Aerogels are the type of nanoporous materials possessing extremely low density, low thermal conductivity, high porosity, and high pore volume, and are potentially used in various fields such as environment, energy, cosmetic, transportation, building, catalyst, and biomedicine (Koebel et al., 2012; Ziegler et al., 2017; Vareda et al., 2018; Qie et al., 2020; Wei et al., 2020; Wang and Wang, 2021). Among a variety of aerogels, silica aerogels are the most frequently used aerogels owing to their relatively low price, ease of scale-up, and excellent overall properties (Du et al., 2013; Wang and Zhang, 2015; Chen et al., 2017; Wang et al., 2017, 2018; Li T. et al., 2018; Li X. et al., 2018). However, silica aerogels are very brittle, which inhibits their wide applications (Wong et al., 2014; Liu et al., 2019). Strategies to improve the mechanical property of silica aerogels have been proposed by several research groups. As illustrated in Figure 1A, one of the most effective methods is the modification of silica networks with different organic groups or polymers (Liu Z. et al., 2020). For example, hybrid silica aerogels prepared from organic group-bridged silica precursors (Boday et al., 2011; Wang et al., 2013; Wang Z. et al., 2015; Wei et al., 2015; Yun et al., 2015; Zhang et al., 2017) were robust and could be compressed up to 80% without any damage. Silica aerogels synthesized from vinyl-functionalized siloxanes were build up with organic–inorganic double cross-linking networks, and they exhibited remarkable mechanical strength, flexibility, and is capable of ambient pressure drying (APD; Shimizu et al., 2016, 2017; Zu et al., 2018a,b,c). However, these approaches involve expensive monomers, complicated synthetic processes, and their scale-up productions are yet to be reported. Another strategy is introducing fibers to synthesize silica aerogel blankets (Figure 1A-2), which results in the flexible aerogel blanket that could be bent using polyester and glass fibers as a matrix (Berardi and Zaidi, 2019; Berthon-Fabry et al., 2017; Huang et al., 2019; Talebi et al., 2019; Lakatos, 2019). Though this approach is able to scale-up and commercially available, the interaction between silica aerogel and the fibrous matrix is weak, and the silica aerogel in the blanket can still be cracked and becomes ineffective.

To solve the problems, a silica–polyimide (PI) foam aerogel blanket is proposed as shown in Figure 1A-3, where the brittle silica aerogel is encapsulated in the three-dimensional (3D) framework of PI foam. Even though the interaction between the silica aerogel and the PI network is weak, the silica aerogel is less possible to drop out from the blanket owing to the encapsulation effect. Indeed, various silica–PI composite aerogels have been synthesized via different approaches, and all of them show excellent flame-retardant performances (Fan et al., 2019; Sun et al., 2020; Zhang et al., 2020; Wu et al., 2021), which indicated that the combination of PI and silica would be an effective way to improve the performance of native silica aerogels. Thus, the proposed method to synthesize the silica–PI aerogel foam blankets cannot only significantly improve its properties, but it is also possible to synthesize the aerogel blanket via a one-pot process (Figure 1B). In this work, the synthetic parameters, physical properties including mechanical strength, thermal stability, hydrophobicity, etc., will be investigated. The relationship between structure and property will also be highlighted. Considering the scale production and excellent overall properties, the silica–PI aerogel blanket will be an ideal candidate for the construction of artificial islands (Figure 1C), which is self-floating, waterproof, flame-retardant, lightweight, and thermal insulation, leading to a safe and comfortable floating platform for emergency event application.

Silica–PI aerogel blankets were synthesized through a one-step process as shown in Figure 1B. Ammonia aqueous solution was mixed in the CS solution with a predetermined concentration. PI foam was then immersed in the CS solution. Aerogel blankets were obtained after gelation, hydrophobization, and APD. As presented in Figure 2A, a large PI foam with a size of 30 cm × 30 cm × 4 cm can be successfully converted into the silica–PI aerogel blanket (Figure 2B) with no visible volume shrinkage. The density of the PI foam was 9 mg/cm³, while that of aerogel blankets were 36, 69, 81, 118, and 196 mg/cm³, respectively. The densities of the blankets showed a direct correlation with CS concentration. The thermal conductivity of the aerogel blankets ranged from 0.018 to 0.033 W/mK, which was dependent on density, in which SAF-3 with a density of 81 mg/cm³ exhibited the lowest thermal conductivity of 0.018 W/mK and SAF-5 with a density of 196 mg/cm³ had the highest thermal conductivity of 0.033 W/mK. The SEM image of the PI foam (Figure 2C) revealed an interconnected porous structure with a 3D framework. The diameter of the pores ranged from 400 to 500 μm. Both the silica aerogel and PI framework in the blanket could be observed (Figure 2D), and they were closely contacted without a large crack (Supplementary Figure 1), indicating that the silica aerogel had been completely filled in by the pores. Thus, the remarkable reduction of the thermal conductivity of PI foam (0.044 W/mK) may be due to the silica aerogels, in which the air in the macropores of the foams was replaced by silica aerogels. Consequently, the thermal conductivity attributed by air convection was suppressed.

