The adhesion of splats formed by impact of molten metal droplets was studied experimentally. Tin droplets (550 µm diameter) were produced using a drop-on-demand generator. To achieve high impact velocities the stainless steel coupons used as substrates were mounted on the rim of a rotating flywheel and heated using cartridge heaters. To hit a falling droplet with the substrate and photograph its impact, a timing circuit was used to synchronize three events with the position of the substrate: ejection of a droplet, triggering of the camera and a flash to provide illumination. The impact velocity was varied from 10 – 40 m/s whereas the substrate average roughness (2.0 µm) and the droplet diameter (~550 µm) were kept constant. We measured the adhesion strength of splats by a simple pull test. A wire was attached to the upper surface of each splat using epoxy and the force required to separate the splat from the substrate was recorded. A significant increase in adhesion strength was observed as the impact velocity was increased. Coatings were produced by depositing many droplets sequentially. Substrate temperature and impact velocitiy were the main parameters varied. SEM images of cross-sections through coatings showed that increasing impact velocity and substrate temperature produced better adhesion between the coating and substrate.

Thermal sprayed coatings are influenced by the wettabillity of the molten particles to substrates. The contact angles of several molten metal droplets are measured in air and vacuum. The flattening ratio is evaluated by taking into account of wettability. Comparing with the experimental data of the freefall droplet, it was concluded that the wettability to the roughened substrates influenced on the flattening characteristics.

The particle parameters including particle size, velocity and temperature influence significantly splat formation process in thermal spraying. The flattening degree of subsequent splat determines the coating structure and properties. Both theoretical analysis and simulation of splatting process indicate that the flattening degree depends on Reynolds number (Re) of spray particles. The experimental correlations suggest that the theoretical models overestimate the flattening degree. In the present study, with careful control of particle size and measurement of particle velocity and temperature, the relationship between the flattening degree and particle Reynolds number is examined experimentally. Copper powders of small size range are used to ensure valid of mean particle size. Plasma spraying is carried out under different conditions to change particle velocity and temperature. The particle velocity and temperature are measured using DPV- 2000. Splats were deposited on preheated polished stainless substrate surface. The diameter of individual splat was measured. The flattening degree was estimated using average diameter of splats and spray particles for individual spray condition. Using the exponential formula of Re with a power of 0.2, it was found that experimental correlation yielded a coefficient about half of that given by Madjeski’s model.

The quality of thermal sprayed coatings greatly depends on the individual flattened splats. Generally they are characterized by two main factors, flattening ratio and splat profile. The flattening ratio is specified straightforward, but it is difficult what factor should be selected from the various properties for the splat profiles. Fractal geometry is introduced to the profile evaluation of the splats which are formed by the free-fall metal droplets and by the alumina coatings using plasma spraying method. The fractal dimension of each splat was measured by SIA (Slit Island Analysis) technique which measures the ratio of area to contour length of the island appeared on the horizontally sliced plane of the splat. The obtained fractal dimension was related to with Re and We numbers in a bi-linear form, which is the same characteristics as the unevenness ratio obtained previously. It was concluded that fractal dimension is an effective measure to evaluate the splat profile.

A thermal spray coating is formed through successive impact, flattening, rapid cooling and solidification processes of a stream of spray droplets. Splashing may occur during droplet flattening process. Recent studies suggested that splashing can be suppressed when a molten droplet impacts on a preheated flat substrate. In this study, the splatting behavior in plasma spray is examined using molten spray droplets of different Reynolds number. Splats are deposited on preheated flat stainless steel surface. The morphology of splats is examined using optical microscopy and scanning electron microscopy. To adjust Reynolds number of spray droplets, copper droplets are produced using both Ar-H 2 and Ar-He-H 2 plasma jets under different operating conditions. As a result, the Reynolds number of spray droplets have been varied from about 18,000 to 90,000. It has been found that Reynolds number will influence splashing phenomena during splatting and consequent splat morphology. At low Reynolds number, splats present a regular disc morphology. However, when Reynolds number was increased up to about 5x104, the severe splashing around periphery of splat droplet was clearly observed despite the preheating of substrate. Based on the morphology of splats, a model for the spreading of molten droplet is proposed to explain the effect of Reynolds number on the flattening behavior of molten spray droplet.

