**3. Results and Discussion**

It is vital to study the high-temperature stability of polymer-derived ceramics. First, the microstructural developments of pure BN and Si3N<sup>4</sup> pyrolyzed at different temperatures from PBN and PCS were investigated by XRD (Figure 1). For BN (Figure 1a), after being pyrolyzed at 1000 ◦C, two broad diffraction peaks at 2θ = 26.7◦ and 41.6◦ appeared. With the pyrolysis temperature rising from 1000 ◦C to 1600 ◦C, these diffraction peaks sharpened a little because of ongoing BN crystallization. At 1600 ◦C, the diffraction peaks were still broad, and no resolutions of the (100) or (101) doublet were displayed, which indicated the formation of BN nanocrystallines. For Si3N<sup>4</sup> (Figure 1b), it can be seen that

the as-pyrolyzed ceramics were amorphous below 1400 ◦C, and the crystallization process started at 1500 ◦C, which would affect the high temperature stability of ceramics or ceramic fibers.

**Figure 1.** XRD patterns of (**a**) BN and (**b**) Si3N<sup>4</sup> pyrolyzed at different temperatures from PBN and PCS.

To elucidate microstructural evolution of these composite polymer-derived ceramics, six hybrid precursors with different PBN/PCS ratios were fabricated and pyrolyzed at 1600 ◦C. The calculative mass contents of BN and Si3N<sup>4</sup> in the composite ceramics (P0-7-C) are listed in Table 1. The chemical environment of B, N and Si atoms in the composite ceramic (P3-C) was studied by XPS (Figure 2). The B1s peak at 190.8 eV and the N1s at 398.1 eV confirmed the presence of BN. Moreover, the binding energy centered at 102.5 eV for Si2p and 399.7 eV for N1s demonstrated the existence of Si–N bonds, indicating that the composite ceramic was composed of a mixture of BN and Si3N4.

The elemental compositions of these polymer-derived composite ceramics obtained at 1600 ◦C (P0–7-C) are listed in Table 2. Additionally, XRD patterns of these composite ceramics are shown in Figure 3. Obviously, when the content of the BN was over 75 wt% in the composite ceramics, no silicon nitride peaks were detected, indicating that the BN phase restrained the decomposition of Si3N<sup>4</sup> and limited the grain size of Si3N<sup>4</sup> crystals. Likewise, with the increase of Si3N<sup>4</sup> mass content in these composite ceramics, the crystallization process of BN was also hindered, and there existed no BN peaks when the mass content of Si3N<sup>4</sup> exceeded 15 wt%. Noticeably, the P3-C containing 75 wt% BN and 25 wt% Si3N<sup>4</sup> was totally amorphous, which indicated that the clusters of BN and Si3N<sup>4</sup> totally hindered the crystallization of each other. Its amorphous state at 1600 ◦C would guarantee

reliable performance for the composite ceramics or ceramic fibers in high-temperature environments. However, unlike the results of Tan et al. [14], who used PSZ as the raw materials rather than PCS, no h-BN crystals were found in this system. After analyzing the oxygen contents of PSZ and PCS, which were 3% and 0.6%, respectively, it was concluded that the introduction of oxygen could promote the crystallization process of h-BN by formulating the oxides with low melting points.

**Figure 2.** XPS spectra of the composite ceramic (P3-C): (**a**) B1s; (**b**) N1s; (**c**) Si2p.


**Table 2.** Elemental content of different BN/Si3N<sup>4</sup> composite ceramics.

**Figure 3.** XRD patterns of the composite ceramics pyrolyzed at 1600 ◦C: (**a**) P0-C; (**b**) P1-C; (**c**) P2-C; (**d**) P3-C; (**e**) P4-C; (**f**) P5-C; (**g**) P6-C; (**h**) P7-C.

The composite ceramics pyrolyzed at 1600 ◦C were further annealed at 1700 ◦C under N<sup>2</sup> for 2 h, and the corresponding XRD patterns are shown in Figure 4. Except for P1-C, the composite ceramics all exhibited Si3N<sup>4</sup> crystals, which was not beneficial for the hightemperature stability of the composite ceramics. In summary, the BN/Si3N<sup>4</sup> composite ceramics could only keep stability below 1700 ◦C with appropriate ratios.

**Figure 4.** XRD patterns of the composite ceramics pyrolyzed at 1700 ◦C: (**a**) P0-C; (**b**) P1-C; (**c**) P2-C; (**d**) P3-C; (**e**) P4-C; (**f**) P5-C; (**g**) P6-C; (**h**) P7-C.

Based on the investigation of the microstructural evolution of BN/Si3N<sup>4</sup> composite ceramics, the crystallization of Si3N<sup>4</sup> could be totally restrained when its content in the composite ceramics was below 25% at 1600 ◦C. Then, the composite ceramic fibers were fabricated from hybrid precursor P1, P2 and P3 through melt-spinning, curing and decarburization in NH<sup>3</sup> under 1000 ◦C and pyrolysis at 1600 ◦C in N2. The elemental compositions of these obtained fibers (P1-F, P2-F and P3-F) are listed in Table 3. Additionally, the XRD spectra of these fibers are shown in Figure 5. All these fibers only showed two broad diffuse peaks, revealing the low crystallinity of BN [28,29]. Noticeably, no diffraction peaks of Si3N<sup>4</sup> in the composite fibers were detected, indicating that Si3N<sup>4</sup> existed in an amorphous state, which was beneficial to the high temperature stability of the composite fibers.