Figure 3A shows the FTIR spectra of the aerogel blankets. The vibration peaks of Si–O–Si and Si–C can be observed at 1100 and 840 cm^–1, respectively, which indicated the presence of silica aerogels. Besides, vibration peaks located at ca. 2,980 and 2,982 cm^–1 can be clearly observed, which could be ascribed to the vibration of C–H bonds from the –CH₃ and –CH₂CH₃ groups. The results confirmed the successful hydrophobization of the silica gel, which was a critical point toward APD (Prakashi et al., 1995; Yun et al., 2014; Li Y. et al., 2018). In addition, the characteristic peaks of PI located at 1,714 and 1,771 cm^–1 (C=O), 1,370 cm^–1 (C–N), and 1,496 cm^–1 (C=C) could be observed in all the blankets, with no change compared with PI foam, suggesting that the in situ synthetic process (including the hydrophobization) did not influence the chemical structure of the PI matrix. Figures 3B,C show the N₂ adsorption–desorption isotherms and the pore size distribution curves of the aerogel blankets. All the blankets exhibited hysteresis loops, which were due to the characteristic features of mesoporous materials (type IV isotherms). The BET surface areas of the samples ranged from 573 to 728 m²/g depending on density. The average pore size of the aerogel blankets (7–20 nm) also showed a reliance on density, which decreased with the increase in density.

Figure 4A and Supplementary Figure 2A show the compressive–stress curves of PI foam and the silica–PI aerogel blankets. The PI foam could be compressed up to 85% and the recovery from the deformation, modulus of PI foam was 0.36 KPa. The blankets with lower densities, e.g., SAF-1, SAF-2, and SAF-3 could be compressed up to 80%, while those with higher densities were broken with a small strain of less than 40%. The modulus of the blankets ranged from 0.4 to1.8 KPa, which was remarkably increased when compared to the pure PI foam (0.36 KPa). Pure silica aerogel monoliths have been reported to be broken at a strain of 10% (Jiang et al., 2017). Thus, the encapsulation of silica aerogels with a 3D foam was an effective method to improve their mechanical properties. In addition, the robustness of the blankets can be reflected from the loading–unloading compression study. As shown in Figure 4B, SAF-3 could be reversibly compressed up to 40% strain, and the curves during loading were overlapped when the strain was lower than 40%. When the blanket was compressed to 50%, the blanket can also be completely recovered by suggesting that the silica aerogels were stable in the blankets, though the modulus decreased. The time–strain and time–stress curves (Figure 4C) of the compression test during 50 loading–unloading cycles at 30% strain for SAF-3 further indicated that the silica aerogel blankets using a PI foam matrix showed a significant improvement in the mechanical properties. The stress of the blankets did not decrease even after 50 cycles of compression. The results confirmed that using a 3D porous structure as the matrix was a robust strategy to synthesize mechanically stable silica-based aerogel blankets. Interestingly, as shown in Supplementary Figure 2B, SAF-3 after burning at 600°C for 30 min could be reversibly compressed up to 20% strain, and the curves during loading were overlapped when the strain was lower than 30%, which indicated that the mechanical property of the blanket could be preserved to a certain extent. Nevertheless, they could not be reversibly compressed higher than 30% strain, which indicated that the PI matrix would be destroyed, as will be discussed in the following section.

