Читать книгу Engineering Physics of High-Temperature Materials - Nirmal K. Sinha - Страница 74

Nickel is the fifth most abundant element on Earth, but most of it is located almost 3000 km below the surface (United States Geological Survey 2012). In Earth's crust, magmatic sulfide deposits (principal ore mineral pentlandite, (Ni,Fe)₉S₈) and laterite deposits (ore mixtures of nickeliferous limonite, (Fe,Ni)O(OH), and garnierite – a mixture of hydrous nickel and nickel‐rich silicates) (United States Geological Survey 2012). Roughly 3000 nickel‐base alloys are in use, forming products for numerous industries, including energy, chemical, petrochemical, and aerospace industries. Ni‐base systems have risen as systems of choice for tailored superalloys largely due to the nature of the precipitates and resulting properties that can be achieved in the system. As such, a deeper look into nickel‐base alloys is critical to frame later discussion in the text.

Nickel‐base alloys tend to have a fully austenitic (FCC) structure and the ability to maintain good tensile, rupture, and creep properties to much higher temperatures than (BCC) systems. They are often used in load‐bearing structures and nickel‐base superalloys have been reported being used to highest homologous temperature (up to T _m = 0.9) of any common alloy system (Bowman 2000). Like Fe‐base alloys, the properties of Ni‐base alloys can be tailored through the addition of many other elements, including chromium, iron, cobalt, molybdenum, tungsten, tantalum, aluminum, titanium, zirconium, niobium, rhenium, yttrium, vanadium, carbon, boron, and hafnium. Nickel‐base alloys can be solid‐solution or precipitation strengthened, but the latter is generally used for more demanding applications.

The properties of Ni‐base superalloys have largely arisen through their unique phases and the ability to tailor the properties through the addition of various elements. Ni‐base superalloys are generally composed of a gamma (γ) phase, which forms an austenitic solid‐solution matrix composed of the alloying elements. As the alloys are cooled from the melt, gamma prime (γ′) phases begin to precipitate within the γ‐matrix (Figure 2.16) (Betteridge and Heslop 1974; Ross and Sims 1987). γ′ phase is a GCP phase generally composed of Ni₃(Al,Ti) in Ni‐base superalloys. It is the main strengthening phase in the alloy and has strong coherency with the matrix, which allows for ductility. The precipitates can take on several different geometries, including cuboidal, spheres, platelets, or combinations thereof. However, to decrease their energy states, they often align along the <100> directions and form cuboidal structures (Sabol 1969). Many Ni‐base superalloys can thus be described as having ordered γ′ particles within a disordered γ‐matrix.

The γ–γ′ system forms the basic structure of Ni‐base superalloys, but several other phases can also be present:

 A gamma double prime (γ″), often with the composition of Ni3Nb or Ni3V, can form small disk‐like precipitates with a GCP body‐centered tetragonal structure. These precipitates form with the (001) planes of the precipitate (γ″) parallel to the {001} family in the matrix (γ) as a result of lattice mismatch. They can strengthen the superalloy at lower temperatures compared to the γ′ phase, but often dissolve at higher temperatures.Figure 2.16 Schematic of cuboidal γ′ (Ni3(Al,Ti)) precipitates within a γ [FCC‐structured Ni, Co, and Fe solid solution]. In this schematic, the characteristic precipitate size is shown as dγ′.Sources: Betteridge and Heslop (1974); Ross and Sims (1987).

 In addition to GCP phases, several phases can form, and many of these can be undesirable due to poor mechanical properties, incoherency with the matrix, and morphologies, such as needle‐like structures that can nucleate cracks as well as depleting the materials ability to form the desired γ′ particles (Belan 2016).

 Carbide phases are also formed from carbon and reactive and refractory elements, such as titanium, tantalum, and hafnium (e.g. TiC, TaC, or HfC). With heat treatment (and/or service), the carbides can decompose to lower metal carbides, such as M23C6 and M6C, and often aggregate to the grain boundaries. Many carbides have an FCC structure and are known to support strengthening at grain boundaries in polycrystalline materials (Sabol 1969). However, their role in single crystal material properties is still under debate (Bowman 2000).

