Welding Metallurgy and Weldability

1

Introduction

This textbook addresses the topics of welding metallurgy and weldability. The two topics are inextricably intertwined since the weldability of a material is closely related to its microstructure. While the term welding metallurgy is universally accepted as a subset of physical metallurgy principles, the term weldability has been subject to a wide range of definitions and interpretations. In its broadest context, weldability considers aspects of design, fabrication, fitness for service, and, in some cases, repair. This broad treatment is reflected in the definitions for weldability that are provided by both the American Welding Society and the ISO Standard 581:1980. Thus, weldability can be used to describe both the ability to successfully fabricate a component using welding and the capacity for that component to perform adequately in its intended service environment.

AWS Definition of Weldability

The capacity of a material to be welded under fabrication conditions imposed into a specific, suitably designed structure and to perform satisfactorily in the intended service.

ISO 581:1980 Definition of Weldability

Metallic material is considered to be susceptible to welding to an established extent with given processes and for given purposes when welding provides metal integrity by a corresponding technological process for welded parts to meet technical requirements as to their own qualities as well as to their influence on a structure they form.

In a Weld in g Journal article published in 1946 entitled This Elusive Character Called Weldability, W.L. Warren from the Watertown Arsenal in the United States stated, That word (weldability)…has grown to mansize in stature and importance in respect to its significance in modern welding application. This term has been and is used with such a variety of shades of meaning that one may rightly conclude weldability to possess a value as changeable as a chameleon [1].

Henri Granjon (Fig. 1.1) in his text Fundamentals of Welding Metallurgy defined weldability as …the behavior of (those) joints and the constructions containing them, during welding and in service… [2] R.D. Stout in Weldability of Steels states that the term weldability has no universally accepted meaning and the interpretation placed upon the term varies widely according to individual viewpoint [3]. At a conference held at The Welding Institute (TWI) in 1988 entitled Quantifying Weldability [4], Trevor Gooch from TWI (Fig. 1.2) stated that …the concept of weldability of a material is complex. At the same conference, A.D. Batte of British Gas Corporation is quoted as saying that …it is incongruous to find that the definition of weldability is still an active area of debate, and W.G. Welland from BP International stated that the concept of weldability is of little interest to the builders and users of most welded fabrications. Despite the many papers published by Warren F. Savage (Fig. 1.3) and his students at Rensselaer Polytechnic Institute and Fukuhisa Matsuda (Fig. 1.4) and his students at Osaka University, there are no definitions of weldability attributed to them (perhaps for good reason).

Figure 1.1 Henri Granjon, Institut de Soudure.

Figure 1.2 Trevor Gooch, The Welding Institute, 1992.

Figure 1.3 Warren F. Doc Savage, Rensselaer Polytechnic Institute, 1986.

Figure 1.4 Fukuhisa Matsuda, Osaka University, 1988 (W.A. Bud Baeslack III in the background).

In this text, weldability will be considered from the standpoint of materials' resistance or susceptibility to failure. From a fabrication standpoint, this relates to the ability to produce welds that are defect-free. There are multiple weld defects that can occur during fabrication, as described in Section 1.1, and these can be separated into those that are related to the welding process and procedures and those associated with the material. For example, defects such as lack of fusion, undercut, and slag inclusions are related primarily to the welding process and can usually be avoided by changes in process conditions. Defects such as solidification cracks and hydrogen-induced cracks are primarily related to the metallurgical characteristics of the material and are usually difficult to eliminate by changes in process conditions alone.

The term weldability also describes the behavior of welded structures after they are put into service. There are many examples of welded structures that are free of fabrication defects that later fail in service. These include failure modes involving corrosion, fatigue, stress rupture (creep), or complex combinations of these and other failure mechanisms. The service-related failure modes are perhaps the most serious of the weldability issues discussed here, since failure by these mechanisms can often be unexpected and catastrophic. As an example of this, consider the catastrophic Liberty ship failures (see Fig. 1.5) during World War II that led to the sinking of many transport ships and the loss of many lives.

