Cold Welding of Cast Iron

N. D. Tyutëva

Introduction

Purpose of the Study

The author set herself the goal of obtaining a high-quality weld without signs of a chilled zone, the appearance of which is the principal complication in cold welding of cast iron. The research thinking of most welders working in this field was directed either toward the use of hot welding of cast iron, or toward the effort to use costly welding methods with acutely scarce electrodes.

If one takes into account the relative cheapness of the material that has to be welded, it would be logical to settle on the cheapest welding method. Such a relatively inexpensive welding method is arc welding of cast iron using coated cast-iron electrodes.

The method of welding defects in small cast-iron castings and cast-iron castings of medium weight—gas welding, with preheating of the part—is the most expensive method, since it requires the use of acetylene and oxygen and, at the same time, is rather slow.

A comparatively inexpensive welding method is arc welding of cast iron. The relative simplicity of arc welding of cast iron was the reason that led the author to focus her attention precisely on this welding method.

Further efforts were directed toward finding a definite chemical composition of cast-iron electrodes and a composition of coatings. The chemical composition of the electrodes and the chemical composition of the coatings, in combination with a definite working procedure, were to ensure the production of a sufficiently high-quality weld.

The present work sets forth the results obtained when using the above-mentioned welding method, i.e., arc cold welding of cast iron, or welding with preheating to 200–300° using coated cast-iron electrodes.

It should be noted that we settled on the use of alternating current in welding cast iron.

Research Methodology

Until now, many experiments on welding cast iron have been carried out in the following manner: so-called “gray cast iron” of any chemical composition was taken, and an attempt was made to select an electrode for it by varying the compositions of the electrodes and the coating. Therefore, unsuccessful experiments in welding cast iron can be explained by the extreme complexity of the physicochemical foundations of cast iron, and sometimes by the fact that

because researchers engaged in welding carried out welding without taking the structure of the material into account.

In order to approach the welding of cast iron rationally, it is necessary to weld not random cast irons, but cast irons that have a constant composition with respect to manganese and carbon and a variable composition with respect to silicon.

It is most rational to take synthetic cast irons, which is what the author did.

Cast Iron as a Material for Welding

The general concept “gray cast iron” unites a whole series of alloys. Alloys bearing the name gray cast irons have common features in composition and structure—namely, the presence, in various combinations, of iron, carbon, and silicon. In terms of structure, the common feature is the presence of graphite inclusions. Cast irons containing in their microstructure simultaneously bound carbon in the form of cementite and free carbon in the form of graphite are called mottled or transitional.

The name white cast iron is retained for cast irons whose structure contains all the carbon only in bound form, in the form of iron carbide.

Cast irons that have free carbon in the form of graphite in one or another form of separation and do not have free cementite in their structure are called gray cast irons.

Microscopic examination of gray cast irons indicates the presence in them of four principal structural constituents: graphite, pearlite, ferrite, and steadite (a binary or ternary phosphide eutectic). In addition to the indicated structural constituents, gray-iron castings contain a small quantity of inclusions of manganese sulfide.

The principal metallic mass of a gray-iron casting consists of the following structural constituents in various combinations: ferrite, pearlite, and steadite, plus certain nonmetallic inclusions in small quantity.

Ferrite in free form occurs in cast irons with a low content of combined carbon. In the case of a coarse-grained structure of cast iron, ferrite can be observed in the form of a rim around large graphite plates. With a fine-grained structure of cast iron, the small graphite inclusions are located in ferrite (in such cases we have the so-called graphite eutectic).

Gray cast irons with a large amount of ferrite are relatively soft and readily lend themselves to machining. Ferritic gray cast irons, owing to the plasticity of ferrite, possess greater toughness than cast irons with a nonferritic structure of the principal metallic mass.

Iron carbide $\mathrm{Fe}_3\mathrm{C}$ does not occur in free form in gray cast irons, entering only into the composition of pearlite.

Cementite—a metastable phase—under certain temperature conditions decomposes into ferrite and graphite.

Under certain cooling conditions, gray-iron castings may, in structure, consist exclusively of ferrite and graphite.

Pearlite in cast irons, like pearlite in steels, has the appearance of a lamellar eutectoid, and the mechanical properties of pearlite depend to a considerable degree on the size of the ferrite and cementite plates in it.

Thus, for example, finely lamellar pearlite is distinguished by greater strength in comparison with coarsely lamellar pearlite. The formation of pearlite from austenite of eutectoid composition in industrial cast irons occurs at a temperature of 719–732° with a positive thermal effect and a change in magnetic properties.

Steadite. All the phosphorus in cast iron is present in the form of steadite. Steadite is a binary or ternary eutectic of the composition:

\[ \mathrm{Fe} + \mathrm{Fe}_3\mathrm{P} \]

or

\[ \mathrm{Fe} + \mathrm{Fe}_3\mathrm{P} + \mathrm{Fe}_3\mathrm{C} \]

The melting temperature of steadite is about \(982^\circ\).

Graphite. As a result of investigations by a number of authors [1; 2], it has been established that the mechanical properties of gray foundry cast irons are determined mainly by the quantity, size, shape, and nature of the distribution of graphite inclusions.

By type, graphite precipitates are divided as follows: a) lamellar, uniformly distributed; b) rosette-like, cluster-like; c) primary graphite (extra-coarse); d) dendritic; e) dendritic with preferential orientation.

Graphite precipitates are also classified by size: a) coarse; b) medium; c) fine; d) very fine; e) point-like [3].

When silicon is deficient, after cooling and crystallization of the cast iron, carbon remains in bound form as cementite.

The presence of a sufficient quantity of silicon can ensure the decomposition of iron carbide into ferrite and carbon in the form of graphite.

Cementite is a metastable phase; ferrite and graphite are stable within the range from room temperature to the temperature of the eutectoid transformation. There are indications of complete graphitization of a low-carbon steel containing 0.15% carbon. This steel was taken from a pipe of a cracking-plant distillation still and had been heated to a temperature somewhat below the eutectoid temperature for three years.

Principal Methods of Welding Cast Iron

Welding of cast iron is divided into cold and hot welding. Hot welding is commonly understood to mean welding with preheating of the part being welded within the range of \(500—900^\circ\). Cold welding is carried out entirely without preheating of the part being welded; a third variant is then possible—semi-hot welding with preheating to \(500—700^\circ\), and the so-called “welding with preheating,” when the part is slightly preheated within the range of \(200^\circ\). In both cold and hot welding, in turn, several variants are possible depending on the electrode material. As applied to the welding of cast iron, these variants may be as follows:

1. The electrode material is identical in composition to the cast iron being welded.
2. The electrode material is cast iron, but has a chemical composition different from that of the material of the part being welded.

The electrode material is usually richer in the graphite-forming element—silicon—and poorer in the element that impedes graphitization—manganese.

1. Cast iron may be welded with a steel electrode.
2. Welding with copper, nickel-silver electrodes, or electrodes made of Monel metal.

Hot welding with cast-iron electrodes is a process complicated by the necessity of uniformly preheating bulky parts to comparatively high temperatures, on the order of \(500—900^\circ\). The process is extremely inconvenient and expensive.

The technological process of cold welding is considerably simpler and cheaper, but in cold welding the cooling rate of the weld is much greater than the cooling rate of the part during its manufacture by casting.

and, as a consequence, extremely undesirable structural changes may be obtained in the weld—a chilled weld; the latter imparts considerably reduced mechanical properties (brittleness) to the weld. The presence of cementite in a gray-iron weld is an entirely inadmissible phenomenon, since under impact loading cracks will proceed first of all along the cementite inclusions, and only afterward along the ferrite inclusions.

In the case of small overall dimensions of the part, correction of the structure of a chilled weld by subsequent heat treatment of the weld is possible; this latter circumstance also argues for the advantage of cold welding over hot and semi-hot welding. The application of various methods of welding cast iron in recent years has been aimed chiefly at welding up defects in iron castings and at repair work on cast-iron articles.

Graphitization of the Weld

Graphitization of a cast-iron weld depends on the combination of the following circumstances: its composition, melting conditions, pouring conditions, and cooling rate. As a first approximation it is possible to draw an analogy between the cooling of a weld and a casting in a permanent metal mold (chill). Of course, there is a certain difference, namely: between the wall of the mold and the surface of the casting there is close contact only at the first moment, while the poured metal has not yet solidified. As soon as a crust forms, a thin air gap immediately forms between the surface of the casting and the walls of the chill as a result of shrinkage, and the transfer of heat into the chill is greatly slowed. When a weld solidifies, until the very last moment a very close contact remains between the weld and the material of the part, and heat transfer proceeds faster than in casting into molds. But as a first approximation one may use experimental data on permanent-mold casting to judge what composition the weld should have so that, during cooling, it will not become chilled as a result of the rapid removal of heat into the cold part being welded.

