A Developmental Alloy for use in Industrial and Marine Gas Turbine Environments
The initial phase in the development of a new class of high temperature materials based upon the incorporation of platinum group metals as alloy constituents in nickel based superalloys has been reported here previously. It was shown that the new alloys provide enhanced resistance to environmental corrosion when compared with their conventional base metal counterparts, without detriment to high temperature strength. This paper describes the results of more detailed studies of the constitutional and structural aspects of these complex alloys which led to the design of a developmental alloy, RJM2012, specifically tailored for operation under the aggressive environmental conditions found in the industrial and marine gas turbine.
Advanced nickel-based superalloys are essentially precipitation strengthened austenitic alloys. They are complex in the sense that they may contain up to fifteen elements, each contributing some specific property to the material. The typical phases encountered within the general structure are: [i] the matrix, γ, [ii] the coherent γ′ (Ni3Al) precipitate and [iii] various refractory element carbides/borides. The latter phases have a wide spectrum of chemical composition and can be compositionally manipulated by control of casting conditions and/or heat treatment schedules. The overall properties with respect to creep strength, thermal and stress fatigue resistance, stress rupture life, oxidation resistance and hot corrosion resistance of this class of alloy depend critically upon the distribution of alloying elements between the phases and upon the relative proportions of these phases. Similarly, the operational temperatures and environments encountered by these alloys when subjected to service stresses for many thousands of hours do cause a significant redistribution of alloying elements throughout their structures.
The ideal superalloy is one with a combination of very high strength together with very high environmental corrosion resistance. Unfortunately these two properties have traditionally been mutually exclusive as a result of inherent structural instability associated with alloy chemistry, which limits the extent to which the stable oxide formers such as chromium can be utilised. By first considering simple alloys and then proceeding to more complex systems this paper demonstrates that previously unconsidered platinum and platinum group metal additions to this class of materials can be beneficial in producing a more desirable combination of properties. Such improvements have the potential to consolidate and extend the commercial use of these nickel-based superalloys in aggressive engineering environments (1).
Individual Phase Constitution and Properties
The Matrix Phase γ
The effect of various elements on the mechanical properties of the solid solution matrix phase (γ ) in nickel-based superalloys has been studied in some depth. Essentially it is now generally accepted that solid solution hardening may be attributed to the degree of lattice expansion, and changes in stacking fault energy (SFE) occasioned by the presence of solute elements. The effectiveness of solute elements in modifying the SFE of nickel at room temperature has been shown to be a function of the solute element electronic characteristics or electron vacancy number (Nv), and has been presented by Decker (2) from data published by Beeston and France (3) and by Beeston, Dillamore and Smallman (4). A linear correlation between the 0.2 per cent yield stress per unit lattice parameter change and Nv for various solute elements in nickel at room temperature has also been proposed by Pelloux and Grant (5) using some data published by Parker and Hazlett (6). Using these two latter correlations as a basis for predicting the effects of the platinum group metals on solid solution hardening of nickel at room temperature, it can be inferred that they should only provide a moderate improvement in yield stress characteristics.
This is also true for a model nickel-20chromium binary alloy containing ternary additions of gold, platinum, palladium and ruthenium as presented in Figure 1, and those for the base metal additions, namely niobium, tantalum, tungsten, molybdenum, titanium and cobalt which are presented above, in Figure 2.
However, an examination of the 0.2 per cent yield stress characteristics of the binary nickel-20chromium alloy with additions of selected ternary elements at temperatures from 25 to 1200°C, suggests that the platinum group metals can provide effective solid solution strengthening at temperatures exceeding 800°C.
A correlation between yield stress and element periodicity for temperatures up to 1200°C, is presented in Figure 3. The curves at each temperature follow a similar profile showing strong interaction effects at Nv values of 5.66 and 1.66. This indicates that the effectiveness of the traditional strengtheners such as niobium and tantalum reduces more rapidly than that of platinum and palladium, to such a degree that the latter elements become more effective strengtheners above 800°C.
