The influences of sulfur impurities and Pt incorporation on the scale adhesion behavior of aluminide coatings were studied and compared. Low-sulfur NiAl coatings were prepared on a desulfurized, yttrium-free, single-crystal Ni-based superalloy by a modified version of a conventional aluminizing procedure based on chemical vapor deposition. The sulfur level in the resulting NiAl coatings was measured to be less than ϳ0.5 ppmw by glow-discharge mass spectroscopy. Platinum-modified aluminide coatings were synthesized by first electroplating a thin layer of Pt (ϳ7 m) on the superalloy, followed by the same low-sulfur aluminizing procedure. The measured sulfur content in the (Ni, Pt)Al coating was substantially higher than that of the low-sulfur NiAl coating due to contamination during the Pt electroplating process. A very adherent ␣-Al 2 O 3 scale formed on the grain surfaces of the low-sulfur NiAl coating during cyclic oxidation testing at 1150 ЊC, but scale spallation eventually occurred over many of the NiAl grain boundaries. In contrast, despite the higher level of sulfur in the (Ni Pt)Al coating, a very adherent scale was formed over both the coating grain surfaces and grain boundaries during thermal cycling. These results suggest that Pt additions can mitigate the detrimental influence of sulfur on scale adhesion.

The effects of Pt incorporation on the isothermal oxidation and diffusion behavior of low-sulfur aluminide bond coatings were investigated. Aluminide (NiAl) coatings and Pt-modified aluminide (Ni,Pt)Al coatings were synthesized on a low-sulfur, yttrium-free single-crystal Ni-based superalloy by a high-purity, low-activity chemical vapor deposition (CVD) aluminizing procedure. The isothermal oxidation kinetics and scale adhesion behavior of CVD NiAl and (Ni,Pt)Al were compared at 1150 ЊC. Compositional profiles of alloying elements in the NiAl and (Ni,Pt)Al coatings before and after isothermal oxidation were determined by electron microprobe analysis. Platinum did not reduce oxidescale growth kinetics. No significant differences in bulk refractory metal (W, Ta, Re, and Mo) distributions were observed as a result of Pt incorporation. Spallation of the alumina scale and the formation of large voids along the oxide-metal interface were commonly observed over the NiAl coating grain boundaries after 100 hours at 1150 ЊC. In contrast, no spallation of Al 2 O 3 scales occurred on (Ni,Pt)Al coating surfaces or grain boundaries, although the sulfur content in the CVD (Ni,Pt)Al coatings was higher than that of the CVD NiAl coatings. Most significantly, no voids were observed at the oxide-metal interface on (Ni,Pt)Al coating surfaces or cross sections after 200 hours at 1150 ЊC. It was concluded that a major beneficial effect of Pt incorporation on an aluminide coatings oxidation resistance is the elimination of void growth at the oxide-metal interface, likely by mitigation of detrimental sulfur effects.

A laboratory-scale chemical vapor deposition (CVD) reactor was used to perform "continuous" Hf doping experiments while the surface of a single-crystal Ni alloy was being aluminized to form an aluminide (␤-NiAl) coating matrix for 45 minutes at 1150°C. The continuous doping procedure, in which HfCl 4 and AlCl 3 were simultaneously introduced with H 2 , required a high HfCl 4 /AlCl 3 ratio (Ͼϳ0.6) to cause the precipitation of Hf-rich particles (ϳ0.1 m) at grain boundaries of the coating layer, with the overall Hf concentration of ϳ0.05 to 0.25 wt pct measured in the coating layer by glow-discharge mass spectroscopy (GDMS). Below this ratio, Hf did not incorporate as a dopant into the growing coating layer from the gas phase, as the coating matrix appeared to be "saturated" with other refractory elements partitioned from the alloy substrate. In comparison, the Hf concentration in the aluminide coating layer formed on pure Ni was in the range of ϳ0.1 wt pct, which was close to the solubility of Hf estimated for bulk NiAl. Interestingly, the segregation of Hf and the formation of a thin ␥Ј-Ni 3 Al layer (ϳ0.5 m) at the coating surface were consistently observed for both the alloy and pure-Ni substrates. The formation of the thin ␥Ј-Ni 3 Al layer was attributed to an increase in the elastic strain of the ␤-NiAl phase, associated with the segregation of Hf as well as other refractory alloying elements at the coating surface. This phenomenon also implied that the coating layer was actually growing at the interface between the ␥Ј-Ni 3 Al layer and the ␤-NiAl coating matrix, not at the gas/coating interface, during the early stage of the coating growth.

A chemical vapor deposition (CVD) procedure was developed for preparing a high-quality ␣-Al 2 O 3 coating layer on the surface of a single-crystal Ni-based superalloy using AlCl 3 , CO 2 and H 2 as precursors. A critical part of this procedure was a short-time preoxidation step (1 min) with CO 2 and H 2 in the CVD chamber, prior to introducing the AlCl 3 precursor. Without this preoxidation step, extensive whisker formation was observed on the alloy surface. Characterization results showed that the preoxidation step resulted in the formation of a continuous oxide layer (ϳ50 nm) on the alloy surface. The outer part of this layer (ϳ20 nm) appeared to contain mixed oxides, whereas the inner part (ϳ30 nm) mainly consisted of ␣-Al 2 O 3 grains with -Al 2 O 3 as a minor phase. We observed that the nucleation of ␣-Al 2 O 3 in the preoxidized layer was promoted by (1) rapid heating (10 seconds) of the alloy surface to the temperature region where ␣-Al 2 O 3 was expected to nucleate; (2) the low oxygen pressure environment of the preoxidation step, which kept the rate of oxidation low; and (3) contamination of the reactor chamber with HfCl 4 . The preoxidized layer served as an effective diffusion barrier for mitigating the interaction with some of the alloying elements such as Co and Cr with the CVD precursors and eliminating whisker formation on the alloy surface.Al 2 O 3 (s) ϩ 6HCl (g) ϩ 3CO (g) 2AlCl 3 (g) ϩ 3H 2 (g) ϩ 3CO 2 (g) →

The (100) surface of a single-crystal Ni alloy was aluminized as a function of time to study the development of the resulting coating microstructure. A chemical vapor deposition (CVD) reactor, which was specially configured for short-term aluminizing experiments, was used to prepare coating specimens at 1150 ЊC. After 5 minutes, ␥ Ј-Ni 3 Al particles ϳ100 nm in size randomly nucleated on the alloy surface. Within 20 minutes, a coating layer consisting of preferentially oriented, columnar ␤-NiAl grains was formed with the segregation of refractory elements (i.e., Ta and W) from the alloy to the coating grain boundaries. The lateral growth of the columnar grains was observed to be relatively rapid for up to 45 minutes, but slowed considerably between 45 and 180 minutes. While the columnar nature of the coating did not change significantly after 20 minutes, the surface features continually evolved, with the appearance of a small amount of the ␥ Ј phase, which coincided with the segregation of the refractory elements to the coating surface.