Materials

Polyvinyl alcohol (PVA, AR, 87–89% hydrolyzed) was purchased from Macklin (Shanghai, China). Sepiolite (Sep, Purity: 99%) was purchased from Sigma-Aldrich, (Newark, New Jersey, USA). Hydrochloric acid (AR, 36–38%) was purchased from Guoyao, Shanghai, China. The crosslinking agent (KH-550, (C₂H₅O)₃-Si (CH₂)₃NH₂, AR, M.W 221) was purchased from Aladdin (Shanghai, China). All materials were used as received, without purification.

Preparation of Aerogel Samples

Sep and a certain volume of aqueous solution with adjusted pH were placed in a 100 mL beaker and the beaker was placed under the high-speed disperser (IKA T25, Suzhou Sainz Instrument Co., Ltd., Suzhou, China) at 12,000 rpm (400π ω/s) for 30 min to obtain a 3 wt.% Sep suspension. 3 g of PVA was dissolved in water at 90℃ by mechanical stirring for 4 h to obtain a 3 wt.% PVA solution. Then PVA solution (or PVA solution and cross-linker KH-550) was added to the Sep suspension at 80℃ and stirred for 2 h to obtain a uniform mixture. The mixture was poured into Teflon molds placed on a copper rod in contact with a liquid nitrogen bath. The mixture was frozen while the liquid nitrogen level was kept up to the height of the copper rod, then freeze-dried at –55℃ and 15 Pa for 48 h in a freeze dryer (Labconco FD5-3, Kansas City, Missouri, USA). The dried samples were placed in an oven at 110 °C for 30 min in order to promote cross-linking (Cheng et al., Reference Cheng, Li, Wang, Mao, Xu, Zhang, Zhong and Sui2018; Liu et al., Reference Liu, Zhang, Cheng, Luo, Wang, Mao, Sui and Feng2021) (Fig. 7a compares the stress–strain curves of P3S5K5 heated in a 110°C oven for 30 min with those without heat treatment to prove that heating promotes cross-linking). As shown in Table 1, the samples obtained were labeled as P3S10, P3S5, P5S5, P3S2, and P2S1 (or P3S5K3, P3S5K4, P3S5K5, P3S5K6.5, P3S5K8 if cross linked). The nomenclature is based on the relative mass of PVA, Sep, and KH-550 in the aerogel sample. The specific preparation of the composite aerogels is shown in Fig. 1.

Table 1 List of aerogel components

Fig. 1 Preparation process of the composite aerogel

Characterization

The apparent densities of the aerogels were calculated from ρ = m/v (where ρ is the density, m is the weight, and v is the volume). The porosity was measured by a gravimetric method. The sample size was 20 mm × 20 mm × 20 mm and an average value was taken from three measurements. The specific calculation formula was as follows (Sanguanwong et al., Reference Sanguanwong, Flood, Ogawa, Martin-Sampedro, Darder, Wicklein, Aranda and Ruiz-Hitzky2021b):

$\begin{array}{c}P=\left(1,-,\frac{{\rho }_a}{{\rho }_s}\right)\times 100\%\\ \begin{array}{c}{\rho }_s=\frac{1}{\frac{\omega 1}{\rho 1}+\frac{\omega 2}{\rho 2}+\frac{\omega 3}{\rho 3}}\\ {\rho }_a=\frac{m}{V}\end{array}\end{array}$

where P is the porosity; m, mass; V, apparent volume; ρ, density fractions; ρ_a, density of aerogel; ρ_s, skeleton density.

The functional groups of the aerogel composites prepared were analyzed by attenuated total reflectance Fourier-transform infrared spectroscopy (ATR-FTIR, Avatar380, Thermo Fisher Scientific, Waltham, Massachusetts, USA). The spectral analysis range was 400 to 4000 cm^–1. The ATR attachment allowed the sample to be placed directly under the test probe without further processing.

The morphology of the aerogel samples was examined by scanning electron microscopy (SEM, TM3030, Hitachi Co., Ltd., Tokyo, Japan) at an acceleration voltage of 15 kV. In the sample-preparation process, the aerogel was first frozen in liquid nitrogen and then sliced with a blade to obtain a cross section. The cross section of the sample was then stuck on the section observation platform and observed after spraying with gold.

The crystal structure of the aerogel samples was analyzed by X-ray diffraction (XRD D/max 2550PC, Rigaku, Tokyo, Japan). The aerogel samples were taped flat on a glass plate, stuck into the sample slot, and scanned in the test range 5–90°2θ.