The morphology of sprayed splat raises the coatings adhesion and the properties which are determined by the spraying parameters. A lot of studies in this field show that the substrate surface temperature is a very relevant factor for the splat shape: the hypotheses of substrate surface wetability and contamination or absorption layer on the surfaces are supported by the fact that the near disk-shaped splat can be obtained in increasing the substrate temperature. In the PROTAL process, a short duration pulse laser is used to ablate the substrate just before powder spraying. This ablation is powerful enough to eliminate the contaminations on the substrate surface and to improve the adhesion. In this study the analyses of NiAl splat morphology on polished TA6V substrate were carried out using PROTAL process with different substrate temperatures and different heating modes: the flame and another laser. Results show that the temperature at which the disk shaped splat can be obtained was decreased dramatically by PROTAL process and PROTAL process combined with another laser has increased the adhesion strength of the coatings.

A new thermal spray process is under development in order to produce thin coatings (thickness lower than 50 µm), with a fine microstructure (grain size smaller than 1 µm). It consists in injecting in a Direct Current (D.C.) plasma jet a suspension containing submicronic particles of the material to be deposited. To study the interaction between the plasma jet and the suspension, a system based on a pendulum allows the collection of particles at different distances from the injection point. In this paper the effect of substrate temperature upon the formation of micrometric zirconia splats was studied, glass and stainless steel were used as substrate materials.

This work deals with a 3-D transient simulation of the air plasma spraying of ceramic powders using a C.F.D. commercial code ESTET v3.4 that has been adapted to thermal plasma conditions. The mathematical model computes the distribution of particle velocity, temperature, molten state and size at impact and predicts the heat transfer to the substrate by plasma jet and particles. It incorporates the conversion from electrical to thermal energy in the torch nozzle as well as coating formation on the substrate. It makes it possible to predict the shape of the coating footprint when the torch and the substrate are fixed. The projections of the model are compared with experimental results that involve flow characteristics, time-dependant particle behavior in the flow and heat flux to the substrate.

The melting behavior of in-flight particle and its impact on splat morphology are studied. A group parameter, “melting index”, has been derived to correlate the melting status of inflight particles with particle size, velocity, and temperature which can be measured experimentally. Numerical simulations have been used to determine the unknown parameters in the melting index. The effect of particle size on its melting behavior has been investigated.

The system B-C-N contains the hardest known materials like diamond, cubic boron nitride and boron carbide, which also show excellent chemical resistance. The oxidation resistance is shifted to higher temperatures in comparison to pure diamond. But pure BCN coatings cannot be produced by conventional thermal spray processes, as the materials lack both a liquid phase and sufficient ductility to permit deposition. Conventional VPS equipment is successfully applied in Thermal Plasmajet CVD processes for high deposition rate synthesis of diamond coatings. The feasibility of SiCN or boron carbide synthesis by this method has also been proven. The use of liquid precursors results in outstanding deposition rates and improved operational safety. Methylized borazine is applied for synthesis of BCN coatings in thermal plasma jets. The use of single source precursors is advantageous with concern to the homogeneity of the coating forming species stoichiometry. For long-term storage cooling is necessary, but also under ambient conditions the precursor shows sufficient stability. Plasma gun nozzles with different diameter and design are applied and evaluated with concern to the resulting coating properties. Deposition rates of up to 1,500 µm/h have been achieved with homogeneous coating thickness and morphology on areas with 50 mm diameter. No porosity is detected in SEM investigations on cross sections and fracture surfaces show a fine columnar coating morphology. XRD investigations point at an amorphous structure. Only for very high substrate temperatures the formation of crystalline boron carbide B8C and h-BN or graphite phases is detected. Oxygen contamination results in boric acid formation and therefore has to be avoided carefully. During coating deposition on mild steel substrates the formation of boride and nitride reaction zones is observed. VPS sprayed nickel or molybdenum interlayers permit to inhibit the evolution of reaction zones. Thereby BCN coatings with thicknesses of up to 10 µm are deposited without local delamination. Space resolved emission spectroscopic analyses are carried out in order to detect coating forming and intermediate species. As Thermal Plasmajet CVD is a pure gas phase deposition process, the control of the space resolved emission permits easy process control.