N. P. Dubinin [4], in order to determine the influence of the chemical composition of cast iron poured into cast-iron molds on the structure of the casting, poured rectangular specimens 5 to 40 mm thick, 70 mm wide, and 210 mm long into cast-iron molds heated to 170–200°, with a wall thickness of 25 mm. It turned out that a ferritic structure of the casting (whose thickness was 10 mm) is obtained with a silicon content of 4.5% and a carbon content of 2.5%. As the carbon content increases, the required silicon content decreases; thus, with 4.6% carbon, only 2.5% silicon is required. Pearlitic cast iron can be obtained, respectively, with 4% silicon and 2.5% carbon, and with 2% silicon and 4.6% carbon.

Of course, these experiments give only an approximate idea of the chemical composition required for a weld with one structure or another.

As was already stated at the time, the cooling rate of a weld is considerably greater than the cooling rate of a permanent-mold casting.

N. P. Dubinin [4] gives, for a casting 5 mm thick, a formula for the cooling rate in the interval 1150–100° (after the lapse of \(t\) sec.):

\[ v = 2{,}348\, l^{-0{,}00255 t} \]

\(t\)—cooling time in seconds.

Our observations of the cooling rate of welds 3–5 mm thick, deposited on cold cast-iron plates 25 mm thick, showed that the cooling rate of the welds is at least 2–3 times greater than that calculated by this formula. The cooling rate was determined on the surface

ness in the temperature interval 1000–700° by observing the change in temperature with an optical pyrometer.

Therefore the silicon content in our electrodes (with an average carbon content of 3%), which give ferritic welds when deposited on ferritic plates, proved to be about 5%, and when deposited on pearlitic cast irons—about 6%.

In the latter case the pearlitic cast iron of the part, mixing with the ferritic electrode (provided that the silicon in the mixture is less than 5%), can give, even if very small, nevertheless extremely undesirable inclusions of ledeburitic cementite. This explains why it is so difficult to weld with a steel electrode, which in the remelting process is alloyed from the coating with silicon and carbon. The carburization of such an electrode can with difficulty be brought up to 2.5% C, and alloying with silicon to 3%, which, under rapid cooling, gives a pearlite–graphite structure in which, with the slightest acceleration of cooling or a change in composition toward a decrease in the silicon and carbon content, cementite appears.

Conclusions

1. The weld metal has a higher temperature at the beginning of solidification than ordinary cast iron.

2. Owing to the higher heating temperature, there is less reason to expect a large number of crystallization centers in the form of graphite nuclei.

3. Therefore it may be assumed that graphitization proceeds at the expense of decomposition of the solid phase.

4. Carburization of a steel electrode to cast iron probably proceeds through the action of the gas phase, and not of solid graphite, which also excludes the presence of ready-made crystallization centers.

Materials and Method of Work

Synthetic and Gray Foundry Cast Irons

The question of cold electric-arc welding of cast iron, in view of the specific properties of the latter, is among the most difficult problems of welding technology. Chilling during the cooling of the weld, low strength and toughness create the prerequisites for the formation of cracks in welds and zones of increased hardness.

We posed the following questions for study:

1. The possibility of using steel electrodes with an alloying coating for cold welding of cast iron.

2. The possibility of using coated cast-iron electrodes with various contents of the principal alloying element (silicon) for cold welding of cast iron.

The experimental part of the work was begun with the preparation of cast-iron specimens for subsequent experimental surfacing. Plates of synthetic cast iron measuring 15×70×200 mm were cast.

For the preparation of synthetic cast iron, U9A grade steel was melted in a graphite crucible with a capacity of 5 kg.

At a temperature of 1500° the melt was carburized while being stirred with electrode carbon. At first the dissolution of the carbon rod in the melt proceeded very rapidly, but then the carburization process slowed down, so that stirring had to be carried out for 1–2 hours.

In addition to the indicated method of producing synthetic cast iron, the latter was also prepared from good Ural roofing iron. The iron cuttings were first carburized to a carbon content of 1.5%. Carburization was carried out in a carburizer of ordinary composition, made of charcoal and potash. The iron cuttings carburized in this way were then melted and carburized in a graphite crucible in the same manner as in the preceding method of producing synthetic cast iron from U9A steel. As the analysis showed, in both cases white cast iron of almost identical composition was obtained. Several melts produced by the methods described above were, in turn, remelted in a large crucible with a capacity of 35 kg.

The synthetic cast iron obtained in this way had the following chemical composition (in percent).

$C_{\text{total}}$ %	Si %	Mn %	S %	P %
3.70	0.52	0.40	0.02	0.04

The cast iron, cast into small pigs, served as the starting material for making electrodes and plates for experimental welding. In addition, we made electrodes and welding plates from gray foundry (coke) cast iron of grades 0 and 1.

Taking the synthetic cast irons as a basis, we prepared, by remelting in a small crucible (capacity 5 kg), welding plates of the following dimensions: thickness 15–20 mm, width and length 70 × 200 mm. This thickness of the plates was chosen because, in practice, parts of precisely this thickness are encountered (ordinary machine castings). The plates were cast in a vertical position in green sand molds. Ferrosilicon was used for alloying the plates and electrodes obtained in this way.

Chemical analysis of the melts cast for experimental welding gave the results presented in Table 1.

Table 1

No.	$C_{\text{total}}$ %	Si %	Mn %	S %	P %
1	2.80	2.12	0.40	—	—
2	2.53	4.23	0.36	0.016	0.006
3	—	0.88	0.07	—	—
4	2.68	1.33	0.31	—	—
5	3.35	0.93	0.08	0.030	0.013
6	2.28	1.76	0.32	—	—
7	2.53	3.00	0.27	—	—
8	2.48	4.53	0.32	—	—
9	3.16	1.82	0.07	—	—
10	2.62	2.89	0.39	—	—

The electrodes were made from the same synthetic and coke gray cast iron of grade 0. Ferrosilicon was used as the alloying element. The electrode rods were cast in a dry clay mold. The rod dimensions were \(d = 5\) mm and length 250 mm. As the experiments showed, it is quite possible to cast electrodes in metal molds. The analysis of the electrode material and the description of the structures are given in Table 2.

Table 2

Microstructure	Melt No.	Si%	Al%	Ni%
Ferrite, a little pearlite	1	4.9	1.0	—
Ferrite	2	8.7	1.5	—
Ferrite	3	8.0	—	—
Ferrite + pearlite	4	4.8	—	—
Pearlite, a little cementite	5	4.6	—	1.8
Pearlite, a little ferrite	6	6.2	—	1.3
Ferrite	7	6.7	—	—

To improve the conditions of graphitization (obtaining fine suspended graphite), not only silicon, but sometimes aluminum and nickel as well, were introduced into the electrode melts.

The structures of some electrodes are shown in Figs. 1 and 2.

Fig. 1

Fig. 2

The microstructure of the electrode material shown in Fig. 1 is a ferritic base metallic mass with inclusions of coarse lamellar graphite.

The microstructure of the electrode material shown in Fig. 2 is ferrite with inclusions of a small amount of pearlite and graphite.

A total of eight electrode melts were cast. The chemical composition and microstructure of electrodes of various melts are given in Table 2.

Owing to the use of a dry mold for casting the electrodes, their structures change very little across the section; the nature of the graphite precipitates and of the basic metallic mass also differ very little across the section. The structure at the edge and in the middle of the specimen was almost identical. Electrodes cast in metal molds had a sharply variable structure across the section. Electrodes with a low silicon content were slightly chilled in this case, but the latter circumstance had no effect on the welding process. The available selection of plates of different chemical composition and of electrodes of several variants in composition and microstructure for experimental welding made it possible to determine the feasibility and results of welding with various combinations of electrodes and base material.

Theoretical prerequisites for selecting the electrode structure

The quality of the seam, the welding results, and above all lack of fusion depend on the nature of melting of the electrode on the one hand and on the type of arc on the other.

A rapidly melting electrode, flowing off in large drops, as is known, does not give good penetration. Insufficient penetration is also produced by a long arc and an arc of low power.

If the polarity of the arc does not correspond to the character of ionization of the arc space (for example, in the case of direct current), lack of penetration is also possible.

Taking into account the influence of the character of electrode melting on penetration, it is very important first to determine how a cast-iron electrode of various structures should melt.

The theory of the melting of gray cast iron was given by N. V. Gutovskii [5]. In brief it amounts to the following: “In the melting of gray cast iron three periods should be distinguished:

1. Carburization of the solid metal.
2. Formation of a liquid phase by carburization.
3. Melting proper.”

The process of carburization of austenite by graphite in the first period is precisely what determines the melting temperature of cast iron, which is variable for one and the same cast iron depending on the rate of heating. The temperature at the end of melting obviously depends on the course of the second and third periods and therefore is also not constant, but depends on the rate of heating and on the character of the structure.

