Synthesis of polyimide cross-linked silica aerogels with good acoustic performance

Abstract

Polyimide cross-linked silica gels with different weight percentages of solid precursors and different volumes of MTMS were prepared by using bis(trimethoxysilylpropyl)amine (BTMSPA) cross-linked polyamide acid, synthesized by 4,4′-oxydianiline (ODA), p-phenylenediamine (PPDA) and biphenyl-3,3′,4,4′-tetracarboxylic dianhydride (BPDA), compounding with a silica network which was hydrolysed by MTMS, then supercritically dried to produce a hybrid aerogel. The obtained aerogels exhibit striking properties such as low density (as low as 90 mg cm⁻³), a large surface area (up to 582 m² g⁻¹), lower shrinkage (up to 20%) and are more easily accessible than those previous reported. Formulations made of 8.7% and 11.72% solid precursors never collapsed under compression. Furthermore, the latter shows the maximum modulus (up to 17 MPa) and a higher sound absorption coefficient (more than 0.8) with a thickness of 3 cm in a specific range of frequencies.

---

1. Introduction

Silica aerogels are featured by fascinating properties such as extremely low density (3–350 mg cm⁻³), remarkably low thermal conductivity (0.01–0.03 W mK⁻¹), high porosity (80–99%), and high surface area^1–6 and were first reported by Kistler in 1931.⁷ However, because of their native fragility,⁸ there has recently been growing interest in improving the mechanical properties and durability of silica aerogels. To surmount the native weakness that is the obstacle of the widespread application of silica aerogels,^9–14 Leventis, Meador et al. have made an outstanding development in reinforcing the silica aerogels by using different kinds of silanes or siloxanes and reinforcement with polymers such as epoxy,^15,16 styrene^17,18 and isocyanate^19–21 to form polymer conformal coatings²² over the silica skeletal structure. Considering the low thermal stability of the above mentioned polymers, polyimide is steadier in higher temperatures, has good chemical resistance and outstanding mechanical ability and has quite a wide usage in many applications under high temperatures.^23–26

Unlike the linear polyimide aerogels, which have a strong tendency to shrink during processing and unsatisfactory mechanical performance,²⁷ the cross-linked polyimide aerogels were produced by polyamide acid (PAA) oligomers which were fabricated from combinations of rigid and flexible aromatic diamines and capped with anhydride cross-linked with octa(aminophenyl)silsesquioxane (OAPS) and have been reported previously to not only display low shrinkage but also have a good combination of being lightweight, strong and flexible.²⁸ However, using this approach, formulations made with p-phenylenediamine (PPDA) as the rigid diamine exhibited high density (up to nearly 0.45 g cm⁻³), significant shrinkage (almost 30–40%), and a slightly reduced surface area. Moreover, the expensive price and difficulty in obtaining OAPS and 2,2′-dimethylbenzidine (DMBZ) limits the large scale preparation of polyimide aerogels.

In this study, robust monolithic cross-linked polyimide cross-linked silica aerogels were prepared, with a cross-linked silica particle network and surrounded by a polyimide network, which should function as the skeleton to enhance the mechanical properties. The silica particle network was prepared by reacting MTMS and water in alkaline conditions, followed by condensation polymerization with bis(trimethoxysilylpropyl)amine (BTMSPA) which had been capped with polyamide acid oligomers. Comparing with the OAPS cross-linker’s expensive price and the difficulty in commercially obtaining DMBZ cross-linker, BTMSPA is commercially available and its price is cheap. To synthesize a fibrous three-dimensional network, acetic anhydride and pyridine were used as dehydrants for the chemical imidization of polyamide acid at room temperature to obtain the polyimide cross-linked silica gels. After aging and exchanging the organic solvent, the aerogels were obtained by means of supercritical CO₂ drying. In this paper, we present a systematic study of the influence of different mass percentages of solid precursors and the volume of MTMS on the microstructure, mechanical and acoustic properties, about which few reports have been published for polyimide cross-linked silica aerogels, and overcome the cracking that is often observed in monolithic aerogels during the aging process.

