Influence of air plasma spraying process parameters on the thermal barrier coating deposited with micro- and nanopowders

Inconel 625 Ni superalloy with the following chemical composition (wt%) was chosen as the base material: Al: 0.4; C: 0.1; Cr: 22; Co: 1; Fe: 5; Mn: 0.5; Mo: 9; Nb: 3.75; P: 0.015; Si: 0.5; S: 0.015; Ti: 0.4; Ni: balance. This material is characterized by good tensile strength (R_m = 900 MPa) and good heat resistance; therefore, Inconel 625 offers an optimal combination of corrosion resistance and creep, which enable its application in many industries, especially aeronautical technology [26]. In the research, flat samples with dimensions of 25 × 50 × 1 mm were used.

The TBC bond coat was made of a multicomponent NiCoCrAlY alloy plasma sprayed in the APS process on the Inconel 625 superalloy substrate. To deposit the bond coat, Amdry 386 powder (Oerlikon Metco, Wohlen, Switzerland) with the following chemical composition (wt%) was used: Co: 21.5; Cr: 16.3; Al: 14.8; Y: 0.1; Hf: 0.2; Ni: balance.

The outer ceramic layer of the TBC was also produced by APS using three types of Oerlikon Metco YSZ oxide powders. Metco-6609 powder with the smallest particle size D₁₀ = 0.1 μm, D₅₀ = 0.3–1 μm, and D₉₀ = 0.8–2 μm has so far been used mixed with ethanol in suspension in the SPS process [27, 28]. In this investigation, it was fed with a carrier gas in the form of a powder during the APS process. Metco-6700 powder with a particle size of D₁₀ = 2 μm, D₅₀ = 8 μm, and D₉₀ = 18 μm is developed for the deposition of the outer ceramic layer of the TBC with columnar structure in the PS-PVD process [29]. To perform a comparative analysis of the ceramic coatings deposited from micro- and nanopowders, a standard Metco-204NS powder with large particle sizes (D₁₀ = 20–28 μm, D₅₀ = 49–59 μm, and D₉₀ = 90–109 μm) was also examined [30].

The APS process was carried out using the thermal spraying equipment in the Research and Development Laboratory for Aerospace Materials at the Rzeszow University of Technology, Rzeszów, Poland. An A60 single-electrode torch (Thermico, Dortmund, Germany) with radial powder injection into the plasma plume was used. The powder was fed with a Thermico CPF2 feeder (Thermico, Dortmund, Germany). It allows gravimetric control of the flow rate of the powder independently of the grain size.

The data [31] and the results of previous research [25] clearly show that the properties of the deposited ceramic layer in the TBC are most influenced by the current and the chemical composition and flow rate of the plasma gas mixture, especially H₂. Therefore, these parameters were changed in the research.

Characterization of the temperature and velocity of the particles moving in the plasma stream was done using the DPV eVolution (Tecnar, Saint-Bruno-de-Montarville, Canada). Measurements were made at a distance of 100 mm from the plasma torch nozzle (the spraying distance used in the experimental processes). The DPV system enables the displacement of the measuring point in the x- and y-axes, perpendicular to the axis of the torch nozzle, within the range of ±50 mm. Thus, it is possible to measure the temperature and velocity of particles in the zone of the plasma plume with the highest flow rate. The distribution of temperature and particle velocity values on the cross section of the plasma plume with dimensions of 20 × 20 mm for 5 × 5 measurement points, at 4-mm intervals, was also determined. For each point, measurements were made of a fixed number of particles moving in the plasma stream (n=400) or for a duration of 7 s, whichever occurs earlier.

The characterization of the powder particles moving in the plasma plume using a DPV system was the basis for determining the range of conditions for ceramic layer deposition in the APS process using powders of different grain sizes, as provided in Table 1.

Microscopic examination of the fabricated TBC was performed using a Phenom XL scanning electron microscope (SEM) (Thermo Fisher Scientific, Waltham, MA USA) to determine the thickness and porosity of the outer ceramic layer. An electron beam acceleration voltage of 15 kV and a backscattered electron (BSE) mode were used. The samples were cleaned in isopropyl alcohol solution using an ultrasonic cleaner.

