The plasma-spray process features complex plasma-particle interactions that can result in process variations that limit process repeatability and coating performance. This paper reports our work on the development of real-time diagnostics and control for the plasma spray process. The strategy is to directly monitor and control those degrees of freedom of the process that are observable, controllable, and affect resulting coating properties. This includes monitoring of particle velocity and temperature as well as the shape and trajectory of the spray pattern. Diagnostics that have been developed specifically for this purpose and integrated with a closed loop process controller are described.

This paper investigates the advantages that an advanced control system can provide for plasma spray deposition in terms of variation reduction and greater ability to engineer coating structure. We report our work on implementing a feedback control system that automatically adjusts system inputs to maintain the desired particle states in spite of process variations. The limitations and performance capabilities of both feed forward and real-time control are evaluated. Important system characteristics needed to develop such controllers are discussed including dominant nonlinearalities, dynamics, cross-coupling, distributions, and sensor issues. Performance is evaluated in terms of engineering coating structure and production objectives.

An advanced closed loop control system that enables tight control of the particle and substrate states has been developed. This unique capability allows deposition of coatings under very controlled conditions. This enables the construction of detailed process/property maps that can lead to a fundamental understanding of the formation mechanisms of key microstructural features during the plasma deposition process. The microstructural development during processing is discussed in light of the physics of microcrack formation during plasma deposition, including the effect of particle and substrate states on splat solidification, thermal gradients and residual stresses.

Recently it has been suggested that the carrier gas jet interaction with the plasma can have a large effect on the resulting particle temperature. The postulated interaction is through deflection of the main plasma jet and by delaying the heating of particles by the formation of a "cold" gas bubble. We have examined the effect of the gas jet itself on the temperature of the particles by attempting to artificially form a cold gas bubble using a separate, closely oriented gas jet. The effect of the "twin" co-flowing jet was evaluated by measuring its effect on the mean and standard deviation of the particle injection velocity and the resulting spray pattern and particle temperature. Additionally we have used alternative carrier gases with similar density but with specific heats that are higher than argon by a factor of two. A measurable but minor effect on particle temperature is observed.

Computational modeling is used to systematically examine many of the sources of statistical variance in particle parameters during thermal plasma spraying. Using the computer program LAVA, a steady-state plasma jet typical of a commercial torch at normal operating conditions, is first developed. Then, assuming a single particle composition (ZrO2) and injection location, real world complexity (e.g., turbulent dispersion, particle size and density, injection velocity and direction, etc.) is introduced "one phenomenon at a time" to distinguish and characterize its effect and enable comparisons of separate effects. A final calculation then considers all phenomena simultaneously, to enable further comparisons. Investigating each phenomenon separately provides valuable insight into particle behavior. For the typical plasma jet and injection conditions considered, particle dispersion in the injection direction is most significantly affected by (in order of decreasing importance): particle size distribution, injection velocity distribution, turbulence, and injection direction distribution or particle density distribution. Only the distribution of injection directions and turbulence affect dispersion normal to the injection direction, and are of similar magnitude in this study. With regards to particle velocity and temperature, particle size is clearly the dominant effect.

A model for oxidation of molybdenum particles during plasma spray deposition is developed. The diffusion of metal an-ions or oxygen cat-ions through a thin oxidized film, chemical reactions on the surface, and diffusion of oxidant in gas phase are considered as possible rate-controlling mechanisms with controlling parameters as the temperature of the particle surface, and local oxygen concentration and flow field surrounding the particle. The deposition of molten particle and its rapid solidification and deformation is treated using a Madejski-type model, in which the mechanical energy conservation equation is solved to determine the splat deformation and one-dimensional heat conduction equation with phase change is solved to predict the solidification and temperature evolution. Calculations are performed for a single molybdenum particle sprayed under the Sulzer Metco-9MB spraying conditions. Results show that the mechanism that controls the oxidation of this droplet is the diffusion of metal/oxygen ions through a very thin oxide film. A higher substrate temperature results in a larger rate of oxidation at the splat surface, and hence, a larger oxygen content in the coating layer. Compared to the oxidation of droplet during m-flight, the oxidation during deposition is not weak and can become dominant at high substrate temperatures.

Thermal spray processing of functionally graded materials requires that the spray patterns of different particle types coincide at impact and that each particle type arrives with the appropriate temperature and degree of melting. Measurements of particle velocity, temperature, and size along with spray pattern characteristics have been obtained for co-injected NiCrAlY and zirconia powder. The plasma and particle flow fields were also simulated with a pseudo 3-D model using the LAVA computer code. The model assumes that the gas flow is axisymmetric while the particles are treated in a fully 3-D manner. A stochastic discrete-particle model that includes turbulent dispersion dictates particle behavior. The simulation produced reasonably accurate velocities and particle trajectories, although, particle temperature is consistently over predicted. Comparisons between the calculated and measured velocity and temperature statistical distributions and calculated molten fractions are discussed.

