2020 Volume 61 Issue 7 Pages 1390-1395
It is predicted that the operating temperatures of next-generation aircraft engines will exceed the present critical temperatures of any of the conventional metallic materials (<1400°C). As a result, it is likely that the application of SiC/SiC ceramic matrix composites (CMC) as primary materials. When exposed to an oxidative environment, SiC forms a protective silica scale. This layer provides additional protection from problems that may occur upon oxidation. It will, however, react with the water vapor formed during the combustion process, creating gaseous silicon hydroxides that will reduce its thickness. This problem has hindered the practical realization of CMC engines in aircrafts. Consequently, environmental barrier coatings (EBCs) are required to protect CMC components from oxidative degradation and thus ensure the reliability of CMC engines. In this study, a new deposition process, namely, the suspension plasma spray (SPS) process, is proposed to form the EBC. It produces much denser coatings by feeding a suspension of particles that are a single micron in size. It was confirmed that the coating structure and composition resulting from the SPS process were largely influenced by the residence time of the suspensions in the plasma flame. Subsequently, the optimum spray conditions were examined and discussed.
This Paper was Originally Published in Japanese in Japan Thermal Spray Soc. 56 (2019) 2–7.
Due to the global increase in aviation demand, there is a significant incentive to lessen the fuel consumption of aircraft engines, achieving energy savings, and reducing their environmental load. In this respect, the reduction of the weight of the aircraft fuselage the weight of the high-temperature engines, is a key objective. Ni-base superalloys are commonly used materials for turbine blades, which are high-temperature components within these engines. Highly efficient next-generation aircraft engines, however, comprise blades with densities approximately three times higher than the traditionally used metal materials. SiC ceramic matrix composites (CMCs) are one such promising next-generation material with high heat resistance.1) Practical application of these CMCs, however, is challenged by the reduction in thickness that occurs due to oxidation under a high-temperature steam environment. To protect these components from this deterioration and damage, and ensure their reliability, it is necessary to apply environmental barrier coatings (EBCs). These EBCs are generally composed of multiple layers of ceramic coatings.2–4) They must be stable and dense in a high-temperature steam environment and provide improved environmental shielding. Furthermore, electron beam physical vapor deposition (EB-PVD) methods, which may be used to apply these coatings, experience problems in terms of both their coating formation speed and their coating density.
In this study, suspension plasma spraying (SPS)5,6) has been proposed as a new EBC-deposition method. Using this method, the powder material is applied as a suspension, making it possible to apply particles of the order of a single µm, unattainable by conventional plasma spraying due to its rheology.7) Because fine particles are supplied by the SPS method, the powder material can be melted easily, reducing both the pore size and microcracks, and forming a dense coating.8) Additionally, this method allows for the formation of a columnar structure, which presents the potential to reduce thermal stress resulting from the usage environment by tuning the coating-formation conditions.9) As CMCs are expected to be used at a higher temperature in future applications, it is likely that the EBC layer will require a thermal barrier.10) Therefore, it is beneficial that the SPS method can form both dense EBC layers and columnar TBC layers. It is also desirable that it can form such coatings with varying functions in one single process. While there have been many studies on the application of the SPS method to obtain TBCs, there are limited research results regarding the SPS method in the formation of EBCs. In this study, the applicability of Yb2Si2O7,11) which has been proposed as a topcoat material in EBCs in recent years, has been examined. Both SPS and APS methods have been used to form coatings and the variations in structure and composition have been evaluated. Furthermore, for the SPS method, the impact of the spraying parameters on the coating structure and composition has been evaluated. Finally, based on the results obtained in this study, a pathway for future improvement of the SPS EBC method has been proposed.
In this study, a 2 mm thick Aluminum plate and SiC sintered body were prepared as the substrate materials. The Aluminum plate was used in as-sprayed coating observations, while the SiC sintered body was used in high-temperature oxidation testing. In this high-temperature oxidation test, a Silicon bond coat, applied by the APS method, was introduced to improve adhesion. Yb2Si2O7 powder, prepared via granulation sintering, was used as the feed-stock material. The average powder particle size was 3.40 µm for the SPS method and 35.2 µm in the case of the APS method. Figure 1 presents an SEM image of the powder used in the study. When using the SPS method, the suspension concentration was 10 wt.%, with ethanol used as the suspension dispersion medium. The coating was applied to the substrate, fixed to the jig, using the Praxair (SG-100) and Progressive Surface (100HE) spray guns for the APS and SPS methods, respectively. The APS and SPS construction conditions are detailed in Tables 1 and 2, respectively. The output of the APS-method spray gun was controlled by the current, and Ar and He gases were used. Conversely, Ar, N2, and H2 gases were used in the SPS process.
