EXTERNAL COEFFICIENT OF FRICTION OF WELD METAL ON CAST IRON, STEEL, AND BRONZE

V. I. Strokopytov.

Introduction

The method of electric-arc welding, as a means of surfacing metallic surfaces in machine building, is widely used. It is enough to cite several examples to confirm this statement. Surfacing by electric welding is used as the principal method for repairing worn rolling rolls, wagon tires, shaft journals, rails, etc. Up to the present time, there have been no indications in the literature, either Soviet or foreign, as to how metal deposited by the electric-arc method with a metal electrode behaves with respect to abrasion. In particular, up to the present time there have been no indications as to the external coefficient of friction between the deposited metal and metals not subjected to welding.

The purpose of the present work is to determine the external coefficient of friction between metal deposited by the electric-arc method with a metal electrode, on the one hand, and steel, cast iron, and bronze, on the other.

Taking into account the complexity of the question of choosing a suitable method for determining the external coefficient of friction, in this first part of the work we limited ourselves to a laboratory method.

Research Method and Material

As an apparatus for determining the external coefficient of friction between metallic surfaces, a modified design of the apparatus first used by Yakobi was employed.

The setup with which the present work was carried out is shown in Fig. 1. A horizontal disk A, made of steel, cast iron, bronze, or of metal deposited by the electric-arc method with a metal electrode, 200 mm in diameter, was set in rotation by a shunt electric motor through a gear reducer and belt drive. On the vertical axis B, fixed rigidly in the disk, a bushing C was freely fitted, held from above by a nut and locknut. On the bushing C, which had a small flange, a lever D was fastened; in it there was an opening E, into which a specimen of the metal being tested was inserted and clamped with a thumb screw. The lever D could rotate not only in the horizontal plane about the axis B, but also in the vertical plane YY. At point K, with the aid of a hook, a load was suspended, pressing the specimen against the disk. To measure the friction force, the lever D was connected at point H with a dynamometer by means of a practically inextensible thread. The dynamometer made it possible to measure the tensile force with an accuracy of up to 5 g, which gave an error in calculating the external coefficient of friction of up to 0.05%. Calibration of the dynamometer was carried out at the place of installation after each test, which ensured the obtaining of sufficiently accurate determinations of the force.

Amontons’ mathematical law, usually attributed to Coulomb, is formulated as follows:

\[ F=\mu P, \tag{1} \]

where \(F\) is the friction force, \(P\) is the normal load (load per unit area), and \(\mu\) is the coefficient of proportionality, called the coefficient of external friction.

A distinction is made between static friction (friction at rest) and kinetic friction (friction of motion). In the first case, in relation (1), \(F\) is understood as the force that must be applied in order to set one body in motion relative to another. In the case of kinetic friction, \(F\) is understood as the force needed to maintain the relative motion of the rubbing bodies (sliding) at a given velocity. Amontons’ law is generally considered the fundamental law of both static and kinetic friction, for “dry” (unlubricated) as well as lubricated surfaces.

For our case, let us denote the load applied to the end of lever \(K\) by \(P_0\); the force with which the dynamometer spring is stretched by \(F_0\); the load on the specimen due to the dead weight of the lever, hook, etc., by \(P_1\); and the pressing force on the specimen from the load \(P_0\) by \(P_2\). Then, according to Fig. 2, we may write:

\[ Fc=F_0d \tag{2} \]

\[ P_0a=P_1b \tag{3} \]

\[ P=P_1+P_2 \tag{4} \]

On the basis of these equalities, expression (1) then takes the following form:

\[ \mu=\frac{ad}{bc}\cdot\frac{F_0}{P_0+\frac{a}{b}P_1} \tag{5} \]

Substituting the values of the lever-arm dimensions and \(P_1=425\ \text{g}\), found experimentally, we obtain the formula for determining the external coefficient of sliding friction:

\[ \mu=0.771\,\frac{F_0}{P_0+118.2}. \tag{6} \]

In calculating by formula (6), we neglected friction in the bushing, since it had practically no effect on the dynamometer readings.

To prepare specimens from metal deposited by electric welding, a series of beads was deposited on a boiler-iron plate measuring \(100\times100\times15\ \text{mm}\). Standard welding wire of \(d=5\ \text{mm}\), with a chalk coating, was used as the electrode. Electric welding was carried out with direct current on an Elin machine at a current of 180 A and a voltage of 35 V.

