TMS QUICK LINKS:
  • TMS ONLINE
  • MEMBERS ONLY
  • MEMBERSHIP INFO
  • MEETINGS CALENDAR
  • PUBLICATIONS

JOM QUICK LINKS:
  • JOM HOME PAGE
  • MORE HTML ARTICLES
  • COMPLETE ISSUES
  • BOOKS REVIEWS
  • SUBSCRIBE

As structural materials, nickel-based superalloys possess outstanding properties at elevated temperatures and play an important role for high-temperature applications, especially in the development of gas turbine engine components. The continual demand for improved performance of jet engines has pushed the usage of these alloys to even higher temperatures, which requires better mechanical properties, such as creep strength. The high-temperature strength of modern superalloys depends heavily on the use of refractory alloying elements. However, the composition of superalloys must be carefully controlled since the refractory alloying elements promote the formation of topological closed packed (TCP) phases which are detrimental to the mechanical properties. On the other hand, while focusing on the superior mechanical properties, the environmental stability of these alloys, such as oxidation and corrosion resistance, must be addressed due to the harsh operating environment of gas turbine engines. It is therefore desirable to develop nickel-based superalloys, which are inherently strong, and at the same time possess good oxidation and corrosion properties. This is challenging since optimal alloying compositions for strength often contradict those needed for oxidation and corrosion resistance. This leads to the inevitable use of environmental barrier coatings (EBCs) and thermal barrier coatings (TBCs) for high-temperature applications.

It was found in the 1970s that platinum and the other platinum-group metals (PGMs) can be used as major alloying constituents in nickel-based superalloys to promote enhanced resistance to both oxidation and corrosion at elevated temperature in various aggressive environments, especially the cyclic oxidation resistance.² This makes them attractive for the development of innovative TBCs for nickel-based superalloys. In recent years, the technology of depositing platinum onto nickel-based superalloys followed by treatments to produce an aluminum-rich layer has been used to produce improved protection at elevated temperature as compared to the conventional aluminide coating. Gleeson et al.³ showed that platinum plus hafnium modified γ+γ' coatings may be viable superior alternatives to current B2 type bond coatings. On the other hand, the PGMs are also very effective solid solution strengtheners to nickel-based superalloys comparable to the refractory metals,² which allows for the development of inherently strong nickel-based superalloys.

The purpose of this paper is to understand the properties of PGM-modified nickel-based superalloys from a thermodynamic point of view. Phase diagrams, which provide insight into phase stability, are road maps for alloy design and development. Traditionally, phase diagrams have been determined purely by experimentation, which is costly and time consuming. While an experimental approach is feasible for the determination of binary and simple ternary phase diagrams, it is less efficient for the complicated ternaries, and becomes practically impossible for higher order systems over a wide range of composition and temperature. On the other hand, commercial nickel-based superalloys are mostly multi-components in nature, and a more efficient approach is therefore needed to understand the phase equilibria when PGMs are used as alloying elements. The CALPHAD (i.e., CALculation of PHAse Diagram) approach will be adopted in this study. This approach has been used to understand binary, ternary, and even quaternary phase equilibria of nickel-based superalloys including PGMs.^4–7 In this study, it will be used to understand the thermodynamics and phase equilibria of PGM-modified multi-component nickel-based superalloys. In this paper, we will first give a brief introduction to the CALPHAD approach and the use of this approach in the development of a multi-component thermodynamic database for nickel alloys containing PGMs. The developed database will then be used to calculate phase transformation temperatures, phase fractions, and other properties for the multi-component nickel alloys containing PGMs and will be validated by the experimental data. Finally, the validated database is used to understand the effects of alloying elements, such as Al, Cr, Pt, Ir, and Ta, and their interplay on the properties of nickelbased superalloys.

