Pseudo-binary and Pseudo-ternary diffusion couple methods for understanding the growth of interdiffusion zone between bond coat and superalloys

Abstract

Ni-based superalloys are widely used in jet engine applications. Bond coats are applied to these alloys for protection during service in a very harsh environment at high operating temperature. The addition of Pt to the -NiAl bond coat increases the cyclic oxidation resistance and the service life by reducing sulphur segregation between the bond coat and thermally grown oxide (TGO). However, it also increases the rate of interdiffusion of Ni and Al, leading to an increase in the thickness of the interdiffusion zone (IDZ) developed between the bond coat and superalloy. This results in an unwanted loss of Al from the bond coat, leading to a decrease in the service life of the turbine blade. Hence, it is important to understand the diffusion-controlled growth and microstructural evolution of IDZ at elevated temperature based on quantitative diffusion analysis. The composition-dependent interdiffusion coefficients (not the average over a composition range) could not be estimated experimentally following the conventional diffusion couple technique in multicomponent systems. The equations for the estimation of the interdiffusion coefficients in multicomponent systems are established based on the Onsager formalism. However, the number of independent equations required for the estimation of these data could not be achieved experimentally in a system with more than three components. Therefore, all the quantitative diffusion analyses are conducted in binary and ternary systems. In a binary system, both interdiffusion and intrinsic diffusion coefficients can be estimated. However, we can estimate only the interdiffusion coefficients in a ternary system. The intrinsic diffusion coefficients cannot be estimated following the Kirkendall marker experiments because of stringent requirements, which cannot be fulfiled experimentally. Since most of the inhomogeneous bulk materials in applications are multicomponent in nature, the lack of an experimental method for estimation of these diffusion coefficients in systems with more than three components makes it impossible to understand many physicomechanical properties based on quantitative analysis. To circumvent this problem, the pseudo-binary diffusion couple in the multicomponent system is explored. In a pseudo-binary diffusion couple, only two components develop the diffusion profiles while maintaining the compositions of other components as constant making it possible to estimate the diffusion coefficients in a multicomponent system by removing various complications. However, this method is still at the nascent stage and we cannot estimate the cross interdiffusion coefficients, which are essential for understanding the influence of thermodynamic parameters on diffusional interactions between the components. Therefore, we have developed the concept of the pseudo-ternary diffusion couple method in which only three components develop the diffusion profiles keeping other components as the constant facilitating the estimation of both main and cross interdiffusion coefficients. The benefits of using both pseudo-binary and pseudo-ternary diffusion couple methods in inhomogeneous multicomponent systems are demonstrated in this study by estimating different types of composition-dependent diffusion coefficients. The steps and analyses for utilizing the pseudo-binary and pseudo-ternary methods are first described in the Ni-Co-Fe-Mo system by producing ideal diffusion profiles fulfiling the concepts behind these methods. Subsequently, the discussion is extended to the Ni-Co-Cr system for the estimation of the tracer diffusion coefficients. Steps are also suggested to avoid the complications of developing non-ideal pseudo-binary diffusion profiles. These methods are easy to adopt for research engineers. Most importantly, these allow validating the data calculated following newly proposed numerical methods by different groups, which otherwise would not be possible earlier. The tracer diffusion coefficients estimated following the radiotracer method could compare the data estimated indirectly following the newly proposed method. However, these are available now only in very few multicomponent systems. Therefore, this study is extended to the Ni-Co-Fe-Cr system for the sake of comparison of the data estimated following the pseudo-binary method with tracer diffusion coefficients measured by the radiotracer method. Along with the interdiffusion coefficients, the intrinsic diffusion coefficients are also estimated for all the components by designing the pseudo-binary couples such that Ni and Co develop the diffusion profiles keeping Fe and Cr constant in one couple, and Fe and Cr develop the diffusion profiles keeping Ni and Co constant in another couple. We have proposed the relations for calculating the tracer diffusion coefficients considering the vacancy wind effect by utilizing the thermodynamic details. A very good match was found with the data estimated directly following the radiotracer method at the equiatomic composition. Since the PB method cannot explain the diffusional interactions between the components, which plays an important role in the evolution of diffusion profiles, the concept of the pseudo-ternary diffusion couple is demonstrated in the Ni-Co-Fe-Mo system. The pseudo-ternary diffusion profiles can be intersected since the diffusion paths are restricted in two dimensions compared to the situation in which all the components develop the diffusion profiles and difficult to intersect even two diffusion paths in a multicomponent space. The modifications of the equations for estimating the diffusion coefficients in this type of diffusion couple is proposed for estimating the main and cross interdiffusion coefficients, which was otherwise impossible during the last many decades. As a next step, we have utilized these methods to estimate the composition-dependent diffusion coefficients in the B2-Ni(CoPt)Al system to understand the diffusion process based on quantitative analysis, which was not possible earlier. We found that the PB-interdiffusion coefficients of Ni and Al decrease in Co's presence but an increase in the presence of Pt. The thermodynamic driving forces show an opposite trend with respect to the composition as compared to the changes in the interdiffusion coefficients. The PT-methods indicate that the main interdiffusion coefficients increase significantly in the presence of Pt. On the other hand, the cross interdiffusion coefficients change marginally, causing a minor change in these components' diffusional interactions. These indicate a dominating role of the Pt(Co)-induced modifications of point defect concentrations. The estimated diffusion coefficients are utilized to explain the increase in the thickness of the IDZ between the B2-Ni(Pt)Al bond coat and the single crystal superalloy René N5. Further, to understand the nature of the precipitates formed in IDZ, The electron probe micro analyser (EPMA) and transmission electron microscopy (TEM) analyses are carried out. It reveals that the precipitates are rich in refractory elements. The partial or complete replacement of Pt with Pd in B2-NiAl is examined by studying the interdiffusion zone's growth rate and the spallation of the oxide layer during thermal cycling. The quantitative diffusion analysis following the PB method indicates that the PB interdiffusion coefficient of Ni and Al decreases in the presence of Pd. This reflects the decreased growth rate of the interdiffusion zone and, therefore, can be considered beneficial compared to the Pt modified bond coat. On the other hand, the spallation of the oxide layer is high in the presence of Pd. However, a partial replacement of Pt with Pd leads to superior spallation resistance against thermal cycling along with the absence of deleterious PtAl2 phase during pack cementation. The combination of Pt and Pd reduces the Al loss caused by the growth of IDZ and forms a stable oxide layer. So, this combination of Pt+Pd modified bond coat is beneficial for gas turbine applications.