After each layer of deposited metal, the weld metal was thoroughly cleaned of slag and scale, and only then were the subsequent beads applied. The plate thus built up was planed flush, and from the remaining deposited metal round specimens \(5\ \text{mm}\) in diameter and \(40\ \text{mm}\) high were cut. One end of the specimen was machined into a cone. The sharp end of the specimen was slightly blunted, polished with No. 0000 emery paper, and then polished on cloth to a mirror finish.

As the sliding surfaces, disks were taken: of gray cast iron, structural steel, bronze, and metal deposited by the electric-arc method. The last disk was made in the following manner. Onto a steel disk, three layers of metal were deposited around the circumference, with

Fig. 1.

Fig. 2.

Fig. 3.

Fig. 5.

Fig. 4.

under the same conditions and regime as in making the specimens. The upper layer of the deposited metal was turned off on a lathe. The surfaces of the disks were carefully ground. The Brinell hardness of the material used for making the disks is given in Table 1.

Table 1.

The microstructure of the specimens and of the wheel made of deposited metal is shown in Fig. 3, and consists of a ferritic field with not quite clearly expressed boundaries of the ferrite grains. The ferrite field is covered with a considerable number of excess nitride needles, $\mathrm{Fe}_3\mathrm{N}$, differently oriented with respect to direction and of various sizes both in length and in thickness, precipitated from the solution of nitrogen in $\alpha$-iron. The dark areas visible in the photomicrograph are the eutectoid of iron–nitrogen alloys—braunite.

The microstructure of the metal of the wheel made of structural steel is shown in Fig. 4 and consists of pearlite grains surrounded by a thick, discontinuous network of ferrite. The pearlite is finely lamellar. The microstructure of the metal of the wheel made of cast iron is the characteristic structure of gray cast iron. The matrix is pearlitic-ferritic with inclusions of veins of lamellar graphite. In places, islands of globular graphite separation and inclusions of phosphide eutectic are observed.

The microstructure of the metal of the wheel made of bronze is a sharply expressed dendritic structure of an $\alpha$-solid solution. At higher magnifications, in addition to this dendritic structure of the $\alpha$-phase, islands of eutectoid are also noticeable in the form of mottled areas, sharply separated from the principal $\alpha$-phase by clearly delineated boundaries and a lighter shade.

The experiments for determining the external coefficient of sliding were carried out as follows. The specimen under test was inserted into the fastening point $E$ (Fig. 1), then the lever together with the specimen was placed on the disk, the corresponding load was suspended from hook $K$, and the motor was switched on; for ten minutes the disk moved, during which time the specimen became sufficiently well lapped to the disk and the dynamometer showed only small fluctuations in the force readings. After this time had elapsed, the force was read from the dynamometer, the motor was stopped, the correctness of the dynamometer readings was checked, and the area of contact of the specimen with the disk was measured. The area was measured under a microscope with a micrometric eyepiece (the contact area in most cases was a regular circle).

The determination of the coefficient of external friction at the start of motion was carried out in an analogous manner. The specimen, fixed on the lever at point $E$, was lowered onto the plane of the disk, a load was suspended from hook $K$, and then the wheel was slowly turned by hand. During rotation, the wheel carried the specimen with the lever along with it until the tensile force of the dynamometer spring became equal to the friction force reduced to point $H$ of the lever; at this point the specimen slipped over the surface of the disk, and the maximum

the deflection of the dynamometer pointer showed the force corresponding to the frictional force at the moment of starting from rest.

Influence of Specific Pressure and Speed on the External Coefficient of Friction

Before proceeding to determine the external coefficient of friction of weld metal on other metals, and in order not to fall into error, preliminary investigations were carried out to ascertain the influence of the specific pressure and the speed of rotation of the rubbing surfaces on the external coefficient of friction.

The data available in the literature on the dependence of the kinetic friction of “dry” surfaces on the speeds of their relative motion are highly contradictory. Coulomb notes that the speed of motion has very little influence on the kinetic friction of metal. According to the data of Hond and Yamod (2), the coefficient of friction \(\mu k\) remains constant when the speed of relative motion is varied in the interval from 15 to 40 m per minute.

Galton and Westinghouse (3), as well as Smith (4), indicate, however, that as the speed of relative motion of the rubbing surfaces decreases, the external coefficient of kinetic friction increases.