The mechanical properties of the model nickel-20chromium alloy at temperatures above 800°C will depend primarily upon the relative diffusion characteristics of the solute additions and therefore the effect of the change in SFE with temperature. This can be considered a critical factor in the relative performance of these alloys, and ultimately of the more complex γ -phase compositions in nickel-based superalloys. It would seem likely from the data presented in Figure 3, that the platinum group metals, notably platinum and palladium, maintain a low SFE in the γ matrix at temperatures above 800°C.
The Precipitate γ′
The intermetallic phase γ′ based upon Ni3Al forms the principle basis for high temperature strengthening of the nickel matrix in most superalloy systems. The composition is complex, containing most of the alloying elements to varying degrees. The mechanical properties and environmental corrosion resistance of γ′ are therefore closely related to alloying behaviour and a great deal of work has been undertaken to gain an understanding of these effects. Much of the effort, however, has been directed at resolving the anomalous strengthening behaviour of γ′ with increase in temperature. Little if any attention has been given to the influence of environmental conditions on the corrosion properties of simple and complex γ′ compositions, despite the fact that this phase is often considered to be the weak link in the corrosion resistance of the bulk superalloy.
The strengthening and corrosion behaviour of γ′ alloys are therefore now considered in the following three sub-sections.
1. Strengthening in γ′ Precipitates
Although it was of primary interest to examine the effects of platinum group metals on the corrosion properties of simple and complex γ′ alloys, a number of comments can be made regarding the possible strengthening effects the platinum group metals may have on the mechanical properties of these alloyed precipitates.
The γ′ phase (Ni3Al) together with many other L12-structured A3B compounds show a significant increase in yield stress with increase in temperature (7–18). The inclusion of platinum as a substitutional element for nickel modifies this anomalous yet valuable behaviour but maintains, nevertheless, a degree of positive temperature dependence.
Overall the effect of platinum would be to maintain the yield stress values at a high level over the temperature range of interest. This inference may be drawn from published hot hardness data (18). Even at the limit of platinum solubility in γ′, positive strengthening has been observed with increase in temperature.
The results of published data on the strengthening behaviour of modified Ni3Al alloys would suggest a beneficial effect of substituted platinum group metal additions.
2. Corrosion of Ternary γ′ Alloys
Although Ni3Al has particularly good oxidation characteristics, the presence of sulphur species and/or alkaline earth metal salts within an oxidising environment causes severe degradation, commonly referred to as hot corrosion. It is known that the simple stoichiometric compound retains the L12 ordered structure at temperatures up to at least 1320°C, and has a narrow range of compositional stability but sufficient solubility for other elements to allow improvements to be made to both mechanical properties and corrosion resistance. Indeed, γ′ has been considered sufficiently interesting in terms of high temperature strength and stability to be used in its own right as a basis for a number of high temperature alloys (19–22). However, unless the environmental corrosion properties of γ′ can be improved, particularly within the 700 to 900°C hot corrosion regime, there is little possibility of commercial success. The inherent resistance of platinum to corrosive environments led to a general examination of platinum group metals as ternary additions to Ni3Al (γ′ ) in an attempt to improve this situation.
The elements which confer improved mechanical properties on the L12-structured γ′ phase, such as titanium, tungsten, niobium, vanadium, hafnium and tantalum (23), generally decrease resistance to hot corrosion attack. The data presented in Figure 4 for molybdenum, tungsten and titanium modifications to stoichiometric Ni3Al, where molybdenum and tungsten were substituted directly for nickel, and titanium for aluminium, provide evidence for this behaviour. Chromium and cobalt have no significant effect on corrosion resistance of γ′, as shown in Figure 5. Ruthenium and palladium are similarly neutral in their effect whereas platinum reduces the corrosion rate almost to zero and the alloy maintains a stable protective oxide throughout the period of the test.