Fig. 3

The finer-grained the graphite, the more uniformly it is distributed in the principal metallic mass, the faster, all other circumstances being equal, carburization will proceed, and the lower the melting temperature will be.

Obviously, with infinitely rapid heating, the melting temperature of pearlitic cast iron will correspond to the melting temperature of pearlitic steel, and that of ferritic cast iron—to that of pure iron. Ferritic cast iron, with very rapid heating, will melt at 1530°.

In addition, graphite may float up in molten cast iron. This will retard carburization and raise the melting temperature. Again, the more entangled and branched the graphite (graphite eutectic, Fig. 3), the more difficult it is for it to float up, the longer it will remain in contact with the unmelted parts of the structure, and the more vigorous carburization will be. In the words of N. V. Gutovskii [5]: “The temperatures of the beginning and end of melting of gray cast iron depend on the amount of combined carbon in the cast iron, on the rate of heating, and on the influence—difficult to take into account—of spel formation.”

Adopting N. V. Gutovskii’s point of view, it must be recognized that pearlitic electrodes will melt fastest of all, and purely ferritic ones slowest of all. This is confirmed by direct experiment. We tried welding with cast-iron electrodes of various compositions without coating. To make arc burning possible on alternating current, the electrodes and the welding site were covered with a thin layer of potash. The experiment showed that a ferritic electrode melts approximately 20 percent more slowly than a pearlitic one, gives drops of smaller size, and provides better penetration.

From these considerations, electrodes having a ferrite + graphite microstructure may be recommended.

Electrode Coatings

Practical testing of electrode coatings recommended in the literature

We tested several variants of electrode coatings recommended in the literature for welding cast iron, as well as coatings whose compositional variants were proposed by us (Table 3).

In the process of welding with electrodes having these coatings, we paid attention, on the one hand, to the behavior of the coating during melting of the electrode and, on the other, to the results of surfacing. We were especially interested in the microstructure of the deposited metal. The coating formulations, the microstructure of the electrodes used, and the results of the deposits are brought together by us in Table 3.

Table 3

Coating compositions	Quantity, wt. parts	Microstructure of the electrode	Results of surfacing
1. Wood charcoal Graphite Ferrosilicon Aluminum Chalk Feldspar Sodium water glass with an addition of 10% potassium water glass.	8 10 15 1 9 15	The coating was used with electrodes of the 1st, 4th, and 5th melts 1. Ferrite, a little pearlite 4. Ferrite + pearlite 5. Pearlite, a little ledeburite	The coating is refractory, but it is quite possible to work with it
2. Wood charcoal Borax Chalk Ferrosilicon	25 8 12 20	The coating was used with the electrode of the 4th melt 4. Ferrite + pearlite	The coating is refractory Welding is impossible The deposited metal, in structure, is gray cast iron
3. Wood charcoal Ferrosilicon Scale Aluminum Chalk	20 15 6 9 12	The coating was used with the electrode of the 4th melt 4. Ferrite + pearlite	The coating melts poorly Welding is almost impossible The deposited metal, in structure, is gray cast iron
4. Graphite Carborundum Barium carbonate Chalk	40 60 5 5	Coating for steel electrodes Structure—ferrite, a little pearlite	The structure of the deposit contains 1.7% C; traces of graphite and ledeburite are present; the deposit is unsatisfactory in structure
5. First layer of coating Wood charcoal Ferrosilicon Over the first layer, up to 1.0 mm thick, a wrapping of iron wire (0.6) with gaps equal to the diameter.	50 50	Result of welding with steel electrodes wrapped with wire, using the indicated three-layer coating

Coating compositions	Number of constituent parts by weight	Electrode microstructure	Surfacing results
6. Second coating layer Charcoal Ferrosilicon Chalk Water glass. Third coating layer Pure graphite on water glass	5 5 5		The deposit is steel with a carbon content of 1.7%; sometimes white-gray cast iron with 2.2% C. Microstructure of the steel electrode: ferrite + a small amount of pearlite. The arc held well
7. Coating for electrodes wrapped with wire First coating layer Charcoal Ferrosilicon Chalk Aluminum Second coating layer Charcoal Ferrosilicon Aluminum Scale Third layer Pure graphite on water glass	5 5 3 4 5 5 1 4	Steel electrodes wrapped with wire	White-gray cast iron with 2.0–2.2% carbon. After annealing—gray pearlitic-ferritic cast iron
8. Coating for electrodes wrapped with wire. First coating layer Charcoal Ferrosilicon Chalk Aluminum Scale Second coating layer Charcoal Ferrosilicon Aluminum Scale Third layer Pure graphite on water glass.	5 5 3 1 4 5 5 1 4	Steel electrodes	Refractory coating One of the deposited seams showed lack of penetration. The other seam showed entirely satisfactory penetration. In some cases, welding with these electrodes produced white-gray cast iron

Conclusions for § 1

1. Into coating 1 we introduced aluminum, as a strong deoxidizer and graphitizer, and partly potassium water glass to provide good ionization.

Feldspar was introduced to increase the amount of slag (protection of the weld from cooling). In subsequent coatings, feldspar was discarded as needlessly reducing the concentration of the most important constituents of the coating—silicon and carbon.

1. A coating with borax is attractive because of the easy adjustability of the melting temperature and the possibility, if necessary, of lowering it to 600–700°.

Borax conducts current poorly; the ionization of the arc is poor. Welding with an ordinary transformer is impossible.

1. Coatings composed according to the type of the Järko coatings (6; 7) produced no effect.

2. Coatings with carborundum did not give good results.

3. In two-layer coatings, the reactions take place in the first layer; the second layer protects these reactions from the influence of atmospheric oxygen and nitrogen.

Potassium water glass was added to the first layer, where good ionization is needed. The thickness of the coatings of the first and second layers is the same. Feldspar should not be used in coatings.

1. Three-layer coating.

In order to improve the alloying of the steel, we tried to use a three-layer coating, with an inner winding of 0.5% C steel wire.

The first layer provides alloying and ionization (for ionization, wood charcoal is boiled in a solution of potash).

The second layer—protection and alloying; the third—only protection, as in the two-layer coating.

Ultimately, we adopted coatings for cast-iron electrodes on the basis of experiments to determine the transfer of carbon and silicon from the coating into the weld (electrode material).

Practice of Mixing the Constituent Parts of Electrode Coatings

The quality of the electrode coating depends to a large extent on the quality of preparation of the dry mixture. All constituent parts included in the dry mixture were crushed and sifted through appropriate sieves. For ferroalloys (ferrosilicon), chalk, and organic substances, a sieve with 1200 openings per square centimeter is recommended.

After grinding, the constituent parts are mixed, and then water-diluted liquid glass is poured in, and the whole mass is again stirred and ground in a large porcelain mortar.

The coating was applied in two layers by dipping. After the first layer, the coated electrode was dried in air until the coating hardened; then the second layer was applied and subjected to air drying for 20 hours in the laboratory at \(+20^\circ\mathrm{C}\), after which the electrodes were calcined at a temperature of 100–120°C for 1 hour.

Classification of Coatings and Requirements Imposed on Them

The basic purpose of all coatings is to improve the quality of the deposited metal. Coatings are divided into thin, or ionizing, and thick, or quality, coatings.

Thick, or quality, coatings change the chemical composition, structure, and mechanical properties of the deposited metal; moreover, their action is manifested fully at a certain definite layer thickness, which is why they are called thick coatings.

High-quality coatings must meet the following requirements:

1. They must produce a slag that protects the molten metal from the action of atmospheric air and slows its cooling. Practically all high-quality coatings satisfy this requirement.

2. During melting they must produce a layer of gas necessary, on the one hand, to protect the molten metal from the action of atmospheric air and, on the other hand, to increase the stability of the arc; this requirement is not met by coatings that contain no gas-forming components.

3. The melting temperature of the coating must lie within limits close to the melting temperature of the electrode; otherwise, with an easily fusible coating, exposure of the end of the electrode and nonuniform distribution of slag over individual sections of the weld are possible, while with an excessively refractory coating the electrode metal melts away and the projecting part of the coating forms too deep a cup. This leads to lengthening of the arc and ultimately to frequent interruptions of it.

4. The coating must ensure stable burning of the arc when welding with alternating current.

5. The coating must adhere tightly to the electrode, covering it with a uniform layer, and must not crack.

The slag obtained as a result of melting the coating must meet the following requirements:

1. Flowing over the deposited metal, it must cover it with a uniform layer; consequently, the slag must not be too viscous. Practically all high-quality coatings meet this requirement.

