Optimization of small-particle plasma-sprayed alumina coatings using designed experiments

Abstract

Statistically designed experiments and multiple regression analysis have been used to determine the effects of processing parameters on three properties of small particle plasma sprayed alumina coatings – permeability, hardness, and thickness. First, an initial screening study was conducted with six variables at two levels each. Based on the results from this preliminary set, a second set of experiments was then conducted to probe the effects of four variables at three levels each. Multiple regression analysis was used to create models describing the effects of the variables on the three coating properties. The models were then used to identify the optimum conditions for producing coatings with the desired properties. It was found that the optimum conditions for low permeability did not correlate with those for high hardness due to defect structures that formed in the hardest coatings. The morphology of the defect structures was such that they did not affect indentation measurements, but did affect permeability. These defect structures formed as a result of copious splashing during splat formation and correlated to coatings that had high surface roughness.