3.1. Thermal analysis, crystallization and morphology of the Al2TiO5 fibers
Figure 1A shows the FT-IR spectra of the PAAT precursor and Al2TiO5 fibers heat-treated at different temperatures. The infrared absorption peaks located at 1200–1600 cm− 1 in the precursor belong to the characteristic vibration peak of acetylacetone [50]. With the increase of heat treatment temperature, the absorption peak in this range gradually decreased, indicating that the organics in the fibers gradually decomposed. The assignments of the FT-IR spectral peaks of PAAT precursor and Al2TiO5 fibers heat-treated at different temperatures are shown in Table S1 [51–52]. The thermal decomposition of PAAT precursor fibers in air was further characterized by TG-DSC in Fig. 1B. From the figure, the weight losses of the sample were mainly divided into three stages. In stage one, the obvious weight loss of approximately 53% below 300°C was mainly caused by evaporation of the solvent and removal of adsorbed water [53], and the mass loss rate was fast. In the second stage, from 300–600°C in PAAT, weight loss of 18% was observed, with two broad exothermic peaks in the DSC curve in the range of 300–400°C. The peaks are due to carbonization and combustion of organics [54]. In the third stage, above 600 ºC, the PAAT sample almost had no weight loss, and the TG curve tended to be flat. At the same time, two sharp exothermic peaks at 710 ºC were attributed to the crystallization of rutile and corundum phases. Furthermore, the endothermic peak at 1300 ºC was ascribed to the beginning of the transformation of rutile and corundum into Al2TiO5, which could be confirmed by the XRD spectra in Fig. 1C. From the XRD spectra, only Al2O3 (PDF:71-1123) and TiO2 (PDF:76–0319) phases existed at 1200 ºC; Al2TiO5 (PDF:26–0040) began to crystallize at 1300ºC. As the temperature increased, the peaks of Al2O3 and TiO2 got increasingly smaller until they disappeared, while the peak strength of Al2TiO5 became more and more intense. The Al2TiO5 peak gradually formed the main crystal phase. When the temperature reached 1500 ºC, single-phase Al2TiO5 formed and no peaks for Al2O3 and TiO2 remained. This result is in agreement with the TG/DSC curve, which demonstrated the well crystallized and wholly obtained Al2TiO5 phase.
Figure 2A-D shows the SEM images of Al2TiO5 fibers sintered at 1200 ºC, 1300 ºC, 1400 ºC, and 1500 ºC, respectively. As the heat-treating temperature increased, the grain size also increased. The fiber surface was worm-like, and each fiber had numerous Al2TiO5 grains, which were closely packed. There were large amounts of grain boundaries between grains, along which the particles were arranged irregularly. The fibers had been significantly sintered at 1500 ºC. Nevertheless, doping into the ZrO2 fiberboards might make the ZrO2 fiberboards denser and increase the strength of the fiberboards. However, the fibers heat-treated at 1500 ºC remain maintained good fiber morphology. For one thing, Al2TiO5 fibers used Ti-Al polymer precursor as the reactant, and the oxide content of Ti-Al polymer precursor was relatively high (oxide content was close to 31.5%). Additionally, we analyzed the chemical structure of PAAT by 1H NMR and XPS (Fig. S3 and S4). The attribution of chemical shift peaks in 1H NMR spectra are listed in Table S2 [50, 52, 55]. Al and Ti were, respectively, stabilized by enol-like acetylacetone groups (-acac) and subsequently formed polymers through hydroxyl group (-OH) bridging (Fig. S5A and B). Among them, -acac and four -OH groups coordinated with Al ion, while -acac and six -OH groups coordinated with Ti ion [56]. Two aluminum acetylacetonate and a titanium acetylacetonate chelated together to form a monomer, which was polymerized in solution to form an Al-Ti polymer precursor. Therefore, the precursors exhibited excellent spinnability. The mapping images of a selected area (Fig. 2E) of the Al2TiO5 fibers obtained by heat-treating at 1500°C are presented in Fig. 2F-H, showing that the fibers only contained Al, Ti, and O elements. From the spectrograms, Al and Ti elements of the fibers were uniformly distributed without a segregation region of any element, indicating that Al3+ and Ti4+ were uniformly dispersed during the polymerization process. Representative EDS results in Fig. 2I show that the atomic ratios of O, Al, and Ti were 61.2%, 26%, and 12.8%, respectively. The Al:Ti atomic ratio was close to 2:1 in the initial raw material, which indicates the successful polymerization of Al3+ and Ti4+.