2. The slag must cover the deposited metal and thereby slow the rate of its cooling so that the gases dissolved in the metal and the nonmetallic inclusions are removed from it more completely.

3. The slag must reduce or absorb the metal oxides, i.e., react with the deposited metal, deoxidizing it and thereby improving its mechanical properties.

4. The solidified slag must be easily removable from the deposited metal.

Coatings Recommended by the Author

As the binding component for the electrode-coating mixture we use water glass.

According to the provisional technical specifications, soluble glass used as the binding component for electrode coatings must have a modulus of 2.8 (9), and it is still better if the modulus is equal to 3.0.

The analysis of the glass was carried out by the method described by P. K. Grigoryev.

As the alloying component, ferrosilicon (80% Si) was introduced into our coatings. As is known, ferrosilicon reacts with excess, uncombined Na₂O, SiO₂. The reaction is accompanied by the evolution of hydrogen (in the presence of water)

$$ 2\mathrm{NaOH} + \mathrm{Si} \longrightarrow \mathrm{Na}_2\mathrm{SiO}_2 + 2\mathrm{H}_2 $$

This reaction proceeds slowly, accelerating somewhat with an increase in temperature; the latter is especially troublesome, since the hydrogen, being liberated during drying of the thickened coating, makes it porous. During drying the coating increases in volume by 50–100%. The gas bubbles, when the coating melts, increase severalfold; the liquid slag has the character of a foam, and alloying of the molten metal proceeds very weakly. Porous

the coating has low thermal conductivity and does not have time to be heated through by the heat of the arc; therefore part of it remains unmelted and does not participate in the reactions of the welding zone.

By passivating the ferrosilicon we attempted to impede the interaction of the ferrosilicon with the liquid glass, but this measure, while reducing the porosity of the coating, did not completely eliminate the evolution of hydrogen. Passivation consisted in heating, with access of air, the prepared fine ferrosilicon powder to various temperatures. As is known, on the surface of individual crystals of the ferroalloy, even at room temperature, an oxide film 20–30 Å thick forms very rapidly (within hundredths of a second).

Experiments on passivation of ferrosilicon powder at 600–700° for one hour showed that hydrogen evolution in a coating with passivated ferrosilicon decreases rather strongly, although it does not cease completely. Raising the passivation temperature and increasing the duration did not reduce the activity of the ferrosilicon to the extent necessary for complete elimination of hydrogen evolution, and we were forced to follow another route—increasing the modulus of \( \mathrm{Na_2O} \, m \, \mathrm{SiO_2} \).

During passivation, apparently, the \( \mathrm{SiO_2} \) film increases in thickness only very slightly; at least, it was not possible to detect any thickening of the film from the increase in weight of the powder after passivation.

As experiments showed, when glass with a modulus of 2.9–3.2 is used, hydrogen evolution is not observed at all. Increasing the modulus of the liquid glass under laboratory conditions was carried out as follows: from the glass, by the action of hydrochloric acid on it, a gel was separated. After washing, it was dissolved in liquid glass (with heating in a water bath) in the required amount. In this way it was possible to obtain glass with a modulus of 3.4–3.6.

In order to increase ionization, in some experiments we used potassium soluble glass (\(\mathrm{K_2O} \, m \, \mathrm{SiO_2}\)) by dissolving silica gel in a solution of caustic potash. From the phase diagram \(\mathrm{Na_2O—SiO_2}\) it follows that the melting point of \(\mathrm{Na_2O \cdot SiO_2}\) is 1090°, and that of \(\mathrm{Na_2O \cdot 2SiO_2}\) is 874°. A completely different regularity is observed in the \(\mathrm{K_2O—SiO_2}\) system. \(\mathrm{K_2O \cdot SiO_2}\) melts at 980°, and \(\mathrm{K_2O \cdot 2SiO_2}\) at 1030°. But the eutectics \(\mathrm{Na_2O \cdot 2SiO_2 + SiO_2}\) and \(\mathrm{K_2O \cdot 2SiO_2 + SiO_2}\) melt at fairly low temperatures, 790 and 570°. The latter circumstance is very important, since the coating for cast-iron electrodes must be more readily fusible than that for steel electrodes. On the other hand, this makes it possible, in coating cast-iron electrodes, to rely on a slag of very simple composition, since during oxidation of the part of the ferrosilicon that has not passed into the molten metal, the amount of \(\mathrm{SiO_2}\) in the liquid coating increases, and the melting point of the slag decreases up to a content of 84% \(\mathrm{SiO_2}\) (the eutectic \(\mathrm{Na_2Si_3O_5 + SiO_2}\)), remaining low even at a content of 87% \(\mathrm{SiO_2}\) (880°).

Of course, in addition, part of the oxidized iron also passes into the molten coating (from the ferrosilicon and, probably, from the electrode), but its influence on the change in the melting point of the slag will not be decisive.

In the ternary system \(\mathrm{Na_2O \cdot SiO_2 — SiO_2 — Fe_2O_3}\), the melting point of the eutectic consisting of \(\mathrm{Na_2O \cdot SiO_2}\), aegirine \((\mathrm{Na_2O \cdot Fe_2O_3 \cdot 4SiO_2})\), and quartz, corresponding to the weight percentages:

\[ \begin{aligned} \mathrm{SiO_2} &= 72.9\%,\\ \mathrm{Fe_2O_3} &= 4.2\%,\\ \mathrm{Na_2O} &= 23.7\%, \end{aligned} \]

is 760°.

In ferrosilicon the ratio between the contents of SiO₂ and Fe is 4 : 1. Taking into account the great activity of silicon with respect to oxygen, the Fe₂O₃ content can hardly be expected to exceed 4–5% (this was also confirmed by analysis of the slag), and it should be considered that iron does not greatly raise the melting temperature of the slag formed during melting of our coating.

Taking the foregoing into account, we chose soluble sodium glass with a modulus of 3.0 for making all coatings for steel and cast-iron electrodes.

One of the alloying constituents of the coating is ferrosilicon. It alloys the steel electrode with silicon and makes up for the loss of silicon burning out in the cast-iron electrode.

The second alloying element, performing an analogous role, is carbon, which may be taken in the form of graphite, charcoal, or in the form of carborundum. In the latter case carborundum (SiC) will alloy the weld simultaneously with silicon. In both cases part of the silicon passes into the alloy, while part, burning to SiO₂, dissolves in the slag. Carbon, burning, escapes into the atmosphere in the form of CO or CO₂, while another part alloys the alloy.

Oxides of manganese, iron, and silicon may pass into the slag from the burning electrode. But the use of thick coatings almost completely excludes any noticeable oxidation of the electrode.

Therefore, finally, the slag of our coatings consists of Na₂O, Fe₂O₃, and SiO₂. This is very good, since such a slag is easily calculated on the basis of available experimental data. As the carbon-alloying constituent of the coating, as indicated, graphite or charcoal may be used. Graphite would be entirely suitable for this purpose, but its weak reactivity in some cases makes it preferable to use charcoal powder. This powder contains enormous quantities of adsorbed gases, which may be released during drying and melting of the coating and cause the same difficulties as hydrogen, of which we have already spoken. In order to deprive the charcoal powder to some extent of its ability to absorb gases and then release them, we applied the following treatment to this powder: the powder is boiled for 3 hours in a 20% solution of potash (K₂CO₃) in water. After boiling down, the powder is transferred to a filter and then dried in a drying cabinet at 120° and then ground again. Experience has shown that such a powder contains very little gas and almost does not absorb it. On the other hand, the potash remaining in the pores of the charcoal considerably increases its activity (its ability to carburize).

The interaction of potash and charcoal ensures the production of an atmosphere containing gas that very effectively cements iron, in accordance with the following reactions:

\[ \begin{aligned} 1)\;& \mathrm{K_2CO_3} \rightleftarrows \mathrm{K_2O} + \mathrm{CO_2} \\ 2)\;& \mathrm{CO_2} + \mathrm{C} \rightleftarrows 2\,\mathrm{CO} \\ 3)\;& 2\,\mathrm{CO} + \mathrm{C} \rightleftarrows \mathrm{CO_2} + \text{austenite}. \end{aligned} \]

The presence of K₂CO₃ also has a favorable effect on the ionization of the arc space, since potash is a very good ionizer.

Coatings containing charcoal treated in this way do not require the addition of ionizers (chalk, marble, etc.). The arc burns quite steadily and is very easily struck. If graphite is used instead of charcoal, marble must be introduced as an ionizer. Marble is better than chalk, since a coating with marble

less porous than with chalk. If possible, it may be recommended that part of the sodium silicate in the coating be replaced by potassium silicate, which ensures exceptionally good ionization. At the same time, the presence of \(K_2O\) in the slag lowers the melting temperature of the latter, which, as we noted in due course, is of very great importance in welding cast iron.