Figure 1.5 Liberty ship failure.

This text will focus primarily on the aspects of weldability that are influenced by the metallurgical properties of a welded structure. As such, chapters addressing various fabrication cracking mechanisms are included. These chapters are designed to not only describe the underlying mechanisms for cracking but to provide insight into how such forms of cracking can be avoided. Similarly, the various forms of service cracking are described, particularly those associated with corrosion, brittle fracture, and fatigue. In order to provide the reader with sufficient metallurgical background to interpret the contents of these chapters, a chapter on welding metallurgy principles has been included.

1.1 Fabrication-Related Defects

Fabrication-related defects include cracking phenomena that are associated with the metallurgical nature of the weldment and process- and/or procedure-related defects. A list of common fabrication defects is provided in Table 1.1. The defects associated with the metallurgical behavior of the material can be broadly grouped by the temperature range in which they occur.

Table 1.1 Fabrication-related defects

Hot cracking includes those cracking phenomena associated with the presence of liquid in the microstructure and occurs in the fusion zone and PMZ region of the HAZ. Liquid films along grain boundaries are usually associated with this form of cracking.

Warm cracking occurs at elevated temperature in the solid state, that is, no liquid is present in the microstructure. These defects may occur in both the fusion zone and HAZ. All of the warm cracking phenomena are associated with grain boundaries.

Cold cracking occurs at or near room temperature and is usually synonymous with hydrogen-induced cracking. This form of cracking can be either intergranular or transgranular.

A number of nonmetallurgical defects that can occur during fabrication are also listed in Table 1.1. These are generally associated with poor process/procedure control and include lack of fusion, undercut, incomplete joint penetration, and geometric defects. Such defects can usually be remedied by careful attention to process conditions, joint design, material preparation (cleaning), etc. This text will not address the nature or remediation of these types of defects.

1.2 Service-Related Defects

Welds are subject to a wide range of service-related defects. Since welds are metallurgically distinct from the surrounding base metal and may contain residual stresses, they are often susceptible to failure well in advance of the base metal. These defects are usually manifested as cracks that form under specific environmental and/or mechanical conditions. A list of service-related defects is provided in Table 1.2.

Table 1.2 Service-related defects

Corrosion of welds is often a problem due to both the microstructural and local mechanical conditions of welded structures. The presence of fabrication-related defects can often accelerate service failures, particularly by fatigue. Welds in many engineering materials may contain softened regions that can promote mechanical overload failures. Conversely, local hard zones can result in reduced ductility and possible brittle failure.

1.3 Defect Prevention and Control

Although the understanding of the mechanisms leading to various forms of cracking is important, developing a methodology to prevent cracking is the ultimate goal of the welding engineer. Preventative measures can usually not be developed until the nature of the failure is understood. In some cases, changes in welding technique, or procedure, may be effective. For example, simple changes in heat input and bead shape can sometimes prevent weld solidification cracking. Another example is the use of preheat and interpass temperature control to prevent hydrogen-induced cracking.

Before such preventative measures can be implemented, the nature of failure must be determined in order that the cure does not lead to other weldability problems. Many Ni-base weld metals are susceptible to both solidification cracking and ductility-dip cracking, but the remedy for each defect type is different.

This textbook provides the necessary background to understand and interpret weld failures and recommends possible remedies for such failures. It should be noted, however, that the solution for many weldability problems will require a change in material rather than a tweaking of composition or process parameters. For example, reheat and strain-age cracking are significant problems when welding thick-section or highly restrained Cr–Mo steels and Ni-base superalloys, respectively. Again, knowledge of the precise mechanism of failure is required before remedial measures can be implemented.

References

[1] Warner WL. This elusive character called weldability. Weld J 1946;25 (3):185s–188s.