The porosity of the ceramic layer was determined by digital image analysis. For a single layer, 15 images were taken at a magnification of 9,000× with a size of 1,024 × 1,024 pixels. The layered microstructure was binarized using Nikon NIS-Elements (Nikon Instruments Inc., Melville, NY, USA) image analysis software, as presented in Figure 1.

Fig. 1

The hardness measurement was carried out using the Vickers method with the Innovatest Nexus 4303 (Innovatest, Maastricht, Netherlands) micro-hardness tester. A measured force load of 200 gf and a spraying time of 10 s were selected to measure the coatings of various thicknesses. For each plasma-spraying condition, five measurements were taken for the deposited coating.

Measurements of the average temperature and velocity of the ZrO₂ × 8Y₂O₃ oxide powders moving in the plasma stream were carried out using the DPV measuring equipment. It enables one to establish how the conditions of the APS process influence the properties of the molten particles and, as a result, the deposited coating. As shown in Figure 2, the mean values of the average temperature and velocity of the melted particles have a linear dependence on the flow rate of H₂ for their radial feed of the Metco-204NS powder into the plasma stream.

Fig. 2

Increased plasma energy at a flow rate of H₂ > 6 NLPM (normal liter per minute) was found to cause the melting of powder particles. The measurement results also show that as the flow rate of H₂ increases, the mean particle velocity decreases because a higher energy of the plasma stream enables the melting of particles with larger sizes that are characterized by lower velocity. However, the lack of H₂ in the plasma gas mixture leads to a lower temperature. Thus, it is impossible to deposit the ceramic layer with good properties. As a result, it is possible to exclude these processes in experimental APS processes.

The results show that the average temperature of the Metco-6700 powder particles that radially feed into the plasma stream in the A60 plasma torch depends more on the current than the flow rate of H₂, as shown in Figure 3A. Increasing the flow rate of H₂ has little or no effect on the temperature of the melted particles. This effect can occur due to the small difference in the mean particle diameter of the Metco-6700 powder and sufficient plasma stream energy to melt them at the plasma gas flow rate of H₂ ≥ 6 NLPM. Higher H₂ flow rates could cause difficulties with the radial feed method of highly dispersed powder particles injected into the plasma plume zone with the highest temperature, due to increased plasma enthalpy [32]. In Figure 3B, a linear dependence of the melting particle velocity on the H₂ flow rate was also found in the range of 6–18 NLPM. What is noticeable here is that both the highest temperature (approximately 2,900°C) and the highest particle displacement velocity (370 m/s) were obtained for the H₂ flow rate of 6 NLPM and the current intensity of 800 A. Similarly for Metco-204NS powder measurement, the lack of H₂ in the plasma gas mixture resulted in temperature values that were too low to deposit a ceramic coating with good physical and functional properties. Consequently, for the Metco-6700 powder, these processes were excluded in the experimental tests.

Fig. 3

In Figure 4, the comparison of example DPV measurement results can be seen for standard ceramic powder Metco-204NS and micropowder Metco-6700 with H₂ flow rate accordingly of 18 NLPM and 6 NLPM and the current intensity of 800 A (for both powders), which results in the highest-obtained temperature. The temperature was the highest (>2,900°C) in the center of the plasma plume, which proves that Metco-6700 powder particles are transported and melted in the center of the plasma plume despite their average size (approximately D₅₀ = 10–12 μm). The results also show that for Metco-6700 ceramic powder, the velocity was almost 100 m/s higher than for the standard powder, indicating a possible lower density of the deposited coatings.

Fig. 4

In the investigation, an attempt was also made to measure the melted powder particles with the smallest diameter D₅₀ <1 μm (Metco-6609) using the radial feed method of injecting particles into the plasma stream in the A60 plasma torch. It was found that the particle diameter is below the detection threshold of the DPV eVolution measuring system. Therefore, it is not possible to measure the temperature and velocity of these powder particles. Thus, powders must have a particle size >4 μm.