The melting temperature of a cast-iron electrode is \(1200\)—\(1300^\circ\).

The melting temperature of the coating should be somewhat lower \((1000^\circ)\).

It is rather difficult to say which reacts more strongly in the process of carburizing the metal—the liquid metal–liquid slag, or the liquid metal–gas phase. Incidentally, the temperature of the liquid metal passing through the voltaic arc is not so high; for a steel droplet it is \(1600\)—\(1700^\circ\).

On Alloying

Alloying elements from the coating pass into the metal of the electrode. We have no firm experimental, still less theoretical, data on the quantitative (and sometimes even qualitative) side of the phenomenon.

No sound studies deserving full confidence on this question have been published either in Soviet or in foreign literature. The available foreign data and the reported compositions of coatings are unreliable. Certain authors give some indications of the coefficient of transfer of alloying elements from coatings into the weld, but they are unsystematic, accidental, and therefore extremely difficult to use.

It was not part of our task to conduct systematic investigations determining the coefficient of transfer of the elements most important to us—silicon and carbon—from the coating into the weld; nevertheless, it is necessary to present some theoretical considerations and experimental data on this question.

The transfer of carbon from the coating into the metal of the melting electrode is possible both from the solid phase (graphite, charcoal, soot) and from the gaseous phase.

Direct contact of the carbon of the coating with the liquid metal of the molten electrode is possible, but it is impeded by the liquid slag, in which the carbon is present in the form of a more or less fine suspension. The individual carbon particles are thus evidently isolated from the metal by molten slag; therefore transfer of carbon into the weld from the gaseous phase is much more probable.

Carbon has a high boiling point \((4500^\circ K)\). Therefore, in the course of the blast-furnace process, at the temperatures characteristic of this process, of the order on average of \(1500^\circ\), the vapor pressure of carbon saturating the space is negligibly small, and gaseous carbon plays no significant role. The temperature of certain places in the gas phase of the arc-welding focus reaches \(6000^\circ K\), but the temperature of steel, molten and passing through the gas phase in the form of droplets, does not exceed \(2100^\circ K\), and in any case cannot exceed the boiling point of iron \((3500^\circ K)\). Therefore, in the case of arc welding, the transfer of carbon from the coating into the metal through the gas phase, which includes carbon vapors, must by no means be ignored.

Pieces of graphite, passing through zones of the welding focus having a temperature of \(6000^\circ K\), evaporate, and the carbon vapors condense at the coldest places of the welding focus—the droplets of molten iron. True, carbon has a very considerable heat of vaporization \((10800\ \text{cal}/\text{g})\); its evaporation will lower the temperature of the gas phase, and it is hardly possible to expect a large amount of evaporated carbon, especially since, with the lower boiling point \(3500^\circ K\) and the smaller heat of vaporization \(1320\ \text{cal}/\text{g}\), the role of iron vapor in the formation of the weld

is small, and the steel weld is formed not by condensation of the gas phase, but by transfer of iron in droplets.

There remains the carburization of the liquid metal by the reaction:

\[ 2\mathrm{CO} + \text{dissolved Fe solution} \rightarrow \mathrm{C\ in\ Fe} + \mathrm{CO}_2 \]

In a blast furnace the reaction

\[ \mathrm{C}+\mathrm{CO}_2 \rightarrow 2\mathrm{CO} \tag{n} \]

is established, when operating on charcoal, at \(t^\circ = 750—800^\circ\), and when operating on coke at \(900—950^\circ\).

In the surface layers of the molten coating one reaction proceeds:

\[ \mathrm{C}+\mathrm{O}_2 \rightarrow \mathrm{CO}_2, \]

and in the depth another:

\[ \mathrm{CO}_2+\mathrm{C}\rightarrow 2\mathrm{CO}. \]

Carbon monoxide, entering the gas phase, carries out carburization of the molten iron.

Graphite, in comparison with coke, has a lower reactivity with respect to oxygen. The activities (reactivities) of charcoal, coke, and graphite are related as: \(90:20:5\), or \(18:4:1\).

As a measure of activity, the degree of reduction of oxides at \(1000^\circ\) after one hour was taken [8].

Owing to the reduced reactivity of graphite, reaction (n) will probably proceed at a temperature higher than \(950^\circ\).

Any additional introduction into the welding focus, simultaneously with carbon, of substances that give off carbon dioxide (upon heating)—namely potash and soda—will promote the production of additional quantities of carbon dioxide and facilitate the carburization process.

At high temperatures, if the gas phase contains even a small amount of carbon dioxide, the reaction proceeds:

\[ \mathrm{Fe}_3\mathrm{C}+4\mathrm{CO}_2 \rightarrow 3\mathrm{FeO}+5\mathrm{CO}. \]

The equilibrium curve for this reaction runs parallel to the temperature axis at zero content of carbon dioxide in the gas phase.

It is difficult to expect one hundred percent carbon monoxide in the gas phase of a welding focus surrounded on all sides by atmospheric oxygen; therefore it must be assumed that carburization takes place not in the zone of very high temperatures, but in the liquid pool of the weld itself, at temperatures close to the crystallization temperature of the weld.

The question of the transfer of silicon from the coating into the weld is very interesting. The transfer of silicon is evidently possible mainly from the liquid phase (molten ferrosilicon). Silicon has a great affinity for oxygen and forms \(\mathrm{SiO}_2\) with oxygen, with the liberation of an enormous amount of heat. Experience shows that a very large amount of silicon from the coating passes into the slag in the form of \(\mathrm{SiO}_2\).

But, on the other hand, there is a reaction of direct reduction of silicon:

\[ \mathrm{SiO}_2+2\mathrm{C}\rightarrow \mathrm{Si}+2\mathrm{CO}-137560. \]

In a blast furnace this reaction proceeds weakly, since the temperatures are too low.

Calculating the temperature in the gas phase at which this reaction can proceed, we obtain:

\[ \lg P_{\mathrm{CO}}=-\frac{14912}{T}+1.75\lg T-9.66\cdot 10^{-4}T+3.5, \quad \text{assuming } P_{\mathrm{CO}}=1 \text{ atmosphere} \]

\[ T \cong 2050^\circ. \]

At this temperature the reduction reaction begins; it proceeds the more vigorously, the higher the temperature. Thus, it may be considered that the transfer of silicon from the coating into the weld is effected directly by dissolution of ferrosilicon crystals in the molten steel, and also from the silica of the coating reduced by carbon. In the latter case, the carbon content in the coating and the composition of the gas phase with respect to the content of carbon monoxide and carbon dioxide are of substantial importance.

This circumstance makes it difficult to develop a method for determining the coefficient of transfer of carbon and silicon from the coating into the weld, since the transfer of silicon into the metal is substantially influenced by carbon—its content both in the molten coating and in the gas phase.

Carbon dissolved in the metal also has an effect, since it reduces the silicon oxide that comes into contact with it. On the other hand, silicon, by absorbing oxygen in the welding arc, protects carbon from burning out.

The coefficient of transfer of silicon from the coating into the weld varies depending on fluctuations in the carbon content of the coating; conversely, the coefficient of transfer of carbon depends on the silicon content of the coating.

Nevertheless, we made an attempt to determine how the silicon and carbon content in the coating affects the content of these same elements in the weld.

We prepared a series of coatings with varying ferrosilicon content at a constant carbon content, and with an unchanged carbon content and a variable ferrosilicon content. In addition, a coating with graphite was prepared, as well as another almost identical in composition but with charcoal boiled in a potash solution.

In the welding arc focus produced by an electrode coated with a coating containing charcoal boiled in a potash solution, carbon dioxide should be evolved as a result of the decomposition of potash.

This latter circumstance will lead to an increase in the carbon-oxide content in the gas phase and to better carburization of the electrode metal.

These coatings, like all the coatings finally adopted by us later, were prepared on sodium water glass with a modulus of 3.0–3.2. Marble was introduced into the coatings as an ionizing additive.

For chemical analysis, the upper layers of the weld were taken (to a depth of 3 mm from the surface of the weld). This was done in order to exclude the influence of the material of the cast-iron plate on which the surfacing was performed.

The steel electrode had a thickness of 4 mm; the coating layer was applied in two passes and had a thickness of 1 mm per side. The alternating-current strength during welding was kept at about 180 A.

The cast-iron electrode had a diameter of 5 mm; the coating layer was 1.2 mm thick per side. The alternating-current strength fluctuated around 200 A.

During the experiment it was ensured that the arc was of constant length and that the welding speed did not change.

