Creep rupture properties of bare and coated polycrystalline nickel-based superalloy Rene ® 80

Creep deformation is one of the life time limiting reasons for gas turbine parts that are subjected to stresses at elevated temperatures. In this study, creep rupture behavior of uncoated and platinum-aluminide coated Rene ® 80 has been determined at 760°C/657 MPa, 871°C/343 MPa and 982°C/190 Mpa in air. For this purpose, an initial layer of platinum with a thickness of 6µm was applied on the creep specimens. Subsequently, the aluminizing were formed in the conventional pack cementation method via the Low Temperature-High Activity (LTHA) and High Temperature-Low Activity (HTLA) processes. Results of creep-rupture tests showed a decrease in resistance to creep rupture of coated specimen, compared to the uncoated ones. The reductions in rupture lives in LTHA and HTLA methods at 760°C/657 MPa, 871°C/343 MPa and 982 °C/190 MPa were almost (26% and 41.8%), (27.6% and 38.5%) and (22.4% and 40.3%), respectively as compared to the uncoated ones. However, the HTLA aluminizing method showed an intense reduction in creep life. Results of fractographic studies on coated and uncoated specimens indicated a combination of ductile and brittle failure mechanisms for all samples. Although, the base failure mode in substrate was grain boundary voids, cracks initiated from coating at 760°C/657MPa and 871°C/343. No cracking in the coating was observed at 982°C/190MPa.


Introduction
Cast polycrystalline nickel based superalloy Rene ® 80 (a well-known turbine blade alloy-GE-trademark), is extensively used for manufacture of rotating blades of aero-turbo jet engines.
This superalloy is employed at temperatures between 760°C and 982°C. Many design applications depend on alloys that are subjected for a long time to high-temperature conditions. High temperature creep under elevated centrifugal stress leads to cracking of grain boundaries perpendicular to the principal stress axis that limits the life of the rotating blades. The temperature regime, for which creep is important in alloys is 0.5Tm < T < Tm, where, Tm is the melting temperature of the alloys [1,2].
Although the superalloys have comparatively favorite mechanical properties at elevated temperatures, their oxidation and corrosion resistance are not satisfactory at high temperatures.
The aluminide diffusion coating is deposited on the surface of airfoil turbine blades, in order to enhance the oxidation resistance of the base alloys at high temperatures. Modification of diffusion aluminide coatings by platinum has been known to significantly improve their resistance to high-temperature oxidation [3][4][5]. To form the platinum-aluminide (Pt-Al) coatings, an initial layer 5-10 μm of platinum was electroplated, and then heat-treating to diffuse the platinum into the substrate prior to aluminizing. Diffusion coatings play a remarkable role in the mechanical behavior of the substrate at the different temperatures. Effect of Pt-Al coating on the tensile properties of a nickel-base, single crystal superalloy was evaluated by Parlikar et al. [6]. The results of mentioned work indicated that application of Pt-Al coating on the superalloy led to a decline in the strength properties of the alloy (yield and ultimate tensile strength) at temperatures between ambient and 1100 °C. On the other hand, the coated superalloy exhibited more ductile behavior than bare alloy. Influence of diffusion coatings on the creep life of the superalloy is equivocal and there are a few reports about the creep behavior of coated superalloys. Detrimental effect on creep life in Pt-Al coated AM1 (a nickel-based superalloy) at the temperature range of 850-1100 °C has been reported by Ref [7].
While Haynes 75 and CMSX4 coated by Pt-Al, lean in Al, showed an increase in creep strength at lower temperatures, because of a higher volume fraction of the spread γ´ (Ni3(Al,Ti)) phase (strengthening precipitates) than types that are richer in Al [8]. The contents of Al and Pt in Pt-Al coatings have a significant effect on The Ductile/Brittle Transition Temperature (DBTT).
Depending on working temperature, below or above DBTT, negative or positive effects of the coating on mechanical properties of the base alloys are reported [9,10].
Today, Turbine Inlet Temperature (TIT) has been increased by designers for improving the performance of turbojet engines. As a result, the creep behavior of coated rotary/stationary components, such as turbine blades and vanes, has been affected by temperature variations. It should be mentioned that creep tests measure the magnitude of creep strain as a function of time, while creep rupture tests measure the time to fracture for an assumption temperature and stress levels [11]. Although there are some investigations of the mechanical behavior of the bare nickel-based superalloy Rene ® 80 [12][13][14], no research has been done to study the effect of Pt-Aluminide coating on the creep behavior of this superalloy. In light of the above, in this study, the influence of the high and low activity Pt-Al coatings on the creep rupture life of Rene ® 80 has been evaluated.

