Thermal Barrier Coatings (TBC) are used to shield hot sections of gas turbine engines, helping to prevent the melting of metallic surfaces. TBC is a sophisticated layered system that can be divided into top coat, bond coat, and the super-alloy substrate. The highly heterogeneous microstructure of the TBC consists of defects such as pores, voids, and cracks of different sizes, which determine the coating’s final thermal and mechanical properties. The service lives of the coatings are dependent on these parameters. These coatings act as a defensive shield to protect the substrate from oxidation and corrosion caused by elevated temperatures. Yttria Stabilized Zirconia (YSZ) is the preferred thermal barrier coating for gas turbine engine applications. There are a certain number of deposition techniques that are used to deposit the thermal coating layer on the substrate; commonly used techniques are Air Plasma Sprayed (APS) or Electron Beam Physical Vapour Deposition (EB-PVD). The objective of this thesis is to model an optimized TBC that can be used on next-generation turbine engines. Modelling is performed to calculate the effective thermal conductivity of the YSZ coating deposited by EB-PVD by considering the effect of defects, porosities, and cracks. Bruggeman’s asymmetrical model was chosen as it can be extended for various types of porosities present in the material. The model is used as an iterative approach of a two-phase model and is extended up to a five-phase model. The results offer important information about the influence of randomly oriented defects on the overall thermal conductivity. The modelled microstructure can be fabricated with similar composition to have an enhanced thermal insulation. The modelling results are subsequently compared with existing theories published in previous works and experiments. The modelling approach developed in this work could be used as a tool to design the porous microstructure of a coating.

The aim of this research project was to investigate the microstructural changes of MCrAlY coated Ni-based superalloys that are routinely used in industrial gas turbine engines for power generation. One of the main aims of the characterisation was to understand ageing time and temperature effects on the microstructural evolution so that a methodology could be developed where the characterisation of a thermally exposed microstructure can be used to estimate unknown service exposure conditions.

The aim of this project was to investigate the potential of fused-slurry coatings for improving the oxidation resistance of wrought alloys. Slurry-aluminised coatings were deposited on Alloy 800H (Fe-33Ni-20Cr), Alloy HCM12A (Fe-12Cr-2W), Alloy 214 (Ni-16Cr-4Al-3Fe), Fe-27Cr-4Al and Fe-14Cr-4Al alloys. The slurry contained a cellulose-based binder in an aqueous carrier and spherical aluminium powder, with a particle size below 20 microns. The slurries were applied with a paint-brush, dried in air and heat treated in either hydrogen or argon at temperatures between 700 and 1150C. The slurries were characterised by thermogravimetry, differential scanning calorimetry and viscometry. The coatings were characterised by optical microscopy, scanning electron microscopy, energy-dispersive X-ray analysis, X-ray diffraction and Vickers hardness measurements. The oxidation resistance of selected slurry-coated specimens was assessed in air at 1000 and 1100C in tests lasting up to 1000 hours.Slurry-aluminising was found to be a simple, effective way of forming protective coatings that were similar in composition and microstructure to chemical vapour deposits. However, it was difficult to control the amount of slurry applied to the substrate and produce coatings of uniform thickness.The coatings on Alloy HCM12A and the Fe-Cr-Al alloys contained cracks in the brittle FeAl phase due to tensile stresses arising from differences in the thermal expansion coefficients of the substrates and the coatings. Rapid interdiffusion between the coatings and the ferritic substrates resulted in the appearance of Kirkendall voids.Coatings on Alloy 214 required a two-stage heat treatment to convert the brittle δ-Ni2Al3 to β-NiAl. Cracking along the coating/substrate interface was prevented by limiting the coating thickness to a maximum of 250 microns. During oxidation at 1100C, the β-NiAl was converted to γ'-Ni3Al. After 1000 h, the centre of the coating consisted chiefly of γ'-Ni3Al and bands of austenite (γ-Ni) were present at the inner and outer edges of the coating. The aluminium content at the coated surface was higher than the original aluminium content of the alloy, the protective alumina scale was improved and the oxidation life of the substrate was extended. An additional life of 1250 h at 1100C is estimated from a slurry coating before the aluminium content returns to that of the original alloy (4%), providing a potential improvement in oxidation resistance.Microstructural changes such as grain growth, sensitisation and formation of aluminium nitride particles near the coating/substrate interface, were detected in the alloy substrates after forming the slurry coatings. However, these microstructural changes did not detract from the good performance of the coatings during oxidation tests at 1100C.The work in this study has demonstrated a low-cost method of coating high-temperature alloys providing coatings with microstructures, densities and modes of degradation similar to those obtained by other coating methods. The coatings are potentially applicable to a wide range of high-temperature substrates.

