% C. Relief polishing.

Fig. 35. Troostite at high magnification. The continuous mass of troostite (white fields on the polished section) has on its surface rudiments of granules (or) martensite, into which, upon further heating, the troostite also passes. The steel contains 0.45% C, was heated to 825° C and slowly cooled to 690° C, and then quenched in water at room temperature.

Table VIII.

Figs. 36–41 show the successive change in the structure of very soft iron during the rolling of wire in the first three grooves-

them, with even-numbered figures referring to transverse sections of the specimens, and odd-numbered figures to sections along the rolling direction. The ferrite grains of the observed specimens have a typical oval shape, and no noticeable change in this shape under the influence of rolling is observed, since the rolling temperature is still considerably high.

Table IX.

Figs. 42–47 show the change in the structure of soft iron in the 4th, 5th, and 6th rolling passes. The character of the change in the ferrite grains in the 4th and 5th rolling passes remains the same as before, but in the 6th pass, owing to the noticeable cooling of the metal, the oval ferrite grains begin to be drawn out along the rolling direction, which is quite clearly noticeable on the polished section of Fig. 47.

Table X.

Figs. 48–53 show the further change in the structure of soft iron during the rolling of the wire in the 7th, 8th, and 9th passes. The ferrite grains continue to be drawn out in the direction of rolling; the direction of this elongation of the ferrite grains is very clearly visible on polished sections taken along the rolling direction. As the ferrite grains are drawn out, they are broken up into parts, with the prior formation of constrictions (see Fig. 51).

Table XI.

Figs. 54–59 represent the change in the structure of the same soft iron during further rolling of the wire in the 10th, 11th, and 12th passes. The general character of the change in the shape of the ferrite grains fully recalls the preceding picture of such change in the 8th and 9th passes, but only, under the influence of the longitudinal forces of rolling, the fibers do not have time to be drawn out sufficiently and begin to tear noticeably, with the formation of voids inside the metal. (See the black spots on the polished sections of Figs. 55, 56, 57, 58, and 59).

Table XII.

Figs. 60–65 refer to the 13th, 14th, and 15th rolling passes of the iron into the above-mentioned wire. The character of the change in the shape of the ferrite grains under the influence of rolling is entirely analogous to the general picture of the preceding table, differing only in that the ferrite grains tear noticeably more strongly along the rolling direction, as a result of which more voids are observed inside the metal in comparison with the preceding case.

T. I. Tikhonov.

Table XIII.

Figs. 66–69 show the change in the structure of soft iron during the rolling of wire in the last two passes, the 16th and 17th. The temperature at the end of rolling is very close to the beginning of the critical interval, as a result of which the ferrite grains, under the influence of the mechanical forces of rolling, are very strongly crushed and destroyed along the planes of cohesion with one another, which is quite distinctly visible in Fig. 67. Then, in the last transverse section of the wire (Fig. 68), owing to the rupture of the ferrite fibers (grains), large voids are visible inside the metal (black spots on the polished section).

Figs. 70 and 71 show the structure of wire received from the works after rolling and before drawing. The wire was first annealed and then drawn through a die in order to give it a round sectional form. The shape of the ferrite grains is almost the same both in the longitudinal (Fig. 71) and in the transverse (Fig. 70) sections.

Table XIV.

Figs. 72 and 73 show the structure of wire received from the works after rolling, but without annealing and without drawing through a die. The ferrite grains, both in the transverse section (Fig. 72 and, subsequently, even numbers) and in the longitudinal section (Fig. 73 and, subsequently, odd numbers), are noticeably smaller in size in comparison with the preceding phototypes (Figs. 70 and 71, Table XIII), which indicates that the annealing temperature of this wire was very considerable. The presence of voids within the metal, owing to rupture of the grains, is very visible on the field of the polished section.

