ZrSi 2 -SiC/SiC Gradient Coating of Micro-Structure and Anti-Oxidation Property on C/C Composites Prepared by SAPS

: A ZrSi 2 -SiC/SiC gradient coating system was designed to reduce the thermal stress of anti-oxidation coatings for C/C composites and prolong their anti-oxidation time at a high temperature. The SiC transition layer was prepared by pack cementation and the gradient ZrSi 2 -SiC outer coatings with different ZrSi 2 contents added were deposited by supersonic air plasma spraying (SAPS). The micro-morphologies and phase compositions of the coatings were studied by SEM, EDS, XRD, and TG/DSC, and their anti-oxidation performances were tested by a static oxidation experiment. The ﬁndings suggested that the gradient coating with 30 wt.% ZrSi 2 content displayed the optimum and dense microstructure without obvious pores and microcracks, compared with the other three proportional coatings. During the oxidation test, because of the oxidation reaction of ZrSi 2 and SiC phases, a large amount of silica was formed in the coating to ﬁll the pores and microcracks and densify the coating further. Oxidation products ZrO 2 and ZrSiO 4 , having a high melting point and outstanding anti-oxidation property, were embedded in the SiO 2 glass layer to reduce the layer volatilization rate and improve the ability to block oxygen. Therefore, the specimen with 30 wt.% ZrSi 2 still kept mass gain after 188 h oxidation time at 1500 ◦ C. However, when the oxidation time was increased to 198 h, it had a mass loss of 0.1%, because the coating compactness was destroyed by the escape of the oxidation gases.


Introduction
Outer coating technology is an effective method to improve the oxidation resistance of C/C composites in a high temperature aerobic environment in air [1][2][3]. The core of the technology is to choose reasonable coating materials, which generally must have excellent anti-oxidation property, mechanical capacity, and low oxygen permeability at a high temperature [4,5]. Among them, SiC ceramic has been successfully applied to the anti-oxidation coating for C/C composites, because it can not only meet the abovestated requirements, but also generate dense oxidation film SiO 2 at middle and high temperatures on the coating surface [6][7][8]. SiO 2 film could effectively prevent oxygen from entering C/C composites for a long time and then prolong their high-temperature service life [9][10][11]. However, with the rapid development of aerospace technology, an SiC single coating has not been able to protect the C/C composites from being oxidized in extreme environments [12][13][14]. In order to further improve the oxidation resistance of the C/C composites, some materials possessing low volatility, high melting point, and thermal stability, such as ZrB 2 [15], ZrC [16], TaC [17], YSZ [18], and so on, are added into the SiC coating to optimize its microstructure and enhance the interfacial bonding strength between anti-oxidation coatings and the C/C composites. Related studies show that the into a centrifugal spray dryer to form ZrSi 2 -SiC particles with fine fluidity for spraying. The related plasma spraying parameters are listed in Table 1 and [29] can provide a more detailed preparation process. Table 1. Spraying parameters of the ZrSi 2 -SiC coatings prepared by SAPS.

Anti-Oxidation Property Test
The oxidation resistance experiments were carried out in an electrical furnace in air at 1500 • C. Before the test, the furnace was heated up to 1500 • C until the temperature remained stable. The specimens were divided into four groups and placed into the furnace respectively according to the different proportion of ZrSi 2 . After a given time, they were taken out directly and cooled from 1500 • C to RT in air. Meanwhile, the specimens also underwent thermal shock resistance evaluation. Their mass changes were measured by an electronic balance with a sensitivity of ±0.1 mg. The mass loss and oxidation time curves of the four kinds of specimens were used to evaluate their anti-oxidation behaviors. The weight loss rates could be calculated according to Formula (1): ∆w = (m 0 − m 1 )/m 0 (1) where ∆w is the oxidation weight loss rate of the specimen; m 0 is the mass of the specimen before oxidation; and m 1 is the mass of the specimen after oxidation.