2. Experimental section

2.1 Materials

N-Methyl-2-pyrrolidinone (NMP), acetone, hydrochloric acid, ammonium hydroxide (NH₄OH), deionized water, 4,4′-oxydianiline (ODA), methyltrimethoxysilane (MTMS), pyridine and anhydrous acetic acid were purchased from Sinopharm Chemical Reagent Co. Ltd., China. BTMSPA was purchased from J&K Chemical, China. PPDA was purchased from Aladdin, China. Biphenyl-3,3′,4,4′-tetracarboxylic dianydride (BPDA) was purchased from Beijing InnoChem Science & Technology Co. Ltd., China. All reagents were used without further purification except dianhydride which needed to be dried at 125 °C for 24 h under vacuum before being used.

2.2 Preparation of the aerogels

The process of synthesizing polyimide cross-linked silica aerogels is illustrated in Scheme 1. Polyamide acid (PAA) oligomers were prepared as previously reported.

Scheme 1 Preparation of polyimide cross-linked silica aerogel.

The molar ratio of dianhydride to diamines = 26 : 25. Diamines were constituted by 50% PPDA and 50% ODA respectively in molar percentage. A sample procedure using 11.72% of the solid precursors in solution A and 3 ml MTMS in solution B is as follows: MTMS (3 ml), H₂O (0.15 ml) and aqueous ammonia (25% NH₃ basis) (0.15 ml) were added to NMP (2 ml) and stirred for 4 hours, yielding silica sol as solution A. The molar ratio of MTMS : H₂O : aqueous ammonia = 5 : 2 : 1. BPDA (2.556 g) was slowly added to a solution of ODA (0.836 g) and PPDA (0.452 g) in 27 ml NMP. The mixture was stirred for nearly 30 minutes until all the solids were dissolved, then a solution of BTMSPA in NMP was added and stirred for another 10 minutes, forming solution B. Solution A was slowly poured into solution B and stirred vigorously after which acetic anhydride (6.55 ml) and pyridine (4.62 ml) were added. The reaction mechanism of the synthesis process is shown in Scheme 2.

Scheme 2 Synthesis of cross-linked polyimide cross-linked silica gel.

The mixture solution was continually and vigorously stirred for 3 minutes, after which it was poured into a syringe (1.96 cm in diameter), which was cut off at the needle end and then sealed using polystyrene film. The hybrid gel was formed within 5 minutes or less, as it depended on the percentage mass of polyamide acid oligomer. Hybrid gel was removed from the mold after aging overnight, and soaked in fresh NMP to remove the acetic anhydride and pyridine. The solvent within the obtained gel was exchanged gradually over 24 hours intervals as follows: 75/25 vol% NMP/acetone, 50/50 vol% NMP/acetone and 25/75 vol% NMP/acetone, before being exchanged with 100% acetone three times and then supercritically dried by CO₂. The structural changes of the hybrid aerogel during the preparation are demonstrated in Fig. 1.

Fig. 1 The structural changes process during the preparation procedure.

3. Characterization

The bulk density, ρ, of the aerogels was estimated by measuring the weight and volume of the cylindrical samples. The diameter shrinkage D_s of the aerogels was calculated from eqn (1), where D₀ and D_a are the diameter of the gel and the cylindrical sample, respectively.

Infrared spectroscopies were obtained in KBr pellets using a Bruker Tensor-27 FT-IR spectrometer.

The specific surface area, pore size distribution and average pore size of the aerogels were obtained from nitrogen adsorption–desorption isotherms at 77 K, analyzed by using a Quantachrome Autosorb-1 analyzer. All samples were out-gassed by heating at 80 °C for 8 hours under a vacuum before collecting adsorption and desorption isotherm data by using nitrogen as the adsorbent at 77 K. The surface area of the samples was calculated by the Brunauer–Emmett–Teller (BET) method, and pore size distribution was determined by modeling using the BJH theory.