Figs. 74, 75, 76, and 77 present the structure of the wire in transverse and longitudinal section after the first two drawings. The voids between the grains of the metal, under the influence of the lateral pressures during passage through the die of the drawplate, have somewhat decreased in comparison with those for the rolled wire (Figs. 72 and 73), but at the same time the ferrite grains begin to be drawn out intensively in the direction of drawing of the wire (compare Figs. 75 and 77).

Table XV.

Figs. 78–83 show the general picture of the change in the form of the grains of the wire during further drawing. Here the gradual elongation of the grains and their decrease in size in the transverse section is already sharply visible. The ferrite grains, as is evident, cannot retain their previously acquired oval form, owing to the strong lateral pressure during passage through the die of the drawplate. In addition, the ferrite grains, when passing through the drawplate, do not have time to be drawn out

…are drawn out and torn, forming voids inside the metal (black areas on the polished sections). Noticeable slipping occurs between the grains of the metal, accompanied by some destruction of the grains themselves, which is very clearly visible on polished sections in the direction of drawing.

Table XVI.

Figs. 84 to 87, 90 and 91 show the further change in the grains of the wire in the following holes of the draw plates. The general picture of the change in the ferrite grains of the wire remains entirely the same as in the preceding table, but the destruction of the grains, which had only just begun earlier, noticeably increases with each new hole of the draw plate, as a result of which the number of voids (black spots on the polished sections) inside the metal gradually increases.

Table XVII.

Figs. 88 and 89. In examining the section of the wire after its passage through a hole of the draw plate, an uneven size of the ferrite grains was found at the periphery and at the center of the wire section, namely: the ferrite grains in the middle of the section are smaller than at the periphery. For a more detailed illustration of this picture of the change in the ferrite grains across the wire section, photographs 88 and 89 were made, the first being taken closer to the surface of the wire, and the second from the middle of its section. Both of the latter photographs relate to one and the same section of the wire from Figs. 86 and 87. In the photographs of Table XVII the general picture of the nonuniformity of the grains in size at the middle and periphery of the wire is clearly visible, and, in addition, the general picture of deformation and destruction of the ferrite grains themselves in the wire during drawing is presented rather vividly.

Tables XVIII and XIX.

Figs. 92–103 depict the further change in the structure of the iron wire in the following holes of the draw plates, up to the last one inclusive. The general picture of the deformations inside the metal illustrated during its further drawing will be the same as that considered earlier. The only difference will be that in the last holes of the draw plates there will occur a noticeable compression of the previously formed voids inside the metal (cf., successively, Figs. 95, 97, 99, 101 and 103), although complete compaction of the metal is not achieved even in the last hole of the draw plate, as is very clearly indicated by the voids present in the transverse section of the wire from the last hole of the draw plate (black spots in Fig. 102).

Table XX.

Figs. 104 and 105. In drawing, wire, after passing through several holes of the draw-plate, acquires considerable stiffness and brittleness, and for further drawing it is necessary first to anneal it. To illustrate the picture of the change in the structure of the wire after such annealing in the middle of the drawing process (after the 5th hole of the draw-plate), Figs. 104 and 105 are given; the first phototype shows the structure of the wire after slight annealing, and the second—without annealing. Annealing caused an increase in the grains, but owing to the slight height of the heating, the nonuniformity in the size of the grains in the central parts and those close to the periphery of the section was preserved even after annealing, which is clearly noticeable in Fig. 104.

Figs. 106–109 show the structure of the body of a rivet; moreover, Figs. 106 and 108 were taken along the axis of the rivet: the first in the middle of the head of the rivet itself, and the second in its cylindrical shank. Figs. 107 and 109 relate to the same places of the rivet to which Figs. 106 and 108 relate, but were taken in a direction perpendicular to the length of the rivet. On comparing Figs. 108 and 109 with 106 and 107, one notices a strong decrease in the size of the ferrite grains in the head of the rivet, as compared with its shank, and besides this, in the head of the rivet there is also noticeable and significant fragmentation of the grains of the metal, with the formation of voids inside the latter (see the black spots in Fig. 106).