**Table 3.** Elemental content of BN/Si3N<sup>4</sup> composite fibers.

**Figure 5.** XRD patterns of the composite ceramic fibers: (**a**) P1-F; (**b**) P2-F; (**c**) P3-F.

Figure 6 shows the morphologies of the obtained BN/Si3N<sup>4</sup> composite fibers pyrolyzed at 1600 ◦C. The diameter of the fibers was roughly 12 µm, and the surface was smooth and compact, without any apparent voids. The cross sections were nearly circular without inter-fusion, exhibiting a glass-like fracture feature, which demonstrated that the curing and pyrolysis process could meet the preparation requirements.

In order to clarify the distributions of these elements (Si, B, N), the fibers were embedded in epoxy resin, with further polishing and spraying carbon, and then characterized by EPMA (Figure 7). Obviously, the distributions of each atom for P1-F and P2-F were nearly homogeneous. For P3-F, the atomic Si aggregated in the core, while the concentration of atomic B was higher in the outside shell, revealing the phase separation of Si3N<sup>4</sup> and BN. It was concluded that during the spinning process, the PCS tended to aggregate in the core under the shearing pressure owing to the huge differences of the viscosity and softening point of PBN (80 ◦C) and PCS (210 ◦C), which led to the unique structure of the final composite fiber; this structure could only be obtained when the Si3N<sup>4</sup> content of the fibers reached 25 wt%.

*Materials* **2021**, *14*, x FOR PEER REVIEW 8 of 11

**Figure 6.** SEM images of the surface and cross sections of the composite fiber: (**a**,**d**) P1-F; (**b**,**e**) P2-F; (**c**,**f**) P3-F. ture of the final composite fiber; this structure could only be obtained when the Si3N4 content of the fibers reached 25 wt%.

**Figure 7.** *Cont*.

**Figure 7.** Elemental concentration of B, N and Si along the composite fiber diameter: (**a**) P1-F; (**b**) P2-F; (**c**) P3-F. **Figure 7.** Elemental concentration of B, N and Si along the composite fiber diameter: (**a**) P1-F; (**b**) P2-F; (**c**) P3-F.

The Weibull plots of failure strength of these fibers are illustrated in Figure 8. The tensile strength, Young's modulus, Weibull modulus and dielectric properties (f = 10 GHz) of BN/Si3N4 fibers are listed in Table 4. The tensile strength of BN/Si3N4 fibers rose dramatically with the increase of Si3N4 mass content, and that of P3-F reached 1360 MPa with the Young's modulus of 117 GPa. Compared with the results of Tan et al. [24], the composite fiber (P3-F) showed no h-BN crystals, but a higher tensile strength. Except in the case of the slightly higher mass content of Si3N4 in P3-F, the micro-cracks were caused by the crystallization process of h-BN. Therefore, in order to enhance the tensile strength of BN/Si3N4 composite fibers, the crystallization process of h-BN crystals should be avoided and the mass content of Si3N4 should be enhanced as much as possible in an appropriate ratio range of BN/Si3N4, where Si3N4 could remain amorphous in the final composite fibers. Apart from the excellent tensile strength, the composite fibers (P3-F) showed a low dielectric constant of 3.34 and loss tangent of 0.0047 at 10 GHz. The excellent dielectric properties could be ascribed to the low carbon content [30], which was less than 0.1 wt%. The combination of improved mechanical properties and excellent dielectric behavior demonstrated the potential for wave-transparent applications. The Weibull plots of failure strength of these fibers are illustrated in Figure 8. The tensile strength, Young's modulus, Weibull modulus and dielectric properties (f = 10 GHz) of BN/Si3N<sup>4</sup> fibers are listed in Table 4. The tensile strength of BN/Si3N<sup>4</sup> fibers rose dramatically with the increase of Si3N<sup>4</sup> mass content, and that of P3-F reached 1360 MPa with the Young's modulus of 117 GPa. Compared with the results of Tan et al. [24], the composite fiber (P3-F) showed no h-BN crystals, but a higher tensile strength. Except in the case of the slightly higher mass content of Si3N<sup>4</sup> in P3-F, the micro-cracks were caused by the crystallization process of h-BN. Therefore, in order to enhance the tensile strength of BN/Si3N<sup>4</sup> composite fibers, the crystallization process of h-BN crystals should be avoided and the mass content of Si3N<sup>4</sup> should be enhanced as much as possible in an appropriate ratio range of BN/Si3N4, where Si3N<sup>4</sup> could remain amorphous in the final composite fibers. Apart from the excellent tensile strength, the composite fibers (P3-F) showed a low dielectric constant of 3.34 and loss tangent of 0.0047 at 10 GHz. The excellent dielectric properties could be ascribed to the low carbon content [30], which was less than 0.1 wt%. The combination of improved mechanical properties and excellent dielectric behavior demonstrated the potential for wave-transparent applications.

**Figure 8.** Weibull plot of failure strengths of three composite fiber diameters.


**Table 4.** Tensile strength, Young's modulus and dielectric properties (f = 10 GHz) of BN/Si3N<sup>4</sup> composite fibers.