Chemical composition of the steel electrode

C%	Si%	Mn%
0.11	0.08	0.45

Physical composition of the cast-iron electrode

C%	Si%	Mn%
3.2	6.7	—

When the content of one of the constituent parts of the coating is changed (say, carbon), the content of the other (silicon) also changes. Their action, however, is manifested jointly.

It would be more correct, without changing the carbon content, to change the silicon content, and to maintain the overall balance by introducing inert additions. But then, according to the law of acting masses, the alloying action of an addition present in a constant amount will still be variable. Most important, the activity of both active alloying elements will be artificially reduced as a result of the decrease in their concentration in comparison with a normal coating.

This will be especially important in our case, when in the ordinary coating adopted by us the content of ferrosilicon and graphite totals about 70–80% of the weight of the dry coating. In order not to reduce the activity of the coatings in our experiments, in some cases we were compelled also to change the content of the second component of the coating under investigation. The results of the experiments are given in Table 5 and in Figs. 4, 5, 6, 7.

After a number of preliminary experiments, when it became clear that carborundum is less active than ferrosilicon, and it was found that wood-charcoal powder treated with carbonate is more active than graphite, nine coatings of different composition were prepared. Three of them represent different variants in the ratios between wood charcoal and ferrosilicon; three between graphite and ferrosilicon; and three between carborundum and graphite (Table 4).

Table 4

Variants of coating compositions (in percent)

No.	Component	%	No.	Component	%	No.	Component	%
1	Graphite	50	4	Graphite	20	7	Charcoal	50
1	Ferrosilicon	40	4	Carborundum	70	7	Ferrosilicon	45
1	Marble	10	4	Marble	10	7	Marble	5
2	Graphite	30	5	Graphite	10	8	Charcoal	30
2	Ferrosilicon	60	5	Carborundum	80	8	Ferrosilicon	65
2	Marble	10	5	Marble	10	8	Marble	5
3	Graphite	60	6	Graphite	30	9	Charcoal	70
3	Ferrosilicon	30	6	Carborundum	60	9	Ferrosilicon	25
3	Marble	10	6	Marble	10	9	Marble	5

In addition to these two constituent parts, marble was introduced into all nine coatings in amounts of 10 and 5%. Five percent marble was introduced into the coatings with wood charcoal treated, by our method, with potassium carbonate. The wood charcoal so treated was an additional ionizer, and therefore the amount of marble in some coatings was reduced by half.

In the course of the experiments it was found that the coating with ferrosilicon and wood charcoal gave good results both as a protective and partially alloying coating for cast-iron electrodes and as an alloying coating for steel ones.

The amount of soluble glass in the coatings is 25% of the weight of the dry coating.

During welding it is recommended to maintain a short arc.

The thickness of the coating layer is 1.2 mm per side for a cast-iron electrode 5 mm in diameter and 1 mm per side for a steel electrode 4 mm in diameter.

Fig. 4

Fig. 5

Fig. 6

Labels in the graph: \(C\%\); \(\circ Si\%\).

Mixture composition along the horizontal axis:

Graphite	Ferrosilicon
30%	60%
50	40
60	30

Fig. 7

Labels in the graph: \(C\%\); \(Si\%\).

Mixture composition along the horizontal axis:

Carborundum	Graphite
60%	30%
70	20
80	10

Deposits that were made specially for chemical analysis were three-layered, and the samples for analysis were taken from the third layer. Several control specimens were deposited on copper plates in order to obtain a sufficiently clean deposit and to eliminate the influence of the base material.

Table 5 gives the data of the chemical analysis of deposits made when applying a coating of one or another variant of composition.

Table 5

	Coating	Coating	Coating	Weld	Weld
A	Ferrosilicon	Graphite %	CaCO$_3$ %	Si %	C %
1	60	30	10	6,5	3,3
2	40	50	10	6,7	3,3
3	30	60	10	6,1	2,9
B	Carborundum %	Graphite %	CaCO$_3$ %	Si %	C %
1	60	30	10	5,5	3,3
2	70	20	10	6,7	3,2
3	80	10	10	6,8	2,8
V	Ferrosilicon	Charcoal %	CaCO$_3$ %	Si %	C %
1	25	70	5	5,0	3,5
2	45	50	5	6,7	3,2
3	65	30	5	6,8	2,3

Three variants of coatings (1, 2, 3), consisting of ferrosilicon, graphite, and marble, are grouped under the letter A.

Three variants of coatings made of carborundum, graphite, and marble are grouped under the letter B.

Three variants made of charcoal, ferrosilicon, and marble are under the letter V.

A coating of the composition: graphite, carborundum, marble in the three variants of component parts indicated in Table 5 was used in welding with cast-iron electrodes.

On the basis of the coating compositions and the results of chemical analysis of the deposits corresponding to these coatings, curves were plotted, presented in Figs. 4, 5, 6, 7.

1. The curves, Fig. 4, characterize the transition of the principal alloying elements—carbon and silicon—from a coating of the composition: charcoal 70%, ferrosilicon 25%, marble 5% into the weld when welding with cast-iron electrodes of composition 6,7% Si and 3,2% C.

2. The curves, Fig. 5, characterize the transition of carbon and silicon from a coating of the same composition: charcoal 70%, ferrosilicon 25%, marble 5% into the weld, but now when welding with steel electrodes of composition: C—0,10%, Si—0,08%.

1. The curves, Fig. 6, show the transition of carbon and silicon from a coating of the composition: graphite 30%, ferrosilicon 60%, marble 10%, into the weld when welding with cast-iron electrodes of the above-indicated composition (the cast-iron electrodes in all three cases had one and the same chemical composition).

2. The curves, Fig. 7, correspond to the transition of carbon and silicon from a coating of the composition: carborundum 60%, graphite 30%, marble 10%, into the weld when welding with cast-iron electrodes.

On the basis of plotting these curves, we found the optimum for the composition of the coatings, in all four cases taking as the characteristic of the composition of a given coating the abscissa of the point of intersection of the curves characterizing the change in the carbon and silicon content in the weld.

Thus, for the third case, Fig. 6, the optimum coating composition is: graphite 55%, ferrosilicon 36%, marble 9%.

From the curves presented in Fig. 4, the optimum coating composition will accordingly be: charcoal 53%, ferrosilicon 42%, marble 5%.

As a result of analyzing the curves of Fig. 7, the optimum composition of the coating is obtained as follows: graphite 22%, carborundum 67%, marble 11%.

As a result of analyzing the curves of Fig. 5, the optimum coating composition for steel electrodes is the same as for cast iron: charcoal 53%, ferrosilicon 42%, marble 5%.

The results of a metallographic study of welds deposited with a cast-iron electrode of composition—6.7% Si and 3.2% C on a base material of the following chemical composition:

C%	Si%	Mn%
2.80	2.12	0.40

Microstructure of the base material: pearlitic metallic matrix plus inclusions of flake graphite.

1. The surfacing was carried out with a cast-iron electrode using coating grade A.

Microstructure of the deposited metal: ferritic basic metallic matrix and inclusions of coarse flake graphite.

1. The weld was deposited with a cast-iron electrode using coating grade B.

Microstructure of the deposited metal: ferrite + a small amount of pearlite + graphite of flake structure.

1. Result of surfacing with a cast-iron electrode coated with coating grade V.

The microstructure of the deposited metal consists of a ferritic basic metallic matrix of dendritic structure; the outlines of the ferritic dendrites are visible due to inclusions of eutectic graphite.

1. Result of surfacing with a steel electrode with coating grade V. The microstructure of the deposited metal is pearlite + ledeburitic cementite (the weld is chilled).

Conclusions

If a high-silicon cast-iron electrode is taken, the possibility of welding with such an electrode without coating is very probable, since the burning silicon gives a sufficiently large amount (7 calories per gram) of heat. If care is taken for ionization, then welding with a bare electrode is entirely possible. We came to the conclusion that coatings for welding cast iron with a cast-iron electrode must protect the main ele-

elements (silicon and carbon), ionize the arc and, with sufficient fusibility, give a slag that to some extent protects the weld from heat loss. But the latter circumstance (the cooling-protective action of the slag) cannot and should not be overestimated.

The basis of the slag obtained as a result of the burning of the coating is sodium oxide and silicon oxide.

Taking this circumstance and a number of other factors into account, we arrived at the coating formulations set forth below.

To facilitate better protection against hydrogen penetration, it is recommended to introduce organic substances into the coatings (starch, dextrin, slightly charred wood sawdust).

