Microstructure and Properties of Thermal Electrode Material Si3N4–MoSi2 Composite Ceramics

With good high temperature and corrosive resistance performance, ceramic based composites can be used as promising materials to replace metal thermocouple materials. In this study, Si3N4–MoSi2 composites were prepared via hot pressing technology. X-ray diffraction (XRD), optical microscopy (OM), and scanning electron microscopy (SEM) were used to analyze the microstructure of the composites. The mechanical properties and electrical conductivity were tested. The results showed that the composites were composed of β-Si3N4, MoSi2, a small amount of Mo5Si3, and an amorphous glassy phase. The MoSi2 phase was evenly distributed in the matrix. The percolation network was formed with increasing MoSi2 content. The strength of the composites reached its maximum value when the MoSi2 content reached a critical point. The electrical conductivity behaved like a typical percolation phenomenon. The percolation threshold was about 30% to 45%.


Introduction
The thermocouple is the most widely used temperature testing device in many industrial areas [1][2][3]. To realize accurate temperature testing in a high temperature and corrosive environment, a corrosive and high temperature resistance tube has to be used to protect the metallic thermocouple material [4][5][6][7][8]. However, the application of the protection tube will reduce the stability and accuracy of the thermocouples due to the change of thermo-electrical potential resulting from the variation of the physical status of the metal electrode materials.
With increasing requirements from metallurgy, chemical, and aerospace engineering, the stable properties of thermocouple materials, which can be used in high temperature and corrosive environments, is still a big challenge in materials science, especially when they are used in molten salt electrolysis as the temperature testing device. Apart from the temperature range, the testing accuracy is another important factor that needs to be considered. As a result, a traditional metallic thermocouple is not able to satisfy the requirements when they are exposed to the above-mentioned working environments. Based on this condition, ceramic based composites, which have good high temperature and corrosive resistance performance, have been used as candidate materials to replace metallic thermocouple materials [9][10][11][12][13].
Disordered ceramic composites are composed of at least two types of ceramics, which combines the advantages of each component. Previous studies have demonstrated that Si 3 N 4 -MoSi 2 composites have good mechanical, anti-oxidation, and corrosive resistance performances [14][15][16][17][18]. As a result, they are able to be used as candidate materials to fabricate thermocouple electrode materials. In this study, the mechanical and electrical properties and microstructure of Si 3 N 4 -MoSi 2 composites were tested and analyzed to shed light on their potential application as new high temperature and corrosive resistance thermocouple materials.

