Le procédé de dépôt physique en phase vapeur assisté par plasma thermique (PS-PVD) consiste à évaporer le matériau sous forme de poudre à l’aide d’un jet de plasma d’arc soufflé pour produire des dépôts de structures variées obtenus par condensation de la vapeur et/ou dépôt des nano-agrégats. Dans le procédé de PS-PVD classique, l’intégralité du traitement du matériau est réalisée dans une enceinte sous faible pression, ce qui limite les phénomènes d’évaporation ou nécessite d’utiliser des torches de puissance importante. Dans ce travail, une extension du procédé de PS-PVD conventionnel à un procédé à deux enceintes est proposée puis explorée par voie de modélisation et de simulation numérique : la poudre est évaporée dans une enceinte haute pression (105 Pa) reliée par une tuyère de détente à une enceinte de dépôt basse pression (100 ou 1 000 Pa), permettant une évaporation énergétiquement plus efficace de poudre de Zircone Yttriée de granulométrie élevée, tout en utilisant des torches de puissance raisonnable. L’érosion et le colmatage de la tuyère de détente peuvent limiter la faisabilité d’un tel système. Aussi, par la mise en oeuvre de modèles numériques de mécaniquedes fluides et basé sur la théorie cinétique de la nucléation et de la croissance d’agrégats, on montre que, par l’ajustement des dimensions du système et des paramètres opératoires ces deux problèmes peuvent être contournés ou minimisés. En particulier, l’angle de divergence de la tuyère de détente est optimisé pour diminuer le risque de colmatage et obtenir le jet et le dépôt les plus uniformes possibles à l'aide des modèles susmentionnés, associés à un modèle DSMC (Monte-Carlo) du flux de gaz plasmagène raréfié. Pour une pression de 100 Pa, les résultats montrent que la barrière thermique serait formée par condensation de vapeur alors que pour 1 000 Pa, elle serait majoritairement formée par dépôt de nano-agrégats. / Plasma Spray Physical Vapor Deposition (PS-PVD) aims to substantially evaporate material in powder form by means of a DC plasma jet to produce coatings with various microstructures built by vapor condensation and/or by deposition of nanoclusters. In the conventional PS-PVD process, all the material treatment takes place in a medium vacuum atmosphere, limiting the evaporation process or requiring very high-power torches. In the present work, an extension of conventional PS-PVD process as a two-chamber process is proposed and investigated by means of numerical modeling: the powder is vaporized in a high pressure chamber (105 Pa) connected to the low pressure (100 or 1,000 Pa) deposition chamber by an expansion nozzle, allowing more energetically efficient evaporation of coarse YSZ powders using relatively low power plasma torches. Expansion nozzle erosion and clogging can obstruct the feasibility of such a system. In the present work, through the use of computational fluid dynamics, kinetic nucleation theory and cluster growth equations it is shown through careful adjustment of system dimensions and operating parameters both problems can be avoided or minimized. Divergence angle of the expansion nozzle is optimized to decrease the clogging risk and to reach the most uniform coating and spray characteristics using the aforementioned approaches linked with a DSMC model of the rarefied plasma gas flow. Results show that for 100 Pa, the thermal barrier coating would be mainly built from vapor deposition unlike 1,000 Pa for which it is mainly built by cluster deposition.

Thermal barrier coatings (TBCs) are advanced material systems used in the hot sections of gas turbines. The TBCs are designed to provide insulation against hot gases by a ceramic top coat and to provide oxidation and corrosion resistance by a metallic bond coat. As the operating environment is harsh and complex, the TBC often requires stricter material properties. Failure of TBCs can limit the longevity of the turbine severely. In this study, failure caused by erosion has been the main focus. The erosion behaviour of TBCs processed by atmospheric plasma spay (APS), electron beam physical vapour deposition (EB-PVD), and plasma spray physical vapour deposition (PS-PVD) has been studied by an experimental investigation and a literature study. The erosion performance of different TBCs was studied by conducting erosion tests under 90° and 15° alumina particle impact (50 μm) and measuring the weight loss and thickness loss of the ceramic top coat. Variables affecting the erosion behaviour were studied by means of scanning electron microscopy (SEM), investigating the microstructure, the erosion damage, porosity content, and column density. Hardness tests were also conducted to investigate a potential correlation between hardness and erosion performance. It was evident that the 8YSZ top coat processed by EB-PVD had higher erosion resistance than APS, which in turn had higher erosion resistance than PS-PVD. Their microstructures are significantly different, resulting in different erosion failure mechanisms. APS TBCs have a splat-on-splat lamellar microstructure, and the failure mechanismsare ploughing of furrows, splat boundary failure, and tunneling via pores. In contrast, EB-PVD TBCs have columnar microstructure and fail by near-surface cracking. The investigated PS-PVD TBC had a feathery columnar microstructure, containing many large grain boundaries and flaws, making grain boundary failure the governing mechanism. The APS and EB-PVD TBCs impacted at a 90° angle had significantly higher erosion rates than those eroded at 15°, which also was reported in literature. However, the opposite was observed for the PS-PVD TBCs. The level of porosity and hardness of the TBC top coat was found to affect the erosion rate, even though no evident correlations could be observed in this study. No factor alone was found to dictate the erosion behaviour of the investigated TBCs. Based on the literature study and findings in the experimental study, a TBC with good erosion performance has, in general, low porosity, few defects, high hardness and high fracture toughness. Specifically for APS TBCs, good splat bonding is favourable and for EB-PVD and PS-PVD it recommended to have high column density, columns orthogonal to the substrate, and low gap width between the columns.