SEM images of the feedstock Yb2Si2O7 powder: (a) APS and (b) SPS.
Each deposited specimen was subjected to a high-temperature oxidation test, under atmospheric pressure, using a high-temperature electric furnace (Nitto Kagaku Co. Ltd., NHK-170AF). The testing temperature was 1300°C, and the test time was varied from 0 to 100 h. After the test, the furnace was left to cool to room temperature, at which point the specimen was removed.
2.3 Measurement of in-flight particle velocity and temperatureThe velocity and temperature of the in-flight particles were measured to clarify their relationship with various SPS method spray parameters. To obtain these measurements, a thermal spray process measuring system (TECNAR, Accuraspray-G3C) was used. This system involves detection of light emissions, caused by powder passing through the measurement area, by an optical sensor comprising two optical fibers in the center head and a CCD camera. The average particle temperature was calculated using a two-color thermometer by quantifying the phase shift of the detected light emission waveform.
2.4 Coating evaluation 2.4.1 X-ray fluorescence (XRF) analysisAs depicted in Fig. 2, SPS spraying was performed over a two-minute period in a water-filled stainless container, with the subsequent precipitated particles collected. X-ray fluorescence (XRF) analysis was carried out to evaluate the material composition before and after thermal spraying by means of Shimadzu XRF-1700.
Technique for collection of flying particles.
The effect of the different spray parameters and influencing factors on the degree of SiO2 volatilization was discussed. Due to operational limitations, the distance from the spray gun to the water surface, for the purpose of particle recovery, was 700 mm. The actual spray distances for coating sample preparation were 80–120 mm and 76 mm for the SPS and APS methods, respectively. This resulted in a minor discrepancy and a subsequent variation in the degree of volatilization.
2.4.2 X-ray diffraction (XRD) analysisX-ray diffraction (XRD) was performed on the feed-stock powder and on each coating using Shimadzu Maxima XRD-7000. The material composition and the crystallinity before and after thermal spraying were evaluated, and the impact of the key spray parameters and influencing factors on these characteristics was discussed.
2.5 Cross-sectional evaluation of the coatingsEach coating was cut and embedded with resin. The cross-sectional area was mirror-finished by wet polishing using a diamond disk, followed by buffing with a single crystal diamond suspension (9 µm and 3 µm), and an alumina liquid abrasive. SEM images were then captured and EDX analysis performed on the finished cross-section.
The XRF measurement results for each specimen are presented in Fig. 3. It was confirmed that the proportion of the SiO2 component within the collected powder was smaller in all of the prepared coatings than in that of the as-supplied powder. This decrease in SiO2 content was also noted to be significantly greater in the SPS coatings when compared to the APS coatings. It has been reported previously that in the case of Rare earth Element (RE) silicates containing Yb2Si2O7, the partial pressures of SiO and SiO2 in the material are significantly higher than those of the other components at 2000°C or higher than these silicate melting point of each material.12) This decrease in the SiO2 component ratio suggests that the SiO2 component in Yb2Si2O7 will volatilize when the flying particles are exposed to a high-temperature plasma flame that exceeds the melting point of the material. A volatilization phenomenon such as this may occur remarkably near the powder surface, where the temperature is higher. The average particle size of the powder used in this study was approximately 30 µm for APS and 3 µm for SPS. This phenomenon was noticeable in the SPS coatings with a larger specific surface area than that of the supplied powder. In the case of the SPS coatings, there was a correlation between the gas flow rate and the ratio of the contained SiO2 components. It was observed that a lower gas flow rate led to a greater degree of SiO2 volatilization.
XRF analyses results for feedstock powders and coatings.
The relationship between the gas flow rate and the particle velocity, measured during coating formation, is presented in Fig. 4, while the relationship between the gas flow rate and the particle temperature is shown in Fig. 5.
Relationship between the gas flow rate and particle velocity.
Relationship between the gas flow rate and particle temperature.