3. Corrosion of Complex γ′ Alloys
In order to verify the corrosion benefits attributed to selected platinum group metal modifications of stoichiometric Ni3Al, the high temperature hot corrosion resistance of platinum-modified multi-element gamma prime phases was examined. An example of the performance of such an alloy modification is presented in Figure 6 where 2.5 atomic per cent of nickel was replaced by platinum. As distinct from the previous model ternary γ′ alloys, shown in Figure 5, the overall corrosion performance of stoichiometric Ni3Al has been improved, suggesting some synergistic effect between platinum and the other solute elements. An experimental γ′ alloy designed particularly for strength improvements over conventional γ − γ′ alloys, (19) was also modified in composition to include 6.2 weight per cent platinum. This alloy, along with the non-platinum containing γ′ alloy, was subjected to a gas turbine simulation test operating at a gas temperature of 900°C with sulphur dioxide and sea salt contaminants injected at levels of 1.56 litres per hour and 24.2 ppm, respectively. The photomacrographs presented in Figure 7 show the dramatic effect the modification has had upon the hot corrosion resistance of the original platinum-free alloy.
Phase Interaction and Composition
γ − γ′ Alloys
The ternary gamma (γ ) and gamma prime (γ′ ) model alloys described earlier are partial analogues to the true phases found in complex precipitation strengthened nickel-base alloys. The composition and volume fraction of the γ′ precipitate, which ultimately dictates high temperature strength in γ − γ′ alloys, are controlled by overall alloy composition and heat treatment and in order to obtain full benefit from alloying additions it is important to establish the chemical relationships between the individual phases. Examples of the latter exercise can be seen in the publications of Kriege and Baris (24) and also Donachie and Kriege (25).
Establishing the γ and γ′ constitutional relationships in γ − γ′ alloys containing platinum group metals has proved to be valuable in designing corrosion resistant alloys and much of the chemical data produced has been obtained using the electrolytic phase extraction and chemical analysis techniques established by Kriege and Baris (24).
Traditionally, partitioning data produced by the latter technique are represented as ratios of the weight percentages of elements found in each phase, and when a series of values for different elements are averaged the error is such as to allow these ratios to represent either weight or atomic percentage data. This type of representation allows comparison of element partitioning for alloys with a wide range of precipitate fraction, but is not readily applicable to the design of novel alloy systems.
In dealing with the partitioning effects of elements in γ − γ′ alloys containing single platinum group metals, partitioning coefficients have been established by identifying the relative proportions of atoms of a particular element associated with each of the two major phases and comparing these values with the atomic percentage levels in the alloy composition. Partitioning coefficients generated by this technique are valid for alloys containing similar volume fractions of γ′ precipitate (∼50 to 70 per cent). A series of γ′ partitioning coefficients are presented in Figure 8 for alloys which contain a single platinum group metal.
The composition of the matrix γ in alloys containing a fixed γ′ precipitate volume fraction is essentially predetermined by the solubility limits of alloying elements in γ′ and the elemental partitioning coefficients. It is essential in the design of any new precipitation strengthened nickel-based superalloy, particularly where different elements such as the platinum group metals are being used, that the γ composition is controlled so that detrimental phase precipitation, notably sigma but also other topographically close packed phases, does not occur during extended periods of elevated temperature service.
The partitioning coefficient data generated from the work on complex γ − γ′ alloys containing single or multiple platinum group elements were used in a modified “phacomp” programme calculation to produce appropriate alloy average electron vacancy numbers (Nv) for the design of sigma-safe alloys. Nv values for γ − γ′ alloys containing one or more of platinum, palladium or ruthenium can be determined using the formula:
where element compositions are in atomic per cent.
The maximum Nv for avoidance of sigma phase has been determined experimentally as 2.22.
