Abstract
Flattening of an individual particle on the substrate is a unit cell for the coating formation in plasma spraying. In this study, individual Cu splats and coatings were collected on the AISI304 substrates at different temperatures by plasma spraying. Microstructures of the splats and coatings were observed by scanning electron microscope, and the adhesion strength of each coating was measured by tensile test. The coating hardness was also evaluated. The splats collected on the unheated substrate performed as splash shape, while most of the splats collected on the substrate at high temperatures exhibited as regular disc type. In particular, less nanopores could be found at the bottom surface of the disc shaped splats. The coating adhesion strength increased dramatically from 22 to 63 MPa with the increase in substrate temperature; the coating hardness also increased with the increase in substrate temperature. Accordingly, the coating performance has a close relation with the splats flattening behaviour on the substrates surface.
Introduction
Plasma spraying is a typical surface treatment technique that can provide thick coatings on the roughen substrate by the impingement of the fully or partly melted particles. This method has been used in a wide range of industrial applications. With the continuing development of industry, materials are being subjected to severer and more demanding environments, and coating quality requirements must significantly increase. However, the controllability and reliability of this process have not been established yet, and better knowledge of the basic scientific phenomena is necessary for process optimisation. It was proposed that as many as 50 parameters control the quality of plasma coatings, 1 and substrate temperate was recognised as one of the dominations in plasma spraying. Many studies have been devoted to the effect of substrate temperature on coating properties during the last years.2–11 Either metallic or ceramic coatings were deposited on the substrates at different temperatures. In relation to the coating requirements, the dependence of coating properties such as hardness, porosity, thermal expansion coefficient, elastic modulus, compositions, ionic conductivity, etc. on substrate temperature was systematically carried out. The results proposed that the substrate temperature has a profound effect on the coating properties.
Actually, the coating was fabricated by the individual splat deposited on the substrate or on the previously deposited one, the flattening and solidification of the single splat could be treated as the fundamental process of the coating fabrication; thus, it is meaningful to investigate the dependence of coating performances on the formation process of single splat. The effect of substrate temperature on the splat flattening behaviour has been intensively studied for metallic and ceramic powders onto different substrates.12–22 The feasible factors, such as adsorption/desorption of adsorbates and condensates on the substrate surface, wetting of molten droplet by substrate surface, surface characteristic and thermal contact resistance, etc., might affect the splat formation process jointly.
In general, the adhesion strength between the coating and the substrate is one of the most important properties from which the lifetime and the quality of a specific application can be estimated;23–25 high bond strength of a coating is associated with higher erosion, corrosion and abrasion resistance. 26 With the purpose to predict the coating performance and optimise the spray process though observation on the flattening behaviour of the final depositions, the effects of substrate temperature on splat formation process and coating adhesion strength in plasma spraying have been systematically investigated in this study. The dependence of coating hardness on the substrate temperature was also evaluated.
Experimental
The feedstock powder was commercially available Cu powder with diameter <75 μm. The substrate material was AISI304, and substrate temperatures were room temperature (300 K), 373, 473, 573, 673 and 773 K respectively. The spraying was carried out by atmospheric plasma spraying operated at 40 kW. The spraying distance between the gun and the substrate was kept at 120 mm. The powders were injected at a feeding rate of 6 g min−1 for splat collection, and 30 g min−1 for coating fabrication experiments. Argon and hydrogen were used as operating gas with flowrate of 60 and 6·5 L min−1 respectively.
Mirror polished AISI304 substrates were heated by the heater, and the heated substrate was located perpendicular to the spray gun. Once the substrate was heated to the preset temperature, the spraying gun was moved rapidly in one direction, and some individual splats could be caught on the substrate surface. The top surface morphologies of the splats were observed by scanning electron microscope (SEM). After the top surface observation, the splats were pulled off using carbon tape, the bottom surface morphologies of splats caught on the carbon type were observed using SEM as well.
During coating fabrication process, the grit blasted substrates were heated by the plasma flame, and the substrate temperatures were measured by thermocouple. Once the temperature reached the preset value, the plasma torch was moved immediately by a robot and the coating deposited on the substrate gradually. The top surface morphologies of the coatings were observed by SEM. Vickers hardness was measured using microhardness tester on polished sample surface using a load of 100 g for 15 s. The adhesion strength measurements were carried out using universal testing machine in accordance with the standard American Society for Testing and Materials C633-79, entitled ‘Standard Test Method for Adhesion or Cohesion Strength of Thermal Spray Coatings’. Coatings were firstly sprayed onto the cylindrical specimen, the coating surface is bonded to the end of another cylinder of the same material and geometry using epoxy adhesive film (FM 1000, Cytec Industries, West Paterson, NJ) with a tensile strength of 90 MPa is recorded. The bond strength is calculated by dividing the force to pull the cylinders apart by the cross-sectional area of the cylinder.23–26 The fractured surface after tensile test was examined using SEM.