Microstructure and Phase Composition Analysis
The crystalline structures of specimens were obtained by X-ray diffraction (XRD, Bruker D2 Phaser, BRUKER, Karlsruhe, Germany with Cu Kα radiation in the range of 0-90 • [30]. A scanning electron microscope (SEM, JSM-6460, JEOL Ltd., Mitaka, Japan) equipped with energy dispersive spectroscopy (EDS) was used to study the morphologies and element distributions of the specimens [31]. The thermogravimetric (TG, METTLER, Zurich, Switzerland) test was carried out on a Metter Toledo Star TGA/SDTA 851 thermal analyzer in air from room temperature to 1500 • C with a heating rate of 10 K/min.

Results and Discussion
The XRD pattern of the 20 wt.% ZrSi 2 -80 wt.% SiC mixed powders is shown in Figure 1. The result indicates that the powders did not have any other impurity. Figure 2 shows the SEM images and EDS results of the 20 wt.% ZrSi 2 -80 wt.% SiC particles prepared by the centrifugal spray dryer. The particles with a spherical shape with good mobility were able to be uniformly melted during the spraying process to form a denser coating on the SiC surface. Particles included mainly C, Zr, and Si elements, and were consistent with preformed coating components. O element was incorporated into the particles in the process of centrifugal spray granulation.    Figure 3a-d, the coatings changed from porous to dense and then to porous, accompanied by the change in visible pores. The porosities of the ZrSi2-SiC outer coating with different ZrSi2 contents were analyzed by means of Image-Pro Plus software. Their porosities were 18.9%, 11.7%, 10.6%, and 15.4%, respectively. In Figure 3a, some big grooves can be seen in the coating surface, indicating that most of the flying ZrSi2-SiC mixed particles were blown away by plasma jet and not deposited onto the substrate. The main reason was that, because the density of SiC was lower than that of ZrSi2 (ρ = 4.88 g·cm −3 ), the flying velocity of SiC particles was higher than that of ZrSi2 after ejecting from the nozzle. Therefore, the mixed particles with a greater SiC content did not melt sufficiently and produced a splash upon impact with the substrate surface [26]. With the increase in the ZrSi2 content in the ZrSi2-SiC particles, their flying velocity could decline and their deposition rate could     Figure 3a-d, the coatings changed from porous to dense and then to porous, accompanied by the change in visible pores. The porosities of the ZrSi2-SiC outer coating with different ZrSi2 contents were analyzed by means of Image-Pro Plus software. Their porosities were 18.9%, 11.7%, 10.6%, and 15.4%, respectively. In Figure 3a, some big grooves can be seen in the coating surface, indicating that most of the flying ZrSi2-SiC mixed particles were blown away by plasma jet and not deposited onto the substrate. The main reason was that, because the density of SiC was lower than that of ZrSi2 (ρ = 4.88 g·cm −3 ), the flying velocity of SiC particles was higher than that of ZrSi2 after ejecting from the nozzle. Therefore, the mixed particles with a greater SiC content did not melt sufficiently and produced a splash upon impact with the substrate surface [26]. With the increase in the ZrSi2 content in the ZrSi2-SiC particles, their flying velocity could decline and their deposition rate could   Figure 3a-d, the coatings changed from porous to dense and then to porous, accompanied by the change in visible pores. The porosities of the ZrSi 2 -SiC outer coating with different ZrSi 2 contents were analyzed by means of Image-Pro Plus software. Their porosities were 18.9%, 11.7%, 10.6%, and 15.4%, respectively. In Figure 3a, some big grooves can be seen in the coating surface, indicating that most of the flying ZrSi 2 -SiC mixed particles were blown away by plasma jet and not deposited onto the substrate. The main reason was that, because the density of SiC was lower than that of ZrSi 2 (ρ = 4.88 g·cm −3 ), the flying velocity of SiC particles was higher than that of ZrSi 2 after ejecting from the nozzle. Therefore, the mixed particles with a greater SiC content did not melt sufficiently and produced a splash upon impact with the substrate surface [26]. With the increase in the ZrSi 2 content in the ZrSi 2 -SiC particles, their flying velocity could decline and their deposition rate could increase. Therefore, the ZrSi 2 -SiC coating presented a dense and lamellar structure in Figure 3c. However, when the content of ZrSi 2 was increased to 40 wt.