The main objective of this research was to establish the influence of APS process parameters on the thickness and porosity of ceramic coatings. To evaluate the properties of the coatings deposited with micro and nanopowders (Metco-6700 and Metco-6609), a comparison with the standard Metco-204NS YSZ powder is required.

Figure 5A shows a strong correlation between the thickness of the ceramic coatings deposited with the Metco-204NS powder and the increase in H₂ flow rate. The thickest ceramic coatings (100–120 μm) were obtained only for the flow rate of H₂ ≥ 12 NLPM and current >700 A. Thermal energy has been shown to be sufficient to melt powder particles with average diameter D₅₀ = 49–59 μm. These results are consistent with the DPV measurements shown in Figure 3A. The highest temperature (2,900°C) results in the thickest coating. As seen in Figure 5B, the porosity of the coatings depends more on the current than on the H₂ flow rate. As visible in Figure 5C, the hardness of the obtained coatings slightly increases with the H₂ flow rate from 500 HV0.2 at 6 NLPM to 570 HV0.2 at 18 NLPM, which may be caused by the increase in the thickness of the coating. An exception is for the process with the lowest current (600 A) and H₂ flow rate (6 NLPM) when the coating had only 320 HV0.2, whereby the low temperature and velocity of the particle and the low thickness of the coating result in weak bond strength. The application of high current (800 A) and H₂ flow rates (12 NLPM) simultaneously leads to the lowest porosity (10%) of the deposited coatings. In Figures 6A and 6B, we see that the outer ceramic layer of the TBC produced from the largest particle-diameter powder, Metco-204NS, introduced radially into the plasma plume in the A60 torch, is characterized by a lamellar structure, typical of plasma-sprayed coatings, with visible spherical pores, interlayer porosity, and microcracks resulting from thermal stress relaxation. The surface of the outer layer of the ceramic heat barrier coating exhibits the morphology consisting of (i) areas with well-melted particles and (ii) areas with recrystallized small-diameter particles with a characteristic spherical shape, which is visible in Figure 6C.

Fig. 5

Fig. 6

Until now, in the aerospace industry, ZrO₂ × 8Y₂O₃ oxide powder with a smaller average particle diameter of D₅₀ = 10–12 μm (Metco-6700) has been used mainly in PS-PVD processes [33]. The results show that it is possible to deposit a ceramic coating with the radial powder feed method in the plasma stream (using an A60 single-electrode plasma torch) using a powder with small particles, Metco-6700. As can be seen in Figure 7A, the ceramic coatings with the highest thicknesses of 82 μm and 92 μm were obtained for the process with a current of 700 A and 800 A, respectively, and flow rate of H₂ 6 NLPM, which is consistent with the highest temperature and velocity measured on the DPV system for this set of parameters. For the other APS process parameters, the thickness of the deposited coatings was only 50–64 μm. These process parameters also result in the smallest porosity of the coating (<8%), which can be seen in Figure 7B. Furthermore, it was found that the high H₂ flow rate (12 NLPM), regardless of the current (in the examined range: 600–800 A), leads to problems with the radial introduction of the powder with a large particle dispersion, Metco-6700, into the highest temperature zone of the plasma jet. Therefore, the ceramic layers for these process conditions were characterized by a small thickness (50–64 μm) and high porosity (9%–12%). Furthermore, it also affects the hardness of the ceramic coatings, which is shown in Figure 7C. The thickest coating with the smallest porosity obtained at 6 NLPM H₂ flow rate and 800 A results in the highest hardness of the coating (542 HV0.2) – similar to the hardness of the ceramic coatings deposited with the standard Metco-204NS powder, discussed above. Increasing the H₂ flow rate to 12 NLPM caused lower hardness values (400–470 HV0.2), which could be caused by lower coating thickness (50–64 μm). For this reason, the radial feed method should be used for a powder with micrometric particle sizes, such as Metco-6700, with a high current and a low flow rate of H₂ of 6 NLPM. It can be seen in Figures 8A and 8B that the use of the smaller particle-diameter powder, Metco-6700, leads to the formation of a lamellar-structured layer (similar to that of standard powder), but with visibly smaller pore sizes and a noticeably larger number of pores in the areas of the interlayer boundaries formed by the YSZ powder particles. The use of Metco-6700 in the APS process also increases the area with visible recrystallized particles compared to areas with well-melted particles, which is evident in the surface morphology shown in Figure 8C.