The CALPHAD approach, which has been discussed a great deal in the past several decades,^8–11 is a phenomenological approach. The essence of this approach is to obtain self-consistent thermodynamic descriptions of the lower order systems—binaries and ternaries—in terms of known thermodynamic data measured experimentally and/or calculated theoretically, as well as the measured phase equilibria. The advantage of this method is that the separately measured phase diagrams and thermodynamic properties can be represented by the same “thermodynamic description” of a materials system in question. More importantly, on the basis of the known descriptions of the constituent lower order systems, a reliable description of a higher order system can be obtained via an extrapolation method.¹² This description enables us to calculate phase diagrams of the multi-component systems that are experimentally unavailable. Development of a thermodynamic description (usually called “thermodynamic database” or “database) of a multi-component system therefore starts with the development of the descriptions for all the constituent binaries and ternaries. For example, a quaternary system consists of six binaries and four ternaries and the database for the quaternary can be built up by combining the descriptions of these six binaries and four ternaries using a geometric model. It is found that the binary interactions are strong, the ternary interactions are less strong, and the quaternary and higher order interactions are several orders of magnitude smaller than those of the binaries. It is for this reason that a multicomponent system can be well predicted when the thermodynamic descriptions for the constituent binaries and ternaries are well developed. However, a 20-component system consists of 190 binaries and 1,140 ternaries. Not only are the assessments of so many binaries and ternaries not realistic, the lack of experimental data for some of these subsystems has made the development of a complete 20-component thermodynamic database impossible. As a result, a reasonable alternative is to focus on some key systems that are important for industrial applications and have abundant experimental information.

Nickel-based superalloys are very complex, and typically contain ten or more components in the system. The alloying elements usually belong to one of the three groups according to their effects. The first group of elements comprises those which partition to the γ matrix and impart solid solution strengthening, such as Co, Cr, Fe, Mo, and W. The second group consists of those which partition to the γ' precipitate and promote precipitation strengthening, such as Al, Ti, Nb, and Ta. The last group of elements comprises those which segregate to grain boundaries and aid ductility, such as boron and zirconium. Previously, we developed a database for nickel alloys, PanNi,¹ which has 19 components and includes elements from all three groups. In this work, we add Pt, Ru, and Ir into this database to study the properties of PGM modified nickel-based superalloys. To thoroughly understand the interaction of these three elements with all the 19 components currently included in the PanNi database would involve development of thermodynamic descriptions for 60 binaries and 571 ternaries, which is not a realistic task. In this study, we have focused on the Ni-Al-Cr-(Pt, Ru, Ir) system due to key roles played by Al and Cr in the Ni-based superalloys. The key binaries and ternaries for this subset are: Ni-Pt, Ni-Ru, Ni-Ir, Al-Pt, Al-Ru, Al-Ir, Cr-Pt, Cr-Ru, Cr-Ir, Ni- Al-Pt, Ni-Al-Ru, Ni-Al-Ir, Ni-Cr-Pt, Ni-Cr-Ru, and Ni-Cr-Ir. Thermodynamic descriptions for these systems have been developed and published by the authors of this work.^4–7 Readers are referred to these papers for details on the phase equilibrium and thermodynamic properties of these key subsystems.

In addition to the thermodynamic database for the Ni-Al-Cr-(Pt, Ir, Ru) system, interactions of the PGMs with other elements, such as Re, Ta, W, and Hf, need to be considered as well in order to understand the PGM-modified nickel-based superalloys. In this study, such interactions were developed only for the key phases, such as liquid, γ, and γ'. The database thus developed was then used to study the phase relationship for the PGM-modified nickelbased superalloys with special focus on these three phases. In particular, we will use this database to calculate liquidus, solidus, γ'-solvus, phase fraction and phase composition of γ and γ', and other properties for a variety of nickel alloys containing PGMs.

The database was first applied to the Ni-15Al-5Cr-0.1Hf-2.5Pt (at.%) and the Ni-15Al-5Cr-0.1Hf-2.5Ir (at.%) alloys. The fraction of γ' as a function of temperature is calculated for these two alloys and compared with the experimental data¹³ as shown in Figures 1 and 2. The calculated phase composition of γ and γ' and the partition coefficients for these two alloys at 1,000°C are listed in Table I and compared with the experimental data¹³ (in parentheses).

As can be seen from these two figures and the table, the calculated phase fractions of γ' agree with the experimental data very well for both alloys, while the calculated phase compositions are also in reasonable accord with the experimental data. The calculated hafnium composition in the γ phase is too low for both alloys, which leads to the strong partitioning of hafnium to the γ' phase. This is because the overall concentration of hafnium is very low and a small variation in either calculation or experimentation will lead to a big difference. For the Ni-15Al-5Cr-0.1Hf-2.5Pt alloy, the measured chromium concentration in the γ' phase is more than 1 at.% higher than that calculated, while the measured chromium concentration in the γ phase agrees with that calculated value very well. The measured chromium concentration in the γ' phase is believed to be too high when mass balance is applied. As can be seen from Figure 1, the calculated and measured phase fraction for the γ' phase match each other perfectly (0.34) at 1,000°C. The overall concentration of chromium is then calculated by the following equation:

By taking the experimental measured concentration of chromium in the two phases into the above equation, we get the overall composition of chromium to be 5.35 at.% which is higher than its nominal composition of 5 at.%. On the other hand, the calculated phase composition meets the mass balance and gives the perfect overall composition of 5 at.% Cr.