[2] Granjon H. Fundamentals of Welding Metallurgy. Cambridge, UK: Woodhead Publishing Ltd.; 1991 (Translated from French edition published in 1989).

[3] Stout RD. Weldability of Steels. 4th ed. New York: Welding Research Council; 1987.

[4] Pargeter RJ, editor. Quantifying Weldability. Cambridge, UK: The Welding Institute; 1988.

2

Welding Metallurgy Principles

2.1 Introduction

The purpose of this chapter is to review the basic principles that govern microstructure evolution in welds. Since the cracking susceptibility of welded structures is a function of microstructure, environment, and applied stress, it is essential to understand the basic principles that govern the evolution of microstructure during welding. This chapter will focus primarily on fusion welds, but a section specific to solid-state welds is also included. This is not meant to be an exhaustive review of welding metallurgy principles. For a more detailed treatment of this topic, the reader is referred to textbooks by Kou entitled Welding Metallurgy [1] and Easterling entitled Introduction to the Physical Metallurgy of Welding [2].

There are a number of metallurgical processes that control the microstructure and properties of welds. Melting and solidification are important processes, since they are the key to achieving acceptable joints in all fusion welding processes. Coupled with solidification are segregation and diffusion processes resulting in local compositional variations that influence both weldability and service performance.

Many metallurgical processes occur in the solid state, including phase transformations, precipitation reactions, recrystallization, grain growth, etc. The extent of these reactions may significantly alter the microstructure and properties of the weldment (weld metal and heat-affected zone (HAZ)) relative to the base metal. Many of these reactions, or complex combinations of reactions, can result in embrittlement, or cracking, of welds. This embrittlement can occur due to liquation, the presence of liquid films in an otherwise solid matrix, or in the solid state due to a loss in ductility.

Thermal expansion during heating and contraction during cooling can result in complex stress patterns in and around welds. These stresses can subsequently affect the microstructure and properties of the weldment and may promote cracking in regions where the tensile strain resulting from these stresses exceeds the ductility of the material.

The nature of the weld microstructure for a given material results from the combination of the weld thermal cycle and the material composition. In general, the heating and cooling rates associated with welding are quite high (10–1000°C/s) and usually prevent prediction of microstructure based on equilibrium thermodynamic principles. All of the metallurgical processes that influence the weld microstructure are temperature and heating/cooling rate dependent, and thus, the welding thermal cycle plays a key role in the evolution of microstructure and, ultimately, the weldability of the material, as shown schematically in Figure 2.1.

Figure 2.1 Block diagram for weld microstructure evolution and performance.

2.2 Regions of a Fusion Weld

Examination of a welded joint reveals distinct microstructural regions. The fusion zone is associated with melting. The HAZ, though not melted, is affected by the heat from the joining process. Beyond the HAZ is the unaffected base metal. The fusion zone and HAZ can be further subdivided, as described in this section.

The fusion zone is described as such because it is the region where melting and solidification occur to form the joint, or weld. Since all metals are crystalline in nature, many possessing cubic crystal lattices, there are general solidification phenomena common to all metals. In many materials, solidification behavior is very sensitive to composition. For example, the addition of small amounts of carbon and nitrogen to some steels can change their solidification behavior from ferritic (bcc) to austenitic (fcc). Minute additions of sulfur to steels can promote severe solidification cracking in the fusion zone. Aluminum alloys that are otherwise crack susceptible can be welded with a filler material containing more than 6% of silicon in order to avoid cracking.

The microstructure and properties of the HAZ are solely controlled by the thermal conditions experienced during welding and postweld heat treatment (PWHT). Aluminum alloys are routinely precipitation hardened or work hardened to increase strength; welding can completely eliminate these strengthening effects in the HAZ. Steel undergoes a phase transformation, which can result in a HAZ that has a radically different microstructure and properties than either the base metal or the fusion zone.