The elements of the aerogel samples were analyzed by X-ray photoelectron spectroscopy (XPS Escalab 250Xi, Yuzhong Industrial Co., Ltd, Shanghai, China). The sample size was 3 mm × 3 mm × 3 mm.

The thermal conductivity of aerogels was measured using the transient planar heat source method (Hot Disk TPS 2500S, Gothenburg, Sweden) at room temperature (25°C). The sample size was 50 mm × 50 mm × 10 mm.

Aerogel composites were tested using a BS EN ISO 604 universal testing machine (HY 940F, Yuzhong Industrial Co., Ltd, Shanghai, China) at 45% RH and 23°C (this instrument measures the compression properties of an aerogel or the tensile properties of a film and the stress–strain curve can be derived directly after instrument testing, without additional calculation and process). The tester speed was 20 mm/min and the test end-condition was 50% strain. The compression modulus of four specimens was measured and the average value was calculated. The sample size was 20 mm × 20 mm × 20 mm.

The limiting oxygen index (LOI) values of the samples were tested on an oxygen index instrument (ATS 1,004,050, ATS FAAR S.p.A, Shanghai, China) with the samples at a size of 100 mm × 10 mm × 10 mm.

Thermogravimetric analyses (TG 209F1, Netzsch, Shanghai, China) of the 5 mg samples were done in a nitrogen atmosphere at temperatures ranging from 30 to 900°C at a rate of 10°C/min.

A vertical combustion test for aerogel samples according to the UL-94 experiment was conducted (Wu et al., Reference Wu, Niu, Liang and Xia2019). The sample size was 125 mm × 10 mm × 10 mm.

The relevant combustion values of the aerogel samples were measured using a microcombustion calorimeter (MCC FAA-PCFC, British Fire and Combustion Characteristics Test Technology Co. Ltd, Manchester, UK) according to the ASTM D7309-2007 standard (Luo et al., Reference Luo, Ning, Zhou and Yuan2022). A ~5 mg sample was heated from room temperature to 750°C for pyrolysis at a nitrogen flux of 80 cm³/min, and the temperature-increase rate was maintained at 1°C/s. The main purpose of this test was to evaluate the flame retardancy of aerogels. In this paper, the flame retardancy of aerogel was characterized by the MCC test.

The excitation wavelength of 532 nm of the Raman spectrometer (DXR2xi, Thermo Fisher Scientific, Waltham, Massachusetts, USA) was used to determine the residual carbon for the aerogel. The aerogel residual carbon was attached directly to the slide for testing without further processing.

Optimization of Preparation Conditions

The Effect of pH on Sep Suspension Stability

The surface charge of Sep was sensitive to pH, which directly affected the three-dimensional arrangement of the Sep skeleton structure in the aqueous dispersion (Baldermann et al., Reference Baldermann, Mavromatis, Frick and Dietzel2018). Therefore, controlling the pH of the Sep dispersion environment was necessary.

Sep was dispersed in a series of aqueous solutions with varying pH values that were adjusted by using 0.1 mol/L hydrochloric acid or 0.1 mol/L sodium hydroxide solution to obtain Sep suspensions. The stratification of Sep suspensions after 24 h is shown in Fig. 2, the viscosity of the Sep suspensions was measured, as shown in Table 2. The viscosity of Sep suspensions in an acidic environment was obviously greater than in an alkaline environment. In an ideal state or model, Mg²⁺ will occupy each octahedral position of Sep, but in the actual environment, Mg²⁺ and Si⁴⁺ in Sep tetrahedra will be replaced, resulting in the generation of negative charges and structural changes. Therefore, the surface charge of Sep is very sensitive to pH, and it is easy to precipitate in an alkaline environment (Wu et al., Reference Wu, Hong, Zhan, Tang, Cui and Kou2022). At a pH value of 4.62, the viscosity of the Sep suspension peaked at 1.08 Pa s. After 24 h of resting, Sep suspensions in the alkaline environment revealed an obvious delamination (Fig. 2b). Evident solid precipitation appeared at the bottom of the alkaline suspensions in the inverted glass bottle, and the delamination of Sep intensified as the alkalinity of the suspensions increased. In addition, after 24 h of resting in a neutral or acidic environment, the fluidity of suspensions decreased significantly. This was due to the fact that the structure of the Sep sol–gel network in neutral or acid environments is more stable, whereas the Sep sol–gel network could not form in an alkaline environment (Baldermann et al., Reference Baldermann, Mavromatis, Frick and Dietzel2018). Therefore, a pH value of 4.62 was selected to prepare Sep suspensions for subsequent experiments.