The database is then applied to alloys containing Re, Ta, Ru, and W. The compositions of the four groups of alloys studied in this work are listed in Table II. The purpose is to understand the effects of a variety of elements on the phase transformation temperatures, the amount of γ', and the partitioning of elements in the γ and γ' phases.

Group I alloys are based on Ni-15Al-1Re-0.1Hf (at.%) with Pt varying from 2.5 to 5 at.%, Ir 0 to 2.5 at.%, Cr 2.5 to 5 at.%, and Ta 0 to 2 at.%. Figure 3 shows the comparison between the calculated and measured14,15 liquidus and solidus for this group of alloys, and Figure 4 compares that of the γ' solvus for the same group of alloys. It is seen that the calculated liquidus temperatures agree with the experimental measured values very well. The calculated solidus temperatures for this group of alloys tend to be higher than those measured experimentally, while the difference is less than 20°C for the majority of these alloys. Group II alloys are based on Ni-15Al-2.5Ir- 5Cr-1Re-2Ta-0.1Hf (at.%) with Pt varying between 2.5 and 5 at.%, Ru between 1 and 2 at.%, and W between 1 and 2 at.%. The major difference between Group II and Group I is the use of Ru and W in Group II. Group III and IV alloys reduce the Al content while increase the Cr content as compare to Group II. Figure 5 compares the calculated and measured^14,15 liquidus and solidus for Group II-IV alloys. It is interesting to see that the calculated solidus are lower than those experimentally determined values for Group II-IV alloys, which is contrary to the Group I alloys. Yet, the agreement between the calculated and measured phase transformation temperatures for all these alloys is in general quite satisfactory. A similar plot for the γ' solvus was not shown for Group II-IV alloys due to the lack of experimental data. Figure 6 compares the calculated and the measured¹⁴ percentage of the γ' phase for several alloys from Group II-IV.

While Figures 1 to 6 serve as a validation of the database, the ultimate goal of this study is to predict the effects of a variety of elements on the phase relations of multi-component nickel alloys containing PGMs. In particular, it would be interesting to see which elements enhance the γ' precipitation (i.e., increase the fraction of the γ' phase) and how the phase transformation temperatures vary with the alloy composition. Heat treatment window (HTW), which is defi ned as the temperature range between the solidus and the γ' solvus, plays a key role in the selection of heat treatment temperature. A very narrow HTW will certainly lead to the difficulties in solutionizing the alloy. It is therefore of great help if the effect of different elements on the HTW can be predicted.

The first example is to study the effect of the chromium content on the properties of the Ni-15Al-5Pt-0.1HfxCr based alloy. As listed in Table III, the liquidus, solidus, γ' solvus, heat treatment window, and the amount of γ' at 1,000°C are calculated for this alloy with different amounts of chromium. Both liquidus and solidus decrease with the increase of the chromium content, while γ' solvus increases to the maximum at ~5 at.% Cr, then starts to decrease with higher chromium content.

The heat treatment window decreases at first, and stays at almost constant value with chromium content greater than 10 at.%. The amount of γ' reaches the maximum at ~10 at.% Cr, and starts to decrease as more chromium is added. The trend can also be illustrated by the phase diagram shown in Figure 7. As is seen in this figure, the liquidus and solidus keep going down as more chromium is added to the Ni-15Al-5Pt-0.1Hf-based alloy, while the γ' solvus reaches the maximum at ~5 at.% Cr. A similar trend is also observed for the Ni-Al-Cr alloys as shown in Figure 8, which compares the calculated and experimental measured16 γ' solvus lines for a variety of Ni-Al-Cr ternary alloys.

The effect of small amounts of Pt and Ir on the liquidus and solidus of the Ni- 15Al-0.1Hf-xCr based alloys is shown in Figure 9. It is seen that 2.5 at.% Pt slightly increases the liquidus but marginally decreases the solidus of the Ni- 15Al-0.1Hf-xCr based alloy, while 2.5 at.% Ir increases both the liquidus and solidus of the Ni-15Al-0.1Hf-xCr-based alloy. In general, depending on the concentration of chromium, the liquidus is ~5–20°C higher when 2.5 at.% Ir is used rather than platinum.