Through examination of Figs. 4 and 5, it was estimated that the gas flow rate and the in-flight particle velocity/temperature were proportional to one another. Regarding the in-flight particle temperature, when the gas flow rate was increased from 200 to 320 SLM (Volume flow rate L/min in 0°C, 101.325 kPa), the temperature rose from 3600 to 4000°C, which is approximately twice the melting point of Yb2Si2O7 (1850°C).13) This suggests that a change in temperature in this high-temperature environment does not significantly impact the volatility of SiO2. On the other hand, the particle velocity was observed to increase from 700 to 900 m/s with the previously described change in gas flow rate. In this instance, the distance from the gun to the substrate was set at approximately 80 mm, implying that, when the particle velocity rose from 700 to 900 m/s, the particle residence time in the gas flow decreased by approximately 30%. For this reason, the change in particle velocity, caused by altering the gas flow rate, has a significant effect on the volatilization degree of SiO2. With a lower gas flow rate, the residence time of the flying particles in the high-temperature plasma flame was extended. The resulting in-flight particle velocity was then reduced and an increased amount of SiO2 was volatilized. Based on these findings, it is assumed that a larger extent of SiO2 volatilization occurs close to the central location of the plasma generation, where the temperature gradually decreases as the particles move further away.14)
3.1.2 X-ray diffractionFigure 6 presents the XRD results from the as-sprayed coating. It was observed that this coating has a broader peak on the low-angle side when compared to the supplied powder for both the SPS and APS coatings. Furthermore, almost no Yb2Si2O7 peak was observed. It was concluded that the as-sprayed coating was primarily amorphous material. The SPS coating was found to be slightly more crystalline than the APS coating, with this variation in crystallinity thought to be largely due to the difference in particle size. The smaller particle size in the SPS coating allows for easier heat transmission, which causes the crystallization of these particles to progress further than those in the APS method.
XRD patterns of the as-sprayed coatings.
Figure 7 depicts the XRD results obtained after the high-temperature oxidation test. It was confirmed that, after this test, the broad amorphous peak on the low-angle side disappeared and crystallization progressed in both the SPS and APS coatings. It was also confirmed that the coating contained Yb2Si2O7 and Yb2SiO5, suggesting that, due to SiO2 volatilization, an amount of Yb2Si2O7 became Yb2SiO5. When contrasting the peak intensities of two coatings, the composition ratio was higher of Yb2SiO5 in the SPS coating and of Yb2Si2O in the APS coating. This may be due to the difference in the SiO2 component volatilization between the two processes, as confirmed by the XRF analysis results described earlier. By varying gas flow rates in the SPS method, the Yb2Si2O7 peaks disappeared under a flow rate of 200 SLM, with Yb2O3 peaks confirmed around 29°, 34°, 49° and 58° of Fig. 7.
XRD patterns of the coatings after 100 h at 1300°C.
Therefore, the coating composition after the high-temperature oxidation test depends on the volatilization amount of the SiO2 component during the coating formation. This is considered to correlate with the residence time of the particles within the plasma flame.
It is predicted that a coating with a high ratio of Yb2Si2O7 can be achieved when the coating is formed under a proportionately higher gas flow rate.
3.2 Cross-sectional SEM observation and EDX analysisFigure 8 shows the cross-sectional backscattered electron image of the as-sprayed coating. It was observed that the SPS coating has a significantly smaller pore size, a denser structure, and a smaller number of microcracks than the APS coating. In terms of the uniformity of the coating structure, it was observed that both coatings exhibited multiple regions with different contrasts. The results of the EDX analysis on the coating cross-section are presented in Fig. 9. The increase and decrease of the signal intensity of Si and Yb elements in both coatings were confirmed by the contrast brightness. It is considered that the dark region, with the high signal intensity of Si, is a region close to Yb2Si2O7, and the bright region is a region with components close to Yb2SiO5.
Cross-sectional BSE images of the as-sprayed coatings: (a) APS; (b) SPS 200 SLM; (c) SPS 260 SLM; and (d) SPS 320 SLM.
EDX line analysis of the as-sprayed coatings: (a) APS and (b) SPS 260 SLM.
Figure 10 shows the cross-sectional backscattered electron image of the coating after the high-temperature oxidation test. It was confirmed that, after this test, the coating structure changed substantially when compared to the as-sprayed coating. In the APS coating, a lamellar structure consisting of bright and dark areas was confirmed. From EDX analysis, shown in Fig. 11, it is considered that the lamellar structure is composed of Yb2SiO5 and Yb2Si2O7, since bright and dark regions of the contrast are composed of components close to Yb2SiO5 and Yb2Si2O7, respectively. This phase is produced under equilibrium cooling conditions. Due to the rapid solidification of the droplets, the metastable phase was observed in the as-sprayed coating. However, it is assumed that such microstructural changes occurred after the high-temperature oxidation test, because the equilibrium cooling condition was achieved by furnace cooling.
Cross-sectional BSE images of the coatings after 100 h at 1300°C: (a) APS; (b) SPS 200 SLM; (c) SPS 260 SLM; and (d) SPS 320 SLM.
EDX line analysis of the coatings after 100 h at 1300°C: (a) APS and (b) SPS 260 SLM.
Figure 1215) displays the Yb2O3–SiO2 binary system phase diagram. When the content of SiO2 is 50 mol% or less, the phase is Yb2O3 + Yb2SiO5. However, XRD results (Figs. 6 and 7) showed that Yb2SiO5 + Yb2Si2O7 phase is main component in coatings with gas flow rates of 320 SLM and 260 SLM. In these cases, the SiO2 component ratio is higher than it is at 200 SLM. Even though the same feedstock powder was used, different phases were obtained by changing the gas flow rate. It can be observed from the phase diagram that the SiO2 component ratio increased after the high-temperature oxidation test. The reason for this is unclear; however, it is possible that Si, used as a bond coat, has diffused into the coating by heat treatment. This is a topic for further study.