Table XXI.

Figs. 110 and 111 present the structure of a machine-made iron nail of \(1/8''\) in longitudinal and transverse section near the head. Between the grains of the metal there is observed a considerable number of voids (black spots on the polished sections).

Figs. 112 and 113 illustrate the structure of a machine-made nail of \(1/16''\) in longitudinal and transverse sections near the head of the nail. The general character of the structure of the grains of the metal is quite the same as that for the nail of \(1/8''\), but the grains are considerably smaller, and therefore in Fig. 112 the direction of the change of the grains during the formation of the cap (head) of the nail is distinctly noticeable in relief.

Fig. 114. Macrostructure of the riveted seam for a rivet of \(22\,m/m\) and a thickness of the sheets being joined of \(8\,m/m\), with hydraulic riveting.

Fig. 115. Macrostructure of the riveted seam for a rivet of \(18\,m/m\) and a thickness of the sheets being joined of \(8\,m/m\), with hand riveting by experienced boilermakers of one of the large factories in the south of Russia. On comparing the last two macrostructures, it is noticeable that with hand riveting the stratification of the metal inside the rivet is considerably deeper than with hydraulic riveting. The general character of the dispo-

the arrangement of the grains in the body of the rivet head fully resembles the structure of the heads of the nails considered earlier.

Table XXII.

The polished section in Fig. 116 shows the structure of soft iron after considerable smith forging. Under the action of the hammer blows, the oval ferrite grains assumed a wedge-like form in the direction of the greatest action of the forces on the grains, or in the direction of least resistance to deformation of the grains.

Figs. 117, 118, 119, and 120 illustrate the general picture of the deformations of iron sheets when holes are punched in them with punches; the first gives a general picture of such deformations, and the three following ones show them at the entry, middle position, and exit of the punch. The magnitude of such deformations, as is evident, gradually increases toward the exit of the punch, and then the deformed part of the metal, under the influence of annealing, separates from the sound part, which is very clearly noticeable in Fig. 120.

Fig. 121 shows the external form of the punches and their destruction in experiments on the question of the influence of the form of the punch on the magnitude of the deformation of the sheet being pierced. The punch on the right in the phototype had an angle of sharpening at the cutting edges of \(65^\circ\) and possessed a middle guiding cone. The holes pierced by this punch had ragged edges, and then the punch itself was badly damaged during operation (part of its head lies in the middle between the punches). The next punch to the left of the one considered had no guiding cone; the angle at its cutting edges was \(75^\circ\). The punch produced ragged holes and itself broke down. The third punch to the left had an angle at the cutting edges equal to \(87^\circ\) and was provided with a guiding cone; in operation this punch proved the best—it produced smaller deformations of the metal at the edges of the sheets pierced by it. Finally, the fourth punch, the leftmost one, had a small guiding cone on the smooth end face, provided with an annular cutting edge on the cone with a sharpening angle of \(75^\circ\) and a width of the edge itself at its base of about \(2.5\,\mathrm{m/m}\). The cutting edge of this punch quickly broke down.

Note. The conventional designations under the phototypes \(100:1\), \(150:1\), etc., denote magnifications of 100, 150, etc., times.

Noticed Misprints

Pages	Lines	Printed	Should read
7	3 from bottom	Metallography	T. Tikhonov. Metallography
18	1 from bottom	ferritic	cementite
19	1 from top	ferritic	cementite
19	8 from bottom	are changed	is changed
20	20 from top	points	point
21	14 from bottom	fe b	Feβ
25	1 from top	specimen	specimens
26	11 from bottom	cooling	cooling
31	7 from top	consequence	consequence
”	23 from top	8 tabl. XXIV	8 tabl. XXVI
35	14 from top	grains	on
46	9 from top	punch	punch
65	18 from bottom	using	using
70	20 from top	isolation	isolation
72	22 from top	contributes	contributes

Fig. 1.

Fig. 2. 1:1

Fig. 3. 15:1.