 Cr and Al in the systems can also protect against oxidation and corrosion by forming protective oxide phases.

The addition of other elements can further tailor the properties of the alloy. For example, not only does higher Al and Ti result in higher volume fractions of the γ′‐phase, but also the addition of cobalt reduces the solubility of Al and Ti in the NiCr matrix and further promotes precipitation. This improves strength at high temperatures, as well as the workability of these alloys. Other elements are added to improve microstructural stability, strength, and deformation characteristics. The creep‐rupture life of these alloys can be increased significantly by the careful control of composition and microstructure. For different uses, variations in composition and process are critical leading to a variety of different nickel‐base superalloys in use. An example of the chemical compositions of some well‐known nickel‐base alloys is given in Table 2.2.

Table 2.2 Nominal chemical compositions (wt. %) of Nimonic 90 (Betteridge and Heslop 1974), IN‐738LC and René 80 (Balikci and Raman 2000; ASMH 1991), Waspaloy (ASMH 1991), and CMSX‐10 (Erickson 1996).

Sources: Betteridge and Heslop (1974), Balikci and Raman (2000), ASMH (1991), Erickson (1996).

Element	Nimonic 90	IN‐738LC	René 80	Waspaloy	CMSX‐10
Ni	≈54.6 (balance)	≈61.2 (balance)	≈60 (balance)	≈58.3 (balance)	≈69.3 (balance)
Cr	19.6	16	14	19.5	2
Co	18	8.5	9.5	13.5	3
Al	1.4	3.5	3	1.3	5.7
Ti	2.35	3.5	5	3	0.2
W	—	2.6	4	—	5
Mo	0.3	1.8	4	4.3	0.4
Ta	—	1.7	0.05	—	8
C	0.09	0.1	0.17	0.08	—
Fe	1	0.1	0.18	—	—
B	0.003	0.01	0.1	0.006	—
Zr	0.07	0.1	0.06	0.06	—
Nb	—	0.9	—	—	0.1
Si	1.5	Trace	Trace	—	—
Cu	0.2	—	—	—	—
Mn	1	Trace	Trace	—	—
Re	—	—	—	—	6
Hf	—	—	—	—	0.3

Due to economic and environmental demands, the efficiencies of gas turbine engines have been steadily and systematically increased by design modifications, such as air cooling of components, to handle higher turbine inlet to firing temperatures. The gas turbine engines of today are more energy efficient and more reliable, as well as have cleaner burning capabilities, than their predecessors. The increase in operating temperatures, however, has led to the decrease in the life of components and increase in costs of replacement. This is serious because the technologies are becoming obsolete in a very short time, while the world's commercial fleet of aircraft or utilities is getting older. Around 80% of the large frame industrial/utility gas turbines operating in the world today were installed in the mid‐1960s to early 1970s, and they are now old and getting replaced. Consequently, there are now greater opportunities to repair and refurbish older models (Natole 1995; Valenti 1999). It is estimated that 50–70% of the operating and maintenance cost is due to the repair and/or replacement of hot section components made of nickel‐base superalloys. Repair or replacement of damaged engine components is not therefore a question of academic importance.

Figure 2.17 Dependence of strain age cracking susceptibility based “weldability” on the contents of aluminum and titanium. A general increasing trend for γ′ solvus temperature of the alloys is also shown; chemical compositions for alloys shown as solid points are given in Table 2.2.

Source: N. K. Sinha.

The repairability of a superalloy component is determined by a number of factors, such as chemical composition, grain size, grain‐boundary characteristic, substructural details of the matrix, prior heat treatment, history of service life, welding parameters, and welding processes. A full understanding of the interrelationship between the material characteristics, welding parameters, and the performance of the welded junctions requires knowledge of the physical chemistry of the parent and the filler materials involved.

