Figure 1D shows the XRD patterns of the SiC/BBSZA composites sintered at different temperatures. The XRD peaks of 6H-SiC (PDF#97-001-5325) and Bi (PDF#97-005-3797) were observed in all specimens. The presence of bismuth indicated that bismuth oxide had been reduced. It is known that silicon carbide can reduce some metal oxides at high temperatures [33]. In this system, the following chemical reaction occurred: S i C + B i 2 O 3   = B i + S i O 2   + C O ↑ . (2)

This explains the appearance of bismuth. The phase with relatively small intensities of zinc was detected by XRD in the specimens of HP900 and HP1000. The emergence of zinc may be attributed to the decomposition of zinc oxide in the BBSZA glass. According to the Ellingham diagram, zinc oxide can decompose into zinc and oxygen under a certain ultra-low oxygen partial pressure (almost vacuum conditions) above 800°C. It can be comprehended that why the samples of HP700 and HP800 did not contain zinc. The chemical reaction process exhibits as following: Z n O = Z n + O 2 ↑ (3)

Moreover, aluminum oxide and zinc oxide in the BBSZA glass can react with each other to form spinel in the following reaction: Z n O + A l 2 O 3   = Z n A l 2 O 4 (4)

The X-ray feature peaks of ZnAl₂O₄ (PDF#97-000-9559) can be found in samples of HP800 and HP900. The crystal of ZnAl₂O₄ in HP1000 disappears due to the difference in Gibbs free energy between reactions (3), (4). The Gibbs free energy of reaction (3) is lower than that of reaction (4), which makes reaction (3) have more chance to occur at 1,000°C.

A high-resolution transmission electron microscope (HRTEM) can determine the phase of composites by analyzing the diffraction patterns and streaks. Figure 1A shows the HRTEM images of the SiC/BBSZA composite ceramic sintered at 900°C. The interplanar spacing values of .257 nm and .292 nm correspond to d (311) and d (220) of ZnAl₂O₄, respectively, as shown in Figure 1B. Otherwise, in Figure 1C, the interplanar spacing value of .315 nm matches d (012) of the Bi crystal. An amorphous phase can be seen around the ZnAl₂O₄ and Bi crystals, which represent the BBSZA glass. The analysis results of HRTEM verify the conclusion of XRD.

Figure 1E shows the EDS mapping images of the SiC–BBSZA composite ceramic sintered at 900°C. EDS mapping can exhibit the distribution of elements in the composites. The appearing location of the Si element matched well with that of the C element, indicating the existence of the SiC crystal in this area. It is worth noting that the O element almost has the same location as that of the Si and C elements. This could further prove the occurrence of reaction (2). In addition, the distribution of Bi, Zn, and Al elements is both uneven along with the phenomenon of aggregation. The aggregation of the Bi element is because of the generation of bismuth in reaction (2), and the aggregation of Zn and Al elements is contributed by the products of Zn and ZnAl₂O4 in reactions (3), (4). Otherwise, the position of Al element aggregation is partially the same as the Zn element, suggesting the formation of the ZnAl₂O₄ crystal. These EDS mapping images corresponded well with the results of XRD.