Results and discussion
Microstructures of splats
Figure 1 illustrates the top surface morphologies of Cu splats deposited on AISI304 substrates at different temperatures. According to the figures, the splats achieved on the substrate at room temperature were completely splashed, leaving only a small centre splat surrounded by a ring of fragments. With the increase in substrate temperature, the splat deposited as disc shaped ones instead of splashing gradually. On the substrate at 373 K, most of the splats show a uniform morphology with clear flow pattern and projections along the periphery of the splat. In particular, some ripple structures were found at the periphery of the splats, because the rapid solidification at the bottom surface of the splat prevents the complete smoothen flattening before reaching its final diameter.18,22 Increasing the substrate temperature to 573 K, most of the single splats remained on the substrate, with only few radial fingers at the periphery of the majority of the splats. With the continuing increase in substrate temperature up to 773 K, perfectly disc shaped splats with very little evidence of splashing were found on the substrate surface, and some exfoliation induced by the oxidation was found at the central zone. The oxidation might be generated during the splat flattening process and, in particular, after the flattening finish because the splat–substrate system was still very hot on the substrate at high temperature. However, ripple structure could not be found at the periphery of the splat in this case, which might have been covered by the oxidation layer on the top surface. 18
Top surface morphologies of Cu particles sprayed onto AISI304 substrates at different temperatures: a 300 K; b 373 K; c 573 K; d 773 K
The bottom surface morphologies of splats achieved on the substrates at different temperatures are illustrated in Fig. 2. Numerous pores could be found at both the central core and the surrounding ring of the splash splat collected on the substrate without preheating. The formation process of the pores has been reported by Qu et al., 27 which proposed that the desorption of adsorbed water and other gas condensations on the substrate surface during the flattening process of molten droplet induced the formation of nanopores. For more outer region of the central splats, irregular shaped flowing pores can be found. This might be the escaping channel of the evaporated gas and the ‘gas pocket’ induced by the entrapped environmental gas during the splat flattening process. With the increase in substrate temperature, the amount of the nanopores at central zone and the flowing pores at periphery decreased gradually; the solidification structures also became more dense and homogeneous. In particular, for the splat collected on the substrate at 773 K, almost no pores could be found, and solidification structure with smaller grain size was observed near the peripheral region of the splat.
Bottom surface morphologies of Cu particles sprayed onto AISI304 substrates at different temperatures: a 300 K; b 373 K; c 573 K; d 773 K
It has been proposed that the surface roughness increase in nanometre scale due to the oxidation caused by heating could promote the favourable wetting between molten droplet and substrate.28,29 Meanwhile, the adsorbed gas condensations could be gradually removed with the increase in substrate temperature.12,13 Consequently, the increase in substrate temperature can result in increasing of intimate contact between molten droplet and substrate. As a result, the wetting of substrate by molten droplet and the heat transfer between them should be affected, so that the viscosity at splat bottom surface should be influenced, and finally determine the splat shapes jointly.
Microstructures and hardness of coatings
The substrates for coating fabrication were grit blasted before the spraying. Although the surface roughness of substrates for coating fabrication is much rougher than the mirror polished substrate for splat collection, the heating before the spraying should have a similar effect to remove the adsorbed gas condensations on the surface and improve the surface activation. The thicknesses of the coatings slightly increase with the substrate temperature, because less material loss of the individual splats on the hot substrate can improve the deposition efficiency. The top surface morphologies of the coatings fabricated under designated conditions were observed as shown in Fig. 3. Almost all the particles were fully melt; however, the coatings obtained on the substrate without heating seems rougher than deposited on the substrate at high temperatures. Similar with the individual splat observation, fragmentation of the splats containing the small particles was found at the top surface of the coating on the substrate at room temperature before the spraying, while the splats deposited on the top of the coating exhibited as well layered lamellae.