Ytterbia-silica pseudo-binary phase diagram [11].
In addition, the SPS coating presented regions with different contrasts, and the lamellar structure seen in the APS coating was not observed. The disparities in structure are thought to be due to the varying particle size. Further investigation is required to confirm this hypothesis. Under different gas flow rates, the SPS coating exhibited structural changes that were different to those in the APS coating. This was observed under all conditions, and it can be seen that the spray conditions differ. Based on the EDX analysis and XRD results, as depicted in Figs. 6, 7, and 11, it was inferred that the bright region is Yb2O3, the middle region is Yb2SiO5, and the dark region is Yb2Si2O7. Image processing was used to calculate the area ratios of the three contrasted regions. These results are presented in Fig. 13. It was noted that the ratio of each region varies with the gas flow rate. Yb2O3 + Yb2SiO5 were identified at a gas flow rate of 200 SLM, Yb2O3, Yb2SiO5 + Yb2Si2O7 at 260 SLM, and Yb2SiO5 + Yb2Si2O7 at 320 SLM. This agrees with the XRF and XRD results, and is presumed to be due to the variation in SiO2 volatilization. At a gas flow rate of 200 SLM, a significant amount of SiO2 has volatilized, with only Yb2O3 + Yb2SiO5 remaining. Analyzing the results described above, a gas flow rate of 320 SLM appears to be the most appropriate to suppress the volatilization of SiO2 and form the most dense coating possible. Previous research has suggested that this should comprise the composite structure of Yb2SiO5 + Yb2Si2O7. While several oxidation examples do exist,16) it was found to be difficult to form a coating solely composed of Yb2Si2O7 by controlling the gas flow rate, as was done in this study.
Area ratio of each of the SPS coatings.
However, in the case of the Yb2Si2O7 coating sprayed at a gas flow rate of 260 SLM, the area ratio of Yb2SiO5 calculated after the high-temperature oxidation test is approximately 85%; almost single-phase Yb2SiO5. This is highest proportion of Yb2SiO5 within any coating, deposited using Yb2SiO5 powder, in any study that we have conducted thus far. XRF analysis showed that the weight percentages of the SiO2 and Yb2SiO5 components within this coating were 14.87 and 13.03 wt.%, respectively. It is therefore predicted that it is possible to compensate for the volatilized SiO2 lost during construction, by using Yb2Si2O7 powder with a more significant proportion of SiO2 component than Yb2SiO5. In doing so, the effectiveness of the SiO2-rich suspension can be increased. It is predicted that this suspension could be applied to other silicate materials, and that utilizing the SPS method to deposit it can form a dense coating with a uniform composition.
The purpose of this study was to examine the applicability of EBCs deposited via the SPS process. Using Yb2Si2O7, a common EBC topcoat material in recent years, coatings were prepared by the SPS and APS methods. The variations in coating structure and composition as a result of different thermal spraying processes were evaluated. The SPS process was then examined further, and the effect of the spray parameters on the coating structure and composition was evaluated. The control factors were estimated and the ideal deposition conditions for EBC coatings were investigated. The key findings are as follows:
1. Comparison with APS coatings
The XRF and XRD analysis of the coating composition confirmed that the SiO2 component was significantly volatilized in the case of the SPS coating. This was due to the small feed-stock powder size and the large total surface area.
From the cross-sectional SEM images and EDX analysis, it was confirmed that the SPS coating was dense, and the structure change behavior, upon high-temperature oxidation, varied between the different processes.
2. Effect of spray parameters
By means of in-flight particle velocity measurement and composition analysis, using XRF and XRD, a correlation was observed between the in-flight particle velocity and the volatilization volume of SiO2. The degree of volatilization increased with slower in-flight velocity and longer exposure time in the plasma flame.
Through examination of the cross-sectional SEM images and EDX analysis, a correlation was found between the coating density and the gas flow rate. When a Yb2Si2O7 suspension was used to deposit a coating under appropriate conditions, the coating comprised a single phase of Yb2SiO5. The results suggest that the volatilized SiO2 component has been supplemented through incorporation of a powder with a more significant proportion of the SiO2 component.
This work was supported by the Japan Society for the Promotion of Science (JSPS) Core-to-Core Program A. Advanced Research Networks “International Research Center for Intelligent Layer Materials and Layer Structures for Energy Saving,” JSPS KAKENHI Grant Numbers JP16K14109, and the EBC working group of the Japan Thermal Spray Society.