Silica aerogels cannot be burned, but silica aerogel blanket using polymer-based fibers as the matrix showed a poor fireproofing grade, which severely restricts their applications (Berardi and Zaidi, 2019). On the other hand, PI is a kind of polymer with high thermal stability and most of them are non-flammable (Liu H. et al., 2020). Therefore, the silica–PI aerogel blankets designed in this work show excellent flame-retardant properties. As presented in Figures 5A,B, the IP foam could not get burned, but it was damaged and completely distorted under a flame of 1,300°C for 1 min. On the contrary, the silica–PI aerogel blankets showed much more stability at the same condition (a flame of 1,300°C for 1 min) (Figures 5C,D), and all the silica aerogels were kept intact though the PI foam matrix had been thermally degraded (Supplementary Figure 3). When the silica–PI aerogel blanket was presented on an alcohol burner at 600°C, only the surface of the aerogel blanket was burned (Figures 5E,F). The LOI of the aerogel blanket (SAF-3) was 34%. The results indicated that the silica–PI aerogels possessed remarkable flame-retardant properties. The IR image of the blanket and PI foam was shown in Figures 5G,H. Clearly, with the same thickness, the temperature of the upper surface of the blanket was lower than that of the PI foam. Moreover, when the aerogel blanket was presented on a hot plate at 250°C, the temperature of the upper surface of the blanket was ca. 138°C, which is lower than that of the hot plate. However, when the bare PI foam was placed on the hot plate, the temperature of the foam surface was 94°C, which is lower than that of the temperature of the hot plate. It can be seen that aerogel blankets exhibited better thermal insulation capacity than PI foam when compared to pure silica aerogels (Du et al., 2020; Figure 5I). The thermal stability of the silica–PI aerogel blankets was also similar to that of the PI foam, where the decomposition temperature of the blanket and foam at 5% was ca. 500°C (Figure 5J). The mass decline of the silica blankets below 600°C would mainly contribute to the degradation of side groups of PI, while the sharp decline over 600°C must be due to the decomposition of the backbone of PI. Besides, the degradation of Si–CH₃ and Si–OH groups in the silica aerogels also occurred when the temperature was higher than 350°C (He et al., 2019). However, due to the small fraction of the –CH₃ and –OH groups when compared to PI, the mass decline of the two groups was not obvious in the TGA curves. The residue mass of the blanket at 900°C was much higher than that of PI foam. For example, the mass remaining value of SAF-1 and SAF-3 was 74 and 86%, respectively (which suggested that the mass residue increased with the increase in silica aerogels). While that of PI foam was only 50%. The results discussed above indicated that silica aerogels encapsulated in the PI foam can significantly increase both the flame-retardant and thermal stability of the aerogel blanket.

The densities of the silica–PI aerogel blankets are shown in Figure 6A. The direct relationship between density and the CS concentration indicated that the density could be varied by CS concentration. The highest density was 196 mg/cm³ for SAF-5 with a CS concentration of 0.1 g/cm³. Interestingly, the blankets were superhydrophobic with the contact angle higher than 150°, which was a significant increase when compared to the hydrophobic property of PI foam (Figures 6B,C). Owing to the superhydrophobic property and low density, the aerogel blankets could be used as an ideal self-floating platform on the water. In general, an artificial floating island is a eutrophic water purification technology used for decontamination and ecological remediation due to the landscape effects. It works as a natural evolution of artificial wetlands or green filters, which has a history of nearly 40 years (Kong et al., 2019; Largo et al., 2020; Prashant and Billore, 2020; Ren et al., 2021). In this work, the aerogel blankets could be used as self-floating islands and as movable platforms for supporting or transporting goods and materials, rather than constructions for water purification. Considering the complicated condition on water, such as cold lake and sea, burning crude oil that was leaked, the requirement of working platforms on the water for a long-term period makes the silica–PI aerogel blankets as potential artificial islands. It should also be flame-retardant that meets the requirements of excellent thermal insulation properties and long-term stability, which is not only safe and comfortable but also easily movable on water due to their low density and superhydrophobicity (that may reduce the force of friction). As a primary study, an aerogel blanket was presented on a lake, and it can self-float on water (Figure 6D). An IR image was taken after putting a dummy on the blanket (Figure 6E). It could be seen clearly that both the aerogel blanket and the dummy possessed a temperature of ca. 4°C higher than that of the water surface, which indicated that the artificial island not only acts as a floating platform on the water but also shows excellent thermal insulation property due to their extremely low thermal conductivity (<0.03 W/mK).

In sum, a robust silica–PI aerogel blanket is designed and synthesized using an in situ method based on PI foam. The density of the aerogel blanket can be directly controlled using different CS concentrations, and the extremely low thermal conductivity of 0.018 W/mK is achieved and a remarkable temperature difference of 138°C is achieved using the blanket as a thermal insulator on a hot plate of approximately 250°C. Owing to the 3D porous framework of PI foam, the silica aerogel is encapsulated and the blankets exhibit strong mechanical properties, and they can be reversibly compressed for 50 cycles without reduction of strain. Moreover, the silica aerogel possesses a high degradation temperature of 500°C, and is resistant to extremely harsh conditions, e.g., 600°C for 30 min and 1300°C for 1 min. In addition, the contact angle of the blanket is 152°. Combining with the low density, low thermal conductivity, flame-retardant, and strong mechanical strength of the aerogel blanket, the silica–PI aerogel blanket has significant and commercial potential as the artificial islands that are safe, lightweight, comfortable, and easy to move.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding authors.

RL performed the experiments and summarized the data. JW designed the experiment, supervised, and wrote the manuscript. JL and XZ supervised the projects and comments on the manuscript. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.