The same features that improve high‐temperature performance of nickel‐base superalloys also make them vulnerable to structural degradation, such as coarsening of the γ′ precipitates and the oxidation of nickel and carbon producing subsurface grain‐boundary defects during service life, and also make them difficult to repair by welding (Jones and Westerman 1965; Thamburaj et al. 1983). Liquidation zone cracks (LZC) and heat‐affected zone (HAZ) cracks develop due to thermal stresses generated by volumetric changes associated with the dissolution and reprecipitation of γ′‐phase during welding (Thamburaj et al. 1983). The propensity for weld induced cracking (both LZC and HAZ) increases as the concentration of γ′ forming elements, such as aluminum and titanium, increases in the alloy, and hence as the volume fraction of γ′ increases. Figure 2.17 shows a few commonly used nickel‐base superalloys along with the demarcation line (dashed) proposed by Prager and Shira (1968) to differentiate between weldable and nonweldable alloys. They examined the weldability on the basis of strain age cracking susceptibility. Most of the popular “polycrystalline superalloys” being used today were introduced between 1960 and 1985 (Stephens 1989), but observations confirm that alloys containing more than about 4 wt.% total content of Ti and Al are in general difficult to weld (Ikawa et al. 1974; Kelly 1990).

While investigating the weldability improvement techniques of cast nickel‐base superalloys, the present author was frustrated by the fact that it was extremely difficult to keep track of the chemical compositions of different commercial alloys to develop a mental picture of their differences, which are often very small. There are excellent publications and books on phase transformations in metals and alloys showing complex phase diagrams (Betteridge and Heslop 1974; Ross and Sims 1987; Porter and Easterling 1992). The chemical compositions are, however, always given either in rows of elements with wt.% in the brackets and separated by commas or tables with columns of elements and rows of wt.%. In order to develop a visual representation to facilitate quick recognition of the differences, the author tried different approaches. One such graph is given in Figure 2.18. Here, the chemical compositions of three nickel‐base superalloys, Nimonic 90, IN‐738LC, and René 80 from Table 2.2, are compared. The chemical compositions in Table 2.2 were first entered in decreasing amounts of the major elements keeping in mind to enter the common elements first.

Figure 2.18 Chemical composition diagrams of three nickel‐base superalloys. The data are taken from Table 2.1. Whereas Nimonic 90 is weldable, IN‐738LC and René 80 are very difficult to weld as indicated in Figure 2.16.

Source: N. K. Sinha.

The complexity of the chemical compositions of superalloys, with a dozen or more important elements (and several “tramp” elements, such as oxygen, nitrogen, sulfur, phosphorous, and silicon, and “trace” elements, such as bismuth, lead, selenium, and thallium), may at first glance appear to be confusing (Betteridge and Heslop 1974; Ross and Sims 1987; Porter and Easterling 1992). However, the chemical compositions are carefully controlled to impart desired mechanical and service exposure requirements. Although Table 2.2 provides the data quantitatively, Figure 2.18 with the use of logarithmic scale (that may be extended to show the “trace” and “tramp” elements) for the composition brings out the differences in the amount of several elements that are used in small quantities in these alloys. Note the positions of these three alloys in the “weldability” plot (Figure 2.17). This type of compositional representation was appreciated by the colleagues of the author, and the primary objective of this communication is to share this idea with other researchers whose main background may not be metallurgy.

Thus far, we have focused our discussion on polycrystalline material. Single crystal (SC or SX) materials, with (ideally) no grain boundaries in the entire sample, are generally prepared using a modified directional solidification technique. While the process can be time‐ and cost‐intensive, the properties of single crystal material can be quite unique. This will be explored throughout this text.

With proper heat treatment, it is generally assumed that there are very few dislocations in cubes of γ′ as well as in the matrix of γ single crystal materials. However, the microstructure within the grain can still be optimized through the addition of microalloying elements, and several different single crystal Ni‐base superalloys are currently in use. The first generation of SX alloys was essentially based on modifications of the chemical composition of earlier successful polycrystalline nickel‐base superalloys. Rhenium additions improved certain desirable properties (Giamei and Anton 1985), and alloys containing up to about 3% rhenium were developed. These are called the second‐generation SX alloys. The rhenium content was then increased further to about 6%, and these alloys were labeled as the third‐generation SX alloys. For example, CMSX‐10 is a third‐generation SX alloy with a nominal composition provided in Table 2.2.