To prevent the oxidation of ultra-high-temperature ceramic zirconium diboride (ZrB₂) at high temperatures, this study fabricated sintered composites containing silicon carbide and tungsten carbide, and examined the properties related to high-temperature oxidation. Spark plasma sintering was employed for rapid sintering, and a high-temperature torch test was conducted on samples to determine their surface oxidation behaviour. The composites oxidised at high temperature showed different surface oxidation behaviour according to the type of carbide-based additive. Composites containing both carbides, which have different oxidation mechanisms, exhibited better resistance to oxidation than those containing a single carbide.

Keywords: ZrB₂, SiC, WC, Spark plasma sintering, Oxidation resistance

In general, ceramic materials tend to melt or decompose at temperatures higher than 3,000 °C. However, ultra-high-temperature ceramics (UHTCs) are known for their melting points above 3,000 °C, high chemical stability, and excellent resistance to oxidation. UHTCs are group IV, V, and VI transition metals such as boride, carbide, nitride, including TaB₂, HfC, and BN [1, 2]. The term UHTC was proposed in 2004 during a joint workshop between the National Science Foundation of the United States and the Air Force Office of Science Research. UHTCs have been extensively studied as materials for extreme environments such as space shuttles, transonic airliners, and rocket propulsion systems [3].
Zirconium diboride (ZrB₂), a covalent material with hexagonal crystal structure, is a representative UHTC along with TiB₂ and HfB₂. ZrB₂ has a melting point higher than 3,200 °C, Young’s modulus of 490 GPa, and hardness of at least 20 GPa. It also offers excellent thermal and electrical conductivity. Among UHTCs, ZrB₂ has relatively low density and excellent strength at high temperatures, making it an ideal material for aerospace and next-generation energy devices [4, 5]. Despite these advantages, the material has limited applications in extreme environments owing to its poor durability against surface oxidation at high temperatures. When oxidised in air, oxide is produced in two phases on the surface of ZrB₂, the states of which change with temperature. First, a B₂O₃ layer has been observed on the surface at temperatures below 1,000 °C, with a porous ZrO₂ layer beneath it; the pores are filled with B₂O₃. When the temperature increases further, the B₂O₃ glass layer is liquefied and vaporises at approximately 1,300 °C, and the space that was occupied B₂O₃ becomes a path for oxygen from the atmosphere to enter, which accelerates the oxidation of ZrB₂ [6-9]. The chemical equations describing this oxidation process are as follows.
 
2ZrB₂ + 5O₂ → 2ZrO₂ + 2B₂O₃
B₂O₃(l) → B₂O₃(g)
 
To prevent the above oxidation reactions, researchers are studying oxidation-resistant ZrB₂ composites, especially ZrB₂-SiC and ZrB₂-WC. In the ZrB₂-SiC composite, B₂O₃ binds with SiO₂ to form a borosilicate layer which, in contrast to pure ZrB₂, prevents B₂O₃ from becoming porous owing to volatilisation. However, B₂O₃ shifts to the surface when volatilised, and pulls SiO₂ existing in liquid phase to the surface layer. The surface becomes rich with SiO₂, and a layered structure rich in ZrO₂ develops within the composite, leading to separation of the borosilicate surface layer [10-17].
In the case of the ZrB₂-WC composite, WC is oxidised into WO₃, and the particles experience volumetric expansion during this process. The compressive stress acting between the particles blocks the path of oxygen, thus interfering with the oxidation of the composite [18-22]. The purpose of this study is to enhance oxidation resistance by combining the two abovementioned advantages, i.e., blocking the path of liquid-phase SiO₂ due to volumetric expansion of WO₃ oxidised from WC, and preventing oxygen infiltration by the glass layer formed on the surface during the oxidation of SiC composites. This study sintered ZrB₂-SiC-WC and compared its oxidation resistance to those of pure ZrB₂, ZrB₂-SiC, and ZrB₂-WC.

