Interaction between a Thin Non Continuous D.C. Plasma Sprayed Alumina Layer and Preoxidized Low Carbon Substrate

As already shown 3 years ago, the preoxidation of smooth (Ra < 0.05 µm) low carbon steel substrates in a furnace under a CO₂ rich atmosphere at atmospheric pressure allows the formation of a wustite (Fe1-xO) layer which improves significantly the adhesion (> 55 MPa) of alumina coatings in spite of the rather low roughness (0.10 µm < Ra < 1.00 µm) of the oxidized surface. This contribution is devoted to a more precise study of the wustite layer and its interface with the alumina layer by X-ray diffraction (XRD), Mossbauer spectroscopy and scanning electron microscopy (SEM). Firstly the substrate was oxidized under different temperatures and durations in order to control the oxide layer thickness and structure. Secondly the substrate samples were preoxidized during 15 minutes at 1273 K under CO₂ atmosphere and, afterwards, preheated by the plasma jet in air just before coating. In this case the analysis was focused both on the alumina splat formation and the interface between splat and the oxide layer. Only a non-continuous alumina layer (a few splats) was sprayed: this allowed surface analysis down to the substrate through the alumina layer and the interface. This method avoids any modification of the searched information by a complex specimen preparation as required in the case of transmission electron microscopy (TEM) for example. For the steel surface preheated in CO₂ atmosphere, before spraying, SEM observations and XRD patterns showed the presence of a continuous oxide layer formed by wüstite crystals with an average size of 1-5 µm. After deposition, splats consisted of transitional alumina (γ phase) but the underlayer was no longer pure wüstite. XRD and Mossbauer identified magnetite at the surface of the oxide scale in contact with alumina. This can probably be considered as the result of a partial topotactic transformation of wustite into magnetite, since no morphological change of the oxide layer has been observed. It has been established that this transformation is a consequence of the pre-heating treatment, and not due to any reaction with alumina. It is worth noting that, under these conditions, γ alumina has a spinel structure analogous to that of the magnetite phase with which it was in contact: the alumina structure was possibly induced by that of the magnetite underlayer.