| Popis: |
Thermal barrier coatings (TBCs) are critical to gas turbine engines, as they protect the components in the hot section from the extreme temperatures of operation. The current industry standard method of applying TBCs for turbine blades in jet engines is electron-beam physical vapor deposition (EB-PVD), which results in a columnar structure that is valued for its high degree of strain tolerance. An emerging deposition method is plasma-spray physical vapor deposition (PS-PVD), capable of producing a variety of customizable microstructures as well as non-line-of-sight deposition, which allows more complex geometries to be coated, or even multiple parts at once. The pseudo-columnar microstructure that can be produced with PS-PVD is a possible alternative to EB-PVD. However, before PS-PVD can be used to its full potential, its mechanical properties and behavior must be understood. This work contributes to this understanding by characterizing PS-PVD TBCs that have been thermally cycled to simulate multiple lifetimes (0, 300, and 600 thermal cycles). Residual stress in the thermally grown oxide (TGO) layer is characterized by photoluminescence piezospectroscopy as TGO residual stress is correlated with the lifetime of the coating. Residual stress in the top coat is characterized by Raman spectroscopy, because this stress drives cracking in the top coat that can lead to failure. Scanning electron microscopy (SEM) and energy dispersive X-ray spectroscopy (EDS) are performed to observe microstructural and phase evolution to provide context and possible explanations for the stress results. In addition, EB-PVD samples of the same thermal cycling history are characterized in the same way so that PS-PVD can be benchmarked against the industry standard. The compressive residual stress in the TGO in both coatings was relieved with thermal cycling due to the TGO lengthening as well as microcracking. The PS-PVD samples had slightly lower TGO stress than the EB-PVD, which is attributed to the greater extent of cracking within the TGO, whereas cracking in the EB-PVD samples was at the TGO/topcoat interface. The PS-PVD cycled samples had significant cracking within the topcoat near the TGO due to both greater porosity than EB-PVD samples and regions of unmelted particles that provide little resistance to cracking. The greater convolution of the TGO in the PS-PVD samples results in greater out-of-plane tensile stresses that cause crack initiation, as well as diverts cracks away from the difficult-to-follow interface. The TGO stress results agree with existing literature and extend the thermal cycling beyond what has previously been reported for PS-PVD coatings, revealing a trend of stress relief and stress values similar to that of EB-PVD coatings in this study and in the literature. Residual stress in the topcoat for both coating types became increasingly compressive with thermal cycling, indicating loss of strain tolerance by sintering. The trend of the YSZ stress for both coating types to become more compressive with cycling and with depth agrees with the literature, and the thermal cycling is longer than has been previously reported for PS-PVD. The two coating types had quite different microstructures and crack modes as well as different as-deposited residual stresses, but after thermal cycling had similar stresses in both the TGO and top coat. No samples experienced spallation. These results indicate that, while PS-PVD coatings have different properties and behavior from EB-PVD coatings, they had comparable levels of damage to EB-PVD coatings of the same lifetime and are a viable alternative to EB-PVD. Further tuning of the processing parameters may result in PS-PVD coatings with even more similar behavior to EB-PVD coatings. |