The materials used in the experiment were commercial powder ZrB₂ (99.5%, 1-2 µm, APS powder, Alfa Aesar), SiC (99.8%, 1 µm, S.A. 11.5 m²/g beta-phase, Alfa Aesar), and WC (99%, 2 µm, Sigma-Aldrich). Composites with four different compositions were prepared using ZrB₂ as a base. The composites are referred to as Z, ZS, ZW, and ZSW according to their compositions. The four batches are presented in Table 1. Powder samples from each batch were placed in a Nalgene bottle with methanol and subjected to ball milling for 24 h using WC media. The mixed powders were collected via rotary expansion and dried for 24 h in a vacuum dryer, followed by sieving using a 100-mesh sieve. The powders were then subjected to spark plasma sintering. The furnace had a vacuum atmosphere (under 2×10^-2 mmHg), and the sintering temperature was raised at a rate of 100 °C/min. Sintering was performed for 15 min at 1,800 °C and 30 MPa for Z and ZS, and for 15 min at 1,900 °C and 60 MPa for ZW and ZSW. The sintering conditions for each composition are listed in Table 2. During sintering, carbon paper was used to prevent reactions with the mould and punch, and to facilitate separation. The sample surfaces were polished with sandpaper to remove any traces of carbon paper after sintering. The sintered pellets were disc-type with a diameter of 10 mm and thickness of 2 mm. The density of each pellet was measured by Archimedes’ principle, and the relative density was obtained by calculating the theoretical density based on the rule of mixtures. Torch oxidation was performed to determine differences in oxidation among pellets. The gases used in the experiment were propane gas and oxygen gas. The gas flow rate was adjusted to maintain the flame temperature, which was measured with a pyrometer, at 1,600 °C. The pellets were placed on a refractory material to withstand flames at high temperature, and a pellet-shaped holder was made to prevent the pellets from moving about due to the strong flames. The experimental setup for torch oxidation and a photograph of the actual experiment are presented in Fig. 1. Thermal gravimetric analysis (TGA; TA Instruments SDT Q600, USA) was conducted to determine the changes in sample mass during oxidation at high temperatures under atmospheric pressure. To analyse the phase of each pellet, X-ray diffraction (XRD, Rigaku, Tokyo, Japan, using CuKa radiation) was used to scan the samples at 2° per minute in the range of 2Ɵ = 20-80°. The oxidation of composites was examined by applying field-emission scanning electron microscopy (FE-SEM; JEOL JSM-6700, Japan) and energy-dispersive spectroscopy (EDS; BRUKER AXS) to the cross-sections. 

