Introduction
Welding of any member of the family of cast irons is typically considered problematic, particularly when compared to the welding of structural steels. This reaction applies whether the welding of cast irons is to be a structural matter (joining components together) or whether welding is used to salvage the casting in a manner such that it would be suitable for engineering applications. The inherent problems in the welding of cast irons are metallurgical, and must be addressed based on an understanding of welding procedures, of cast iron solidification and phase transformations occurring during heating and cooling of the cast iron. The objective being one of being able to produce a weld and its heat affected zone with the metallurgy and mechanical properties of the base metal, i.e., the casting involved.

The weldability of a particular material is usually considered to be the production of a fusion weld (either as bead on plate, or the joining of components together) in a manner such that the weldment does not contain injurious imperfections (more commonly described as defects when they detrimentally affect metallurgical structure, properties and subsequent manufacturing operations). A broad classification of weld defects in a given material subjected to fusion welding would include the following:

2. Defects from the introduction of foreign inclusions. These may involve the formation of oxide films; the entrapment of slag or metal/gas reaction products; delamination (due to the presence of undesirable inclusion distribution in the base metal as affected by the fusion welding procedure); etc.

3. Geometric effects. This involves the shape of the weld and/or weldment including: undercutting (fusion of base metal without adequate compensation with weld metal); excessive reinforcement (excessive deposition of weld metal – both externally and internally; etc.

4. Metallurgical effects. First, undesirable metallurgical structures related to microsegregation, e.g., hot cracking and fissures; cold cracking and delayed cracking; stress relief cracking; strain age cracking; gas porosity; etc. In addition, problems resulting from metallurgical reactions, e.g.; embrittlement; structural notches; etc.

No reasonable fusion weld is perfect, and all fusion welds can be expected to contain some of these imperfections. But an acceptable weld must not contain imperfections that would be detrimental to the use of the welded product.

The root causes of weld defects can be attributed to one or more of several conditions:

1. Lack of welding know-how and experience.
2. Welding process characteristics.
3. Base metal defects or compositions.
4. Materials selection and properties.
5. Welding environment (joint design, fit up, temperature, support, etc.

In consideration of the welding of a member of the family of cast irons, comparisons are often made to the welding of steels, either for structural purposes or for the upgrading of steel castings. However, steels (even those which have been alloyed for one purpose or another) do not present the complex metallurgical, or solidification phase control, characteristics that are paramount in cast irons. The fusion welding of steels typically entails the melting and solidification of single phase structures, not necessarily homogeneous. On the other hand, the unique feature of the family of cast irons is the formation of two significant phases during solidification (one of which is graphite) and the control of that phase morphology. Fusion welding of cast irons starts with this intricate phase morphology, involves the melting or transformation of those phases (usually a process which is not the reverse of their solidification formation), and the re-solidification of this structure.

Furthermore, the composition of the matrix phase (or phases) in cast irons is affected not only by alloying elements in ferrite and austenite, but by the presence of excess carbon in graphite or carbide phases and the solubility of that carbon in ferrite and austenite.

Phase Changes Occurring During Heating of Cast Irons
It is of interest to summarize the major phase changes that take place during the heating of cast irons in fusion welding. Initially the cast iron microstructure consists of the excess carbon phases (graphite and carbide), the matrix (ferrite and/or pearlite), and the intercellular structure. No significant changes typically occur on heating until austenite starts to form (changes in pearlite lamellae are minor). Pearlite and the intercellular structure will transform to austenite first (compositions near eutectoid), with ferrite transforming later (with carbon diffusion from graphite). Homogeneity of the austenite depends upon time and temperature, although diffusion rates for carbon are appreciably greater than for alloying elements (e.g., Si, Mn, Ni, Cr, Mo, etc.).

The ability of the matrix to withstand stresses introduced during the non-uniform heating of cast irons is a function of the matrix microstructure. Ferritic matrices provide the greatest opportunity for dimensional adjustment (yielding). Finer pearlite lamellae will be more prone to crack initiation, as will intercellular regions which typically contain carbides.

Melting starts to take place where the local composition has the lowest melting point (near eutectic composition). Since the local carbon content is a critical factor in this regard, the first areas to melt are observed at the graphite austenite interface and in the intercellular regions (where there is a lower melting point due to carbon and alloying element concentrations). Carbon is more accessible from the edges of the graphite basal planes so that melting starts at the tips of flake graphite, at the graphite spheroid-austenite interface, at the temper carbon-austenite interface, etc. The process of melting develops over a temperature range so that a zone of partial fusion can be expected outside of the resultant fusion zone.

Within the fusion zone convection insures homogeneity of melt composition. However, a thin region at the edge of the fusion zone remains unmixed. The composition of the fusion zone is, of course a function of the amount of dilution due to different metal compositions or filler metals.