The addition of $\mathrm{CaCO_3}$ to the coating (or, what is the same, $\mathrm{CaO}$, since at $940^\circ$ the reaction $\mathrm{CaCO_3}\to \mathrm{CaO}+\mathrm{CO_2}$ takes place) should not be abused. It is better to use some other ionizer. The fact is that from the $\mathrm{Na_2O}—\mathrm{CaO}—\mathrm{SiO_2}$ diagram it follows that only up to a content of 5% $\mathrm{CaO}$ does the melting temperature of the slag lie within the required limits (not above $900^\circ$). And only with a high $\mathrm{SiO_2}$ content (above 80%) does an addition of $\mathrm{CaO}$ not greatly increase the melting temperature.

Taking the data given into consideration, the following considerations may be offered on the preparation of coatings for cast-iron electrodes.

The best choice for the basis of the slag is the system—sodium oxide—silicon oxide. If it is possible to obtain (or prepare) potassium water glass, it should be added to the sodium glass in an amount of 10–20% in order to lower the melting temperature of the slag and improve ionization. The addition of $\mathrm{CaCO_3}$ as an ionizer in the absence of potassium glass does not in general raise objections, but it is better to use marble rather than chalk. Marble makes the coating (in comparison with chalk) denser and the slag less bubbly.

Thus, in designing the coating it is necessary to base it on the ternary system: $\mathrm{Na_2O}—\mathrm{CaO}—\mathrm{SiO_2}$.

The unavoidable presence of $\mathrm{FeO}$ (or $\mathrm{Fe_2O_3}$) will not have a major influence on the properties of the slag. As analysis shows, the slag contains no more than 1–3% iron.

1. For alloying the weld with silicon it is best to use ordinary ferrosilicon. It (in finally screened powder) should be passivated by heating at $500–600^\circ$ for 1 hour.

Carborundum alloys the weld with silicon and carbon more weakly, as the experiments of other authors with steel-wire electrodes have shown, but at the same time it also reacts more weakly with $\mathrm{NaOH}$. Therefore its use can be recommended in cases where high-modulus glass is not available and one has to be satisfied with glass ordinarily used for making coatings (with modulus 2.8).

1. For alloying the weld with carbon it is best to use charcoal prepared by our method.

Owing to the content of $\mathrm{K_2CO_3}$, such charcoal very effectively carburizes the weld and excellently ionizes the arc without the additional introduction of $\mathrm{CaCO_3}$ as an ionizer.

Graphite, usually used in such coatings, does not possess these advantages, but coatings with graphite are also sufficiently satisfactory.

1. There is no special need to protect a cast-iron weld from nitrogen when using an electrode with a thick coating, but for preventive purposes 1–2% starch or dextrin may be added to the coating.

2. Alloying the weld with nickel by introducing nickel into the coating is possible, but it is more rational to introduce nickel into the composition of the electrodes, since in the latter case the nickel loss will be much smaller.

1. On the basis of a series of experiments we can recommend the following coatings for cast-iron electrodes:

Coating No. 1

Graphite . . . . . . . . . . 6 parts by weight
Ferrosilicon . . . . . . . . 4 parts by weight
Marble . . . . . . . . . . . 1 part by weight

The coating is mixed with water glass having a modulus of 3.0.

Coating No. 2

Graphite . . . . . . . . . . 2 parts by weight
Carborundum . . . . . . . . 6 parts by weight
Marble . . . . . . . . . . . 1 part by weight

The coating is mixed with water glass having a modulus of 2.8.

Coating No. 3

Charcoal . . . . . . . . . . 10 parts by weight (charcoal processed by our method)
Ferrosilicon . . . . . . . . 8 parts by weight
Marble . . . . . . . . . . . 1 part by weight

The coating is mixed with water glass having a modulus of 3.0.

Some Results of Experimental Surfacing

In order to avoid excessive bulkiness in this chapter, descriptions of only some surfacing operations are given here.

Welding conditions and technology

Welding with cast-iron electrodes coated with coatings of the compositions indicated in the preceding chapter was carried out on a welding transformer.

The optimum current was selected experimentally for the given electrode and the given coating and averaged 200 A.

When the electrodes were cast in earthen molds, their thickness along the length varied within the range from 5.0 to 5.5 mm. But in the welding process this had little effect on the average current that had been selected.

The coatings used by us had a sufficiently good stabilizing effect on the arc. As a result, arc ignition occurred without any difficulty, and even an inexperienced welder could melt an entire electrode without interruption. The arc was maintained especially well when partially potassium water glass or charcoal powder boiled in potash was used in the coating.

The plates cast for experimental welding were planed on a lathe. The thickness of the plates was 15–20 mm. For surfacing, V- and U-shaped grooves 12 mm deep were planed.

Description of Experimental Surfacing Operations

Experiment 1. The base material was gray cast iron of the following chemical composition.

Ctotal, %	Si, %	Mn, %	S, %	P, %
2.80	2.12	0.40	—	—

The microstructure of the base material (Fig. 8) is a pearlitic base metallic mass with inclusions of graphite of a coarse-lamellar structure and islands of isolated inclusions of ledeburitic cementite.

Surfacing was carried out with an electrode of the following composition:

\[ \mathrm{C}=2.9\% \qquad \mathrm{Si}=6.7\% \]

The microstructure of the electrode material is ferrite + lamellar graphite (Fig. 1).

Fig. 8

Fig. 9

The electrode was coated with a flux of the following composition:

1. Wood charcoal—5 parts (pretreated with \(K_2CO_3\)).
2. Ferrosilicon—5 parts.
3. Chalk—1 part.
4. Water glass—3 parts (dry residue).

Fig. 10

Fig. 11

The water glass used for preparing this coating had a modulus of 3.0. The microstructure of the deposited metal is ferrite plus eutectic graphite (Figs. 9, 10, 11, 12).

At the transition from one layer to another, which cooled more rapidly than the first, alongside eutectic graphite there is ordinary lamellar graphite.

At low magnification (Fig. 13), clearly defined ferritic dendrites are plainly visible in the structure of the deposited layer; eutectic graphite is embedded between their axes. At the transition from the deposited layer to the layer of the base cast iron (the penetration zone), inclusions of ledeburitic cementite are visible; there is less of it here than in the structure of the base cast iron (Figs. 14, 15).

Fig. 12

Fig. 13

The microstructure of the transition zone (at a magnification of 25) is shown in Fig. 16.

On the basis of structural analysis it may be concluded that the deposited metal and the metal of the transition zone possess higher mechanical properties than the base cast iron, in which, as described above, inclusions of ledeburitic cementite are present.

Fig. 14

Fig. 15

The microstructure of the transition zone is shown in Figs. 13, 17. Fig. 18 depicts the same structure at 25× magnification. Here one can distinguish the lighter part—the ferritic deposited layer—and the darker part—the base material. Fig. 19 shows both the base and deposited metal. Traces of ferritic dendrites are visible in the deposited layer.

Deposition was carried out with alternating current at a current of 250 A and a current density of \(9.9 \frac{A}{mm^2}\).

The arc burned steadily, without interruptions, and was fairly easy to strike.

Analysis of the deposit showed the following:

1) a dense bond between the base metal and the deposited metal;

2) the deposited metal is dense, with an insignificant amount of slag inclusions, without gas bubbles;

Fig. 16

Fig. 17

3) with a well-executed surfacing operation, the place of deposition, in its properties (plasticity), undoubtedly exceeds the quality of the base metal.
Results of hardness determination:

Fig. 18

Fig. 19

base material  $H_B = 150—180$
surfacing    $H_B = 120$.

Experiment 2. The base material is gray cast iron of the following chemical composition:

C%	Si%	Mn%	S%	P%
2.53	4.23	0.36	0.01	0.006

The microstructure of the base material is ferrite and coarse lamellar graphite, with the length of individual graphite flakes reaching up to 1–2 mm (Fig. 20).

Surfacing was carried out with an electrode of the following composition:

$C_{\text{total}}$ %	Si %	Ni %	Al %
3.70	6.2	1.3	0.5

The microstructure of the electrode consists of ferrite with a small amount of pearlite and graphite (Fig. 2).

The increased Ni content in the electrode had almost no effect on the structure of the deposited layer.

Apparently, with a sufficiently high silicon content there is no point in adding nickel; its beneficial effect is manifested only at a moderate silicon content, when graphitization of cast iron proceeds at the expense of both silicon and nickel. In our case, however, graphitization proceeds at the expense of silicon, and nickel had practically no effect on the structure of the deposited layer. Structurally, the deposited metal (Fig. 3) consists of ferrite + graphite + eutectic graphite + flake graphite + graphite nests.

Fig. 20

Fig. 3 shows the structure of the deposited metal at a magnification of 600.

The transition of the base metal into the deposited layer (Fig. 22) clearly shows the difference in the structure of the base metal and the deposited layer. With their fundamental similarity (in both cases we have ferritic cast iron), the distribution and form of graphite precipitation are different.