3.2. Mechanical and thermal properties of ZAT fiberboards
In Fig. 3A, ZrO2-Al2TiO5 fiber products with different shapes were prepared by mixing ZrO2 fibers and Al2TiO5 fibers in different proportions with hydraulic molding mold and heating treatment to 1500 ºC. Figure 3B-F presents the section SEM images of ZAT fiberboards doped with Al2TiO5 fibers of different proportions, indicating that the resultant dense form was perceived, the fibers bonded more tightly to each other, and the liquid phase became more and more obvious with the increase in the proportion of Al2TiO5 contents. More figuratively, the microstructure mechanism diagrams of ZAT fiberboards are given in Fig. 3G-K, which vividly display the above change rule. The bonding degree between grains was fine, which enhanced the thermal stress resistance of the ZAT fiberboards. These phenomena promote the sintering of ZAT fiberboards, and we suspected that this might significantly improve the strength of the fiberboards.
The expansion stress of compressed ZrO2 fibers and Al2TiO5 fibers in the ZAT fiberboards could partially offset compressive loadings and enhance the compressive strength [57]. Therefore, we tested the mechanical properties of ZAT fiberboards after heat treatment at 1500 ºC. As expected, according to the compressive strength curve of ZAT fiberboards in Fig. 4A, the compressive strength of ZAT fiberboards prepared in this study increased with increasing proportion of Al2TiO5 fibers. Among them, the compressive strength of ZAT-0 fiberboards without doping Al2TiO5 fibers was only 1.33 MPa, while the compressive strength of ZAT-8 fiberboards reached 24.35 MPa, which was about 18 times that of ZAT-0 fiberboards. The compressive stress–strain curves of ZAT fiberboards are shown in Fig. 4B. Since Young's modulus is the ratio of the compressive stress to the strain, the slope of the curve at the elastic deformation stage is expressed as Young's modulus [54]. It was observed from the Young's modulus curve in Fig. 4A that the Young's modulus of ZAT fiberboards increased with increasing amount of Al2TiO5 fiber. It is worth noting that the Young's modulus of ZAT-8 fiberboards was determined to be as high as 15.6 MPa, much higher than the 1 MPa of ZAT-0 fiberboards. Simultaneously, Fig. S6A shows the flexural strength and the fracture deflection curves of ZAT fiberboards with sample size of 100*20*20 mm. They show that the flexural strength of ZAT fiberboards doped with Al2TiO5 fibers was 57.8% higher than that of ZAT-0 fiberboards, which was about 0.71 MPa. The bending stress increased near linearly with the displacement under bending loadings, and the bending stress of the ZAT fiberboards with Al2TiO5 fibers was significantly higher than that of ZAT-0 fiberboards (Fig. S6B). These results support doping Al2TiO5 fibers to improve the strength of ZAT fiberboards.
From the test results of the thermal expansion coefficient of ZAT fiberboards doped with different proportions of Al2TiO5 fibers in Fig. 5A, the thermal expansion coefficient decreased gradually with increasing proportion of Al2TiO5 fibers. Among the samples, the thermal expansion coefficient of the ZAT-0 fiberboards reached more than 10 × 10− 6/ºC. Conversely, ZAT-8 fiberboards had a lowest thermal expansion coefficient, and its effect was obvious. In Fig. 5C, it descended from 10.6 × 10− 6/ºC to 8.6 × 10− 6/ºC at 1100 ºC by about 20%. Figure 5B shows the relative rate of change in the length of ZAT samples. The thermal expansion rate of the samples gradually declined with increasing proportion of Al2TiO5 fibers. At 1100 ºC, the expansion rate of ZAT-0 fiberboards reached 1.108%, while the expansion rate of ZAT-8 fiberboards was reduced to 0.887% (Fig. 5C). Clearly, Al2TiO5 fibers could effectively diminish the thermal expansion of ZrO2 fiberboards, imparting excellent low expansion performance of the ZAT fiberboards.