Phase Changes Occurring During Cooling of Cast Irons
On cooling, the most rapid cooling rates are in the fusion zone, with cooling rates decreasing as distance from the fusion zone increases. Cast iron microstructures are highly sensitive to cooling rate variations, both during solidification and solid state transformation. Sensitivity of the fusion zone to cooling rates depends upon the composition of the melt in the fusion zone. If it is a typical cast iron composition, solidification will favor carbide and under cooled graphite formation. If sufficiently alloyed, say with nickel, these undesirable morphologies will be minimized or eliminated.

However, the liquid which formed in the zone of partial fusion remains highly susceptible to solidification as eutectic carbide. If the liquid in the partial fusion zone is interconnected, a region of brittle eutectic carbide will be present to serve as a path for crack formation. The ability to limit the interconnection within the partial fusion zone depends upon the initial microstructure of the cast iron and the welding procedures used.

When permitted to cool at a sufficient rate into or through the Ms-Mf temperature range martensite formation will occur. To avoid martensite formation it is necessary to cool sufficiently slowly through this temperature range (essentially from red to black as the casting cools).

Typical Modifications of Procedures for the Welding of Cast Irons
As noted earlier, the fusion welding of cast irons is usually compared to procedures used in the welding of structural steels, with modifications of those procedures to accommodate the metallurgical characteristics of cast irons. The more significant of these procedure modifications will be discussed.

Avoidance of Martensite Formation. The presence of silicon, and of the possible saturation of carbon in austenite by drawing carbon from the graphite phase, greatly increases the effective hardenability of the austenite matrix – both martensitic and pearlitic hardenability. The propensity for cracking therefore requires that cooling through the Ms-Mf temperature range be sufficiently slow that martensite formation will not occur. This is accomplished by pre-heating the casting, or the region of the casting to be welded to a temperature of at least 500F, maintaining that temperature throughout welding followed by slow cooling to room temperature.

Minimizing the Partial Fusion Zone. Because the eutectic of cast irons solidifies (and melts) over a range of temperatures a zone of partial fusion outside of the weld fusion zone cannot be prevented from forming. Cooling rates within this region are sufficient for the melt in this region to solidify as a carbidic eutectic. One must minimize the extent of this partial fusion zone in order to prevent the development of a cracking path just outside of the fusion zone. This requires that a sufficiently steep thermal gradient be maintained throughout welding, thereby minimizing the thickness of the partial zone and the interconnection of eutectic melt.

The geometry of this partial fusion zone is also a function of the graphite morphology, since the partial fusion zone develops as a result of eutectic liquid formed as a diffusion couple between graphite and austenite forming preferentially where the carbon basal planes are exposed, around the circumference of spheroidal graphite, in certain regions of compacted graphite, a the ends of flake graphite, throughout temper carbon nodules, etc.

Avoiding Carbide Formation in the Fusion Zone. The highest cooling rates in fusion welding are encountered within the fusion zone. If the composition of the fusion zone is typical of a cast iron the fusion zone will solidify with a predominance of carbidic eutectic. To avoid this one must either alter the melt composition of the fusion zone or develop a fusion zone melt composition that has sufficient graphitizing potential to avoid carbide formation. In the first case this is readily accomplished by using a nickel bearing filler metal so that sufficient nickel dilutes the base metal in the fusion zone preventing eutectic carbides . In the second case a technique must be employed that provides the required graphitizing potential.

One procedure that has been developed for the welding of cast irons, particularly ductile irons, is an oxyacetylene welding procedure that was patented in 1967 , but which is generally unknown today to casting producers or designers. This fusion welding procedure is being used to produce metallurgically satisfactory welds over the range of mechanical property requirements of engineering grade ductile iron castings.

Summary of the Welding Procedure

This fusion welding procedure consists of the following steps:

1. Preheating of the casting to a temperature above the Ms temperature of the base metal, and maintaining that temperature until welding has been completed.

2. Use of an oxy-acetylene flame for welding, a suitable flux and a filler rod of a chemistry that matches that of the base metal.

3. Air cooling of casting to ambient temperature.

4. Post weld stress relief or heat treatment as necessary to obtain desired matrix structure.

Welding is conducted using an oxy-acetylene welding torch to which a powder spray device has been added. This device holds a quantity of material (flux) which can be aspirated into the oxygen and acetylene gas streams. A torch having a number 36 tip has been found satisfactory. The flux used is a proprietary blend of metallic compounds, ferrosilicon and magnesium ferrosilicon alloys which has been ground to a fine powder. The essential constituents of this flux are shown in Table I.