Fig. 4. 2:1

Phot. I. Pavlov.

Change in the structure of iron during mechanical treatment. T. I. Tikhonov.

Fig. 5. \(\times 1/2\).

Photo by I. Pavlov.

Change in the structure of iron during mechanical
treatment. T. I. Tikhonov.

Plate III.

Fig. 6. 50:1.

Fig. 7. 90:1.

Fig. 8. 6:1.

Fig. 9. 10:1.

Fig. 10. 10:1.

Fig. 11. 100:1.

Pavlov’s volume II.

Study of the structure of iron during mechanical processing. T. I. Tikhonov.

Plate IV.

Fig. 12. 100:1.

Fig. 13. 1000:1.

Fig. 14. 1000:1.

Fig. 15. 100:1.

Fig. 16. 100:1.

Fig. 17. 100:1.

Phot. P. Pavlov.

Change in the structure of iron under mechanical treatment. G. I. Tikhonov.

Fig. 18. 100 : 1.

Fig. 19. 270 : 1.

Fig. 20. 100 : 1.

Fig. 21. 100 : 1.

Fig. 22. 100 : 1.

Fig. 23. 100 : 1.

Photo by P. Pavlov.

Change in the structure of iron during mechanical treatment. G. I. Tikhonov.

Table VI

Fig. 24. 100 : 1.

Fig. 25. 100 : 1.

Fig. 26. 100 : 1.

Fig. 27. 100 : 1.

Fig. 28. 100 : 1.

Fig. 29. 90 : 1.

Phot. P. Pavlov.

Change in the structure of iron during mechanical treatment. T. I. Tikhonov.

Table VI

Fig. 30. 100:1.

Fig. 31. 100:1.

Fig. 32. 100:1.

Fig. 33. 1000:1.

Fig. 34. 1000:1.

Fig. 35. 1000:1.

Phot. P. Pavlov.

Change in the structure of iron under mechanical treatment. T. I. Tikhonov.

Fig. 36. 100 : 1.

Fig. 37. 100 : 1.

Fig. 38. 100 : 1.

Fig. 39. 100 : 1.

Fig. 40. 100 : 1.

Fig. 41. 100 : 1.

Photo by I. Pavlov.

Change in the structure of iron under mechanical treatment. T. I. Tikhonov.

Table IX.

Fig. 42. 100 : 1.

Fig. 43. 100 : 1.

Fig. 44. 100 : 1.

Fig. 45. 100 : 1.

Fig. 46. 100 : 1.

Fig. 47. 100 : 1.

Phot. P. Pavlova.

Change in the structure of iron during mechanical treatment. T. I. Tikhonov.

Table X,

Fig. 48. 100:1.

Fig. 49. 100:1.

Fig. 50. 100:1.

Fig. 51. 100:1.

Fig. 52. 100:1.

Fig. 53. 100:1.

Phot. P. Pavlov.

Change in the structure of iron during mechanical treatment. T. I. Tikhonov.

Plate XI.

Fig. 54. 100:1.

Fig. 55. 100:1.

Fig. 56. 100:1.

Fig. 57. 100:1.

Fig. 58. 100:1.

Fig. 59. 100:1.

Phot. P. Pavlov.

Changes in the structure of iron during mechanical working. T. I. Tikhonov.

Table X

Fig. 60. 100 : 1.

Fig. 61. 100 : 1.

Fig. 62. 100 : 1.

Fig. 63. 100 : 1.

Fig. 64. 100 : 1.

Fig. 65. 100 : 1.

Photo by P. Pavlov.

Change in the structure of iron during mechanical treatment. T. P. Tikhonov.

Table XIII.

Fig. 66. 100 : 1.

Fig. 67. 100 : 1.

Fig. 68. 100 : 1.

Fig. 69. 100 : 1.

Fig. 70. 150 : 1.

Fig. 71. 150 : 1.

Phot. P. Pavlov.

Change in the structure of iron during mechanical treatment. T. I. Tikhonov.