We carried out experiments to determine the influence of the speed of relative motion of rubbing surfaces on the external coefficient of friction between metal produced by electric-arc surfacing. The experiments were conducted at constant specific pressure and at different speeds of relative motion. The speeds were taken from 0.51 to 27.3 cm/sec. The results of the experiments are given in Table 2.

Table 2.

It is evident from the table that, with a change in the speed of relative motion of the rubbing surfaces, the external coefficient of friction also changes.

However, the experiments showed that the degree of cleanliness of the rubbing surfaces has an extremely strong influence on the magnitude of the external coefficient of kinetic friction. Namely: if the rubbing surfaces were cleaned, washed with alcohol or ether, and dried insufficiently thoroughly, then with the passage of time or with an increase in the relative speed of the rubbing surfaces the external coefficient of friction increased noticeably. If, however, the surfaces were thoroughly cleaned and just as thoroughly dried, then successive measurements at the speeds indicated in Table 2 in all cases gave a constant value of the external coefficient of friction for metal deposited by the electric-arc method—in this case it was equal to 0.540. These experiments testify quite clearly that the external coefficient of kinetic friction does not depend directly on the speed; the increase in the external coefficient of friction observed in the case of insufficiently clean surfaces is due, of course, not to a change in speed, but to the gradual removal, in the process of friction, of various particles contaminating the surfaces, formed during friction, from one surface to the other. This circumstance is also indicated by the experiments of Jacob (5), Biry and Bowden (6), and Kuhn (7).

Measurements of the coefficient of external friction of carefully cleaned surfaces made of metal deposited by the electric-arc method, at different values of the normal load applied to the surfaces and at a definite relative sliding velocity for our experiments, are given in Table 3. The values of the coefficient of external friction for each specific pressure were obtained as the arithmetic mean of 40 observations.

Table 3.

From Table 3 it is evident that, with an increase in the normal load, the coefficient of external friction increases somewhat. But, as will be seen below, the increase in the coefficient of external friction for metal produced by the electric-arc method is not due to an increase in specific pressure. From this it may be concluded that the kinetic coefficient of external friction does not depend on the specific pressure; Amontons’ law thus proves valid also for metal deposited by the electric-arc method with a metallic electrode.

The apparent increase in the external coefficient with increasing specific pressure is explained by the fact that, as the specific pressure increases, more intensive wear of the adsorbed films on the rubbing surfaces takes place, as well as very intensive wear of the hard surface itself. One may suppose that, in contrast to the case of static friction, in kinetic friction this wear of the surface itself plays a very substantial role. In our experiments the following phenomenon was observed. If a specimen is placed on a freshly polished disk well wiped with alcohol, then at the beginning of starting the motor the dynamometer readings have a certain minimum value, which during the first minute increases at first rapidly and then more and more slowly. It turned out that, in the process of friction of one surface against another, intensive wear of the rubbing surfaces themselves occurs. On the rotating disk it was possible to observe a large quantity of hard grains of metal, and on the rubbing plane of the specimen, traces of scratches.

Comparison of the particles, as a product of abrasion obtained in our experiments, with the particles of metal obtained during grinding showed that there is a substantial difference between them. The particles of metal during grinding, under the microscope, have the appearance of shavings with a metallic hue. Those particles that were obtained in our experiments, under the microscope, showed a picture of elongated figures with straight lines. The surface of these particles was completely smooth and had a dark color.

It was assumed that these particles are nothing other than needles of a nitrogen compound with iron (\(Fe_3N\)). Indeed, metallographic investigations of the surface of specimens of deposited metal produced by the electric-arc method and subjected to kinetic friction showed, in this case, an increased content of nitride inclusions in comparison with the quantity that existed before grinding. In Fig. 5 a microphotograph is presented of the plane of the metal subjected to abrasion. In this figure a large number of nitride needles is visible in comparison with the quantity that was present before the experiment (see Fig. 3). It is characteristic to note that the greater part of the nitride needles

Fig. 6.

Visible labels in the figure: \(\Delta T^\circ\); load in grams; curves \(I\), \(II\), \(III\).

Fig. 7.

Visible labels in the figure: \(\Delta T^\circ\); \(v\), cm/sec; curves \(I\), \(II\), \(III\).

Fig. 8.

Visible labels in the figure: \(\mu\); experiment numbers.

is located in the middle of the ferrite grain and is absent along its boundaries. According to Fink’s investigations (1929), the mutual friction of dry iron and steel surfaces is not only a purely mechanical process, but also has a chemical aspect. Fink found that, during rolling friction in the presence of air, iron oxide forms on the rubbing surfaces.