The understanding of regions of a weld has evolved tremendously since the 1960s. Prior to that time, a fusion weld was thought to consist of only two regions, the fusion zone and a surrounding HAZ, as shown in Figure 2.2 from a lecture by E.F. Nippes in 1959 [3]. Considerable research conducted by W.F. Savage and his students at RPI in the 1960s and 1970s revealed that other distinct regions of a fusion weld existed [4, 5].

Figure 2.2 Early schematic of regions of a fusion weld

(From Ref. [3]. © AWS).

In 1976, Savage et al. [4] proposed several changes to the terminology used to describe fusion weld microstructure regions, as shown in Figure 2.3. The fusion zone was considered to consist of two regions. The composite region represented the portion of the fusion zone where base metal and filler metal were mixed in a composite composition. Surrounding this region along the fusion boundary, they defined a region called the unmixed zone (UMZ). The UMZ consists of melted and resolidified base metal that does not mix with the filler metal. In some alloy systems, the UMZ can exhibit microstructures and properties very different from those of the composite region, particularly when dissimilar filler metals are used.

Figure 2.3 Regions of a fusion weld

(From Ref. [4]. © AWS).

The HAZ was subdivided into two regions, the partially melted zone (PMZ) and the true heat-affected zone (T-HAZ). The PMZ exists in all fusion welds made in alloys since a transition from 100% liquid to 100% solid must occur across the fusion boundary. In addition, other mechanisms were identified that resulted in local melting (or liquation) in a narrow region surrounding the fusion zone. These include grain boundary melting due to segregation and a phenomenon described as constitutional liquation that results from local melting associated with a constituent particle. The designation of a T-HAZ was used to differentiate that region of the HAZ within which all metallurgical reactions occur in the solid state, that is, no melting, or liquation, occurs.

Little has changed since 1976 regarding terminology for describing regions of a fusion weld, although considerable research has been conducted on a variety of alloy systems to verify that these regions actually exist in these material systems. Additional refinements have been made to this original terminology. For example, the T-HAZ in steels has been subdivided into various subregions, such as the coarse-grained HAZ (CGHAZ), the fine-grained HAZ (FGHAZ), and the intercritical HAZ (ICHAZ) regions.

The only potential addition to the terminology in Figure 2.3 is a transition region within the fusion zone. In heterogeneous welds, where the filler metal is of different composition from the base metal, this would represent a composition transition from the composite region to the UMZ. In some alloy systems, this transition zone (TZ) can exhibit a microstructure distinctly different from the surrounding regions. For example, in welds between stainless steels and low-alloy steels, a martensitic structure may form in the transition region that does not occur elsewhere in the weld.

A new schematic of the regions of a fusion weld is provided in Figure 2.4 for a heterogeneous weld. It is similar to the illustration in Figure 2.3 but contains a composition TZ that may be present in some systems. The following sections will review the various regions defined earlier in considerable detail and will describe the mechanisms involved in their formation.

Figure 2.4 Modern schematic showing regions of a fusion weld.

2.3 Fusion Zone

The fusion zone represents that region of a fusion weld where there are complete melting and resolidification during the welding process. The microstructure in the fusion zone is a function of composition and solidification conditions. Small differences in composition often result in large variations in microstructure and properties. In some systems, changing the solidification and cooling rates can also alter the microstructure, sometimes dramatically.

The fusion zone is normally very distinct from the surrounding HAZ and base metal when samples are prepared metallographically. This is due to both macroscopic and microscopic fluctuations in composition resulting from the solidification process.

In welds where the filler metal is of a different composition from the base metal, three regions theoretically exist. The largest of these is the composite zone (CZ), consisting of filler metal uniformly diluted with base metal. Adjacent to the fusion boundary, two additional regions may exist. The unmixed zone (UMZ) consists of melted and resolidified base metal where negligible mixing with filler metal has occurred. Between the UMZ and CZ, a transition zone (TZ) must exist where a composition gradient from the base metal to the CZ is present.