Table XIV.

Fig. 72. 150 : 1.

Fig. 73. 150 : 1.

Fig. 74. 150 : 1.

Fig. 75. 150 : 1.

Fig. 76. 150 : 1.

Fig. 77. 150 : 1.

Photo by I. Ivanov.

Change in the structure of iron during mechanical treatment. G. I. Tikhonov.

Plate XV.

Fig. 78. 150 : 1.
Fig. 79. 150 : 1.

Fig. 80. 150 : 1.
Fig. 81. 150 : 1.

Fig. 82. 150 : 1.
Fig. 83. 150 : 1.

Phot. I. Pavlov.

Change in the structure of iron during mechanical treatment. T. I. Tikhonov.

Table X

Fig. 84. 150 : 1.

Fig. 85. 150 : 1.

Fig. 86. 150 : 1.

Fig. 87. 150 : 1.

Fig. 90. 150 : 1.

Fig. 91. 150 : 1.

Phot. I. Pavlov.

Change in the structure of iron under mechanical working. T. I. Tikhonov.

Table XVI

Fig. 88. 800 : 1.

Fig. 89. 800 : 1.

Photo by P. Pavlov.

Change in the structure of iron during mechanical treatment. T. I. Tikhonov.

Table XVII.

Fig. 92. 150:1.

Fig. 93. 150:1.

Fig. 94. 150:1.

Fig. 95. 150:1.

Fig. 96. 150:1.

Fig. 97. 150:1.

Phot. P. Pavlov.

Change in the structure of iron during mechanical working. T. I. Tikhonov.

Table XIX

Fig. 98. 150:1.

Fig. 99. 150:1.

Fig. 100. 150:1.

Fig. 101. 150:1.

Fig. 102. 150:1.

Fig. 103. 150:1.

Phot. P. Pavlov.

Change in the structure of iron during mechanical treatment. T. I. Tikhonov.

Table X

Fig. 104. 150 : 1.

Fig. 105. 150 : 1.

Fig. 106. 150 : 1.

Fig. 107. 150 : 1.

Fig. 108. 150 : 1.

Fig. 109. 150 : 1.

Photo by P. Pavlov.

Change in the structure of iron under mechanical treatment, G. I. Tikhonov.

Table XXI

Fig. 110. 150 : 1.

Fig. 111. 150 : 1.

Fig. 112. 150 : 1.

Fig. 113. 150 : 1.

Fig. 114. 1.75 : 1.

Fig. 115. 1.75 : 1.

Photo by I. Pavlov.

Change in the structure of malleable cast iron during mechanical processing. G. I. Tikhonov.

Table XX

Fig. 116. 100:1.

Fig. 117. 3:1.

Fig. 118. 100:1.

Fig. 119. 100:1.

Fig. 120. 100:1.

Fig. 121.

Photo by I. Pavlov.

Change in the structure of iron under mechanical treatment. T. I. Tikhonov.

Determine the structure of iron in mechanical processing.

Gr. Timofeev.

Table XXIII.

Fig. 1.
Fig. 6.

Vertical axis: filtration of \(C\).
Horizontal axis: grain size.

Fig. 2.
Fig. 5.

Horizontal axis: grain size in \(\mathrm{mm}^2\).

Visible graph annotations:

• Fig. 2: \(C = 0.06\%\), \(Mn = 0.11\%\), \(Si = 0.075\%\)
• Fig. 5: \(C = 0.55\%\), \(Mn = 0.53\%\), \(Si = 0.130\%\)

Visible numeric axis marks:

• Fig. 2 vertical scale: 700, 800, 900, 1000, 1100, 1200, 1300, 1400
• Fig. 2 horizontal marks: 0.0289; 0.1214; 0.2611
• Fig. 5 vertical scale: 700, 800, 900, 1000, 1100, 1200, 1300, 1400
• Fig. 5 horizontal marks: 0.0099; 0.0738; 0.1323

Drawing 5.
Table XXIV

Drawing 4.