The microstructures of SiC/BBSZA composites sintered at different temperatures are shown in Figures 2A–D. The samples of HP700, HP800, and HP900 exhibited dense morphology except that the sample of HP700 formed some obvious pores, owing to the poor liquidity of glass at 700°C. The liquid phase sintering process could be divided into three stages: particle rearrangement, solution and precipitation, and solid-state sintering [34]. The densification of composites occurs quickly in the first stage as pores between grains are filled with the liquid phase. So, good flowability of the liquid phase facilitates the densification of ceramics. At 800°C–1000°C, the microstructure of samples showed homogeneous and dense morphology, which is beneficial to their thermal property. Nevertheless, HP1000 seems to have larger grains than other composites sintered at a lower temperature. So, we etched these composites with strong acids and alkalis to remove the glass, Bi, Zn, and ZnAl₂O₄; as shown in Figures 3A–D, the characteristic of acid and alkali resistance of SiC provides a convenient way to display the SiC grain size. It can be found that the distribution of SiC particle sizes is uneven. Also, compared with other samples, the number of fine grain particles of HP1000 decreased. This phenomenon can be attributed to the occurrence of secondary recrystallization [35]. As we all know, normal grain growth exhibits an increase in the average size, but if the grain boundaries are hindered by a second phase, such as impurities, the normal grain growth will stop. Secondary recrystallization is a process in which large grains grow excessively by consuming a uniform fine grain matrix that is essentially no longer growing. In particular, secondary recrystallization may occur when the particle size of the original material is not uniform or the sintering temperature is high. The driving force of secondary recrystallization is the excess interfacial energy of the grain boundary. The grain boundary phase of the large grains moves under the impetus of interfacial energy, resulting in large grains’ further growth and small grains’ disappearance. After secondary recrystallization occurs, the gas pores enter the grain interior and become isolated pores, making the sintering rate decrease or even stop. Because the pressure of the gas in the small pores is relatively large, it may migrate and diffuse into the large pores at low pressure, making the pores on the grain boundaries larger as the grains grow, i.e., the pore size increases.

Backscattered electron (BSE) images can clearly show the distribution of the different phases of the composites as the brightness of the lining depends on the atomic number. In this study, we take advantage of it to know the distribution of Bi, BBSZA glass, and SiC. Figures 4A–D show the microstructures of the ESB pattern of SiC/BBSZA composite ceramics sintered at different temperatures. The bright areas represent the location of the heavy atom in the samples (referring to Bi and BBSZA glass in this study), while the dark areas show the distribution of light atoms (referring to SiC, Zn, and ZnAl₂O₄ in the samples). The distribution of composites of HP800, HP900, and HP1000 in the bright regions both remained uniform. On the contrary, the HP700 sample exhibited uneven distribution in the bright areas due to the poor liquidity of glass. In addition, with the temperature increasing from 800°C to 900°C, the shape of light areas gradually tends to change from round to elongated, which indicated the flowability of Bi and BBSZA glass became better. This suggests the structure of the sample HP900 may be denser than that of HP700.

Figure 5A shows the curves of relative density and thermal conductivity of SiC/BBSZA composites sintered at different temperatures. The two curves showed a similar trend, which both increased as the sintering temperature increased up to 900°C and then decreased. The thermal conductivity of composites is closely related to the density in liquid phase sintering because pores can impede the propagation of phonons. With the sintering temperature increasing from 700°C to 900°C, the thermal conductivity of the composites varied from 4.056 Wm⁻¹K⁻¹ to 8.660 Wm⁻¹K⁻¹; meanwhile, the relative density increased from 86.1% to 95.5%. The thermal conductivity of the SiC/BBSZA composites reached the highest (8.66 Wm⁻¹K⁻¹) at 900°C, which outdistances that of conventional LTCC (∼2–4 Wm⁻¹K⁻¹). In addition, the thermal conductivity decreased by 3.7% when the sintering temperature varied from 900°C to 1,000°C. There may be two reasons for the decrease in thermal conductivity of HP1000. The first factor is the decrease in the relative density, which may be caused by the increase in pore size. The other factor may be the disappearance of ZnAl₂O₄ at 1,000°C. After all, the thermal conductivity of ZnAl₂O₄ potentially is higher than that of Al₂O₃ [36–38], which is a good thermal conductor in composites.

Figure 5B shows the electric resistivities of SiC/BBSZA composites sintered at different temperatures. The resistivities decreased sharply when the temperature varied from 700°C to 900°C and then have a small rise at 1,000°C. According to the pseudo-percolation model [39], electrons of the composite tend to conduct along the lowest resistance (the thinnest grain boundary). Bismuth has a certain electrical conductivity due to its metallicity. The abundant SiC and bismuth could touch each other to form a conductive network, which is beneficial to electron conduction. So, the density microstructure and uniform distribution of bismuth are the main reasons why the HP900 sample has the lowest resistivity. Otherwise, the high resistivities of HP700 and HP800 samples could be attributed to two reasons: first, the amorphous phase (BBSZA glass), an insulator, hinders electron conduction; second, pores in the composites destroy the contact of SiC particles.