%, it could be excessively oxidized by oxygen to form more ZrO 2 and ZrSiO 4 compared with other three samples, according to the XRD results of Figure 4. These oxidation products were embedded in the coating in the form of granules, leading to the decrease in the interfacial compatibility between the coating and the products and the formation of more pores in the coating in Figure 3d. The related oxidation reactions are as follows: ZrO 2 (s) + SiO 2 (s) = ZrSiO 4 (s) Coatings 2022, 12, x FOR PEER REVIEW 5 of 12 increase. Therefore, the ZrSi2-SiC coating presented a dense and lamellar structure in Figure 3c. However, when the content of ZrSi2 was increased to 40 wt.%, it could be excessively oxidized by oxygen to form more ZrO2 and ZrSiO4 compared with other three samples, according to the XRD results of Figure 4. These oxidation products were embedded in the coating in the form of granules, leading to the decrease in the interfacial compatibility between the coating and the products and the formation of more pores in the coating in Figure 3d. The related oxidation reactions are as follows:   increase. Therefore, the ZrSi2-SiC coating presented a dense and lamellar structure in Figure 3c. However, when the content of ZrSi2 was increased to 40 wt.%, it could be excessively oxidized by oxygen to form more ZrO2 and ZrSiO4 compared with other three samples, according to the XRD results of Figure 4. These oxidation products were embedded in the coating in the form of granules, leading to the decrease in the interfacial compatibility between the coating and the products and the formation of more pores in the coating in Figure 3d. The related oxidation reactions are as follows:   In order to prove the above conclusions, Figure 5 shows the cross-section microtopographies of the four kinds of the ZrSi 2 -SiC coatings with different ZrSi 2 contents. In Figure 5a,b, it can be seen that the particle interfaces were clearly visible, proving that the Coatings 2022, 12, 1377 6 of 12 particles in these two coatings were not well melted. The widths of the thin areas in the two images were only about 2 µm and 4 µm, respectively. Fortunately, when the content of ZrSi 2 was up to 30 wt.%, the coating with an average width of 25 µm featured a lamellar and dense structure owing to fully melted ZrSi 2 -SiC particles. However, resulting from having more refractory oxidation products, the grainy structure appeared again in the coating, accompanied by decreases in the coating thickness, as shown in Figure 5d. In addition, obvious microcracks were not found in the four images, signaling that the gradient coating structure was helpful to form a thermal gradient in the coating to alleviate thermal stress coming from the cooling process from a high temperature to RT. The EDS results of line scanning for the whole cross-section of the sample in the Figure 6 agree with the XRD analysis in Figure 4. The main elements included C, O, Si, and Zr, with no other impurity. Figure 7a illustrates the static oxidation curves of the four samples at 1500 • C. It can be seen that, as the oxidation time increased, the curves of the four samples all sharply decreased at first, and then slowly increased. In the initial stage, named as the mass gain phase, when the whole coating surfaces were exposed to a high temperature oxidizing environment, they could quickly oxidize with oxygen to form many oxidation products. According to the XRD result of the coating with 30 wt.% ZrSi 2 content in Figure 7b, after 21 h oxidation at 1500 • C, it could be confirmed that the increasing mass of the four kinds of the samples was attributed to the formation of ZrO 2 , SiO 2 , and ZrSiO 4 , and the related reactions are (2), (3), (4), and (5). According to the previous relevant results, because of quickly being oxidized and generating Si, ZrO 2 , and SiO 2 , the mass growth rate of pure ZrSi 2 was up to near 45% at 900 • C, and it further increased owing to the formation of ZrSiO 4 when the oxidation temperature increased to 1420 • C [26]. That is to say, the greater the ZrSi 2 content in the coating, the more oxidation it produces. Therefore, as the oxidation time increased to 21 h, the mass gain of the specimen with 40 wt.% ZrSi 2 content was 2.34%, which was more than that of the other three specimens.