Magnetic structure of iron under mechanical treatment.
V. Tikhonov.

Table XXV.

Fig. 7.

Diagram labels: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17.

Fig. 15.

Graph labels: A., B., C.

Vertical axis: reading

Horizontal axis: change

Change in the structure of iron during mechanical treatment

M. [[unclear: signature]]

Stamp: Tomsk Order of the Red Banner of Labor Polytechnic Institute named after S. M. Kirov. Scientific and Technical Library.

Plate XXV.

Trial measurement of the number and areas of the grains.
Specimen 8.

\[ f=\frac{400\,m^{2}}{100000\cdot x}\,? \]

Right-hand notes:
number of grains; \(m_k^2\) — square [[unclear: abbreviated word]]; number of grains; diameter; \(500\) microns; \(0.11\ \mathrm{mm}^2\).

Left-hand scale: number of grains per \(4\ \mathrm{cm}^2\) of alloy.

No.	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17
Number of grains	36	47	77	83	81	90	106	127	155	190	190	197	242	272	285	381	500
\(f\), \(\mathrm{mm}^2\)	1.1	0.85	0.55	0.46	0.47	0.44	0.40	0.31	0.29	0.21	0.21	0.20	0.16	0.15	0.14	0.12	0.11

Change in the structure of iron during mechanical working.
M. [[unclear: surname]]

Fig. 9. Table XXVII.

Vertical axis: \(m/m^2\); number per \(1\) sq. millimeter.

Horizontal axis: row number.

Legend:

— Curve DV — for transverse sections
- - - Curve SZV — for transverse sections
- · - · Curve SZR — for longitudinal sections
- · · - · Curve S — for longitudinal sections

Fig. 10.

Vertical axis: number per \(1\) sq. millimeter.

Horizontal axis: row number.

Legend:

— summary curve of transverse sections
- - - longitudinal sections
- · - · summary curve

Change in the structure of iron under mechanical working.

Eng. Mikheev.

Polarization [[unclear: “satellite”]].
Table XXVIII

Fig. 11.

Fig. 14.

Fig. 14a.

Fig. 12.

Fig. 13.

Change in the structure of the electrode during mechanical treatment.

Art. Tikhonov.

Correlation of \(M, m\) of the areas and masses of meteoritic craters and meteoritic structures of the land of the Earth. XXIX.

Figure labels:

• ejecta
• along the azimuth of the fall
• fall trajectory [[unclear]]
• \(m\)
• \(M\)

Panel labels visible in the diagram:

• bl. I
• bl. II
• bl. III
• bl. IV
• bl. V
• bl. VI

Visible numerical annotations include:

• 15, 25, 35, 45, 55, 65, 75, 85
• 52, 55, 62
• 76, 71, 65, 60, 36, 21
• 100, 89, 91, 80, 80, 95, 92, 96, 110, 110, 105, 102, 122, 149
• 94, 49, 44, 39, 34, 29, 24, 19, 14, 9
• 100, 112, 135, 25, 32, 65, 91, 141, 170, 250, 333
• 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95
• 68, 68, 66, 64, 60, 58, 55, 50, 41, 36, 23, 14, 6
• 3, 8, 13, 18, 23, 28, 33, 38, 43, 48, 53, 58, 63, 68
• 82, 50, 98, 98, 98, 22, 22, 60
• 4, 9, 14, 19, 24, 29, 34
• 100, 120, 120, 110, 110, 115
• 24, 28, 32, 36, 40, 44
• 18, 39, 58, 80, 99, 111, 120
• 35, 40, 45, 50, 55, 60, 65, 70
• 18, 65, 56, 92, 112
• 12, 17, 22, 27, 32, 37, 42, 47, 52, 57, 62, 67
• 82, 22, 60, 48, 14, 14, 44, 54, 41, 74, 95, 38

Change in the structure of iron under mechanical treatment

[[unclear: handwritten notation]]

Table XXX.