Apart from osmium, the platinum group elements strongly partition to the γ′ precipitate thereby exercising a major influence upon strength and corrosion resistance. Based upon the fundamental understanding of the effects of platinum group metals on the properties of γ, γ′ and γ−γ′ alloys, an alloy designated RJM2012 was designed and developed to achieve a strength capability at least equal to alloys of the IN792 type but with a low temperature (700 to 900°C) hot corrosion resistance approaching that of higher chromium alloys, such as IN939. Table 1 presents the nominal composition of commercial and developmental alloys.
A Developmental (γ − γ′ ) Alloy
The alloy RJM2012 was designed on the basis of the fundamental understanding that developed from a consideration of the effect of the platinum group metals upon the primary components of γ − γ′ alloys and was formulated specifically to overcome many of the environmental problems known to cause severe degradation of these systems such as hot corrosion attack.
A comparison of cyclic oxidation data for RJM2012 and IN792 at peak temperatures ranging from 900 to 1100°C is shown by Figures 9 and 10. For temperatures up to 1000°C there is little difference in performance between the two alloys but at 1100°C after 600 temperature cycles RJM2012 shows a distinct and worthwhile improvement in performance.
Current studies being carried out on the isothermal oxidation characteristics of RJM2012 at the University of Liverpool have provided some understanding of this effect by demonstrating a distinct morphological difference in the oxidation scales at 1100°C compared to those at 900°C. Using IN792 for comparison, the corrosion product morphologies of both alloys tested at 900°C over a period of 800 hours show similar characteristics, comprising a duplex non-continuous Cr2O3 and a sub-surface Al2O3 scale. However, at 1100°C, although similar oxidation behaviour is observed up to 100 hours of test, RJM2012 develops a continuous protective Al2O3 scale by the lateral growth of the initial internal oxidation zone after spallation of the initial non-protective duplex scale. In contrast, IN792 continues to develop a non-protective scale, with the formation of NiO becoming more predominant as the exposure time increases.
The influence of platinum on oxide morphology and oxidation kinetics is as yet little understood. However, the implications of using platinum as an alloying element in high temperature materials is quite evident. Work is continuing in order to explain and further develop many of the beneficial corrosion properties of RJM2012.
Stress Rupture Properties
The 1000 hour stress rupture properties of RJM2012 together with published data on other commercial alloys are presented in Figure 11. Depending upon thermal history the properties presented for RJM2012 fall within the limits represented in the diagram. Further stress rupture data is currently being generated together with information on the high cycle fatigue performance of the alloy, as part of the overall optimisation process. In general, however, it is clear that the mechanical properties of this alloy will match very closely those of the current commercial alloy IN792.
RJM2012 in a Simulated Turbine Environment
The Johnson Matthey burner rig, which simulates the corrosive conditions found in turbines operating under severe environmental operating conditions, has been described in detail elsewhere (26).
RJM2012, together with other commercial and developmental alloys investment cast in turbine blade form, were evaluated under the programme conditions set out in Table II. The data presented in Figures 12 and 13 describe the results of a full analysis of blade corrosion after 800 hours at temperatures of 760 and 900°C, respectively. Each blade was sectioned at the tip, midspan and base to determine the degree of sulphide and general oxide penetration. Immediately evident from a comparison of the two sets of data is the degree of hot corrosion attack on all the test alloys under programme A conditions. That is, degradation of blade alloys over the temperature range 740 to 780°C is more pronounced than over the 880 to 920°C temperature regime.
The environmental performance of RJM2012, shows however, in both cases a remarkable improvement over that of the more conventional alloys. Although it is appreciated that many alloys receive a full coating treatment (26) prior to service, the performance of the base alloy both in terms of mechanical properties and corrosion resistance is just as important should localised degradation of the coating occur.
The photomicrographs presented as Figure 14 compare the surface degradation of alloys RJM2012 and IN792 (modification C101) after exposure to the lower temperature conditions for 800 hours, and provide clear visual evidence of the improvements gained.