SiC (s) + 2O 2 (g) = SiO 2 (s) + CO 2 (g) (4) 2SiC (s) + 3O 2 (g) = 2SiO 2 (s) + 2CO (g) Coatings 2022, 12, x FOR PEER REVIEW 6 of 12 In order to prove the above conclusions, Figure 5 shows the cross-section micro-topographies of the four kinds of the ZrSi2-SiC coatings with different ZrSi2 contents. In Figure 5a,b, it can be seen that the particle interfaces were clearly visible, proving that the particles in these two coatings were not well melted. The widths of the thin areas in the two images were only about 2 μm and 4 μm, respectively. Fortunately, when the content of ZrSi2 was up to 30 wt.%, the coating with an average width of 25 μm featured a lamellar and dense structure owing to fully melted ZrSi2-SiC particles. However, resulting from having more refractory oxidation products, the grainy structure appeared again in the coating, accompanied by decreases in the coating thickness, as shown in Figure 5d. In addition, obvious microcracks were not found in the four images, signaling that the gradient coating structure was helpful to form a thermal gradient in the coating to alleviate thermal stress coming from the cooling process from a high temperature to RT. The EDS results of line scanning for the whole cross-section of the sample in the Figure 6 agree with the XRD analysis in Figure 4. The main elements included C, O, Si, and Zr, with no other impurity.  Figure 7a illustrates the static oxidation curves of the four samples at 1500 °C. It can be seen that, as the oxidation time increased, the curves of the four samples all sharply decreased at first, and then slowly increased. In the initial stage, named as the mass gain phase, when the whole coating surfaces were exposed to a high temperature oxidizing environment, they could quickly oxidize with oxygen to form many oxidation products. According to the XRD result of the coating with 30 wt.% ZrSi2 content in Figure 7b, after the ZrSi2 content in the coating, the more oxidation it produces. Therefore, as the oxidation time increased to 21 h, the mass gain of the specimen with 40 wt.% ZrSi2 content was 2.34%, which was more than that of the other three specimens.
SiC (s) + 2O2 (g) = SiO2 (s) + CO2 (g) 2SiC (s) + 3O2 (g) = 2SiO2 (s) + 2CO (g) (5)   the ZrSi2 content in the coating, the more oxidation it produces. Therefore, as the oxidation time increased to 21 h, the mass gain of the specimen with 40 wt.% ZrSi2 content was 2.34%, which was more than that of the other three specimens.
SiC (s) + 2O2 (g) = SiO2 (s) + CO2 (g) 2SiC (s) + 3O2 (g) = 2SiO2 (s) + 2CO (g) (5)   With the increase in oxidation time, the mass losses of all specimens remained negative for a while, among which the sample with 30 wt.% ZrSi 2 content maintained negative when the oxidation time reached 188 h. In order to explain this phenomenon, the surface morphology and elementary compositions of the sample with 30 wt.% ZrSi 2 content after 92 h oxidation were analyzed by SEM and EDS, and the results are shown in Figure 8. It can be seen that the smooth gray phase could properly cover the whole surface, fill the pores of the coating, and effectively prevent oxygen from entering the coating, which was the main reason to improve the oxidation resistance of the sample. The gray phase marked as Spot B was made of O; Si; and a small amount of Zr, Al, and Au elements, thus it could be concluded that the gray phase was mainly SiO 2 phase. Al from the SiC transition layer and Au from the SEM test could be ignored. The light gray phase marked as Spot A contained O, Zr, and Si elements, which had ZrO 2 and ZrSiO 4 , and a small amount of SiO 2 covering ZrO 2 . Therefore, first, the SiO 2 phase could dense the outer coating to keep oxygen out by means of its good fluidity and low viscosity at 1500 • C; second, although SiO 2 had a certain volatility at a high temperature, granular ZrO 2 and ZrSiO 4 were embedded, which could greatly reduce its losses; third, the gradient coating system could effectively reduce CTE mismatch between the SiC layer and outer layer to avoid cracking. The effective combination of the above three factors was able to substantially protect the specimen for a long time from being oxidized. However, CO 2 and CO gases were produced by oxidation in the coating in the initial stage, and could break through the silica layer and leave varying pores, as seen in Figure 9. These pores provided the opportunity for oxygen to enter the coating to react with the substrate, finally leading to the anti-oxidation failure of the coating. When the oxidation time was increased to 198 h, the weight loss of the sample with 30 wt.