Heat treatment	Tempering temperatures / series shown
Annealing	350, 700, 850, 1000, 1100, 1200
Quenching in water, 20°. Quenching temperature 850°. Tempering.	350, 500, 650
Quenching in oil, 80°. Quenching temperature 850°. Tempering.	500, 550
Quenching in water, 20°. Quenching temperature 1000°. Tempering.	350, 550, 650
Quenching in oil, 80°. Quenching temperature 1000°. Tempering.	500, 550

Plot labels and scales visible in the table-figure

• Strength, kg/mm²
• Group I. \(C = 0.84\%\)
• Group II. \(C = 0.39\%\)
• Elastic limit, kg/mm²
• Group III. \(C = 0.20\%\)
• Group IV. \(C = 0.09\%\)
• Elongation in %

Bottom series numbers visible: 1, 2, 3, 4, 5, 6, 7; 12, 13, 14, 15; 23, 24, 25; 16, 17, 18, 19; 26, 27, 28.

Legend:

• Group I. \(C = 0.84\%\) — solid line
• Group II. \(C = 0.39\%\) — dashed line
• Group III. \(C = 0.20\%\) — dash-dot line
• Group IV. \(C = 0.09\%\) — dotted line

1. A. L. Baboshin. Transactions of the I.R.T.S., 1907, Issues IX–X. ↩↩↩

2. Transformation du fer et du carbone dans les fers, les aciers et les fontes blanches par F. Osmond. Paris. 1888. ↩↩

3. Zeitschrift d. Vereins d. Ingenieure. 1900. No. 14. ↩↩↩

4. The specimens of wire examined during rolling belong to the Alexandrovsk plant of the Bryansk Society, while the specimens for wire drawing belong to the neighboring plant of the Society at Gantke, which takes wire from the former plant. ↩↩↩

5. Howe. The Metallurgy of Steel. p. 51. ↩↩↩

6. Tar'ev. Homogeneity of Rails. p. 101. ↩↩↩↩↩

7. Tar'ev. Homogeneity of Rails. p. 188. ↩↩↩

8. Tagéev. Uniformity of rails. P. 112. ↩↩↩

9. T. Tikhonov. Metallography. p. 27. ↩↩

10. M. Considères. L’emploi du fer et de l’acier. p. 118. ↩↩

11. Thorston. Report on cold-rolled iron and steel. Pittsburgh 1878. ↩↩

12. See Metallic Alloys. T. Tikhonov. 2nd ed., 1910. ↩↩

13. Investigation of Rail Steel. Proceedings of the Institute of Ways of Communication, 1906, Art. 3. ↩↩

14. Belyaev. Crystallization of Steel. P. 17. 1909. ↩

15. Belyaev. Crystallization of steel. Table VII. ↩

16. Belyaev. Crystallization of steel. Table VIII. ↩

17. A triangle is obtained when the plane of the section is arranged parallel to the surface of the octahedron. A rectangle or square is obtained when the plane of the section is parallel to the plane of the cube, and a parallelogram when the plane of the section is arranged parallel to the surface of the dodecahedron. See E. Cohen, Meteoritenkunde, p. 79. Belyaev, Crystallization of Steel, p. 23. 1909. The structural elements of Widmanstätten figures in meteorites consist of solid alloys of nickel in iron and iron in nickel, of kamacite and taenite, or of both together in the form of a eutectic mixture under the name plessite. Kamacite, taenite, and plessite in meteorites are analogous in steel to ferrite, cementite, and pearlite. ↩

18. Journal of Iron and Steel Inst. 1903. p. II, p. 66. ↩

19. The Iron and Steel Metal. 1904. No. 2. ↩

20. Journal of the Russian Metallurgical Society. 1910. Col. 979. ↩

21. Ueber die Veränderung der Elasticitätsgrenze. Mittheilungen aus dem Mechanisch. Laboratorium in München Heft, XIII. 1886. ↩