The flux constituents noted in Table I are not present elementally (e.g., Mg, Si, Ca, Ce, etc.) but are present as more convenient alloys of compounds. For example: Mg may be present as a 5% MgFeSi alloy; Si may be present as FeSi, CaSi or MgFeSi alloy; Ce may be present as a halide or incorporated into another alloy. The principle fluxing agent is borax.

Table I. Flux Composition

*Not critical
Examples of two satisfactory flux formulations are presented in Table II.

Table II. Exemplar Flux Compositions

These flux compositions are hydroscopic and must be maintained dry requiring storage in a suitable weld rod hot box. In order to assure free flowing in the welding torch, the particle size of the flux may be as large as 140 mesh, however, the flux is typically ground to a particle size under 200 mesh.

Filler metal is supplied using weld rod cast into rod shape or from machined from ductile iron castings having the same composition as the subject casting. A machined rod is preferable in order to eliminate any influence of the metal/mold interface (e.g., inclusions, metallurgical surface effects, etc.). The dimensions of the weld rod are not critical, but should enable ease of handling during welding. This rod is also stored in a weld rod hot box to prevent surface oxidation.

Fusion Welding of Ductile Iron

As with most fusion welding, but particularly with oxy-acetylene welding the welding procedure is operator sensitive, requiring some skill and experience and requires training over a period of time producing welds that will be examined metallurgically. The goal of this procedure is not merely to excavate a portion of the casting and to adequately fill that excavation with weld metal, but to accomplish this task in a manner such that there is minimal metallurgical difference between the weld metal, the heat affected zone and the base metal. In this way metallurgical integrity (and assurance of mechanical properties) is achieved.

The first step involves the removal of the unacceptable condition from the casting, or suitable edge preparation of components to be joined. This requires preparation down to clean, defect free base metal and a minimum 60° angle. The area to be welded (or the entire casting) is uniformly heated to, and maintained at, 650° F. When using gas torches for pre-heating and maintaining temperature during welding, the flame should be neutral or slightly reducing (avoiding excessive oxidation of the casting surfaces as well as carbon soot deposition). Where possible, welding should be conducted in the horizontal, or flat, position. After pre-heating the weld area is further heated using the oxy-acetylene spray torch, the weld rod is held adjacent to the flame while the heat of the flame is concentrated in the center of the area to be welded. When a molten puddle starts to form, the weld rod is brought to the tip of the inner torch cone where it is melted, and the molten rod is permitted to drop into the puddle. At that time, the lever on the spray torch is depressed with a quick snap introducing the flux into the puddle. Each time a drop of molten weld rod drops flux is added in this manner. This procedure is followed while moving in a circular manner until the region is slightly overfilled. After welding has been completed, the casting is slowly cooled to about 650° C in order to prevent cracking, and then air cooled to ambient temperature. This is of particular concern in larger castings.

Post weld heat treatment depends upon the metallurgical and physical property requirements for the specific casting. For ASTM A-536 ductile irons, the nodule count developed in the weld metal is sufficient to avoid the formation of excessive carbides or intercellular structure. Accordingly a stress relief heat treatment will relieved the residual stresses developed during cooling after weld repair. When the weld metal and base metal matrix structures are to be matched, appropriate heat treatment is necessary. In highly alloyed ductile irons, such as ASTM A-439, a full post weld heat treatment is necessary. That involves heating the casting to 1650° F and holding at temperature for 3 hours, or for one hour per inch of casting section, furnace cooling to 980° F followed by air cooling to ambient temperature.

This fusion weld repair procedure is applicable to a wide range of castings and casting section thicknesses. Typical thicknesses range from 3/8 to 10 inches.

Assessment of Effective Welds

A weld test coupon was devised for the purpose of illustrating the metallurgical characteristics and mechanical properties of ductile irons welded by this process. The geometry of this coupon is shown in Figure 1. This 1 inch thick coupon can be produced from any grade of ductile iron and contains a 3/8 inch deep oblong depression 1.5 in. wide by 3 in. long. The test piece is designed so that 0.5 in. D. tensile bars can be machined from the two legs, with the welded section in the reduced portion of the bar where a minimum of 60% of the test bar cross section is composed of weld metal. The weld test coupon may be heat treated in any manner called for to meet customer

specifications. The microstructure of the weld metal, heat affected zone and base metal may be determined from the weld test coupon or from the machined tensile bars.

Figure 1. Test piece used for the evaluation of welded ductile iron. Depression in cast block filled with weld metal. Test bars (2) machined so that the weld constitutes a minimum of 60% of the tensile test bar cross section in the reduced region.