La résistance à l'écaillage d'un système barrière thermique est fonction de la composition et microstructure des matériaux constituant le système, ainsi que des procédés utilisés pour son élaboration. Cette thèse s'intéresse à l'influence d'une couche de platine déposée à la surface du dépôt NiCoCrAlYTa (sous-couche) sur le comportement en oxydation du système barrière thermique. Une étude approfondie est d'abord menée afin d'identifier les atouts et points faibles en oxydation cyclique d'un système comprenant un revêtement NiCoCrAlYTa. La formation d'une couche d'oxyde composée non exclusivement d'alumine et l'importante rugosité de la sous-couche, favorisant les défauts au sein de la barrière thermique, accélèrent l'écaillage de la barrière thermique. Parallèlement, la présence de carbures de tantale au sein du dépôt ne suffit pas à stopper le titane qui diffuse depuis le superalliage jusqu'à la couche d'oxyde et dégrade le système. Le platine ayant déjà démontré son effet très bénéfique sur les dépôts aluminures de nickel, il apparaît comme prometteur pour améliorer le comportement en oxydation du revêtement NiCoCrAlYTa. L'étude de son influence débute par une analyse fine de deux sous-couches NiCoCrAlYTa modifié platine : la première comprend un revêtement NiCoCrAlYTa obtenu par co-dépôt électrolytique, la seconde un dépôt NiCoCrAlYTa élaboré par projection plasma sous vide. Cette caractérisation, par diffraction des rayons X et microscopie électronique à balayage et en transmission, met en évidence la présence de martensite en surface du revêtement, conséquence de la diminution de l'activité de l'aluminium par le platine. Elle révèle également la forte influence du procédé utilisé pour l'élaboration du dépôt NiCoCrAlYTa sur la microstructure obtenue après le traitement thermique de diffusion. Des essais d'oxydation isotherme et de préoxydation sont ensuite réalisés sur la sous-couche dont le revêtement NiCoCrAlYTa est élaboré par co-dépôt électrolytique. Les couches d'oxydes formées sont analysées par diffraction des rayons X, spectroscopie Raman et fluorescence. Grâce à l'ajout de platine, qui entraîne l'augmentation de la teneur en aluminium dans la zone externe du revêtement, l'oxydation sélective de l'aluminium est favorisée. Cela se traduit par une diminution de la cinétique d'oxydation et une augmentation de la résistance à l'écaillage de la couche d'oxyde. Cependant, les carbures de tantale se décomposent lors du traitement thermique de diffusion puis lors de l'oxydation, laissant le titane libre de diffuser depuis le superalliage jusqu'à l'oxyde. De l'oxyde de titane est en effet détecté par spectroscopie Raman en petite quantité dans de la couche d'oxyde (avec l'AM3 comme substrat). Un autre point important sur la composition du superalliage est la présence d'élément réactif qui permet de diminuer la croissance de la couche d'oxyde. Concernant les essais de préoxydation, les résultats obtenus indiquent la nécessité d'une faible pression partielle d'oxygène afin de promouvoir la formation d'alumine-a. Le platine, quant à lui, ne favorise pas la formation d'alumine de transition. Des essais d'oxydation cyclique sur des systèmes barrière thermique sont ensuite menés. L'effet bénéfique du platine sur l'oxydation sélective de l'aluminium est confirmé, ce qui entraîne une augmentation de la durée de vie en cyclage. Cependant, la décomposition des carbures de tantale est de nouveau observée. Une diffusion très importante de titane depuis le superalliage jusqu'à l'oxyde est ainsi notée pour les systèmes barrière thermique comprenant une sous-couche modifiée platine avec un dépôt NiCoCrAlYTa obtenu par projection plasma sous vide. Dans le cas de système avec une sous-couche modifiée platine comprenant un dépôt NiCoCrAlYTa élaboré par co-dépôt électrolytique, le problème majeur est la présence de pores en surface et d'une certaine porosité à l'intérieur du revêtement. L'oxydation des pores en surface ainsi que le cyclage thermique provoque la pénétration de l'oxyde puis sa propagation catastrophique dans le revêtement. Les résultats obtenus permettent de dégager les points importants de l'élaboration d'un système barrière thermique. Il est alors recommandé que le superalliage contienne un élément réactif mais peu de titane. Le dépôt NiCoCrAlYTa nécessaire à la fabrication de la sous-couche doit être dense et la préparation de surface, avant et après le dépôt de platine, doit permettre d'obtenir une faible rugosité de surface avant le dépôt de la barrière thermique. Enfin, les paramètres (température, pression partielle d'oxygène, sablage) lors