Experimental procedure
The substrate superalloy utilized in this work was a cast nickel-based superalloy (Rene ® 80) with the nominal chemical composition of Ni-0.16C 13.81Cr-9.69Co-4.23Mo-4.02W-3.02Al- Creep specimens were manufactured by machining cylindrical bars of 10 mm in diameter and 60 mm in length obtained from the investment casting according to Figure 1. Where Rz, R, d0, L0, d1, h, Lc and Lt are required surface finish, radius, test piece width, original gauge length, width of gripped ends, length of gripped ends, parallel length and total length of the test piece, respectively (DIN50125). It should be mentioned that to make sure there were no surficial (cracks, etc.) or internal (pores, etc.) defects, all specimens were inspected by the Fluorescent Penetration Inspection (FPI) and X-ray test method. At the next task, all of the rods were solutionized at 1205 °C for 2 h and aged at 1095 °C for 4 h [15]. Both the solution and aging heat-treatments were performed in a vacuum furnace.
Then, specimens were machined from the solutionized rods. Then, some of the creep specimens were subjected to a two-procedure aging heat-treatment, i.e., 2 h at 1050°C and after that 16 h at 845 °C, to attain the favorite γ-γ´ micro-structure (fully heat treatment). The remaining specimens were used for preparation of coated samples. After grit blasting and cleaning the specimens with acetone, an intermediate layer of nickel with a thickness of 1-2µm was applied on the specimens for decreasing the negative effect of the chromium on the lack of adhesive property of platinum. Then, the specimens was electroplated with a layer of Pt. Before electroplating of Pt, the specimens were blasted with alumina grits, and then cleaned with acetone and ethanol. Electroplating of Pt was conducted in an electrolyte solution containing P salt (di-nitro di-amino platinum), calcium carbonate, sodium acetate, and distilled water at 90°C. The electroplating current was 0.2-0.4 A/dm 2 and the deposition rate was about 1 μm/h. Prior to electroplating, the pH value was adjusted about 10.5 [16]. In order to achieve a platinum layer with a thickness of 6 µm, the time of 360 min was considered for the plating process. The platinum layer was subjected to heat treatment at 1050°C for 2 hours to enhance cohesion and distribution of the platinum into the substrate, followed by cooling the specimens in a furnace at 400°C and then air-cooled [17]. The conventional pack cementation method via the LTHA (at 750 °C for 4 h followed by post aluminizing for 2 h at 1050 °C ) and the HTLA method (at 1050 °C for 2 h) used for aluminizing. Aluminizing packs in the current study were composed 2NH4Cl-12Al-86Al2O3 (wt %) and 1NH4Cl-4Al-95Al2O3 (wt %) for LTHA and HTLA methods, respectively. Subsequently, a final aging treatment was performed at 845 °C for 16 h in vacuum [15].
Microstructural characterization and fractography studies were conducted using a light microscope (LEICA MEF4A) and a Scanning Electron Microscope (SEM) Model Zeiss Supra 55 equipped with Energy Dispersive Spectroscopy (EDS) Oxford Model, both prior to performing creep tests (to ensure the quality of the coating) and after the tests (to investigate fracture surfaces). X-ray diffraction analysis was performed using an Inel Equinox 6000 with X'Pert High Score Plus v2.0, Cu Kα1 with Graphite monochromator, 2θ = 16° to 93°, to determine distributions of different phases across the coating thickness and measurement the residual stress of the coating. Micro-hardness testing was performed normal to the alignment of the Pt-Al coating and the substrate according to ASTM E384 using an automatic Akashi micro-hardness tester equipped with the Clemex software through applying 50gf. The creep rupture tests on the Pt-Al coated and uncoated specimens were undertaken uninterrupted per the ASTM E139 in the air and constant load, using an ATM CR-100KN machine equipped with an electrical furnace that could apply temperatures up to 1000°C ± 1. The creep strains in the gauge section were measured by a linear variable differential transformer attached to an 5 extensometer frame which was also attached to the creep specimen. The temperatures were measured by three thermocouples attached to the specimens. The test temperatures and nominal stresses applied were 760 °C/657 MPa, 871 °C/343 MPa and 982 °C/190 MPa. All testing started after a 30 min soak to ensure that the specimens obtained the required temperature.      coated Rene ® 80 that were tested at different conditions are also shown in Fig. 6.

