A critical assessment of particle temperature distributions during plasma spraying: Experimental results for YSZ

Plasma sprayed coatings of Yttria Stabilized Zirconia (YSZ) have been studied extensively through the years to understand variations in coating properties as well as to achieve control on microstructure of the coatings. The requirement for microstructural control and reliability have become all the more important as coatings have now become part of an integrated "prime reliant" design strategy aimed at increasing turbine inlet temperature and associated efficiencies. One of the important thrusts in monitoring and controlling the process has been the application of process sensors that measure spray stream characteristics, notably particle temperature and velocity. Although single particle-based measurements have been available for some time, in general control strategies based on particle state rely on average values of temperature and velocity. In this study, a detailed examination of particle temperature distributions is presented. When systematically examined over a wide range of operating conditions of the resulting range of particle temperatures, a significant structure in the statistical distribution has been observed. A close inspection of the data indicates that this distribution can be interpreted as melting state indicator for YSZ. A characteristic peak at the melting point of ZrO2 (error in absolute T-measurement is approximate to +/- 10%) can be used as an indicator for re-solidified particles. In the past, control strategies based on process diagnostic sensors have been based on average particle temperatures and velocities. Although the average values seem to be promising as control parameters, it has been shown through our results that different melting states could be demonstrated for the same average T and V settings. The melting state in turn has an important bearing on the coating structure and properties. It therefore implies that a process control strategy (to maintain coating quality) based on in flight particle sensors will have to take these findings into account. As an example, one strategy of process control would not only define the process in terms of the average particle temperature and velocity but also include the effect that parameter changes have on distributions.