It is clear from the results of these two programmes that RJM2012 satisfies the original alloy design objective of achieving a superior corrosion resistance without detriment to high temperature strength. While this alloy has been specifically developed for marine and industrial gas turbines, the platinum group metal concept can be and is being utilised to provide alloys for a wide range of industrial applications with the precise compositional requirements dictated to a considerable degree by the environment to which the structural material is exposed during service. Environmental and economic pressures are placing increasingly severe demands upon materials of construction; it is anticipated that the selective use of platinum metals as alloying constituents will provide a cost-effective means of meeting performance targets.
C. W. Corti,, D. R. Coupland and G. L. Selman, Platinum Metals Rev., 1980, 24, (1), 2
R. F. Decker, Proc. Symp. Steel Strengthening Mechanisms, sponsored by Climax Molybdenum Co., Zürich, 5–6th May, 1969, pp. 147 – 170
B. E. P. Beeston and L. K. France, J. Inst. Met., 1968, 96, 105
B. E. P. Beeston,, I. L. Dillamore and R. E. Smallman, Met. Sci. J., 1968, 2, 12
R. M. N. Pelloux and N. J. Grant, Trans. Met. Soc. AIME, 1960, 218, (4), 232
E. R. Parker and T. H. Hazlett, Trans. A. S. M., 1954, 46A, 701
H. Gleiter and E. Hornbogen, Phys. Status Solidi, 1965, 12, (1), 235 and 251
B. H. Kear,, J. M. Oblak and A. F. Giamei, Met. Trans., 1970, 1, (9), 2477
J. R. Nicholls and R. D. Rawlings, J. Mater. Sci., 1977, 12, (12), 2456
S. J. Liang and D. P. Pope, Acta Metall., 1977, 25, (5), 485
R. B. Schwarz and R. Labusch, J. Appl. Phys., 1978, 49, (10), 5174
K. Suzuki,, M. Ichihara and S. Takeuchi, Acta Metall., 1979, 27, (2), 193
T. Suzuki,, Y. Oya and D.-M. Wee, Acta Metall., 1980, 28, (3), 301
T. Suzuki and Y. Oya, J. Mater. Sci., 1981, 16, (10), 2737
B. Reppich, Acta Metall., 1982, 30, (1), 87
B. Reppich,, P. Schepp and G. Wehner, Acta Metall., 1982, 30, (1), 95
D.-M. Wee,, O. Noguchi,, Y. Oya and T. Suzuki, Trans. Jpn. Inst. Met., 1980, 21, (4), 237
D.-M. Wee and T. Suzuki, Trans. Jpn. Inst. Met., 1979, 20, (11), 634
J. E. Restall, Proceedings of the 3rd Int. AIME Symposium on Superalloys, Seven Springs, PA, Sept., 1976, pp. 351 – 357
J. E. Restall and M. J. Weaver, British Patent 1,381,859 ; 1975
J. E. Restall and M. J. Weaver, U.S. Patent 3,922,168; 1975
A. S. Pratt and D. R. Coupland, British Patent 2,029,857 B; 1982
R. D. Rawlings and A. E. Staton-Bevan, J. Mater. Sci., 1975, 10, (3), 505
O. H. Kriege and J. M. Baris, Trans. A.S.M., 1969, 62, (1), 195
M. J. Donachie and O. H. Kriege, J. Mater., 1972, 7, (3), 269
R. G. Wing and I. R. McGill, Platinum Metals Rev., 1981, 25, (3), 94
We would like to acknowledge the assistance given to us by Rolls Royce and NGTE in the preparation of samples for burner rig evaluation.
In addition we would like to thank Dr. G. Tatlock and T. Hurd of the University of Liverpool for valuable comments on aspects of the oxidation work.
The alloy designation Mar-M002 is a trade mark of the Martin Marietta Corporation; IN792, IN738C and IN939 are trade marks of Inco Limited; B1925 is a trade mark of Sorcery Metals.