Microstructural characterization
All of the Curves were divided into three stages: I, primary or transient; II, secondary or constant rate and III, tertiary [20]. During the test procedure, in stage I, when the creep stress was applied, a small loading strain, was created. After that, increasing strain, led to a reduction in the creep rate. A constant value of the creep rate occurs at stage II. The steady state region showed the upper limit of time during the creep test because the strain rate was the lowest.  [20], where T is temperature in degrees Kelvin, tr is time to rupture in hours and the quantity C is specified as an alloy constant, but is roughly 20 for the superalloys [1]. Fig. 7 also shows a reduction in stress rupture properties of the coated alloy compared to the bare specimen, where the profile for the bare Rene ® 80 is located above that for the Pt-Aluminide coated Rene ® 80.  phases of PtAl2 and β-(Ni, Pt)Al (Fig. 4) in Pt-Al coatings, the fracture stress of the substrate is higher than that of the coating. The yield and ultimate strength of Pt-Al coatings have been measured 200 and 300 MPa at a temperature range of 800-900 °C, respectively by Alam et al. [21]. These values are lower than tensile properties of Rene ® 80 (UTS=701MPa and YS= 590MPa [2]). Therefore, applied tensile stress causes some cracks on the coating. This phenomenon led to a decrease in the load bearing cross-section of coated creep specimens. The rest cross section will tolerate lower load and finally, coated specimens will be ruptured earlier than uncoated ones.
As it can be concluded from Fig. 7, the lowest rupture life belongs to sample which was aluminized by HTLA method. This could be related to coating stoichiometric constitution, elemental diffusion in the coating/substrate and residual stresses.
The β-(Ni, Pt)Al phase has an important effect on the strength properties of the coatings [22]. The reason of that is related to the stoichiometry of the β-(Ni, Pt)Al phase and the resultant defect structure of this phase affect its strength characteristics. It is known that if the percentage of Al in the β-(Ni, Pt)Al composition is high, this phase will be hyper-stoichiometric (Al-rich).
On the other hand, if the percentage of Ni in this phase is high, intermediate layer will be hypostoichiometric (Ni-rich). Phase stoichiometry plays an effective role on the type of structural defects. Vacancy imperfections are more evident in hyper-stoichiometric coatings, while the substitutional imperfections become visible in hypo-stoichiometric coatings [6,22]. Vacancy defects show higher hardness in comparison with substitution defects [6]. According to Fig. 5 (d), micro-hardness values of LTHA were greater than those of HTLA. Hence, it can be resulted that in LTHA method, β-(Ni, Pt)Al is hyper-stoichiometric and in HTLA method, this phase is hypo-stoichiometric. The strength of hyper-stoichiometric phases is more than hypostoichiometric [6,21]. Thus, coating which has been formed by LTHA method is stronger than the HTLA coating and higher stresses can be tolerated by it.
Another reason for more reduction of the rupture life in HTLA compared to LTHA, is the diffusion of elements from the coatings to the base alloy (or from the substrate to the coatings) and the formation of deleterious phases. Some of the unfavorable phases the same as TCP (intermetallic phases), are formed across the IDZ. According to Fig. 3, it is seen that the thickness of the IDZ layer is thinner in LTHA. On the other side, as observed in Fig. 5 (c), the percentage of Cr in HTLA method is higher than LTHA. Cr can contribute to Cr61Co39 (the σ phase) formation and can also increase the DBTT. The solubility of elements Cr and Co is much lower in β-(Ni, Pt)Al and γ´ than in γ, and therefore the precipitation of these elements in TCPs is predictable [19].
Residual stress was also measured in two conditions of HTLA and LTHA by XRD technique. XRD peaks were analyzed using Rietveld analysis. By consideration of the crystal properties of the present phases in the coatings, elastic strain of (6µm Pt/LTHA) and (6µm Pt/HTLA) specimens were reached as 1.5 ×10 -3 and 1.9 ×10 -3 , respectively (with the help of X'pert high-score plus software and utilize of stress-less sample of Yttrium oxide). Regarding coating elastic modulus 99 and 93 GPa for LTHA and HTLA, respectively [21] and the application of Hook's Law (σ = EƐ) [23], the value of the residual stresses were calculated as 148.2 MPa for LTHA and 176.7 MPa for HTLA, respectively. As seen, the value of residual stress in HTLA method is greater than LTHA, indicating another reason for the reduction in the rupture life in HTLA. Watanabe et al. [24] also have reported the value of residual stress equal to 140MPa for a high activity Pt-Al coating.

Fractography
The evaluation of fracture surfaces of coated and uncoated Rene ® 80 was performed by    As shown in Fig. 9 the thickness of this layer in HTLA method is higher than LTHA. respectively. According to these results and also line scan EDS report Fig. 5 (c), it is clear that the percentage of Cr in the HTLA method is higher than LTHA. Therefore, the formation of needle-like σ-(Co, Cr) is more possible in the HTLA. This brittle phase may induce negative effects on the mechanical behavior of the superalloy, such as loss of ductility and reduce of creep strength. TCPs also play an effective role in stress concentrations within an alloy because of their structure [27]. Thus, another reason for the reduction in rupture life of HTLA coated Rene ® 80samples can be attributed to the higher thickness and faster degradation in the micro-structure of the IDZ in HTLA method.