A mathematical model was used to describe the fluid flow, electromagnetic and heat flow of the plasma jet in the non-transferred model. A computer program has been developed to simulate the fluid properties of the low energy and high efficiency plasma jet. The numerical results approximate the measurements by using the enthalpy probe.

We used an object oriented finite element numerical method to examine residual stress build-up in a simulated crosssection of coating. The cross-section image of coating was the result of simulation produced by a 3-D stochastic model of thermal spray coatings, which model features of coating microstructure including internal pores and surface roughness. An adaptive meshing technique was used to fit a grid in a section through the simulated coating, and stresses in the coating calculated using a finite element method. The results showed that stresses are sensitive to the thickness and roughness of the coating.

The motion and heating of a single particle in a plasma jet is strongly influenced by the feedstock injection conditions and can be predicted using computational modelling as described in the literature. In practice, however, the size and initial velocity of a given particle are essentially random variables within the fairly wide limits of the injected feedstock population. This is compounded by the non-uniformity of the plasma jet. As a result, the motion and heating of particles take on a substantial random element. There are three independent variables that jointly affect the particle motion and heating in a given jet: the diameter, the radial coordinate and the azimuthal angle of a particle. Nevertheless, these parameters cannot be specified for every particle in the jet in a deterministic manner owing to the above complexities and so a simulation based on a single particle cannot provide a realistic prediction of the deposition process. In the present study, the random element present in practical spraying is simulated using a Monte-Carlo approach. The distributions of the motion and heating of a population of particles are simulated rather than those for a single particle. The statistical method presented in this paper gives more detailed information on the effects of processing parameters and the assessment of process quality. The results show that this is able to provide a more accurate means of simulating the thermal spraying process.

Off-line control of thermal spraying generally implements the quantification of the energetic, environmental, injection and geometric conditions. The interdependence of these parameters with regard to the deposit properties were studied with different statistical methodologies. The aim was to point out the large possibilities of the Artificial Neural Network methodology in the process optimization. In that way, experiments were designed considering atmospheric plasma spraying and the following operating parameters: arc current intensity, powder feed rate, carrier gas flow rate, total plasma gas flow rate, primary hydrogen content and scanning step. These parameters were related to the coating porosity level and hardness (HK). For each case, simple correlations were studied with none linear regressions, design of experiments and a Multilayer Perceptrons (MLP, with two hidden layers). Non linear correlations were compared with MLP based on the Average Sum of Absolute Error (ASAE) and other correlation factors (RSS, R 2 , Ra 2 , STD, RS and AR).

A numerical model is presented for the computation of heat transfers during the APS thermal spray process. This model includes the contributions of both the impinging plasma jet and that of the particle flux on the substrate heating. The contribution of the impinging plasma jet is taken into account using a computational fluid dynamic model describing the impact of the plasma jet on the substrate. For this part of the work, a two-layer extension to the Chen-Kim k-s model was used allowing the description of both the turbulent plasma jet and that of the flow in the viscous sub-layer formed on the substrate surface. The contribution of the sprayed particles is taken into account considering their distribution in the spray jet. Since this is an important parameter that could affect the model accuracy, measurements of the deposit thickness profiles were first performed using the non-destructive acoustic microscopy method and the corresponding particle flux distribution was then deduced. Heat transfers inside the substrate were then computed using a three dimensional in-house code based on a finite volume approach. In the case studied, the results show that the contribution of the sprayed particles forming the coating is much more focalized than that of the plasma flow itself whereas the substrate nature has a strong influence on the thermal flux dissipation (not presented in the following). These elements are expected to provide useful information concerning the coating adhesion mechanisms and the formation of residual stresses during the coating elaboration.

This paper presents the numerical simulation of the plasma flow into a dense atmosphere. The plasma generation is performed by the simple model developed by Eichert. Two models are exposed to take in account the arc fluctuation inside the anode. They permit us to simulate plasma puffs convected into the flow. The aim of this study is to compare these two models with experiments and to determine which one is the most relevant.