Fig. 2 Stability of Sep suspensions under various pH conditions: a glass bottles upright after 24 h of resting; b glass bottles inverted from a group upright bottles immediately

Table 2 Table of viscosity values of 3 wt.% Sep dispersions at various pH conditions

The Effect of PVA/Sep Mass Ratio on Aerogel Properties

The addition of PVA can help to improve the mechanical properties of Sep-based aerogels. A series of aerogels was produced by controlling the PVA/Sep mass ratio, and the physical parameters of the aerogels are shown in Table 3. The density of the prepared aerogels was < 0.5 g/cm³ and the porosity was > 96%. With increasing addition of PVA, the density of aerogel increased, the porosity decreased, the compressive modulus decreased from 240.08 to 134.56 kPa, and the resilience increased to 93.51% after a 50% compressive deformation. This was due to the fact that the addition of PVA results in the formation of hydrogen bonding between C–OH in the structure of PVA and Si–OH on the surface of Sep, which can improve the resilience of the Sep-aerogel. With the increase in the amount of PVA added, the PVA within the system underwent a self-cross-linking reaction, and a local agglomeration phenomenon occurred in Sep, thus reducing the mechanical properties. The compression modulus of P3S5 was 209.28 kPa (Table 3), and the resilience reached 90.70%, which was the best physical performance overall. For this reason, the mass ratio for PVA/Sep selected in subsequent experiments was 0.6.

Table 3 Basic parameters of PVA/Sep aerogel

^a $R=\frac{H2-H1}{H0-H1}\times 100$ = Extent of rebound after 50% compression, where H ₀ is the original height of the sample (mm); H ₁, height after compression to 50% (mm); H ₂, rebound height at 60 s after removal of external force (mm) (Guo et al., Reference Guo, Chen, Lv and Wang2018)

The Effect of the Mass of Cross-Linking Agent on Aerogel Properties

In sample P3S5, the Si–OH on the surface of Sep and the C–OH in the structure of PVA were bound by hydrogen bonds. To improve the mechanical properties of the Sep-based aerogels further, the addition of the cross-linking agent KH-550 was chosen to enable the formation of chemical cross-linking (Si–OH of KH-550 and C–OH of PVA; Si–OH of KH-550 and Si–OH of Sep) and hydrogen bonding (–NH₂ of KH-550 and –OH of PVA; –NH₂ of KH-550 and –OH of Sep; –OH of PVA and –OH of Sep) (Cao et al., Reference Cao, Wang, Xia, Chen, Miao, Yang, Qian and Tang2018; Guo et al., Reference Guo, Hu, Pan, Wu and Jiang2006; Zhao et al., Reference Zhao, Liu, Li, Cao, Hua, Ge, He, Jiang and He2022). By varying the amount of KH-550, a range of cross-linked aerogels was created, and the exact physical characteristics are given in Table 4. The density of the prepared aerogels was < 0.5 g/cm³ and the porosity was > 96%. The compression modulus increased at first and then decreased, rising from 341.50 to 474.43 kPa and then falling to 259.92 kPa as a result of the addition of excessive cross-linking agent, which caused excessive Sep agglomeration, affecting the mechanical properties of the crosslinked aerogels. By combining the data above, the P3S5K5 sample was finally selected for the subsequent testing and characterization.

Table 4 Basic parameters of crosslinked aerogels

^a See footnote to Table 3

Structural and Morphological Aspects of Aerogels

The compositions of the aerogels were investigated using FTIR and XPS (Fig. 3). The FTIR band at 3325 cm^–1 was ascribed to –OH of PVA. The band at 1008 cm^–1 was ascribed to Si–O–Si of Sep. The hydroxyl stretching vibration band of P3S5 at 3325 cm^–1 became wider and weaker, while the Si–O–Si vibrational peak at 1000 cm^–1 shifted slightly to the lower wavenumber, indicating the formation of hydrogen bonds between PVA and Sep (Madyan & Fan, Reference Madyan and Fan2017; Pang et al., Reference Pang, Yu, Chen and Chen2021; Simon-Herrero et al., Reference Simon-Herrero, Peco, Romero, Valverde and Sanchez-Silva2019). In contrast, the FTIR of P3S5K5 showed characteristic absorption peaks of –NH₂ at 1535 cm^–1, and Si–O–C and Si–O–Si at 1146 and 1008 cm^–1, respectively (Sanguanwong et al., Reference Sanguanwong, Flood, Ogawa, Martin-Sampedro, Darder, Wicklein, Aranda and Ruiz-Hitzky2021a; Turhan et al., Reference Turhan, Turan, Dogan, Alkan, Namli and Demirbas2008; Xie et al., Reference Xie, Hill, Xiao, Militz and Mai2010). When compared to pure PVA aerogel, the stretching vibration peak of the hydroxyl group at 3325 cm^–1 widened and weakened, and a new absorption peak from Si–O–C appeared (Zhu et al., Reference Zhu, Zhao, Yi, Xu and Wang2022), indicating that the –OH in the structure of PVA formed a chemical cross-link with the Si–OH of KH-550 (Fig. 3a).