The left part of Fig. 21 shows the base metal, the right—the deposited metal. A smooth transition from one structure to the other is noticeable (the same figure).

Fig. 21

Fig. 22

Fig. 23 shows the base material at a magnification of 25.
Fig. 24 corresponds to the transition zone, photographed at low (25) magnification.

Welding was carried out with alternating current. Current strength 225 A, at a current density of \(8\ \mathrm{A/mm^2}\).

The electrode was coated with a flux of the following composition:

Component	Amount
charcoal	5 parts
ferrosilicon	5 parts
chalk	1 part
water glass (dry)	3 parts

Analysis of the deposited layer showed the following:

1. A sufficiently dense bond between the base metal and the deposited metal.

Fig. 23

Fig. 24

1. The deposited metal is dense. The amount of slag inclusions is extremely limited. Gas inclusions are absent.

Results of hardness determination:

Material	Hardness
hardness of the base metal	\(H_B = 137\)
hardness of the deposited layer	\(H_B = 150\)

Experiment 3. The base metal was gray cast iron of the following chemical composition:

\(C_{\text{total}}\%\)	Si%	Mn%
2.62	2.89	0.39

The microstructure of the base material (Fig. 25) consists of ferrite + coarse lamellar graphite + pearlite, with the length of individual graphite flakes reaching \(0.8–1.6\ \mathrm{mm}\).

Figure 26 shows the structure of the base metal at low magnification (25).

Two beads were deposited on a cast-iron plate of the indicated composition with electrodes differing in chemical composition but coated with the same coating. One of the grooves of the plate, having a U-shaped groove preparation, was welded with an electrode of the following composition: \(Si = 4.6\%\) and \(Ni = 1.8\).

The microstructure of the electrode is pearlite, graphite, and a small amount of ledeburite inclusions. With a silicon content in the electrode material of \(4.6\%\), the addition of nickel has a noticeably beneficial effect on

structure in the direction of refinement of the graphite precipitates, more complete decomposition of the bound cementite, with the formation of a ferritic matrix.

The deposited metal (Fig. 11), structurally, is a graphite eutectic with a ferritic matrix of the metallic mass.

Fig. 25

Fig. 26

Near the heat-affected zone, in addition to eutectic graphite, graphite precipitates in lamellar form occur (Fig. 27).

In Fig. 28 we are at the same time able to see three zones at once:

1. The base-metal zone.
2. The heat-affected zone.
3. The deposited-metal zone.

The second U-shaped groove of the cast-iron plate was welded with an electrode of composition 4.8% Si without nickel.

Fig. 27

Fig. 28

Microstructure of the electrode: pearlite + ferrite + graphite of lamellar form.

The deposited metal, structurally, represents a dendrite-like form of precipitation of three structural constituents: pearlite, graphite, and ferrite (Figs. 29, 30). The transition zone is represented by the microphotograph in Fig. 31 and by the microphotograph in Fig. 32 at a magnification of 25.

In Fig. 32 the structure of the deposit is shown, consisting of pearlite and ferrite (of dendritic structure) and inclusions of spheroidal graphite.

In both the first and the second case, the electrodes were coated with a coating of the following composition:

Wood charcoal	5	parts
Ferrosilicon	5	"
Chalk	1	"

Welding was carried out with alternating current at a current of $240\ A$ and a current density of $8.7\ A/mm^2$.

Fig. 29

Fig. 30

The arc burned steadily, with fairly easy ignition. The melting and flowing of the coating from the electrode were uniform. The fusion was quite satisfactory.

Analysis of the weld showed the following:

1. A dense bond between the base metal and the deposited metal.
2. The deposited metal is sufficiently dense.

Fig. 31

Fig. 32

There was a small amount of slag inclusions, with complete absence of gas inclusions.

Experiment 4. The base material was gray cast iron of the following chemical composition:

C%	Si%	Mn%
2.48	1.53	0.32

The microstructure of the base material is pearlite and coarse-flake graphite.

By its microstructure, the base material is pearlitic-graphitic cast iron.

Surfacing was carried out with electrodes of the following composition:

Co%	Si%	Al%
3.70	4.9	1.0

Fig. 33

The microstructure of the electrode consists of ferrite with a small amount of pearlite and graphite in lamellar form.

The electrode was coated with a coating of the following composition:

Component	Amount
Charcoal	5 parts
Ferrosilicon	4 parts
Chalk	1 part

The water glass used for this coating had a modulus of 3.8. The microstructure of the deposited metal is pearlite, graphite, and a certain amount of cementite.

Fig. 34

Fig. 35

The transition zone is shown in Figs. 33, 34, 35.

Conclusions

1. As a result of this work, it may be recommended to weld pearlitic and pearlitic-ferritic cast irons with an electrode having a high silicon content (5%). Ferritic cast irons can be welded with a ferritic-pearlitic electrode with a lower silicon content (not less than 4%); in this case, the absence of cementite in the transition zone and a smooth transition of structures are obtained.

2. Welding with a steel electrode with a carborundum coating cannot be recommended.

1. It is possible to recommend welding ferritic cast iron with a steel electrode having a coating of ferrosilicon and wood charcoal treated, according to our method, with potassium carbonate.

2. Ferritic–pearlitic and pearlitic cast irons can be welded with a steel electrode, but this depends on the skill of the welder and does not always give positive results.

3. The use of steel electrodes with a thin wire winding and a three-layer coating did not produce the expected effect.

4. When using coatings with ferrosilicon and, in part, with carborundum, it is necessary to pay attention to the modulus of the water glass. We recommend, on average, a modulus of 3.0–3.2.

When using passivated ferrosilicon, a modulus of 2.8–3.0 is possible.

1. As a result of using electrodes coated with the coatings whose composition we recommend, a slag is obtained that is calculated quite easily; moreover, it is necessary to bear in mind that, because of the different metallurgical and physicochemical bases of steel and cast iron, the slag composition for steel and for cast iron must be different.

2. Graphitization in the weld metal, owing to the high temperature of the molten droplet (\(\sim 1800^\circ\)), does not depend on the presence of graphite crystallization centers in the liquid metal, as is the case in ordinary cast iron, where superheating of the order of \(1400\text{–}1500^\circ\) is considered significant.

3. Carburization of the steel electrode to cast iron probably proceeds through the action of the gas phase, in the form of CO and partly in the form of carbon vapors, but not in the form of solid graphite, which also excludes ready-made centers of graphite crystallization.

4. The transfer coefficient of alloying elements (carbon and silicon) from the coating into the weld is unknown, and systematic experiments in this direction have not been carried out. It is possible to determine the transfer coefficient of carbon and silicon separately, but when carbon and silicon are introduced together, the picture will be different in accordance with the law of mass action.

5. Quantitative data have been determined for the transfer of certain alloying constituent parts of the coating (silicon and carbon) into the weld for steel and cast-iron electrodes.

On the basis of these data we recommend the following coating recipes:

1. Graphite 55%
   Ferrosilicon 36%
   Marble 9%
   } for cast-iron electrodes.

2. Graphite 22%
   Carborundum 67%
   Marble 11%
   } for cast-iron electrodes.

3. Wood charcoal 53%
   Ferrosilicon 42%
   Marble 5%
   } for steel and cast-iron electrodes.

In all coatings the amount of soluble glass (dry) is 25% of the weight of the dry coating.

1. The use of thermite in coatings for welding cast iron produced no effect.

2. If inclusions of ledeburitic cementite have appeared in the structure of the weld, then, if possible, we recommend using thermi-

heat treatment of the weld, consisting of annealing at a temperature of 900° for complete graphitization.

1. Welding cast iron with preheating to 200–300° promotes more complete graphitization of the weld and the obtaining, in the deposited metal, of a microstructure of ferrite and graphite.

LITERATURE

1. E. Pivovarsky. High-Quality Cast Iron and the Physicometallurgical Principles of Its Manufacture, ONTI, 1932.
2. K. A. Eremenko. Formation of Graphite in Cast Iron, 1938.
3. N. F. Bolkhovitinov. Metallurgy and Heat Treatment, Mashgiz, 1947.
4. N. P. Dubinin. Casting Cast-Iron Consumer-Goods Articles in Permanent Molds, Gizmestprom, 1940.
5. N. V. Gutovsky. On the Theory of the Iron–Carbon System, Part 1, 1914.
6. V. I. Yarkho. Cold Welding of Cast Iron with Cast-Iron Electrodes, Mashgiz, 1940.
7. V. I. Yarkho. Cold Welding of Cast Iron by the Yarkho Method, “Avtogennoe Delo,” No. 7, 1936.
8. I. A. Pavlov. Metallurgy of Cast Iron, issue 2, 1940.
9. K. K. Khrenov and V. I. Yarkho. Technology of Arc Welding, Mashgiz, 1940.