In Fig. 2, the photographs on the left and right were taken before and after the torch oxidation experiment, respectively. The surfaces of the sintered samples were polished. As such, the surfaces before the torch oxidation experiment had a metallic shine. Fig. 2(a) shows the sample with sintering of only ZrB₂. The sample on the right has a largely oxidised surface with ash-grey colour after the torch oxidation experiment. Separation was observed on the surface layer. Because of oxidation, only porous ZrO₂ remained on the surface layer; the layer separation can be traced to the decrease in strength. Fig. 2(b) shows the ZrB₂-SiC composite. The photograph, taken after the torch oxidation experiment, shows that the sample has lost its metallic shine but appears to have a shiny coating. SiC undergoes oxidation to form SiO₂, while ZrB₂ forms B₂O₃, thus resulting in borosilicate. Fig. 2(c) shows the ZrB₂-WC composite; the sample after the torch oxidation experiment has a surface similar to that of Sample Z; however, the surface layer separation observed for Sample Z did not occur. WC produces WO₃ when oxidised, and WO₃ is believed to have filled the space between the porous ZrO₂ layers in Sample Z. Fig. 2(d) shows the ZrB₂-SiC-WC composite, which is similar to Sample ZS in terms of surface and other properties.
Fig. 3 shows XRD patterns of Z, ZS, ZW, and ZSW composites before and after the torch oxidation experiment. In the XRD patterns in Fig. 3(a), ZrB₂, SiC, and WB peaks are marked with circles, triangles, and squares, respectively. In all four XRD patterns, the phases of ZrB₂ can be determined. The composites share main ZrB₂ peaks at 25.2, 32.6, 41.6, 51.7, 58.2, 62.5, 64.4, 68.3, and 74.0°, with each peak corresponding to (001), (100), (101), (002), (110), (102), (111), (200), and (201) planes (JCPDS #34-0423). The composites containing SiC had main peaks of 35.7, 60.1, and 72.0° corresponding to (102), (110), and (116) planes (JCPDS #29-1131). The composites containing WC showed patterns of WB phase with peaks at 39.3, 42.3, and 47.5° corresponding to (105), (112), and (107) planes (JCPDS #35-0738). For the ZrB₂-WC composite, a solid solution was formed during sintering, and the excess WC reacted with ZrB₂ to form WB [19]. Fig. 3(b) shows the XRD patterns of four composites after the torch oxidation experiment. ZrB₂ was oxidised into ZrO₂, and peaks of ZrO₂ crystal growth can be seen at 24.0, 28.1, 31.4, 34.2, 49.3, and 50.6°, which correspond to (110), (-111), (111), (200), (220), and (-221) planes (JCPDS #37-1484). Since there are many ZrO₂ phases, i.e. not only for Z but also for ZS, ZW, and ZSW, overlapping makes it difficult to detect peaks of the WO₃ phase at 23.2 and 33.0°.
Cross-sections of the ZrB₂ composites after the torch oxidation experiment were examined through FE-SEM and EDS analysis (Fig. 4, Table 3). The oxidation layer thickness of the cross-section of ZrB₂ composites (Fig. 4(a)) was approximately 49 µm. EDS analysis revealed that the oxidation of ZrB₂ produced a porous ZrO₂ scale on the surface. The structure of the ZrB₂ oxidation layer was columnar, which facilitates the penetration of oxygen and accelerates the oxidation of composites [4-7].
Fig. 4(b) shows the ZrB₂-SiC composites after the torch oxidation experiment. Based on EDS analysis, we can see that the SiO₂ layer is distributed on the surface, and a ZrO₂ layer with hardly any Si is also present. The SiO₂ glass layer appears dense, while the ZrO₂ layer is porous. In fact, composites containing SiC were observed to have shiny surfaces after the oxidation experiment (Fig. 2(b)), which indicates that SiO₂ produced from the oxidation of SiC shifted to the surface. In the cross-section of Sample ZS, the oxidation layer was 29.4 µm thick, among which the borosilicate layer was found to make up 5.31 µm.
The porous layer of the ZrB₂-WC composite shown by FE-SEM was expected to be found, through EDS analysis, to contain ZrO₂ and WO₃. Unlike the ZrB₂-SiC composite, it has a porous oxidation layer but no dense glass layer. However, it is denser than the porous layers shown in Fig. 4(a) and (b). This can be attributed to the blocking of the oxygen path by the porous layer, with volumetric expansion due to the oxidation of WC into WO₃. The oxidation layer thickness was approximately 33.4 µm, lower than that of Sample ZrB₂ but with similar columnar form.
FE-SEM images and EDS analysis of the ZrB₂-SiC-WC composite are shown in Fig. 4(d). SiC formed a dense SiO₂ layer on the surface, while WC was oxidised into WO₃ and caused the middle porous layer to be relatively dense. The XRD results in Fig. 3 show that ZS and ZSW after the torch oxidation experiment had amorphous shapes at approximately 20°. Previously, in the photographs of samples after the torch oxidation experiment in Fig. 2, it was mentioned that borosilicate formed on the surface, which is also consistent with the FE-SEM images. The borosilicate layer was 2.57 µm thick, and the total oxidation layer thickness was 16.9 µm, the thinnest among the four composites. SiC and WC prevented oxidation from accelerating by blocking the path of oxygen.
Fig. 5 shows thermal gravimetric analysis results of ZrB₂, ZrB₂-SiC, ZrB₂-WC, and ZrB₂-SiC-WC composites. Changes in mass were measured while maintaining the temperature at 1,500 °C under atmospheric pressure for 30 min. As can be seen in the TGA graphs, the weight gains of Z, ZS, ZW, and ZSW are 8.2%, 6.8%, 5.3%, and 1.2%, respectively. Oxidation occurred when the samples reacted with oxygen in air, and smaller weight gain indicates lower reactivity. Similar to the results for oxidation layer thickness in the torch oxidation experiment, the weight gain was smallest for the thinnest sample, Sample ZSW. The weight gain of ZS and ZW was 1.4% and 2.9% smaller than that of Z, respectively. This shows that WO₃ oxidised from WC is more effective in blocking the path of oxygen compared with borosilicate obtained from the oxidation of ZrB₂ and SiC. The glass layer of ZSW was 2.57 µm thick, about half that of ZS at 5.31 µm. For Sample ZSW, borosilicate first formed on the surface where oxidation began, and the glass layer blocked the oxygen from entering. The formation of WO₃ helped block the path of oxygen below the surface; SiO₂, which can exist in liquid phase, blocked the path leading out to the surface. Sample ZSW had a weight gain of 1.2%, comparable to that of Sample Z at 8.2%. TGA analysis shows that oxidation resistance improved significantly. The results are presented in Table 4.

This study sintered ZrB₂-based composites. Spark plasma sintering was employed to sinter poorly-sinterable ZrB₂ composites into composites having a relative density of 98%. SiC and WC are expected to prevent oxidation when sintered with ZrB₂. When the ZrB₂-SiC composite was oxidised, a borosilicate layer developed on the surface, which blocked the path of oxygen and stopped the progress of oxidation. When the ZrB₂-WC composite was oxidised, WC was oxidised into WO₃, leading to an increase in volume. This volumetric expansion was presumed to have further interfered with oxidation by blocking the path of oxygen. Based on the TGA results, the weight gains were 8.2%, 6.8%, 5.3%, and 1.2% for ZrB₂, ZrB₂-SiC, ZrB₂-WC, and ZrB₂-SiC-WC, respectively. In sum, the ZrB₂-SiC-WC composite developed a glass layer on the surface that prevented oxygen from infiltrating, and the oxidised WO₃ prevented SiO₂ leakage while simultaneously blocking the path of oxygen. These factors contributed to significantly enhancing the oxidation resistance of the composites.

This work was supported by grants (NRF-2018R1A5    A6075959) from the National Research Foundation of Korea (NRF) funded by the Korean Government and project for ‘New business R&D Voucher’ between Industry, Academy, and Research Institute funded Korea Ministry of SMEs and Startups in 2019. (project No. S2778019).