Fig. 7

Fig. 8

The Metco-6609 YSZ ceramic powder is generally used mixed with ethanol in suspension plasma spray [27]. In the present research, this powder was fed into a plasma plume using a carrier gas. Due to the fact that the particle diameter is below the detection threshold of the DPV eVolution measuring system, the influence of APS process parameters on thickness and porosity was studied only by metallographic examination.

The layers produced were <40 μm thick irrespective of the process parameters, as shown in Figure 9A. This phenomenon is likely caused by the large dispersion of Metco-6609 powder particles (D₅₀ <1 μm) and their low weight, which renders them difficult to be feed into the plasma plume zone with the highest temperature. Figure 9B shows that the deposited layers are also characterized by high porosity (16%–40%). One possible explanation for such high porosity is that a large number of small particles of this powder melt and evaporate in a short time in the plasma stream and recrystallize before reaching the surface of the substrate, or the particles are moving in the outer zone of the plasma plume where there is a lower temperature; the particles in this zone of the plasma are not heated to the melting point of the ZrO₂ × 8Y₂O₃ oxide. Either way, small spherical particles are visible in the microstructure in Figures 10A and 10B. Similar effects have been observed in the PS-PVD process [34]. The large spray distance in this process leads to the recrystallization of the oxide vapors of ZrO₂ × 8Y₂O₃ before they reach the surface of the substrate. The deposited ceramic layer of the TBC is then characterized by a quasi-columnar structure: between the columnar grains, there are recrystallized spherical ZrO₂ × 8Y₂O₃ oxide particles. Hence, the deposited coating is characterized by low hardness (<300 HV0.2), as shown in Figure 9C. Due to the differences in porosity between coatings caused by different volumes of well-melted particles and small spherical particles (e.g., visible in Figure 10B), the value of hardness is in the range of 150–300 HV0.2. Therefore, the deposition of nanosized ceramic powder using a carrier gas and the radial feed method leads to a microstructure that is not very promising, compared to the YSZ columnar structure coating produced by SPS using Metco-6608 (a suspension version of Metco-6609 powder) [27, 35].

Fig. 9

Fig. 10

The lowest porosity and the highest hardness of the examined APS process parameters examined were found in the ceramic layer deposit with 12 NLPM H₂ flow and 600 A current, making them the most suitable values when using radial powder fed into the plasma plume. As with Metco-6700 powder, the use of nanometric Metco-6609 powder fed radially into the plasma plume increases the proportion in surface morphology with visible recrystallized particles compared to well-melted areas, as seen in Figure 10C.

In summary, results show that the use of the same ceramic material (YSZ) causes a different lamellar structure depending on the size of the powder particles. The Metco-6700 micrometric powder resulted in a less-porous coating with significantly smaller pore sizes and a higher number in the interlayer boundaries compared to coatings deposited with standard Metco-204NS powder with the same thickness (approximately 100 μm) and only for processes with a low H₂ flow rate ≤ 6 NLMP. The coatings obtained by such a process were characterized by a similar hardness >500 HV0.2. The use of Metco-6609 nanometric powder resulted in a coating with spherical particles and remelted areas; therefore, the porosity of these coatings was higher than that of the other powders used in the study. The coatings were characterized by hardness about twice smaller than that of the Metco-204 and Metco-6700 coatings. Such a result indicates that this powder is not suitable for spraying in the APS process with its radial feed method. In further research, the axial introduction of this powder in the Axial III multielectrode torch will be studied, where problems with introducing a powder particle into the plasma plume can be avoided, and the properties of the coatings, e.g., resistance to oxidation, thermal properties, and bond nature of the coatings, can be tested.