Experimental
The Si 3 N 4 powders were provided by Shinuorui Co. Ltd., Fuzhou, China and the MoSi 2 powder was provided by Yantai Torch Special High Temperature Ceramics Co. Ltd., Yantai, China. The chemical contents are listed in Table 1, and the microstructures are depicted in Figure 1. All materials were used as received without further purification.  As can be seen from the chemical contents and microstructures, the purity of the α-Si 3 N 4 was higher than 92% with an average size of 1 µm. The purity of La 2 O 3 , and MoSi 2 was higher than 99% with size of 10 µm, and the purity of Y 2 O 3 was higher than 99% with an average size of 15 µm, as shown in Table 1.
As listed in Table 2, the contents of MoSi 2 were 30 wt %, 45 wt %, and 60 wt %, with 5 wt % of La 2 O 3 , 5 wt % of Y 2 O 3 and the rest of Si 3 N 4 . Before the hot pressing process, the powders were ball-milled in dehydrated ethyl alcohol for 10 h, and subsequently dried in a vacuum furnace at 100 • C. After drying, the temperature was raised to 1750 • C with a heating rate of 30 • C/min, followed by hot pressing with 25 MPa and a 60 min soak in a graphite mold. The furnace chamber was purged with 0.5 atm of argon gas from the start of the hot pressing procedure. The relative densities of the prepared samples were tested according to the Archimedes method. The flexural strengths were tested through a three point bending method in a universal testing machine (Instron-5569, Instron, Norwood, CO, USA). The size of the samples were 3 × 4 × 36 mm 3 , and the flexural strength were calculated according to the following equation: where F is the fracture loading; L is the space between the two points; and b and h are the width and thickness of the samples, respectively. The loading rate was 0.5 mm/min. The final results were the arithmetic average value of three samples. Before testing, the surfaces of the samples were polished to avoid stress concentration. Thermal shock resistance is one of the most important properties of ceramic materials. It reflects the fracture resistance of a material with exposure to sharp temperature change conditions. In general, the thermal shock resistance can be characterized by the strength remaining rate after thermal shock testing, expressed as [19]: where η rem is the flexural strength remaining rate; σ T is the flexural strength of the material experienced n time thermal shocking at temperature T; and σ 0 is the original flexural strength. The samples for the remaining strength tests were air cooled at various temperatures 10 times. The final results were the arithmetic average value of three samples. The phase structures were determined by X-ray diffraction pattern (XRD) on a Rigaku D/max-rA X-ray diffractometer with Cu Kα radiation (λ = 1.5406 Å) (Rigaku Corporation, Tokyo, Japan). The surface morphologies of the samples were observed by scanning electron microscopy with energy dispersive spectroscopy (SEM/EDS), performed on a Hitachi S-4700 scanning electron microscope (Hitachi Corporation, Tokyo, Japan). After that, the cross sectional optical microscopy was observed on Zeiss-MC80-DX (Zeiss, Oberkochen, Germany).  The densities of the prepared samples are shown in Figure 3. As demonstrated in this figure, the densities of Si 3 N 4 -30 wt % MoSi 2 , Si 3 N 4 -45 wt % MoSi 2 , and Si 3 N 4 -60 wt % MoSi 2 were 4.66 g/cm 3 , 4.31 g/cm 3 , and 3.99 g/cm 3 , respectively. After calculation, the relative densities of the samples were 93.5%, 94.9%, and 95.7%. As can be seen from these results, the samples were all well sintered with a relatively low porosity of less than 6.5%. Previous studies have demonstrated that the densification of Si 3 N 4 -MoSi 2 composites is still a challenge without sintering additives [20][21][22][23]. It was found that with the addition of Y and La, the densification could be significantly improved [24,25]. Figures 4 and 5 show the optical microscopy, backscattering SEM (BSE), and transmission electron microscopy (TEM) images of Si 3 N 4 -30 wt % MoSi 2 , Si 3 N 4 -45 wt % MoSi 2 , and Si 3 N 4 -60 wt % MoSi 2 , respectively. In the OM and BSE, the white areas are the MoSi 2 phase and the dark areas are the Si 3 N 4 phase. As can be seen in these figures, these two phases were evenly distributed, and the interfaces between the two components were clear and smooth. Little microcracks could be observed. The size of the Si 3 N 4 was about 3-8 µm, and the size of the MoSi 2 was about 3-5 µm with a uniaxial shape. In addition, the size of the Si 3 N 4 slightly reduced with the increasing weight contents of MoSi 2 .      Figure 7 shows the SEM morphology of the Si 3 N 4 -MoSi 2 composites. As demonstrated in this figure, the fracture of the composites was a typical inter-granular fracture. The grains of the Si 3 N 4 were rod shaped and the grains of MoSi 2 were equiaxial. In sample #1, the grain size of MoSi 2 was relatively small with a size of about 3-5 µm. With an increase of MoSi 2 content, the grain size of MoSi 2 kept increasing slightly, and small amounts of randomly distributed rod shaped Si 3 N 4 crystals were observed among the MoSi 2 crystals. It can be concluded that with the addition of MoSi 2 , the flexural strength of the composites will start to decrease when the content of MoSi 2 is higher than a critical value due to the increase of porosity and residual stresses.  Figure 8 presents the flexural strength remaining rate of the Si 3 N 4 -MoSi 2 composites at temperatures of 500 • C, 600 • C, and 700 • C after thermal shock testing 10 times. As demonstrated in this figure, the flexural strength remaining rates were 85.2%, 60.5%, and 23.4% corresponding to samples #1, #2, and #3, respectively, when the thermal shock temperature was 500 • C. When the thermal shock testing temperature increased to 600 • C and 700 • C, the flexural strength remaining rate was lower than 20%.  Figure 9 shows the SEM microstructure of the fracture surface of sample #1. As demonstrated in this figure, microcracks could be clearly observed when compared with the samples without thermal shock experience. In addition, the quantity and size of the microcracks increased with increasing thermal shock testing temperatures, which resulted from the internal stresses due to the thermal expansion mismatch. As a result, the flexural strengths decreased with the increasing internal stress and thermal mismatch of the two main constituents.  Figure 10 illustrates the electrical conductivity of the Si 3 N 4 -MoSi 2 composites as a function of filling content of MoSi 2 . As shown in this figure, the addition of the MoSi 2 could considerably increase the electrical conductivity of the composites. When the content of MoSi 2 reached 30 wt %, the electrical conductivity was 1.25 × 10 −5 Ω −1 ·m −1 ; when the content increased to 45 wt %, the electrical conductivity increased seven orders of magnitude and reached 3.45 × 10 2 Ω −1 ·m −1 . After this value, the electrical conductivity increasing rate became flat. This phenomenon can be explained by the percolation theory [13]. Before the percolation threshold, which means that the content of MoSi 2 is relatively low, the MoSi 2 is randomly distributed in the matrix, but no percolation network can be formed. Under this condition, the electrical conductivity is mainly determined by the matrix material, and shows low electrical conductivity. When the MoSi 2 content increased above the percolation threshold, the percolation network of MoSi 2 formed, and as a result, the electrical conductivity was mainly contributed by the MoSi 2 phase rather than the matrix, thus showing high electrical conductivity. In this study, it can be claimed that the percolation threshold is located between 30 wt % to 45 wt %.

Conclusions
In summary, this study successfully prepared Si 3 N 4 -MoSi 2 composites with hot pressing technology. The main constituents of the hot pressed Si 3 N 4 -MoSi 2 composites were β-Si 3 N 4 and MoSi 2 , and a small amount of the Mo 5 Si 3 phase and glassy phase containing La and Y. The SEM morphology of the Si 3 N 4 -MoSi 2 composites demonstrated that the fracture of the composites was a typical inter-granular fracture. The grains of the Si 3 N 4 were rod shaped and the grains of MoSi 2 were equiaxial. The strength of the Si 3 N 4 -MoSi 2 composites decreased with increasing MoSi 2 content. The flexural strength remaining rate of the Si 3 N 4 -MoSi 2 composites at temperatures of 500 • C, 600 • C, and 700 • C after thermal shocking testing 10 times demonstrated that the flexural strength rates were 85.2%, 60.5%, and 23.4% corresponding to Si 3 N 4 -30 wt % MoSi 2 , Si 3 N 4 -45 wt % MoSi 2 , and Si 3 N 4 -60 wt % MoSi 2 , respectively, when the thermal shock temperature was 500 • C. When the thermal shock testing temperatures increased to 600 • C and 700 • C, the flexural strength remaining rates were lower than 20%.
Author Contributions: L.F. and X.Y. prepared the majority of the manuscript and formatted the references. L.F. provided the main inputs of the manuscript structure and the electrical properties analysis section. P.G. contributed other sections of the properties' analysis. Y.H. provided the microstructure analysis section, the applications section and gave some valuable suggestions on the future studies.