The investigations of Shaw and Giltencamp (8) showed that, under severe friction, not only oxygen but also nitrogen from the air enters into combination with iron. The authors explain this by the fact that, at high pressure, which is sometimes accompanied by a considerable rise in temperature, the metal surface becomes active and capable of reactions.

Froster (9) indicates that, in the preparation of metallographic polished sections, the surface of the section becomes saturated with nitrogen. The reason for such absorption of nitrogen by iron should be sought in the activation of atmospheric nitrogen. The splitting of inert nitrogen molecules becomes possible owing to the concentration of the frictional energy that develops during polishing in the thin surface layer of the metal. Experiments carried out in 1936 by Bowden (10) to measure the temperature of rubbing surfaces showed that a sufficiently high temperature develops in the plane of contact. Figures 6 and 7 give curves of the dependence of the temperature of rubbing surfaces on the load \(P\) and the relative velocity \(V\), obtained by Bowden for the case of “dry” friction of a constantan cylinder against a steel plane. The range of velocities was from 4 to 5000 cm/sec, and of loads up to 120 g.

Consequently, nitrogen activated in this way (atomic nitrogen) is intensely absorbed by the metal surface, which has in turn been activated by the polishing process.

Metallographic investigation does not give a clear answer to the question of the form in which nitrogen is absorbed by the metal. Probably, under the influence of the local rise in temperature and of work hardening in the polished planes, a supersaturated solution of nitrogen is formed in the deformed ferrite crystals. A considerable part of the nitride needles, as was shown by Dobrovidov (11), is located directly along the very slip planes of the ferrite crystals, forming something like spikes on a smooth surface, which impede the movement of one plane relative to another. This makes understandable the brittleness and spalling of the surface layer of the rubbing surfaces in our experiments.

Friction of Dissimilar Surfaces.

According to the data of a number of authors who studied the dry friction of dissimilar surfaces under ordinary experimental conditions, the coefficient of static friction of substance \(A\) against substance \(B\), \((\mu_s)_{AB}\), has a value intermediate between the values of the coefficients of friction of pairs of homogeneous surfaces \([(\mu_s)_{AA}\) and \((\mu_s)_{BB}]\). An especially careful investigation of the friction of dissimilar surfaces was carried out by Tomlinson (12), who measured the coefficients of static friction for 55 pairs of different solid surfaces. The data he obtained are given in Table 4. In all cases the inequality is satisfied:

\[ (\mu_s)_{AA} < (\mu_s)_{BA} < (\mu_s)_{BB} \]

In the case of kinetic friction of dissimilar surfaces, the picture is greatly complicated by the circumstance that the coefficient of kinetic friction of substance \(A\) against substance \(B\) has a substantially different value depending on which material constitutes the smaller of the rubbing surfaces and which the larger one (the plane). Thus, for example, in the experiments of Beer and Bowden (6), during the friction of nickel balls against a steel surfac—

Table 4.

Values of \(\mu_s\) according to Tomlinson.

… yielded values of \(\mu_k = 0.66\); for the friction of steel balls on a nickel surface, \(\mu_k = 0.49\). Numerical values for the friction of dissimilar surfaces are given in Table 5. Beer and Bowden note that the external coefficient of kinetic friction \(\mu_k\) assumes larger values in the case when the smaller of the rubbing surfaces consists of a softer material. If the smaller of the rubbing surfaces is made of a soft material, then the coefficient of friction increases as the hardness of the second rubbing surface decreases. If, however, the ball is made of a hard material, then the opposite regularity is observed—\(\mu_k\) increases as the hardness of the second surface increases. Indeed, in the case when the smaller of the surfaces is of a soft metal, it itself is polished in the process of friction; when the hardness of the second surface decreases, it is polished to a lesser degree, which leads to an increase in the external coefficient of kinetic friction \(\mu_k\). In the case when the smaller of the surfaces is of a hard metal, it itself is not polished during friction, but polishes the second surface.

Table 5.

The coefficient of kinetic friction of substance \(A\) on substance \(B\) thus proves to be different from the coefficient of friction of substance \(B\) on substance \(A\). The simple regularities observed in static friction are, in this case, greatly complicated by the accompanying polishing of the surfaces during friction.

Coefficient of external friction of metal deposited by the electric-arc method on other metals.