Top surface morphologies of Cu coatings deposited on AISI304 substrates at different temperatures: a 300 K; b373 K; c 573 K; d 773 K
It is well known that disc shaped splat with good contact with the substrate can result in improved properties. The hardness of coatings deposited on the substrates at different temperatures was summarised in Fig. 4. The result indicates that the coating hardness slightly increased with the increase in substrate temperature. It has been proposed that the hardness of cold sprayed coatings was higher than the bulk materials; 30 the hardening during the striking of the in flight particle onto substrate surface might result in such phenomenon. However, except for the substrate temperature, all the other spraying parameters were kept in constant during the spraying, so that there was no velocity change of the particle before the spraying. The flattening of the individual splat on the substrate at different temperatures might dominate the transition. As the wetting of substrate by molten droplet was enhanced by increasing substrate temperature, the splat spread smoothly on the substrate and few debris from the flattening droplet could be formed, so that less voids exist in the coating achieved on the substrate at high temperature, and the denser structure can prevent the deformation in the hardness test. Second, more particles can be deposited on the substrate at high temperature with the same spraying pass; hence, the peening effect by the subsequent molten droplets should be stronger. Third, a typical thin layer composed of fine grains was found at the splat/substrate interface in the previous study; 22 the thickness of this layer significantly increases with the substrate temperature. The grain refinement strengthening can enhance the hardness of the individual splat and finally promote the increase in coating hardness.
Relationship between coating hardness and substrate temperature
Adhesion strength of coatings
Adhesion strength is the fundamental but most important property of a coating. The coating will never protect a substrate if it does not adhere to it. The images of fracture surfaces after they were pulled apart from substrate at different temperatures are shown in Fig. 5. In each figure, the left specimen was the original test coupons after adhesion measurement, while the right one was the counterpart. According to the figure, the failure happened at the coating/substrate interface when the substrate without heating or held at low temperatures (Fig. 5a and b); most of the coatings was removed, and only a small portion of coating remained on the substrate surface. With the increase in substrate temperature (Fig. 5c and d), the failure happened at the interface between interlayers of the coating rather than the coating/substrate interface. Most of the coatings still remained on the substrate, which indicated that well adhesion might exist between the coating and substrate at high temperature.
Fracture surface of specimen heated to different temperatures: a 300 K; b 373 K; c 573 K; d 773 K
Figure 6 illustrates the dependence of coating adhesion strength on substrate temperatures. The adhesion strength was 22 MPa when the coating was deposited on the substrate without heating. A significant increase in the adhesion strength by increasing the substrate temperature was recognised and reached 63 MPa at the substrate temperature of 673 K. It is, however, found that no continuing increase in adhesion strength when substrate temperature was heated to 773 K. This might result from the severer oxidised interface between interlayer, which was generated by the remarkable oxidation of the individual splat obtained at 773 K.
Relationship between coating adhesion strength and substrate temperature
Precise observations for the microstructures of the fracture surfaces were carried out using SEM as indicated in Fig. 7, which was conducted at the central zone of the test coupons after adhesion measurement. The failure occurred at the interface between substrate and coating, and only few individual splats with pores at the central zone were found on the cylindrical specimen at room temperature (Fig. 7a). Most of the coatings were removed, and only a small portion of coating remained on the substrate surface. For the specimen at temperature of 373 K before the spraying (Fig. 7b), failure also occurred at the coating/substrate interface. Only few individual splats but with less pores remained on the substrate surface, and ripples structure was found at the periphery of the splat. With the continuing increase in substrate temperature to 573 K (Fig. 7c), most of the coatings still remained on the substrate; however, a portion of the splats were cracked, which indicated that there is a very good adhesion between the intersplats. When the substrate temperature was 773 K (Fig. 7d), most of the coatings still maintained on the coupon, significant deformation could be found as shown by the arrow and most of the damages occurred inside the individual disc shaped splat. The cohesion between the splats sometimes is quite favourable. It is not a coincidence that the adhesion strength has a close relation with the shapes of the final individual splats. That is, most of the splats deposited on the substrate without heating exhibited as splash shape, and correspondingly, the adhesion strength is the lowest; the portion of disc shaped splat increased with the increase in substrate temperature, while coating adhesion strength enhanced in the meantime.