In order to determine the cracks nucleation locations and their growth direction in uncoated and coated Rene ® 80, longitudinal sections of ruptured samples were evaluated.
As the creep rupture test was done at an elevated temperature the brittle and hard oxide film on the surface of the uncoated specimens could collaborate in the nucleation and propagation of cracks. Optical micrographs of the fracture surfaces of specimens which were tested at 760 °C/657 MPa and 982 °C/190 MPa are provided in Fig 10(a) and Fig.   10 (b), respectively. As evident from these figures, at both conditions the initial cracks nucleated from brittle oxide layer normal to the applied stress axis, and propagated through cavities, intergranularly. It is clear that the density and width of cracks at 982 °C/190 MPa was higher than that at 760 °C/657 MPa. Raising the thickness and quantity of cracks, led to a decrease in the cross section of load bearing. Consequently, the effective stress experienced by the specimens during the creep test increases and hence the rupture will be occurred earlier.
The longitudinal sections of the failed coated samples, at different conditions, are illustrated in Fig. 11. As shown in Fig. 11   Although the fracture surfaces of theses specimens were similar, the cleavage area of the specimen which coated in HTLA condition and tested at 760 °C/657, was larger than the others. can be seen in the fracture surface of the specimens in both aluminizing methods (Fig. 12 (c) and (d)). A few cleavage facet planes were also noted. Although at 982 °C/190 MPa, both of HTLA and LTHA specimens showed a mixed mode fracture, the ductile area in both of them was greater than the specimens which were tested at 760 °C/657 MPa and 871°C/343 MPa.
As shown in Fig. 11, in comparison with LTHA, the number of cracks were higher in HTLA specimen that tested at 760 °C/657 MPa and 871°C/343 MPa. As mentioned previously, increasing the density of cracks, led to a reduction in the cross section of load bearing, and hence the cracks can easily propagate toward the substrate.
On the other hand, grain-boundary weakening is generally responsible for the creep rupture failure of polycrystalline alloys [11]. As shown in Fig .10, not only cracks initiated from oxide layer, but also some cavities created along the grain boundaries. For polycrystalline superalloys (e.g. Rene ® 80) the addition of few ppm carbon, by producing carbides along grain boundaries, can improve the resistance of grain-boundary sliding [11]. According to the composition measurement by EDS two types of carbides were analyzed in Rene ® 80: (Cr,W)-rich M23C6 (with a chemical composition of 27.34Cr-4.55Co-17.5Mo-8.7W-5.3Al-4.7Ti-32Ni (in at. %)) and (Ti)-rich MC (with a chemical composition of 2.9Cr-0.9Co-11.5Mo-13.5W-2Al-64Ti-4.4Ni (in at. %)). M23C6 was only detected along grain boundaries while MC was common both along grain boundaries and inside of the grains. During the coating process some alloy elements such as Ni, Cr, Ti and W, diffuse from substrate to coating outwardly and Al and Pt diffuse from coating to the base alloy inwardly. This phenomenon will continue during creep test due to high temperature and enough time. Therefore, near the surface some grain boundaries became poor from Cr and W and carbides detached from grain boundary. In this case, grain boundary not only is a suitable site for void nucleation, but a preferential path for crack propagation. Therefore, the creep rupture life in the coated specimen is lower than uncoated ones. As seen in Fig. 5(c), the percentage of Cr in the composition of the coating in HTLA method is more than LTHA. It means that the distribution of Cr23C6 along the substrate grain boundaries, in HTLA method, is lower than LTHA. Thus, it can be expected that the creep rupture life of the substrate which has been coated by HTLA method tends to be more decline.
Despite the reduction in the creep properties of coated Rene ® 80, coatings are necessary to enable the superalloy aviation/industry parts to service at elevated temperatures in a modern gas turbine engine [29,30]. If bare turbine blades are exposed to the operation condition of gas turbine engines, they will damage fast by several of the phenomena such as oxidation, hot corrosion and heat damage. Therefore, investigation of the influence of coatings on the mechanical properties of the base alloy (substrate) at high temperatures, will enable the designer to choose an optimum coating that has the highest life under certain running conditions of turbine engines.

Conclusions
The experimental data of this work showed that the resistance to creep rupture decreased after applying the Pt-Al coating in both methods of LTHA and HTLA.
The fracture surface evaluation indicated that a mixed mode (ductile and brittle) of failure in both uncoated and coated alloy under all creep rupture conditions occurred. In the coated specimens, the dimples (ductile fracture) were observed more and larger in the