Three types of fusion zones have been defined: autogenous, homogenous, and heterogeneous. The classifications are based on whether or not a filler metal is used and the composition of the filler metal with respect to the base material. All three types of fusion zones are commonly encountered.

Autogenous welds are those where no filler metal is added and the fusion zone is formed by the melting and resolidification of the base metal. These are common in situations where section thicknesses are minimal and penetration can easily be achieved by the process selected. In thin sections, autogenous welding can often be applied at high speeds, and normally, a minimum amount of joint preparation is required, that is, butt joints can be used. Welding processes that are, or can be, adapted to autogenous welding include GTAW, EBW, LBW, PAW, and resistance welding. The fusion zone is essentially the same composition as the base metal, except for possible losses due to evaporation or pickup of gases from the shielding atmosphere. Not all materials can be joined autogenously because of weldability issues.

Homogenous welds involve the use of a filler metal that closely matches the base metal composition. This type of fusion zone is used when the application requires that filler and base metal properties must be closely matched. Properties such as heat treatment response or corrosion resistance are examples of such properties. Some common examples include the use of Type 316L base metal joined with 316L filler for matching corrosion properties and the use of E10016-D2 filler metal on AISI 4130 Cr–Mo steel, which is usually given a full PWHT to provide uniform strength.

Heterogeneous welds are fusion welds made with filler metals whose composition is different from that of the base metal. In many situations, matching filler metals may not exist or the weld properties desired may not be achievable with a matching composition. It should also be recognized that many base metal compositions may have inherently poor weldability and that dissimilar filler metals are required to achieve acceptable properties or service performance. Some considerations that would require the use of a dissimilar composition filler metal include strength, weld defect formation (e.g., porosity), weldability/solidification cracking resistance, heat treatment response, corrosion resistance, filler metal cost, and operating characteristics of the consumable.

When using a filler metal that has a composition different from the base metal, dilution effects must be carefully considered or the desired outcome may not be as expected. Common examples of heterogeneous welds include the use of Type 308L filler metal on Type 304L base metal for weldability and corrosion resistance and the use of a 4043 aluminum filler metal with 6061 aluminum base metal for solidification cracking resistance.

As noted earlier, the use of heterogeneous welds often requires close attention to dilution effects. Dilution can be defined as a change in composition of a filler metal due to its mixing with the base metal during the melting process. In many cases, dilution is not desirable and must be carefully controlled. Alteration of the deposited weld metal composition by dilution can negate or lessen the desired weld metal properties that would be achieved by a filler metal in its undiluted condition. One case where dilution is particularly undesirable is in surfacing operations where filler metals are significantly different from the base material and chosen to produce very specific properties such as abrasion resistance, corrosion resistance, or impact properties. For example, if stainless steels are used as cladding on carbon steels for corrosion resistance, significant dilution (~40%) can reduce the chromium content to a level where the clad layer is no longer corrosion resistant.

Dilution is expressed in terms of dilution of the filler metal by the base metal and is shown schematically in Figure 2.5. Mathematically, dilution is the ratio of the amount of melted base metal to the total amount of fused metal. For example, a weld with 10% dilution will contain 10% base metal and 90% filler metal. For most welding processes, dilution is normally controlled below 50%. Cross sections of welds, as shown in Figure 2.5, can be used to estimate dilution based on the original joint geometry, or the actual composition of the weld metal can be determined by analysis, and the dilution calculated if the compositions of the base and filler metals are known.

Figure 2.5 Schematic illustration of the determination of dilution in a heterogeneous weld.

2.3.1 Solidification of Metals

Melting and solidification are primary metalworking phenomena that allow for mixing of various elements to form an alloy that can then be solidified, or cast, into a form that will be used as an as-cast part or subsequently thermomechanically processed into other useful shapes (bar, plate, pipe, etc.). These phenomena are also the basis of the fusion welding processes, and a general knowledge of the solidification of metals is required to understand the metallurgical nature of a fusion weld.