Fig. 3 a FTIR spectra of Sep, PVA, and composite aerogels. XPS of aerogels: b,e,h Sep; c,f,i P3S5; d,g,j P3S5K5

The XPS spectra (Fig. 3b–j) revealed that the elemental content fluctuation in the composites was visible across the entire spectrum, and the aerogel contained N atoms after the cross-linking agent was added (the N element was derived only from KH-550, the crosslinking agent). The O1s peak for Sep (Fig. 3e) corresponded only to Si–O (532.5 eV) (Qiu et al., Reference Qiu, Li, Li, Zhang, Li and Wu2019). In P3S5, the O1s peak was deconvoluted into two peaks (Fig. 3f), one for C–O (531.7 eV) and one for Si–O (532.7 eV). The position of the Si–O peak shifted slightly to lower energy. Deconvolution of the O1s peak from sample P3S5K5 (Fig. 3g) had three peaks which meant that oxygen has three forms of bonding, corresponding to C–O (531.9 eV), Si–O (533.3 eV), and Si–O–C (530.5 eV) (Zhu et al., Reference Zhu, Zhao, Yi, Xu and Wang2022). Additional analysis of the Si2p peaks found that the position of the Si–O peak in the P3S5 sample was shifted significantly to lower energy relative to Sep, from 103.5 eV to 102.1 eV. (Fig. 3h–i). This could indicate that the Si–OH on the surface of Sep formed hydrogen bonds with the C–OH of PVA. In P3S5K5, a new peak corresponding to Si–O–C (102.9 eV) appeared, which could effectively prove that Si–OH in KH-550 and C–OH in PVA could be cross-linked chemically to form Si–O–C after the addition of KH-550 (Fig. 3j) (Jiang et al., Reference Jiang, Zhang, Bourbigot, Chen, Duquesne and Casetta2019).

The crystalline structures of the samples were studied by XRD (Fig. 4a), which produced a strong diffraction peak in sample Sep at 7.38°2θ, corresponding to the crystallographic diffraction of Sep (110). PVA exhibited a strong diffraction peak at 22.71°2θ, corresponding to its (200) crystallographic plane. In the XRD pattern of the P3S5 sample, the position of the strong diffraction peak belonging to Sep did not change, and thus the layer spacing was unchanged at ~1.19 nm, indicating that the PVA molecular chain did not intercalate into the Sep interlayer. Because Sep is different from montmorillonite and has no lamellar structure, intercalation did not occur in the composite (Huang et al., Reference Huang, Wang, Xu and Wang2012). In addition, the peaks of Sep and PVA were still present in the diffraction pattern of P3S5, no new peaks appeared, and peak positions did not change significantly, indicating that the addition of PVA did not affect the crystal structure of Sep. In the P3S5K5 composite aerogel, the strong diffraction peak belonging to Sep moved from 7.38 to 7.79°2θ. According to the Bragg equation, 2dsinθ = nλ, the layer spacing decreased from 1.19 to 1.13 nm, caused by the addition of KH-550 which broke the covalent bond between the Sep lamellae (Aslam et al., Reference Aslam, Kalyar and Raza2018; Lu et al., Reference Lu, Liang, Zhang, Chen and Ma2004).

Fig. 4 a XRD of aerogels; b photos of various sizes of sample P3S5; SEM images of aerogels: c,e P3S5, d,f P3S5K5

As shown in Fig. 4b, aerogels of different shapes and sizes had been fabricated as required. The cross sections of the aerogels (Fig. 4d–h) showed a distinct 3D mesh structure in the P3S5 aerogel, and greater magnification revealed a distinct fibrous entanglement structure in the cell wall (Fig. 4e); this was the skeletal structure formed by Sep (Farooq et al., Reference Farooq, Sipponen, Seppala and Osterberg2018). Compared with P3S5, the addition of a cross-linking agent made the 3D porous structure of the material more compact, which was due to the presence of both physical and chemical cross-linking forces between the components of the system, and the magnification of the scan also showed the skeletal structure formed by Sep (Fig. 4f).