A study of thermal fluxes transferred during the HEATCOOL process is proposed. The concept of this process, specially designed to enhance the residual stresses relaxation, consists in the use of a consecutive three-step procedure during the coating elaboration (heating / spraying / cooling). The present study focuses on thermal exchanges occurring during the heating step. For this, the elaborated experimental equipment incorporates a series of ten holes aligned equidistantly with 5 mm separation. A burning gas mixture (premixed acetylene and oxygen) is injected through these holes and the burning gas jets impinge and heat the substrate. The stand-off distance between the heating device and the substrate may be adjusted between 30 and 90 millimeters. Concerning thermal fluxes transferred using this experimental device, a front work piece incorporating several thermocouples was used to perform heat flux measurements. In a first step, the case of a single hole was considered. Since this method is not able to provide the thermal flux directly, the corresponding thermal fluxes were deduced using an inverse heat conduction problem method that was specially developed. Results obtained using this inverse problem method based on experimental measurements are then compared with numerical predictions obtained using a computational fluid dynamic model representing the system. For this part, the PHOENICS software was used to perform the corresponding computations.

The study of heat transfers in a substrate exposed to an impinging plasma jet is proposed using two different software products. Thermal exchanges between the plasma jet and the substrate were first calculated using the PHOENICS CFD software in which a two-layer extension to the Chen-Kim k-s model was implemented in order to consider both the turbulent nature of the plasma jet and heat transfer phenomena through the viscous sub-layer formed at the surface of the substrate. The model is supposed to provide accurate predictions of thermal exchanges. However this preliminary step is not described since it is part of some previous studies. In a second step, two different commercial software products are used to perform three dimensional transient calculations of the heat conduction inside the substrate. The first approach consists in the use of the finite element based SYSWELD software whereas the second one consists in the use of the finite volume based PHOENICS software. Numerical results are presented and compared for the case of an impinging plasma jet displacing linearly on the substrate. Additionally, the influence of different parameters such as the substrate sample thickness, the stand-off distance, the displacement velocity or the nature of the substrate is also discussed. The results show a good accordance between numerical predictions obtained using the two methods concerning the maximum temperature observed. These results are useful since the substrate temperature is known to have an important influence on the coating adhesion and properties.

The present study was devoted to the optimization of the nozzle exit geometry of a TAFA 9000 wire-arc spray gun. Computational Fluid Dynamic was first used in order to examine the effect of different nozzle configurations, considering small changes operated on the original design. Finally, one of the modified geometries tested numerically was retained in view of the CFD results and the corresponding new nozzle was machined. In a second step, experiments were performed in order to observe the effect of the new design on in-flight droplet characteristics. For this part, a comparison was made between the original design and the modified one for two different type wires (a steel wire and a stainless steel cored wire). These experiments were performed using the well known DPV 2000 device for different operating conditions. The results indicate that the modified design provides a significant increase in the droplet velocity (about 20% higher for the new design) with unchanged thermal spray parameters, indicating that the gun exit design may have a strong influence on inflight characteristics of the droplets. Moreover, droplets were collected in water and the size distribution was analyzed.

It is necessary to cool specimens during spraying in the case APS or HVOF, because process-induced heat rises the specimen temperature and leads to oxidation and spalling of coatings. A reasonable cooling just after spraying improves some properties such as microhardness, adhesion and cohesion of the coating/substrate system. In the modelling of specimen temperature and residual stress, it is necessary to know the flux distribution of the cooling jet like compressed air, CO 2 liquid jet etc. Therefore, the evaluation of the flux becomes important. In order to measure and analyse the distribution of cooling flux imposed on the substrate, the theory of the inverse problem of heat conduction was applied and an experimental apparatus was designed to mesure the transient temperature. Because of its insensibility to the effect of measuring error, the conjugate gradient method, an effective method of inverse problems was chosen among several mathematical optimisation methods. The flux distributions of cooling jet can be estimated by using the measured data and a program written according to the conjugate gradient method.