The performance (particle velocity and velocity distribution) of a typical injector, and the resulting particle spray pattern for metallic (NiCrAlY) and ceramic (ZrO 2 ) particles are examined as a function of carrier gas flow rate and the effect of varying the geometry immediately upstream of the injector. Injector performance is also examined for a 1:1 mixture of ceramic and metallic particles such as is used in the spraying of functionally graded materials. The upstream geometries tested included a 90° "tee," a 90° elbow, and a straight entrance. The elbow geometry was tested in both "up" and "down" orientation to determine the influence of gravity. The upstream geometry can alter the average particle injection velocity by 10-15% influencing both the spray pattern trajectory and width.

In the spraying of functionally graded coatings, the particle ensemble delivered to the substrate varies from a relatively heavy, low-melting-point metallic particle to a significantly lighter, higher-melting-point ceramic particle. The desire is to deliver to the substrate a particle ensemble which has suitable velocity and temperature for the predictable and consistent formation of coatings with mixed particle types. The key to success is a thorough understanding of the relationship between spray gun parameters and the resulting particle condition. The gun parameters examined are powder loading (injection rate), powder mixtures, and secondary plasma gas (H 2 ). The spray characteristics measured were particle velocity, temperature, and spray pattern. The particle temperature and velocity are both significantly influenced by the flow rate of the secondary gas (gun power). The powder feed rate was found to have a small but measurable effect on both the spray pattern and the ensemble average particle temperature. It was observed that a "tight" hot particle spray pattern, unfortunately, does not necessarily minimize the number of cold unmelted particles.

In the spraying of functionally graded coatings the particle ensemble delivered to the substrate can vary from a relatively low melting point metallic particle to a significantly higher melting point ceramic particle. At various stages in the spray process the particle ensemble can be either predominantly metallic, ceramic, or an intermediate combination. For co-injected particles the mixtures do not behave as a simple linear superposition of the spray patterns of the individual particle types. The particle temperature, velocity, size distributions, and pattern characteristics of the resulting spray fields is examined for all ceramic particle sprays (ZrO 2 ), all metallic particle sprays (NiCrAlY), and for a 1:1 mixture. The major particle-particle interaction occurs in the injector itself and results in a modified spray pattern which is different from that of either material sprayed alone. The particle velocity distributions generally exhibit a bimodal nature which is dependent on the size and density of the injected particles.

An investigation into the dependency of the formation of functionally graded materials (FGMs) on process variables was carried out. The initial stage of the investigation involved a complete analysis of the plasma spray parameters used in the fabrication of an FGM constructed of NiCrAlY and partially stabilized zirconia (PSZ). In flight particle temperature, velocity and trajectory data were gathered for individual powders, as well as mixtures of the particle species, over a range of spray parameters. This data was combined with material specific properties such as flowability, apparent density, particle morphology and size distribution. The end result of the studies allowed for size matching of the particle species so as to ensure both species were molten at the nominal spray distance and possessed coincident impact velocities. Following the initial investigation, two spray conditions were selected for further analysis. Individual layers of specific powder mixture ratios were deposited as well as a complete FGM structure. The resulting structures were then compared based on their deposition efficiencies, porosity levels, compositional homogeneity and microstructures.

The variation in microstructure of high power plasma sprayed nickel coatings deposited with particle velocities ranging from 150 to 425 m/s and nominal particle temperatures of 1650 or 2050°C has been characterized. The relative density of coatings produced at the higher temperature is above 99.5% of theoretical regardless of the particle velocity; at the lower particle temperature the relative density is found to increase with increasing particle velocity. The fraction of unmelted particles is also found to increase with increasing velocity at the lower temperature. The relative deposition efficiency is approximately twice as high for the lower temperature particles compared to the high temperature, and for both temperatures the deposition efficiency decreases substantially with increasing velocity. Changes in the morphology of individual splats with changes in particle characteristics are also described.

High power supersonic plasma guns operating in excess of 200 kW can produce molten particles with 3 to 4 times the impact velocity of conventional plasma sprays. With this increased range of particle velocity it is important to understand the relationship between the torch input parameters and the sprayed particle velocity, temperature, pattern and size. Stainless steel particle velocity, temperature, size and relative number are measured for a high power plasma spray system operating at 110 kW. At the same torch operating conditions the plasma and particle flow fields are simulated with a newly developed computational model. It was found that the injection geometry plays an important role in the particle entrainment, heating and acceleration. In spite of the complexity of the system, i.e. supersonic plasma velocity with a high swirl component, the simulation produced reasonalble particle trajectories resulting in good agreement between the calculated and measured particle velocity, temperature and size distributions.

An experimental study of twin-wire electric arc spraying of zinc and aluminum coatings demonstrates the suitability of the process for anticorrosion applications. Experiments were conducted using Box-type full-factorial designs. Operating parameters were varied around the following process parameters: nozzle diameter, nozzle geometry, and system pressure. A systematic design of experiments displayed the range of processing conditions and their effect on the resultant coatings. The coatings were characterized with hardness and deposition efficiency tests, and optical metallography. Coating properties are quantified with respect to roughness, hardness, porosity, thickness, bond strength, and microstructure. The features of the coatings are correlated with the process changes. Selected analytical calculations and process diagnostics of the meltpool dynamics are presented.