% ZrSi 2 content was 0.1%. In addition, the coating still appeared dense and smooth without obvious microcracks after the oxidation test. The XRD results of the gray phase and light gray phase in Figure 9, marked as Spot C and Spot D, respectively, are consistent with those of Figure 8. It is demonstrated that the main reason for anti-oxidation failure was the formation of the pores to provide channels to oxygen into the coating. 92 h oxidation were analyzed by SEM and EDS, and the results are shown in Figure 8. It can be seen that the smooth gray phase could properly cover the whole surface, fill the pores of the coating, and effectively prevent oxygen from entering the coating, which was the main reason to improve the oxidation resistance of the sample. The gray phase marked as Spot B was made of O; Si; and a small amount of Zr, Al, and Au elements, thus it could be concluded that the gray phase was mainly SiO2 phase. Al from the SiC transition layer and Au from the SEM test could be ignored. The light gray phase marked as Spot A contained O, Zr, and Si elements, which had ZrO2 and ZrSiO4, and a small amount of SiO2 covering ZrO2. Therefore, first, the SiO2 phase could dense the outer coating to keep oxygen out by means of its good fluidity and low viscosity at 1500 °C; second, although SiO2 had a certain volatility at a high temperature, granular ZrO2 and ZrSiO4 were embedded, which could greatly reduce its losses; third, the gradient coating system could effectively reduce CTE mismatch between the SiC layer and outer layer to avoid cracking. The effective combination of the above three factors was able to substantially protect the specimen for a long time from being oxidized. However, CO2 and CO gases were produced by oxidation in the coating in the initial stage, and could break through the silica layer and leave varying pores, as seen in Figure 9. These pores provided the opportunity for oxygen to enter the coating to react with the substrate, finally leading to the anti-oxidation failure of the coating. When the oxidation time was increased to 198 h, the weight loss of the sample with 30 wt.% ZrSi2 content was 0.1%. In addition, the coating still appeared dense and smooth without obvious microcracks after the oxidation test. The XRD results of the gray phase and light gray phase in Figure 9, marked as Spot C and Spot D, respectively, are consistent with those of Figure 8. It is demonstrated that the main reason for anti-oxidation failure was the formation of the pores to provide channels to oxygen into the coating. In order to further research the physical and chemical reactions of the ZrSi2-SiC coating at a high temperature, the DSC/TG results are shown in Figure 10. From the TG curve in Figure 10, it can be found that the weight of the coating decreased by 1.65% from 180 °C to 430 °C. Corresponding to the DSC curve, there was an obvious absorption peak at 273 °C. The main reason for this was the volatilization of the residual binder like PVA coming from the preparation process of the ZrSi2-SiC particles. With the increase in the oxidation temperature, the DSC curve tended to be stable. When the temperature was increased to about 519 °C, ZrSi2 started to oxidize at a low rate, which could be proved by a slightly ascendant peak in the TG curve and a weak exothermic peak in the DSC curve. At the same time, SiC was oxidized to generate CO2 and CO gases. When the oxidation rate of SiC was higher than that of ZrSi2 and it had a larger proportion in the ZrSi2-SiC coating, the sample would experience weight loss. Therefore, the TG curve had a slight decrease with the weight loss of 0.07% at 730 °C. With the temperature increase, as the oxidation reaction of ZrSi2 was dominant, the production rate of its products was higher than the volatilization rate of the gases. Therefore, the sample experienced weight gain of 5.89%, corresponding to the increase in the TG curve. DSC shows an absorption peak at 1205 °C, which could be identified as the formation of ZrSiO4.  