There are several requirements for solidification. First, it is necessary to nucleate, or form, solid species within the liquid phase. Once the initial solid forms and the liquid-to-solid transformation proceeds, it is required that heat of fusion generated by the transformation be removed or dissipated. This normally occurs by conduction through the solid away from the solidification front. During the solidification of an alloy, it is also necessary to redistribute solute between liquid and solid, since the composition of the liquid and solid in contact at the solidification front changes continuously as the temperature decreases within the solidification range. This redistribution will result in local variation in composition in the solidified structure if the solid does not have time to reach its equilibrium composition, which is common in most casting and welding processes.

Most pure metals and alloys undergo a negative volume change when they solidify. This shrinkage phenomenon requires special precautions during casting to prevent shrinkage voids from forming. Solidification shrinkage also imparts stresses upon the as-solidified structure that may lead to solidification cracking. This shrinkage also contributes to the residual stress that is associated with fusion welds.

Using a simple phase diagram (Fig. 2.6), the equilibrium solidification behavior of a two-component alloy can be reviewed. For Alloy 1, solidification to solid A begins when the liquid temperature drops below the liquidus and ends when the alloy cools below the solidus. Within the solidification temperature range, the composition of liquid and solid in contact with each other at the solidification front is dictated by the isothermal tie line connecting the liquidus and solidus at a given temperature. At the end of solidification, Alloy 1 is 100% A.

Figure 2.6 Examples of different solidification paths in a simple eutectic system.

For Alloy 2, solidification proceeds as described earlier until the alloy reaches the eutectic temperature (T e). At this point, the remaining liquid, which is of eutectic composition, undergoes a eutectic reaction (L → A + B). The final structure will then be a mixture of A and eutectic (A + B). The relative proportions can be determined using the lever rule.

For Alloy 3, solidification will not proceed until the system reaches the eutectic temperature. At this temperature, the liquid will completely transform to a eutectic structure with the composition of the A and B phase determined by the maximum solid solubility (C Smax) of B in A and A in B at T e.

2.3.1.1 Solidification Parameters

A number of parameters are useful in describing microstructure development and solute redistribution during solidification. These are defined as follows:

Partition coefficient: k = CS/CL

Liquid temperature gradient: GL = dTL/dx

Solidification rate: R = dx/dt

Cooling rate: GL·R = dT/dt

The partition coefficient, k, sometimes called the solute redistribution coefficient, is simply the ratio of the solid and liquid composition in contact with each other at a given temperature within the solidification range. For most alloy systems, k is not a constant and varies as a function of temperature. It can only be constant in systems where the liquidus and solidus lines are straight, which is uncommon. When considering solute segregation during solidification, it is typical to assume an average value of k. When the value of k is less than 1, solute will partition to the liquid. When k is greater than one, solute will be depleted in the liquid. As the value of k approaches 1, solute redistribution during solidification is reduced.

The temperature gradient in the liquid (G L) is also an important parameter since it dictates the nature of the temperature field in advance of the solid–liquid (S–L) interface. In situations where some undercooling of the liquid has occurred prior to solidification, this gradient will be negative. This would be the typical situation for the solidification of a casting. During weld solidification, however, this gradient is normally positive since the weld pool is superheated by the welding heat source.

Solidification growth rate (R) is dictated by how fast the S–L interface is moving during the solidification process. When coupled with the temperature gradient in the liquid, the local cooling rate at the S–L interface can be determined. This latter value (G L·R) will have an influence on the dimensions of the solidification substructure, such as dendrite arm spacing.

2.3.1.2 Solidification Nucleation

In order for the solidification process to begin, it is necessary to nucleate solid within the liquid phase. This can occur either homogeneously or heterogeneously when a nucleating particle or solid substrate is present. Homogeneous nucleation requires that solid of a critical, or threshold, size form within the liquid. The size of this spherical nucleant can be defined by a critical radius size, r*,