Optimierung von APS-ZrO 2-Wärmedämmschichten durch Variation der Kriechfestigkeit und der Grenzflächenrauhigkeit

Gas turbines operate at combustion chamber temperatures up to 1400°C. Therefore the blades and the combustion chamber lining, which consist of Ni-superalloys for highest loads, are coated with a thermal barrier coating (TBC) of zirconium oxide and an underlying oxidation protection coating of MCrAlY-alloys (M=Ni, Co). At high temperature the MCrAlY-coating oxidizes and an Al2O3-scale (thermally grown oxide, TGO) forms between MCrAlY-coating and TBC, what constrains the oxidation of the base material. At plasma sprayed TBCs, the MCrAlY-coating provides a bond coat (BC) for the TBC at the same time and therefore is roughened by sandblasting before the deposition of the TBC. By the growth of the Al2O3-scale and the start up and run down of the gas turbine, stresses arise in the TBC, which lead to lateral crack formation in the field of the TBC-BC-interface and finally to the spallation of the TBC. Thereby other parts of the turbine can be damaged, what causes high costs. Therefrom the aim is to delay the crack growth as strong as possible or rather to maximize the lifetime of the TBC. For this purpose the material properties of the coating components have to be optimized. In the present work, the influence of creep strength of BC and TGO and the influence of TBC-BC-interface-roughness on the lifetime and damage evolution of plasma sprayed ZrO2-TBCs are investigated. To determine the lifetime, cylindrical specimens with plasma sprayed ZrO2-TBC were produced and thermally cycled with a minimum and maximum temperature of 60°C and 1050°C and a dwell time at maximal temperature of 2h. To exclude the interdiffusion and thermal mismatch between BC and Ni-superalloy, a model system was used: The Ni-superalloy was left and the substrate material consists completely of a BC-like FeCrAlY-alloy. The model system was simulated by the project partner TU Braunschweig with the FE-method. The TBC-BC-interface-roughness was abstracted by a 2-dimensional sine-shaped periodic roughness profile. To be able to compare the simulation results with real TBC-systems, a 2-dimensional periodic roughness profile was produced by high speed drawing. Additionally a 3-dimensional stochastic roughness profile was produced by sandblasting. To vary the creep strength of the substrate a conventional (creep weak) and an oxide-dispersion-strengthened (creep strong) FeCrAlY-alloy was used. The creep strength of the Al2O3-scale was varied by depositing a fine crystalline (creep weak) Al2O3-scale on the FeCrAlY-substrates by PVD and a coarse crystalline (creep strong) Al2O3-scale by pre-oxidation of the FeCrAlY-substrates at 1050°C in air. The roughness depth was varied by drawing with different feed rates in the case of the periodic roughness profile and sandblasting with different grain sizes in the case of the stochastic profile. To investigate the damage evolution, infra-red-impulse-thermography-pictures of the specimens were made in regular cycle intervals. The thermography method was optimized in such a way, that delaminations of the TBC are detectable with a detection limit of 1-2 µm lift-off and 0.7-0.8 mm wide. The results show, that the damage evolution and lifetime are significantly influenced by the creep strength of the FeCrAlY-substrate, the TBC-BC-interface roughness depth and the profile type. Based on the results of the investigation of the damage evolution by thermography, a model was created, which gives a good description of the increase of total delamination area on a specimen in principle. The comparison of the FEM-simulation of the present TBC-model-system with the failure mode of the real TBC-specimens with periodic roughness profile showed agreements but also disagreements