In order to further research the physical and chemical reactions of the ZrSi 2 -SiC coating at a high temperature, the DSC/TG results are shown in Figure 10. From the TG curve in Figure 10, it can be found that the weight of the coating decreased by 1.65% from 180 • C to 430 • C. Corresponding to the DSC curve, there was an obvious absorption peak at 273 • C. The main reason for this was the volatilization of the residual binder like PVA coming from the preparation process of the ZrSi 2 -SiC particles. With the increase in the oxidation temperature, the DSC curve tended to be stable. When the temperature was increased to about 519 • C, ZrSi 2 started to oxidize at a low rate, which could be proved by a slightly ascendant peak in the TG curve and a weak exothermic peak in the DSC curve. At the same time, SiC was oxidized to generate CO 2 and CO gases. When the oxidation rate of SiC was higher than that of ZrSi 2 and it had a larger proportion in the ZrSi 2 -SiC coating, the sample would experience weight loss. Therefore, the TG curve had a slight decrease with the weight loss of 0.07% at 730 • C. With the temperature increase, as the oxidation reaction of ZrSi 2 was dominant, the production rate of its products was higher than the volatilization rate of the gases. Therefore, the sample experienced weight gain of 5.89%, corresponding to the increase in the TG curve. DSC shows an absorption peak at 1205 • C, which could be identified as the formation of ZrSiO 4 . oxidation temperature, the DSC curve tended to be stable. When the temperature was increased to about 519 °C, ZrSi2 started to oxidize at a low rate, which could be proved by a slightly ascendant peak in the TG curve and a weak exothermic peak in the DSC curve. At the same time, SiC was oxidized to generate CO2 and CO gases. When the oxidation rate of SiC was higher than that of ZrSi2 and it had a larger proportion in the ZrSi2-SiC coating, the sample would experience weight loss. Therefore, the TG curve had a slight decrease with the weight loss of 0.07% at 730 °C. With the temperature increase, as the oxidation reaction of ZrSi2 was dominant, the production rate of its products was higher than the volatilization rate of the gases. Therefore, the sample experienced weight gain of 5.89%, corresponding to the increase in the TG curve. DSC shows an absorption peak at 1205 °C, which could be identified as the formation of ZrSiO4.   Figure 11 illustrates the cross-section morphology and element distribution of the specimen after oxidation for 198 h at 1500 °C. Although the sample had a 0.1% weight loss, its cross section presented a dense microstructure and had no obvious microcracks and pores. Compared with its thickness before the oxidation test in Figure 5c, it can be found that the ZrSi2-SiC outer coating was thicker after oxidation up to about 100 μm. The reason may be that the SiO2 layer and oxidation products were formed. Meanwhile, the outer coating and SiC coating had a strong interfacial combination without microcracks and disbonding, suggesting that the gradient anti-oxidation coating system largely allevi-  Figure 11 illustrates the cross-section morphology and element distribution of the specimen after oxidation for 198 h at 1500 • C. Although the sample had a 0.1% weight loss, its cross section presented a dense microstructure and had no obvious microcracks and pores. Compared with its thickness before the oxidation test in Figure 5c, it can be found that the ZrSi 2 -SiC outer coating was thicker after oxidation up to about 100 µm. The reason may be that the SiO 2 layer and oxidation products were formed. Meanwhile, the outer coating and SiC coating had a strong interfacial combination without microcracks and disbonding, suggesting that the gradient anti-oxidation coating system largely alleviated the thermal stress to improve its thermal shock resistance. According to the EDS results, the SiC transition coating had a low O element content, proving that SiO 2 layer consumption mainly destroyed the compactness of the outer coating, resulting in the failure of its oxidation resistance. Meanwhile, from the XRD result of the outer coating surface in Figure 12, it can be seen that the surface contained mainly ZrO 2 , further proving the above conclusion. ated the thermal stress to improve its thermal shock resistance. According to the EDS results, the SiC transition coating had a low O element content, proving that SiO2 layer consumption mainly destroyed the compactness of the outer coating, resulting in the failure of its oxidation resistance. Meanwhile, from the XRD result of the outer coating surface in Figure 12, it can be seen that the surface contained mainly ZrO2, further proving the above conclusion.