ASTM A536 Grade 80-55-06. An example of typical microstructures obtained from welding this ductile iron grade (mildly alloyed to insure a pearlitic matrix) is shown in Figures 2a (transition zone) and 2b (weld fusion zone). The weld fusion zone exhibits a uniform dispersion of graphite spheroids with a high nodule count. Many of those spheroids have been enveloped by a shell of ferrite during solid state transformation. No carbides are present, and the nodule count is sufficiently high so that no significant intercellular structure has developed. The transition zone between the weld fusion zone and the base metal reveals the increased nodule count developed in the fusion zone. It is apparent that the alloy content of the filler metal was somewhat higher than that of the casting being welded as indicated by the difference in the ferrite shell thickness in those regions. There is no indication of the zone of partial fusion in this photomicrograph.

Tensile properties of typical as cast ductile irons in the unwelded and welded condition have been presented in Table III. The tensile properties are similar and represent acceptable ductile iron structures, although the elongation values indicate somewhat higher pearlite content.

ASTM A536 Grade 60-40-18. An example of typical microstructures in the base metal, weld metal and transition zone obtained in welding a 60-40-18 grade of ductile iron is presented in Figures 3a and 3b. In this case the welded casting was subjected to sub critical anneal to insure removal of pearlite form the matrix. This procedure,

however, was only partially successful in that a spherodized carbide structure is present in regions that were formerly pearlite. A full anneal followed by suitable slow cooling

Figure 2a (upper). Transition between base metal (upper portion of the photomicrograph) and weld metal (lower portion) of ASTM grade 80-55-06 ductile iron.

Figure 2b (lower). Weld fusion zone of welded grade 80-55-06 grade ductile iron.

would have eliminated these spherodized carbides. As with the pearlitic grade, the weld fusion zone exhibits a rather uniform dispersion of graphite spheroids and a high nodule count. The transition zone exhibits a similar spherodized pearlite structure in the base metal, in which the nodule count is somewhat lower than that of the pearlitic grade shown in Figure 2. Some residual evidence of the zone of partial fusion is present where a carbidic structure had developed around spheroidal graphite.

Table III shows typical tensile properties of welded and unwelded ductile irons that have been normalized (after welding). The tensile properties for the as cast material exceeds the 60-40-18 grade requirements, while the welded material properties represent acceptable ductile iron structures for a 65-45-12 grade with slightly higher pearlite present.

Figure 3a (upper). Transition zone of welded ASTM A539 grade 60-40-18 ductile iron illustrating the structure of the base metal (top portion of the photomicrograph) the transition zone and the weld fusion zone (lower portion of the photomicrograph).

Figure 3b (lower). Weld fusion zone of ASTM A539 grade 60-40-18 ductile iron.

ASTM A536 Grade 120-90-02. Tensile properties of welded and unwelded ductile irons produced to this grade of ductile iron are also shown in Table III. The properties of the unwelded and welded samples are similar and represent acceptable

ductile iron, although one of the welded samples is softer with lower strengths likely due to the high nodule count of the weld metal.

ASTM A439 Modified. Tensile test coupons taken from exemplar welded ASTM A439 Grade D5B modified ductile iron resulted in the data shown in Table III. It is apparent that the tensile properties of these two test bars did not meet the minimum values required in ASTM A439 Grade D5B, having been low in elongation and borderline in tensile strength, but quite adequate in yield strength. This is attributed to the aforementioned carbide network developed, primarily in the zone of partial fusion, due to the introduction of Mo in the modified version of this alloy. The enhancement of this network in the partial fusion zone resulted in decreased elongation and tensile strength which was not evident in either separately cast test bars or in test bars machined from casting sections.

Table III. Tensile Properties of Welded ASTM A536 Ductile Iron2

Table IV. Tensile Properties of Welded ASTM A439 Modified Ductile Iron

Summary
The ductile iron welding procedure described is a process applicable to the salvage of ductile iron castings, or to the joining of ductile iron components, resulting in machinable ductile iron microstructures in the weld metal, zone of partial fusion and base metal. Similar tensile properties in the weld metal, partial fusion zone and base metal could be expected if the microstructures are adequately controlled. However, it is anticipated that most applications of these casting repair procedures will be non-structural and will provide machinable material with a color match of the original casting.

Acknowledgements
The authors wish to thank others at Oil City Iron Works, Inc. for their support and assistance in the development and application of ductile iron welding: Eric Meyers, President; Eric R. Meyers, Vice President; and William Riley, Vice President, Manufacturing.

Figures 4a (upper). Transition zone between the base metal (upper portion of the photomicrograph) and the weld metal (lower portion) of ASTM A439 modified ductile iron. Figure 4b (lower). Weld fusion zone of ASTM A429 modified ductile iron.

References
1. Welding of Cast Iron, American Welding Society, Miami, FL, 1985.

2. D. T. Roberts, New Ductile Iron Welding Process Saves Castings and Dollars, Welding of Cast Iron, AWS, 1985, pp 69-71.