Advanced Thermal Barrier Coating Materials and Design Architectures for Improved Durability

Open Access
Schmitt, Michael Powell
Graduate Program:
Materials Science and Engineering
Doctor of Philosophy
Document Type:
Date of Defense:
July 14, 2016
Committee Members:
  • Douglas Edward Wolfe, Dissertation Advisor
  • Douglas Edward Wolfe, Committee Chair
  • Gary Lynn Messing, Committee Member
  • James Hansell Adair, Committee Member
  • Md Amanul Haque, Outside Member
  • John Richard Hellmann Jr., Committee Member
  • Thermal Barrier Coating
  • YSZ
  • gadolinium zirconate
  • design architecture
  • composite
  • multilayer
  • PS-PVD
  • Gd2Zr2O7
  • rare earth
Thermal barrier coatings (TBCs) are implemented in gas turbine engines to protect the engine components from the harsh combustion environment. The current state-of-the-art TBC is composed of 7 wt.% yttria stabilized zirconia (YSZ) which begins to degrade when employed at temperatures above ~1200 °C for prolonged periods. This dissertation examines advanced materials and coating design architectures which can enable operation above 1200 °C, thereby garnering enhanced engine efficiencies. The primary material of interest for this application is the rare earth pyrochlore gadolinium zirconate (Gd2Zr2O7 – GZO), which is phase stable to ~1500 °C and exhibits very low thermal conductivity (~1.1 W/m-K in coating form compared to YSZ at ~1.5 W/m-K). The drawback of this material is a fundamentally poor durability in terms of thermal cyclic life and particle erosion resistance due to a low fracture toughness. Therefore, novel coating design architectures were created to enhance the durability of TBCs containing GZO or similar analytes via Electron Beam – Physical Vapor Deposition (EB-PVD), Air Plasma Spray (APS), and a novel hybrid technique, Plasma Spray – Physical Vapor Deposition (PS-PVD). The coatings were characterized in terms of microstructure and crystalline phase via scanning electron micrscopy (SEM), x-ray diffraction (XRD), optical profilometry, transmission electron microscopy (TEM), energy dispersive spectroscopy (EDS), and selected area electron diffraction (SAED). Image analysis was performed on the microscopy data while whole pattern fitting was employed for the XRD data to examine phase fractions, lattice parameters, crystallite sizes, and mirostrain. Sin2(ψ) residual stress scans were also performed on select samples to examine the stress state after deposition and heat treatments. Finally, the coatings were tested in terms of thermal conductivity via high heat flux laser technique or laser flash technique, erosion via a custom built erosion rig, and thermal stability via heat treatments and post analysis. The first approach involved utilizing electron beam – physical vapor deposition (EB-PVD) to create a layered design architecture composed of alternating layers of the advanced GZO material and a toughening phase, t’ Low-k (ZrO2 – 2Y2O3 – 1Gd2O3 – 1Yb2O3 mol%), which enhances the durability by resisting fracture. Nanolayering the two materials with individual layer thicknesses on the order of 200 nm resulted in 72% of the maximum possible reduction in the erosion rate. At the same time, the best multilayer coating thermal conductivity was reduced by up to 24% compared to the state-of-the-art YSZ coating in the as deposited state. Thermal conductivity stability also improved, as the nanolayers exhibited a small 7% increase in thermal conductivity over the 20 hour duration of the 1316 °C test, whereas the state-of-the-art YSZ coating increased by 52%. Overall, the nanolayer concept represents a factor of 3 reduction in the erosion rate compared to pure GZO, while maintaining a factor of 2 reduction in thermal conductivity over YSZ. This would potentially enable a thermal gradient of twice that of YSZ (500 °C), or operation at 1500 °C while still maintaining the same temperature of the substrate. The second design architecture approach involved utilizing air plasma spray (APS) to create a composite design architecture composed of primary cubic phase zirconia materials which exhibit higher temperature phase stability, low thermal conductivity, but poor toughness, and a toughening secondary phase which exhibits high toughness. The primary phases were either cubic Low-k (ZrO2 – 6Y2O3 – 2Gd2O3 – 2Yb2O3 mol%) or NZO (ZrO2 – 25Nd2O3 – 5Y2O3 – 5Yb2O3 mol%), while the toughening phase was again t’ Low-k. The best performing cubic Low-k composites reduced the erosion rate by 67% of the maximum possible reduction. Meanwhile, thermal conductivity of the best cubic Low-k composite was reduced by 24% compared to that of pure YSZ, and very close to the value for pure cubic Low-k, indicating the second phase has minimal impact on the thermal conductivity. The reduction in erosion rate did not linearly correlate to the weight fraction of the toughening phase, therefore, mixing rules were applied to describe the degree of parallel and series behavior. It was found that a Halpin-Tsai mixing rule could enable quantification of the composite behavior, and therefore the efficiency of the toughening of the secondary phases via the ξ scaling factor, where ξ=0 indicates pure series behavior, and ξ=inifinity (or a large number such as 100) indicates pure parallel behavior. A ξ value of 0.5 was found for the cubic Low-k composites, indicating some parallel (ideal) behavior was occurring, but that there was significant room for improvement. The NZO – t’ Low-k composites were deposited with a tailored, horizontally aligned, composite microstructure as it was hypothesized that this microstructure would enhance the toughening response of the t’ Low-k phase. For the NZO composites, the erosion rate was reduced by 79% with respect to the pure NZO compared to the pure t’ Low-k. This correlated to a factor of 3 reduction in the erosion rate and ξ values of 1 – 3 confirmed the hypothesis that a tailored design architecture such as an aligned microstructure could improve the toughening response. The thermal conductivity