Figure 5C shows the Vickers hardness of SiC/BBSZA composites sintered at different temperatures. The factors influencing the hardness of materials include the microstructure, composition, crystallographic direction, and grain size. The grain size of SiC/BBSZA composites does not change much when the sintering temperatures are between 700°C and 900°C. Therefore, the difference in the microstructure of these samples may play a vital role in hardness. The hardness showed an increasing trend with the rising temperature, similar to that of relative density. The hardness of the composites of HP900 is the highest up to 330 HV, which corresponds to the densest structure of composites at the sintering temperature. The decrease in pores enhances the hardness of materials. However, hardness tends to decrease when the pore size becomes bigger. This can interpret the decreasing hardness of HP1000 [40].

Figure 5D shows the curves of the dielectric constant and the dielectric loss tangent of SiC/BBSZA composites with different sintering temperatures at 20 GHz. It can be observed that the dielectric constant and dielectric loss tangent represented a similar trend, in which both increased when the temperature varied from 700°C to 900°C and then decreased. The following formula is used for calculating the permittivity of complexes—Lichtenecker’s logarithmic mixture law [41–43]: I n ε = ∑ i = 0 n V i I n ε i , (5) where ε represents the permittivity of composites, ε i is on behalf of the permittivity of each component in the sample, and V i represents the volume fraction occupied by each component, and ∑ V i = 1 . In this composite, the dielectric constant can be presented using this law, and let us take HP900 as an example. I n ε = V S i C I n ε S i C + V B i I n ε B i + V g l a s s I n ε g l a s s + V S i O 2 I n ε S i O 2 + V Z n A l 2 O 4 I n ε Z n A l 2 O 4 + V Z n I n ε Z n + V p o r e s I n ε p o r e s . (6)

By searching the literature, it is found that the dielectric constants of 6H-SiC, Bi, ZnAl₂O₄, and Zn are approximately 6.7, 3.3, 8.5, and 2.1, respectively [44–46]. It is essential to note that the dielectric constant of pores is low to 1.0, the same as that of air. The sample of HP900 obtained the biggest dielectric constant because the composite has fewer pores than other samples. Moreover, the disappearance of ZnAl₂O₄ and the increase of Zn lead the dielectric constant of the composite of HP1000 to decrease even though the sample had a dense structure. Otherwise, the dielectric loss has a close relationship with electrical conductivity because the electric energy can be converted into heat. It can be understood that better electrical conductivity leads to a higher dielectric loss tangent. Accordingly, the resistances’ variation trend of composites sintered at different temperatures is almost the same as that of the dielectric loss tangent; the sample of HP900 holding the smallest resistivity had the biggest dielectric loss tangent, which conforms to the previous rule.

To verify the thermal property of SiC/BBSZA composites sintered at 900°C, we fasten the sample on a chip resistor (JUC-31F) to observe its effect on device heat dissipation, simultaneously using another resistor as a blank control group to compare. The chip resistor can automatically power off when its temperature reaches to a certain value. From Figures 6A–C, it was found that the sintered sheet of HP900 can improve the heat dissipation performance of chip resistors as it can prolong the LED lighting time by nearly 50% more than that of the blank group. The infrared thermal imaging camera was used to record the temperature change of two chip resistors during the heating process, as shown in Figures 6D, E. The composites can effectively hinder the rapid rise of temperature of the device, owing to its high thermal conductivity. The curves of the average and maximum temperature of the two group devices during the heating process are shown in Figures 6F, G. The maximum temperature of the blank group rose to 158°C in a few seconds, while the device possessing composites varied from 27°C to 53°C in 120 s. Moreover, the curves of the highest and average temperature in the chip having composites are closer than that of the blank group. These differences between the two groups suggest the excellent thermal property of SiC/BBSZA composites sintered at 900°C.