de la première oxydation du système doivent être contrôlés de manière à favoriser la formation d'alumine-a. / The resistance to spallation of a thermal barrier coating system depends on the composition and the microstructure of the materials constituting the system, as well as on the processes used for its manufacturing. This PhD is interested in the influence of a Pt layer deposited on the surface of the NiCoCrAlYTa coating (bond coating) on the oxidation behavior of the thermal barrier coating system. A thorough study is first carried out in order to define the assets and the weak points under cyclic oxidizing conditions of a system composed of a NiCoCrAlYTa coating. The formation of an oxide layer not only composed of alumina and the great roughness of the bond coating, favoring defects within the thermal barrier, speed up the thermal barrier spallation. At the same time, the presence of tantalum carbides within the coating is not sufficient to prevent titanium from diffusing from the bond coating toward the oxide layer and from degrading the system. Platinum having already demonstrated its beneficial effect on nickel aluminide coatings, it seems promising in order to improve the oxidation resistance of the NiCoCrAlYTa coating. The study of its influence starts by a thorough analyses of two Pt-modified NiCoCrAlYTa bond coatings: the first one is composed of a NiCoCrAlYTa coating made by composite electroplating, the second one is composed of a NiCoCrAlYTa coating manufactured by vacuum plasma spray. This characterization, done using X-ray diffraction and secondary and transmission electron microscopy, highlights the presence of martensite at the coating surface, consequence of the decrease in the aluminium activity by platinum. It also reveals the strong influence of the process used to manufacture the NiCoCrAlYTa coating on the microstructure obtained after diffusion heat treatment. Preoxidation and isothermal oxidation tests are then carried out on the systems for which the NiCoCrAlYTa coating is made by composite electroplating. The oxide layers that formed are analyzed by X-ray diffraction, Raman spectroscopy and fluorescence. With Pt addition, that leads to an increase in the aluminium concentration in the external part of the coating, the selective oxidation of aluminium is favored. This results in a decrease in the oxidation kinetics and an increase in the resistance to spallation of the oxide layer. However, tantalum carbides decompose during the diffusion heat treatment and then during the oxidation, making the titanium free to diffuse from the superalloy toward the oxide. Indeed, titanium oxide is identified in small quantity in the oxide layer by Raman spectroscopy (with AM3 as substrate). Another relevant point on the superalloy composition is the presence of reactive elements that leads to a decrease in the oxide layer growth. Concerning the preoxidation tests, the obtained results indicate the necessity of a low oxygen partial pressure so as to promote the a-alumina formation. As for platinum, it does not favor the formation of transient alumina. Cyclic oxidation tests on thermal barrier coating systems are then carried out. The beneficial effect of platinum on the selective oxidation of aluminum is confirmed, that leads to longer lifetimes under thermal cycling. However, the tantalum carbides decomposition is observed once again. A great titanium diffusion from the superalloy toward the oxide is noticed for the thermal barrier coating systems composed of a platinum modified bond coating with a NiCoCrAlYTa deposit made by vacuum plasma spraying. In the case of systems composed of a Pt modified bond coating with a NiCoCrAlYTa deposit manufactured by composite electroplating, the main issue is the presence of pores at the surface and of a porosity within the coating. The pores oxidation at the surface as well as the thermal cycling result in the oxide penetration and then its disastrous propagation within the coating. The obtained results reveal the relevant points concerning the manufacturing of thermal barrier coating systems. It is recommended to use a reactive element containing superalloy that has very little titanium. The NiCoCrAlYTa coating required for the bond coating manufacturing has to be dense and the surface preparation, before and after the Pt deposit, has to lead to a surface with a low roughness before the deposition of the thermal barrier coating. Finally, the parameters during the first oxidation of the system (temperature, oxygen partial pressure, grit blasting), has to be done in order to favor a-alumina formation.