Especially low expansion characteristics of a material allowed it to withstand severe thermal shock. Hence, we hold the opinion that with the increase in Al2TiO5 fiber content, the thermal shock resistance of ZAT fiberboards might become better. We thoroughly investigated the thermal shock resistance of different components of ZAT fiberboards at 1100 ºC, with a size of 30*30*8 mm (Fig. 6A), in order to verify this hypothesis. Table S3 depicts the thermal shock resistance times and the apparent morphology after thermal shock of each sample. From Fig. 6B, ZAT-0 fiberboards burst and split instantly after a rapid cooling and heat shock treatment, with extremely poor thermal shock resistance. In contrast, the thermal shock resistance of the ZAT fiberboards doped with Al2TiO5 fibers improved significantly, as shown in Fig. 6C–F. These results indeed support that higher Al2TiO5 fiber content leads to better thermal shock resistance of the fiberboards, and the number of thermal shock resistance also increased. In particular, ZAT-8 fiberboards only cracked slightly after 51 shocks, which was significantly better than ZAT-0 fiberboards (Fig. 6F). It was concluded that the addition of Al2TiO5 fibers could ameliorate the thermal shock resistance of ZrO2 fiberboards. This was mainly related to the low thermal expansion coefficient of ZAT fiberboards, due to the introduction of Al2TiO5 fibers. When the ZAT fiberboards were subjected to rapid cooling and thermal shock, these microcracks could effectively absorb elastic strain energy, reduce the internal thermal stress, and play a role in protecting the fiberboards, on account of large numbers of microcracks inside the Al2TiO5 fibers caused by thermal expansion anisotropy. According to the theory of thermal shock fracture initiation and crack propagation proposed by Hasselman, the lower the thermal expansion coefficient of the material is, the larger the thermal stress stability factor is, the more difficult the crack is to expand, and the better the thermal shock resistance of the material is [58]. In addition, from the microscopic point of view, the interface cracks of multiphase aggregates had a significant influence on the thermal shock resistance. The thermal expansion difference between ZrO2 fibers and Al2TiO5 fibers was considerable, facilitating the formation of grain boundary microcracks in ZAT composite material. The wide grain boundary area and the existence of microcracks provided space for grain expansion, thus improving the thermal shock resistance of ZAT fiberboards.
In addition, we further investigated the high-temperature thermal insulation performance and time-dependent fire resistance of ZAT fiberboards in different proportions exposed to a butane blowtorch flame through infrared camera observation. A detailed schematic of the device is depicted in Fig. 7A, and the optical photo is shown in the inset of Fig. 7C. We used a ZAT fiberboard sample with a length of 3 cm, a width of 3 cm, and a thickness of 8 mm (Fig. 7B). Figure 7C illustrates that the hot surface temperature of the ZAT samples near the butane flame could reach more than 1300 ºC. The surface away from the flame served as the cold surface [59]. In Fig. 7D, the ZAT-0 fiberboards with a thickness of 8mm could reduce the heat source to about 564ºC after 300s. Remarkably, with increasing Al2TiO5 fibers content, the cold surface temperature of the ZAT fiberboards shown an obvious decreasing trend and the thermal insulation effect gradually improved in Fig. 7D-I. Intuitively, in Fig. 7J, when the hot surface temperature was stable at 1250 ºC, the changing trend of the cold surface temperature of the ZAT fiberboards decreased with increasing Al2TiO5 fiber content from 0–8%. In particular, in Fig. 7H, ZAT-8 fiberboards stabilized from 1250 ºC to 332 ºC, indicating excellent thermal insulation performance. These results indicate that ZAT fiberboards could be used as an excellent insulation material with a wide range of applications. The reason is that fibers themselves have robust thermal insulation and fire resistance performance. The intrinsic thermal conductivity of Al2TiO5 is only 0.9–1.5 Wm− 1K− 1, which is significantly lower than that of ZrO2 (2.5 Wm− 1K− 1). The heat insulation performance of Al2TiO5 fibers was observed using an infrared thermal camera (Fig. S7A). Notably, Al2TiO5 fibers with a thickness of 2 mm could withstand 1200°C of the hot surface high-temperature butane blowtorch flame without any visible damage (Fig. S7C). When the spray gun was just turned on, the temperature of the cold surface increased rapidly, reaching 260 ºC within 10 s, and then the cold surface temperature stabilized at about 260 ºC (Fig. S7B). Meanwhile, Fig. S7D-E presents the thermal infrared images of the cold surface of Al2TiO5 fibers after heating for 60 s and 180 s, respectively. The cold surface temperature stabilized at a relatively low temperature of about 260 ± 5 ºC. As a consequence, the addition of Al2TiO5 fibers could markedly enhance the thermal insulation properties of ZAT fiberboards.