Conclusions
Four kinds of compositional gradient ZrSi2-SiC/SiC coating systems with different ZrSi2 contents were prepared by combined pack cementation and supersonic air plasma spraying. The ZrSi2-SiC coating with 30 wt.% ZrSi2 content possessed an optimum microstructure compared with the other three contents. Its morphologies of the surface and cross section appeared to be denser and had no obvious defects. This illustrated that ZrSi2-SiC mixed particles in this proportion were able to be melted fully and by evenly deposited during the process of deposition, which could generate a compact lamellar structure owing to their flattening when impacting the SiC coating surface. Meanwhile, the gradient coating system could effectively reduce the CTE mismatch between the SiC coating and the outer coating to improve their interfacial bonding strength. The compositions included ZrSi2, SiC, and a small amount of oxidation products, proving that the particles were oxidized slightly in the process of plasma spraying. In the static oxidation test at 1500 °C, a large amount of glass phase SiO2 was produced on the coating surfaces and ated the thermal stress to improve its thermal shock resistance. According to the EDS results, the SiC transition coating had a low O element content, proving that SiO2 layer consumption mainly destroyed the compactness of the outer coating, resulting in the failure of its oxidation resistance. Meanwhile, from the XRD result of the outer coating surface in Figure 12, it can be seen that the surface contained mainly ZrO2, further proving the above conclusion.

Conclusions
Four kinds of compositional gradient ZrSi2-SiC/SiC coating systems with different ZrSi2 contents were prepared by combined pack cementation and supersonic air plasma spraying. The ZrSi2-SiC coating with 30 wt.% ZrSi2 content possessed an optimum microstructure compared with the other three contents. Its morphologies of the surface and cross section appeared to be denser and had no obvious defects. This illustrated that ZrSi2-SiC mixed particles in this proportion were able to be melted fully and by evenly deposited during the process of deposition, which could generate a compact lamellar structure owing to their flattening when impacting the SiC coating surface. Meanwhile, the gradient coating system could effectively reduce the CTE mismatch between the SiC coating and the outer coating to improve their interfacial bonding strength. The compositions included ZrSi2, SiC, and a small amount of oxidation products, proving that the particles were oxidized slightly in the process of plasma spraying. In the static oxidation test at 1500 °C, a large amount of glass phase SiO2 was produced on the coating surfaces and

Conclusions
Four kinds of compositional gradient ZrSi 2 -SiC/SiC coating systems with different ZrSi 2 contents were prepared by combined pack cementation and supersonic air plasma spraying. The ZrSi 2 -SiC coating with 30 wt.% ZrSi 2 content possessed an optimum microstructure compared with the other three contents. Its morphologies of the surface and cross section appeared to be denser and had no obvious defects. This illustrated that ZrSi 2 -SiC mixed particles in this proportion were able to be melted fully and by evenly deposited during the process of deposition, which could generate a compact lamellar structure owing to their flattening when impacting the SiC coating surface. Meanwhile, the gradient coating system could effectively reduce the CTE mismatch between the SiC coating and the outer coating to improve their interfacial bonding strength. The compositions included ZrSi 2 , SiC, and a small amount of oxidation products, proving that the particles were oxidized slightly in the process of plasma spraying. In the static oxidation test at 1500 • C, a large amount of glass phase SiO 2 was produced on the coating surfaces and filled with their defects, which could fundamentally prevent oxygen from entering the coating. Granular ZrO 2 and ZrSiO 4 oxidation products pinned in the SiO 2 layer could slow down its volatilization rate. The gradient coating was designed to visibly enhance its thermal shock resistance to avoid cracking. Therefore, the coating with 30 wt.% ZrSi 2 content still kept gaining weight after oxidation for 188 h. However, because of the inevitable consumption of SiO 2 at 1500 • C and the detached film resulting from the escape of the oxidation gases, the coating failed to protect the sample from oxidizing. Finally, it had a weight loss of 0.1% when the oxidation time was increased to 198 h.