Thermal Conductivity and Combustion Performance of Aerogel

Thermal conductivity was an important parameter for evaluating the thermal insulation properties of the material (Table 5). The thermal conductivity of the P3S5 sample was 0.0328 W/(m K), less than reported elsewhere (Table 6) (Di et al., Reference Di, Ma, Wang, Guan, Lian, Qiu and Jiang2022; Guo et al., Reference Guo, Qiang, Chen, Su, Ouyang, Chen and Huang2022; Gupta et al., Reference Gupta, Verma and Maji2019; Li et al., Reference Li, Chen and Zhang2022; Madyan & Fan, Reference Madyan and Fan2018; Sun et al., Reference Sun, Wu, An, Ma, Xu, Zhang, Luo, Li and Liu2021; Wang et al., Reference Wang, Cao, Zhao, Liu, Geng and Wang2020; Zhao et al., Reference Zhao, Tian and Huang2021). The LOI of the P3S5 sample was 26.4% and increased to 30.4% with the addition of the cross-linker KH-550. The results of the vertical combustion test on two samples (Fig. 5c) showed that, after contact with the flame, a black charcoal layer formed on the surface of the aerogel immediately, which self-extinguished immediately after removal of the flame and could not be re-ignited. According to the UL-94 standard rating, P3S5 and P3S5K5 could reach a flame retardancy level of V-0 (according to the ASTMD3801-10 standard, the V-0 grade is achieved when the burning time of two consecutive combustions is within 10 s with no droplet generation; this test can show that the aerogel sample has good flame retardancy). After vertical combustion, the damage length (the maximum length, expressed in centimeters, of the damaged part of the material in the vertical direction under specified test conditions) of the P3S5K5 sample was shorter than that of the P3S5 sample, and the cross-linking agent also increased the flame retardancy of the Sep-based aerogel.

Table 5 Thermal conductivity, compostion properties, and thermal-stability parameters of aerogels

Table 6 Thermal conductivity of aerogels

Fig. 5 High-temperature resistance of aerogels: a TG; b DTG; c aerogels before and after vertical burning tests; d heat-release rate (HRR) temperature curves

The thermal stability of the Sep-based aerogels was investigated by thermogravimetric analysis in nitrogen (Table 5, Fig. 5). The mass loss of the aerogels at ~100°C reflected the removal of adsorbed water (Fig. 5a,b). Compared with P3S5 (taking P3S5 as the standard), the initial decomposition temperature of P3S5K5 decreased to 242.6°C, the maximum decomposition temperature was reduced by 48°C, and the residual carbon content increased significantly to 54.9% at 900°C. This indicated that the addition of KH-550 accelerated the initial decomposition of Sep-based aerogel in the flame and the formation of surface coke, and that the coke layer prevented the aerogel from further decomposition by oxidation.

To evaluate the combustion performance of aerogels further, the heat-release rate (HRR), total heat release (THR), heat-release capacity (HRC), and the peak of heat release rate (pHRR) of aerogels were measured using a microcalorimeter (Table 7, Fig. 5d). The HRR is the most important parameter for characterizing the combustion performance of materials and is the primary basis for evaluating the fire risk of materials. After adding the KH-550, the heat-release capacity of the aerogel was reduced from 141 to 137 J/(g k), the pHRR was reduced from 152.9 to 95 W/g, a reduction of 37.9%, and the temperature of pHRR increased from 314.5 to 386.9°C. These data indicated that the addition of KH-550 improves the flame retardancy of the Sep-based aerogel.

Table 7 Microcombustion calorimeter test results of aerogelsª

^a HRC: heat-release capacity; pHRR: peak of heat release rate; THR: total heat release

Analysis of the Flame-Retardant Mechanism

The residues of P3S5 and P3S5K5 from the vertical combustion tests were investigated by means of SEM imaging and FTIR spectroscopy (Fig. 6) to assess the changes in the morphology and porous structure and the chemical composition after burning. The initial porous structure of P3S5 and P3S5K5 was preserved in the same way as the structure of the samples without combustion (Fig. 6a,d). When the materials were observed under magnification, they showed fibrous structures in the cell walls, indicating that the skeletal support provided by Sep during burning was important in maintaining the aerogel structure (Fig. 6b,e).