Fracture surface microstructures of Cu coatings deposited on substrates at different temperature after tensile test: a 300 K; b 373 K; c 573 K; d 773 K
In general, mechanical bonding, physical bonding and metallurgical bonding are probably the bonding mechanisms involved in plasma sprayed coatings.31,32 Mechanical bonding usually prevails for most coatings. The heat and momentum of the particle promote plastic deformation of both the substrate and the impinging particle, which bond together by mechanical interlocking. As the individual splat performed as the unit cell for the coating build-up, if all the single splats adhered well with the substrate, favourable adhesion strength can be expected for the coatings. On the contrary, the adhesion between the coating and the substrate might be very poor. Accordingly, once the molten particle strikes onto the substrate surface or the previously deposited splats in plasma spraying, a dynamic impact pressure generated by the kinetic energy of the in flight particle should be initially perpendicular to the substrate surface, which keeps the fluid flowing along the substrate surface. This impact pressure could be very high and concentrated at a small contacting area and then spread quickly with droplet flattening. 33 The velocity of the in flight particles before them that strike onto substrate should have no typical difference by changing the substrate temperature only; hence, the initial dynamic impact pressure must be the same. However, the spreading speed of the droplet onto substrate surface should be much faster on the substrate at high temperature due to the favourable wetting of substrate by molten droplet. 27 Most of the molten droplets remained on the substrate and deposited as disc shaped splat with a large flattening degree. The real contact area of the central zone becomes greater than the splash type; hence, the interlocking of splat with substrate could be enhanced, so that adhesion of the coating could be improved.
While the particles were sprayed onto the unheated substrate, the splat does not have contact the substrate over all of its bottom area. The real contact areas is reduced by:
the fragmentations flying with an angle or parallel to the substrate always have no dynamic impact pressure perpendicular to the substrate
the gas pocket induced by the environmental gas and/or working gas dissolved in the molten droplet during the accelerate process in plasma flame
the nanopores at splat bottom surface generated by the desorption of adsorbed gas condensation and environmental gas entrapped by arriving particles on the substrate.
The reduction in the real contact area restrains the formation of interlocking with substrate surface, which can strongly reduce the mechanical bonding with the substrate.
On the other hand, almost all the adsorbed water and gas condensation molecules on substrate surface were removed by heating up to a critical temperature. The substrate surfaces are clean and can produce better intimate contact of particles to the substrate. Surface roughness increase in nanoscale can improve specific surface area, which was defined as the total surface area per unit of mass of the solid material. 34 Accordingly, the intimate contact between the flattening droplet and substrate surface is much closer with each other on the substrate at high temperature than the substrate without heating, which might reach the field of attraction of the atoms; hence, the physical bonding that takes place by van der Waals’ forces should also be enhanced, consequently promoting the better coating adhesion on the substrate at high temperature.
Third, the contact temperature at the particle's interface with the substrate at impact can be approximately treated as the mean value of droplet temperature and substrate temperature at the very initial stage. On the substrate at high temperature, the contact temperature probably higher than the melting point of the AISI304 substrate, diffusion and some chemical reaction between the flattening droplet and substrate could be activated. This phenomenon has been experimentally observed and discussed by the previous researchers.35,36 The metallurgical bonding should be enhanced with the increase in contact temperature by substrate heating. However, the flattening and solidification of the molten droplet on substrate surface are always finished by several tens microseconds. Its effect of metallurgical bonding on the coating adhesion strength probability remains small, but cannot be ignored.
On the basis of the discussion above, it could be summarised that substrate heating can promote the formation of disc shaped splat on substrate surface, and then affect the flattening and solidification behaviour of the splat deposited on the top of the previously deposited splats, with the accumulation of such splats finally resulting in the favourable adhesion with the substrate and cohesion between different splats.
Conclusion
Cu powders were plasma sprayed onto mirror polished and grit blasted AISI304 substrates at different temperatures respectively. Wetting and heat transfer between molten droplet and substrate can be enhanced because most of the surface moisture and other gas condensations were removed by substrate heating. As a result, splats with a better shape could be produced on the substrate at higher temperature. The coating hardness increased gradually with the increase in substrate temperature. By the joint effect of more favourable mechanical interlocking, favourable van der Waals’ forces and probably the diffusion at interface due to the high substrate temperature, the adhesion between the disc shaped splat and substrate or the previously deposited splats is much stronger than the splash one. As the portion of disc shaped splat raised with the increase in substrate temperature, coating adhesion strength increased simultaneously. Accordingly, the coating properties could be predicted, and the spraying process could be optimised through controlling the splat formation process.
Footnotes
Acknowledgements
This research was supported both by the National Nature Science Foundation of China under contract 51301046 and by the National Key Basic Research Program of China under grant no. 2012CB625100.