Microstructural investigation of plasma-sprayed ceramic splats

Alumina and yttria stabilized zirconia (YSZ) (8 wt.% Y2O3) were plasma sprayed on smooth (R-a similar to 0.1-0.4 mu m) substrates preheated to 300 degrees C in order to study the splat formation, This preheating temperature was chosen to obtain very regular disk shaped splats quite adherent to the substrate made either of stainless steel 304 L, plasma sprayed alumina, YSZ polished coatings or pyrex. (C) 1997 Elsevier Science S.A. Thermal contact resistance evaluations suggested that the true contact area between the lamella and the substrate was less than the apparent area covered by the splat because of entrapped gas, surface contamination or oxide film. This resulted in good contact area through which the heat flux was released during crystal growth and poor ones for which the heat flux had to flow through the already solidified parts. In this paper, atomic force microscopy (AFM) of the splats, linked to results obtained by fast pyrometry (50 ns), SEM and XRD, enabled us to show different types of crystallisation fields related to the true splat contacts with the substrate. Different spraying angles and substrates have shown that solidification started before the flattening process was completed which explained the lenticular shaped splats obtained when preheating the substrates to 300 degrees C. A microcolumnar structure (column diameters between 125 and 400 nm depending on the solidification time) orthogonal to the substrate was observed in good contact area. This solidification time could be varied for example for YSZ splats between 0.7 and 7.6 mu s depending on substrate diffusivity and splat thickness linked to particle impact velocity. Much bigger crystals (one to two orders of magnitude) were observed in poor contact area. These poor contact areas were either distributed all over the surface of the splats for alumina or in the splat rim for alumina and YSZ. For the splat rim this poor contact is owing to the curling of the splat because of stress accommodation. For the big crystals formation within alumina splats its seems that it depends on substrate roughness and, for example, they can be observed along a slight scratch at the surface of a 304 L stainless steel substrate. The influence of surface roughness is drastically increased when the liquid material of the splat is still flowing at the end of the flattening stage, i.e. when spraying on an inclined substrate. In this case very small irregularities promote the formation of big crystals for both alumina and YSZ. When spraying on cold substrates the AFM observation is very difficult owing to a very distorted topography of the splat and its poor adhesion to the substrate. However, when possible, AFM shows in the splat central part ''islands'' of good contact surrounded by rims and in the fringes of the splat many ''fingers'', the structure of which is similar to that of the rims. (C) 1997 Elsevier Science S.A.