Fig. 6 SEM and FTIR images of aerogel residue: a,b P3S5 and d,e P3S5K5; c FTIR spectra of residue; f Raman spectra of the carbon layer of residue

FTIR analysis was used to measure changes in bond vibrational energies in the residues (Fig. 6c). While the Si–O–Si absorption peak was retained at 1008 cm^–1, the burned samples P3S5 and P3S5K5 showed bands of C = C stretching and C–H bending vibrations at 1650 and 790 cm^–1, respectively, with no aliphatic absorption between 2800 and 3000 cm^–1. This indicated that a conjugated aromatic carbonaceous structure was formed during combustion (Ghanadpour et al., Reference Ghanadpour, Wicklein, Carosio and Wagberg2018). This was consistent with the formation of thermally stabilized carbon in the samples. In the P3S5K5, the Si–O–C at 1146 cm^–1 was retained after combustion, indicating that the addition of KH-550 prevented effectively the combustion within the aerogel. In addition, Sep could promote the rapid dehydration of the aerogel surface into carbon and inhibit strong blistering during combustion degradation, resulting in flame-retardant protection (Gupta et al., Reference Gupta, Verma and Maji2019; Jiang et al., Reference Jiang, Zhang, Bourbigot, Chen, Duquesne and Casetta2019).

In the characterization of flame-retardant materials, Raman spectroscopy was also used to understand the nature of the residual carbon obtained after the combustion of the material, so that the flame retardant mechanism could be better understood. Raman spectroscopy was performed on the residual carbon of P3S5 and P3S5K5 samples, and the resulting spectra (Fig. 6f) revealed two distinct peaks of the aerogel, corresponding to the D and G peaks of the C atom, respectively. The D band at 1362 cm^–1 represented the amorphous carbon and lattice defects, while the G band at 1596 cm^–1 represented the in-plane stretching vibration of the Sp2 hybridization of the C atom. The peak intensity ratio R (ID/IG) of the two bands of D and G reflected the degree of graphitization of the carbon material tested; the lower the R value, the greater the degree of graphitization of the carbon material, which meant that the carbon-forming performance of the corresponding material was better (Kang et al., Reference Kang, Lu, Qu and Yuan2019). The R-values for P3S5 and P3S5K5 were 1.6 and 1.2, respectively, for which both are considered to be low levels. Furthermore, P3S5K5 had better carbon-formation performance than P3S5, and the addition of a cross-linking agent improved the flame retardant performance of the aerogel.

Mechanical Properties of Aerogels

Excellent mechanical properties are the basis for application in tents. The mechanical properties of aerogels would be affected by the addition of PVA, cross-linking agents, and cross-linking between the aerogel components. The density of the Sep-based aerogels was < 0.5 g/cm³ and the porosity was > 96%. The compressive modulus of the P3S5 sample was 209.28 kPa. The modulus of P3S5K5 increased to 474.43 kPa, which was 2.6 times its original value, following the addition of the cross-linking agent. In addition, the specific modulus increased to 11.08 MPa cm³/g and, after compression to 50% deformation, it reached 89.31% resilience. The stress–strain cycle curves of aerogels (Fig. 7b,c) showed that, in the initial stage of compression, the sample exhibited linear elastic properties, which determined the compression modulus. The stress–strain curve was smooth throughout the test phase with no sudden drop, indicating that the aerogel did not break during the compression process.

Fig.7 a Stress–strain curves for sample P3S5K5 before and after heat treatment to 110ºC for 30ºmin; stress–strain cycle curves for samples b P3S5 and c P3S5K5; images of samples P3S5K5 and P3S5 d before the compressive force was applied, e at the point at which the compressive force was applied, and f 60 s after the compressive force was removed

The loading and unloading cycles of P3S5 and P3S5K5 under 50% axial strain were used to study the elastic properties of the materials. The compression property of sample P3S5K5 was obviously improved after heating, which was caused by the cross-linking promoted by heating (Fig. 7a). After 20 cycles, the two samples showed typical hysteresis loop curves, but still maintained good resilience, indicating that the composite aerogel has good fatigue resistance and long-term stability (Fig. 7b,c). The rigidity of P3S5K5 was better than that of P3S5. P3S5K5 was able to support a 50 g weight without deformation when the 50 g weight was applied to the aerogel, whereas P3S5 showed slight deformation under the pressure generated by the 50 g weight (Fig. 7d). When the P3S5K5 and P3S5 were compared before and after applying pressure (Fig. 7e,f), no significant changes were noted and no cracks appeared.