The influence of particles formed in the grinding process on the coefficient of external friction was noted above. In order to obtain more stable dynamometer readings, in our experiments we continuously removed particles of wear with chamois leather, which throughout the entire duration of the experiment …

was pressed against the rotating disk. Having set ourselves the aim of determining the coefficient of friction under conditions that would be usual for dry friction, i.e., surfaces that had been preliminarily ground were run in for one minute. In this process small scratches appeared on the rubbing surfaces, visible under a microscope and making the surfaces matte. On planes prepared in this way the coefficient of external friction was determined; the linear velocity of the rubbing surfaces was 27 cm/sec, and the specific pressure was 8.3 kg/mm². The experiments were carried out with 15 specimens for each pair of rubbing surfaces, and the coefficient of external friction was obtained as the arithmetic mean of 15 determinations.

The probable error of the mean result is ±5%. The coefficient of external friction at the moment of starting was determined as mentioned above.

The mean values obtained in this way for the coefficient of external friction of metal deposited by the electric-arc method on other metals are given in Table 6.

Table 6.

Mean values do not characterize the coefficient of external friction sufficiently accurately, since in individual cases the latter deviates considerably from the mean value. In Fig. 8 the individual values of the coefficient of friction of weld metal on steel at the moment of starting from rest are shown graphically. Here the circles indicate the mean values from ten for each specimen. From the graph it is clearly seen to what degree the coefficients of friction differ at different points of the disk and for different specimens. The mean value of the coefficient of friction here is 0.510, but in places it reaches maximum values of 0.550 and falls to a minimum of 0.480. Thus, from this example it is evident that it is impossible to indicate a definite value of the coefficient of external friction of weld metal on other metals. Therefore we confine ourselves to giving Table 7, in which the limits of the probable value of the coefficient of external friction are indicated for metal deposited by the electric-arc method on cast iron, steel, weld metal, and bronze.

Table 7.

Comparison of the external coefficient of friction of metal deposited by the electric-arc method and of electrode wire.

In order to determine the effect of the nitride needles present in metal deposited by the electric-arc method on the external coefficient of friction, comparative tests were carried out. For this purpose, 10 specimens were made from weld metal, and the same number of specimens from the wire that served as the electrode for producing the weld metal of the first specimens. The tests were carried out on steel and bronze disks. In exactly the same way as described above, before taking the reading on the dynamometer, the specimen was first run in against the disk after the latter had been ground with emery paper No. 0000. To avoid the influence of damage to the disk surface, the specimens were alternated: first a specimen of deposited metal was used, then one of electrode-wire metal.

Thus, 10 readings were obtained for sliding friction and 100 for the external coefficient of friction at the moment of starting from rest; their mean values are given in Table 8.

Table 8.

The data presented show with complete clarity the effect of nitride needles in metal deposited by the electric-arc method on the external coefficient of friction.

Conclusion.

1. On the basis of the investigations in this work, it has been established that the external coefficient of friction between planes of metal deposited by the electric-arc method does not depend on the specific pressure or on the velocity of relative motion.

2. The external coefficient of friction has been determined between metal deposited by the electric-arc method, on the one hand, and steel, cast iron, and bronze, on the other.

3. The increase in the external coefficient of friction in metal deposited by the electric-arc method is caused by the presence in the weld metal of a large number of nitride needles.

In conclusion, I express my gratitude to student A. V. Streltsov for his participation in the experimental performance of some of the tests in the present work.

Literature

1. Charlotte Iakob, Ann. d. Phys., 32, 126, 1912.
2. Honda a. Iamada, Journ. Inst, Metals 1, 49, 1935.
3. Galtona, Westinghouse. Engineering, 26, 153, 1873.
4. Archbutta, Deeby, Lubrication and lubricauts, 1927.
5. Charlotta Iakob, Diss, Königsberg, 1911.
6. Bearla Bowden, Phil. Trans., A, 234, 329, 1934.
7. Kunin N. F., Journal of Technical Physics, 3, 1099, 1933.
8. H. Schöttk’y und H. Heltenkamp, St. u. Eis. 15, 444, 1936.
9. H. Wiester, Archiv f. d. Eisenhüttenwesen, IV, 525, 1936.
10. Bowden a. Rideer, Proc. Poy. Soc., A, 154, 640, 1936.
11. Dobrovidov A. N., Journal of Applied Physics IV, 1927.
12. Tomlinson, Phil. Mag., 7, 905, 1925.