Summary Of Principal Technical Progress

TBC Sintering

In addition to having a low thermal conductivity, a TBC must maintain strain compliance in the plane of the coating to accommodate the thermal expansion mismatch with the underlying superalloy. With increasing turbine temperatures, there is the natural concern that TBC sintering will decrease the strain compliance and promote failure. Furthermore, there is concern that at higher temperatures the tetragonal-prime phase will transform to the monoclinic phase. Previous investigations have studied TBC sintering by first removing the TBC from the superalloy and then studying its sintering behavior in a traditional ceramics manner, for instance, by measuring length changes or physical density as a function of time and temperature. This approach neglects the effects of the constraint provided by the super-alloy and possibly hides important microstructural changes that can affect TBC failure. This is especially so since TBCs in service are exposed to a temperature gradient.

To investigate the effects of constraint on sintering, EB-PVD TBC coatings on two types of substrates were provided by Howmet Castings and heated isothermally at high temperatures. One substrate type was a N5 superalloy with a standard platinum-modified nickel-aluminide bond-coat. The other substrate type was a sapphire single crystal. These two were selected since the N5 superalloy has a larger thermal expansion coefficient than zirconia whereas sapphire has a smaller coefficient. Thus, at temperatures above the TBC deposition temperature, the differential thermal expansion with the N5 should serve to separate the individual TBC columns whereas the differential expansion with the sapphire should act to bring the columns together. The former would be expected to delay sintering whereas the latter should promote sintering.

Sintering was initially performed at 1200oC; higher temperatures were not used with the PtNiAl coated N5 substrate so as to avoid incipient melting between the bond-coat and the superalloy. Higher temperature sintering will be undertaken on the sapphire substrates in future. Microscopy observations of the coatings at different times indicate that column sintering on both the N5 and sapphire substrates occurs by the same microscopic mechanism. The mechanism involves surface diffusion, the formation of necks between adjoining columns by a Rayleigh instability process and the sintering together of the columns as the necks grow. This process is shown in figure 1 below.

The response of the coating to the sintering process is the formation of a network of cracks, reminiscent of "mud-cracking", as illustrated in figure 2. The cracks, more strictly "gaps", form as a result of the competition between local, in-plane densification caused by inter-columnar sintering, and the constraint of the underlying substrate that fixes the lateral size of the coating. Studies are underway to quantify the spacing of the cracks and their depth as a function of time and temperature. On a flat substrate, such as the sapphire, the onset of cracking has not yet been identified. However, on rougher substrates or on a substrate that undergoes rumpling with thermal cycling the length scale of the cracking appears to correlate with the underlying roughness. This is illustrated by the micrographs in figure 3

where the cracking is associated with the grain boundaries of the bond-coat. It appears that the TBC columns formed on the remains of the grain boundary ridges (left after grit-blasting) are canted and so are slightly closer together along the grain boundaries than away from the boundaries. Consequently, they sinter together faster and a gap opens up across the lines delineating the bond-coat grain boundaries.

At this stage of the work it is not possible to anticipate all the consequences of the sintering processes. However, there are two obvious consequences. One is the "gaps" provide a ready access path deep into the coating for the ingress of corrosive species, such as vanadates and sulphates. The second is that cracks can deflect parallel to the substrate, promoting coating loss, under a thermal gradient.

A detailed description of these findings will be written up as a complete manuscript in the near future.

Figure 1. TBC columns sintering together by neck growth at 1150oC. Left, after 100 hrs. Right, after 850 hrs. Note the surface smoothing of the columns.
Figure 2. Surface of EB-PVD after 100 one-hour thermal cycles at 1150oC. Higher magnification images reveal columns are beginning to sinter together and small gaps are beginning to appear.
Figure 3. Same coating but after 850 one-hour cycles. Periodic gaps, having the appearance of "mud-cracking", are forming as columns are locally sintering together constrained by the fixed size of the underlying superalloy.
0 0

Post a comment