Tantalum has a significant effect on the γ' precipitate of nickel-based superalloys. One example is shown in Figure 10 for the Ni-15Al-5Cr-0.1Hf-2.5Ir- 2.5Pt-1Re alloy. It is seen that 2 at.% of tantalum greatly enhanced the precipitation of γ', which is beneficial in view of materials strength. On the other hand, 2 at.% Ta decreases the liquidus and solidus of the above alloy by ~20°C, while increasing the γ'solvus by ~150°C. This dramatically decreases the heat treatment window, which imposes a practical heat treatment challenge.

Aluminum is found to have an effect similar to tantalum. It is seen from Figure 11 that reducing Al content by 2 at.% for the Ni-15Al-5Cr-0.1Hf-2.5Ir- 2.5Pt-1Re-2Ta alloy increases the liquidus and solidus by ~20°C, while decreasing the γ' solvus by ~50°C. This opens up the heat treatment window by ~70°C, but the strengthening γ' phase is reduced by more than 15% in the temperature range shown in this figure.

The effects of Ru and W on the thermodynamic properties of the Ni-15Al- 5Cr-0.1Hf-2.5Ir-2.5Pt-1Re-based alloys are listed in Table IV. It is seen that 1–2 at.% Ru increases the liquidus slightly, but decreases the solidus slightly; it has almost no effect on the γ' solvus and the fraction of γ'. On the other hand, 1–2 at.% W does not change the liquidus, but somewhat decreases both the solidus and the γ' solvus. Tungsten does increase the fraction of γ' a little. Overall, a small amount of ruthenium and tungsten has a minimal effect on the phase transformation temperatures and phase amount for the Ni-15Al-5Cr-0.1Hf- 2.5Ir-2.5Pt-1Re-based alloys.

Elemental partitioning, which affects the shape of γ' precipitate and the coherence between γ and γ', plays an important role in determining the mechanical properties of nickel-based superalloys. It is therefore interesting to see how the elemental partitioning is affected by a variety of factors, such as temperature and other alloying elements. Figure 12 demonstrates two examples of such calculations. Figure 12a shows the temperature effect on the elemental partitioning of every element in alloy Ni-15Al-5Cr- 0.1Hf-5Pt. It indicates that temperature has significant effect on the partitioning of chromium for this alloy, while its effect on that of nickel is very small. Figure 12b shows the effects of chromium and tantalum on the partitioning of platinum for the Ni-15Al-5Pt-based alloy. It is seen that 5 at.% Cr has a marginal effect on the partitioning behavior of platinum, while tantalum reduces the partitioning tendency of platinum to the γ' precipitate.

The traditional development cycle of nickel-based superalloys using a trial-and-error experimental approach may last for many years due to the difficulty of finding the optimum alloy composition with desired mechanical properties and material stability. Modern material design inevitably involves a computational approach which provides guidance for the selection of alloy chemistry and processing conditions. In this study, thermodynamic calculation is used to understand the phase stability and thermodynamic properties of multi-component nickel alloys containing PGMs, in particular, the calculated liquidus, solidus, and γ' solvus provide the temperature windows for processing. The calculated fraction of γ' precipitate as a function of alloy compositions helps to identify the alloys that have potential for high strength, and the partitioning of elements is used to understand the γ/γ' misfit. These calculations are therefore used to guide the selection of optimum chemistry with balanced properties. For example, alloy chemistry needs to be carefully adjusted so that a high volume fraction of γ' precipitate is obtained for high strength, while a reasonable heat treatment window is maintained. Thermodynamic calculation in combination with key experiments shows great potential for accelerating the process of materials development and optimization.

A portion of this work was conducted due to the PMMS project sponsored by the Air Force Research Laboratory (AFRL) in collaboration with Rolls-Royce, University of Michigan, and Iowa State University. The authors acknowledge D. Ballard (AFRL), B. Gleeson (University of Pittsburgh), A. Heidloff (Iowa State University), T. Pollock (University of California at Santa Barbara), J. Van Sluytman (University of Michigan), A. Bolcavage and R. Helmink (Rolls-Royce) for technical discussions.

REFERENCES

F. Zhang and W. Cao are with CompuTherm, LLC, Madison, WI; J. Zhu is with the Department of Materials Science and Engineering, University of Michigan, Ann Arbor, MI; and C. Zhang and Y.A. Chang are with the Department of Materials Science and Engineering, University of Wisconsin, Madison, WI. Dr. F. Zhang can be reached at fan.zhang@computherm.com.