of the best NZO composite coating was reduced by 55% compared to pure YSZ, again indicating minimal impact of the secondary phase. Plasma spray – physical vapor deposition (PS-PVD) was utilized to deposit coatings in a pseudo-columnar design architecture, with the hypothesis that the hybrid PS-PVD process could combine beneficial aspects of both APS and EB-PVD. YSZ coatings were first deposited to determine an ideal processing window for in-depth studies of process-structure-property relations for the t’ Low-k material. It was found that YSZ coatings deposited by this process yield the t’ structure and often times incorporate metastable reduced zirconia phases due to a high oxygen deficiency. Thermal conductivity of the PS-PVD YSZ coatings after ten hours of testing ranged from a maximum of 2.10 W/m-K to a minimum of 1.05 W/m-K. For comparison, EB-PVD and APS YSZ coatings yielded values of 2.10 W/m-K and 1.47 W/m-K after ten hours of testing at 1316 °C, confirming the PS-PVD process can tailor coatings from EB-PVD-like to APS-like behavior. A strong trend was observed between increasing deposition rates, and decreasing thermal conductivities. Erosion trends suggested increasing erosion rates with increasing gas flow and amperage, while erosion approached a minima at 6.4 g/min feed rates. Study of the YSZ material indicated the ideal processing range would utilize amperages of 1600 A – 1800 A (with 1800 A being the upper limit of the system), gas flows of 80 NLPM – 120 NLPM, and feed rates from 2.6 g/min to 10.6 g/min, which would yield conditions with high deposition rates for low thermal conductivity, as well as conditions for low erosion rates. For the PS-PVD t’ Low-k coatings, the erosion rate varied nearly an order of magnitude, with the highest erosion rate at 0.500 g/kg, less than the APS t’ Low-k studied previously (0.948 g/kg), and the lowest erosion rate at 0.068 g/kg, a significant reduction over the EB-PVD t’ Low-k studied previously (0.118 g/kg). Erosion trends similar to the YSZ were observed for the t’ Low-k coatings, where increasing gas flow yielded high erosion rates while a minima was approached at intermediate feed rates. The erosion mechanism was investigated via post-erosion cross sectional SEM and appears to be similar to that of EB-PVD, however, it is hypothesized that the large columns of the PS-PVD process require more energy to fracture and generally possess larger intercolumnar gaps which prevent crack propagation from one column into neighboring columns. This enables potentially lower erosion rates for the PS-PVD pseudo-columnar microstructures. It was found that if the intercolumnar spacing becomes too large, the porous region around the column erodes quickly, exposing column tips which rapidly erode and yield high erosion rates similar to APS. Finally, this work compared coatings composed of YSZ, t’ Low-k, and the advanced GZO material, all deposited via EB-PVD, APS, and PS-PVD and tested in terms of the erosion durability and thermal conductivity, while quantitatively assessing the crystallography and morphological features. This provided the first direct comparison of the performance of coatings deposited by the various techniques for each of the material systems. In terms of processing, the EB-PVD processing conditions were modified to increase the coating porosity in an effort to improve future thermal cycling results. The PS-PVD conditions were chosen from the t’ Low-k experiments to yield low erosion rates similar to EB-PVD and thermal conductivities intermediate to EB-PVD and APS. The APS coatings were deposited with parameters such that the porosity was sufficiently high to improve the thermal cycling characteristics in future studies. The erosion rates for the YSZ coatings therefore trended from APS (1.026 g/kg) > PS-PVD (0.277 g/kg) > EB-PVD (0.064 g/kg). For the YSZ coatings, the PS-PVD depositions utilized a novel hollow sphere (HOSP) particle feedstock morphology. It was found that this morphology is difficult to fully evaporate since the particles do not break apart when entering the plasma, which then yields very low depositions rates, and subsequently, large intercolumnar gaps and a relatively high erosion rate. The EB-PVD YSZ erosion rate was significantly improved over previous EB-PVD YSZ (0.118 g/kg). It was determined that the higher porosity yields larger intercolumnar gaps which prevent transcolumnar cracking from one column to neighboring columns. This was observed for each of the EB-PVD coatings in the final study. The thermal conductivity was highest for the EB-PVD t’ Low-k (1.56 W/m-K), lowest for the APS t’ Low-k (0.62 W/m-K), and intermediate for the highly porous PS-PVD t’ Low-k (0.83 W/m-K). The erosion rates for the t’ Low-k coatings trended from PS-PVD (8.076 g/kg) > APS (0.789 g/kg) > EB-PVD (0.070 g/kg). The t’ Low-k PS-PVD coating utilized a more ideal spray dry particle feedstock morphology, however, the PSD was sufficiently high that significant portions were not evaporated, and very large intercolumnar gaps were formed. This yielded extremely high erosion rates due to columns fracturing and being removed. The thermal conductivity was higher for the EB-PVD GZO (1.06 W/m-K), and nearly identical for the APS GZO (0.51 W/m-K) and PS-PVD GZO (0.49 W/m-K). The erosion rates for the GZO coatings trended from APS (6.409 g/kg) > PS-PVD (4.050 k/kg) > EB-PVD (0.692 g/kg). The smaller particle size of the spray dried PS-PVD GZO feedstock enables significant vaporization, though the column widths were still slightly too large to take full advantage of the PS-PVD pseudo-columnar morphology, and thus the erosion rate was higher than EB-PVD. To further tailor the PS-PVD GZO coating vapor phase deposition and morphology, coatings were deposited at various incidence angles and assessed in terms of morphology and erosion. It was found that depositing at a 45° incidence can significantly reduce the intercolumnar gaps and reduce the erosion rate to a value of 0.305 g/kg, nearly half that of the EB-PVD GZO.