The morphology of the aerogels was obtained by scanning electron microscopy (SEM). Samples were fractured at room temperature and sputter coated for 1 min with platinum before viewing images by using a Hitachi S-4800 Field Emission instrument.

The sound absorption coefficient and sound insulation of the hybrid aerogels were obtained on SW422/477/499 by the standing wave tube method. The frequency of the sound source varied from 2500 Hz to 10 000 Hz and the specimens were cut into cylinders of 1 cm and 3 cm in length, respectively, and the top and bottom surfaces were polished to 1.6 cm in diameter, which matched the diameter of the instrument. The sound absorption coefficient, a, of materials was calculated using eqn (2), where E_i is the incident sound energy and E_r is the reflection sound energy.

a = (E_i − E_r)/E_i. (2)

A compression test was performed on cylindrical samples with size of about 1.5–1.65 cm in diameter and 3 cm in length, by using an Instron 5982 universal materials testing machine under room conditions with a constant compression speed at 1.27 mm min⁻¹. Samples were polished smoothly using sandpaper to make sure that the top and bottom surfaces were parallel before installation between the two compression plates of the testing machine. The elastic modulus was taken as the initial linear portion of the slope of the stress–strain curve of the first compression.

4. Results and discussion

The detailed properties and the precursor variables of the polyimide cross-linked silica aerogels prepared in this study are listed in Table 1. Polyamide acid (PAA) oligomers were prepared using 5.86%, 8.7%, and 11.72% solutions of solid precursors in solution A, while the volume of MTMS in solution B was in the range of 1 ml to 5 ml. The color of the resulting hybrid aerogels was yellow, while there is a slight difference in visible light translucency, depending upon the amount of MTMS, which can be seen in Fig. 2.

Table 1 Detailed properties and the precursor variables of the hybrid aerogels prepared in this study

w/w solid in solution A	Volume of MTMS (ml)	Diameter shrinkage	Density (g cm^−3)	BET surface area (m^2 g^−1)
5.86%	1	10.8%	0.105	582.7
5.86%	3	11.4%	0.09	534.5
8.7%	1	12.9%	0.113	539.3
8.7%	3	12.9%	0.104	564.8
8.7%	5	13%	0.106	574.2
11.72%	1	19.9%	0.172	471.5
11.72%	3	18.3%	0.141	499.2
11.72%	5	18%	0.150	463.2

---

Fig. 2 Photographs of the polyimide cross-linked silica aerogels with different weight percentages of solid precursors and different volumes of MTMS.

Furthermore, the physical properties of the samples including the density, the specific surface area and the shrinkage ratio from wet gels to the aerogels are summarized in Table 1. The diameter of the samples illustrated in Fig. 2 are beyond 4.2 cm and the densities of the hybrid aerogels are in the range of 90 to 170 mg cm⁻³, increasing with the augment of the weight percentage of solid precursors as well as the shrinkage, while the diameter shrinkage is less than 20% which is almost half of what has been previously reported using PPDA as rigid diamines. The shrinkage mostly occurs in the process of gelation and aging, resulting from capillary force caused by the interaction between the solvent and the skeleton of the gels. Due to the existence of methyl groups on the skeleton surface, capillary force is reduced to a certain extent during gelation and the aging process, whereas the volume of MTMS has little influence on these two properties which can be compared directly in Table 1.

Based on the Brunauer–Emmett–Teller (BET) method, the BET surface areas ranged from over 450 m² g⁻¹ to nearly 600 m² g⁻¹, depending on the weight percentage of solid precursors. The N₂ adsorption–desorption isotherms of the hybrid aerogels with different weight percentages of solid precursors and different volumes of MTMS presented in Fig. 3 are similar. A narrow hysteresis loop with a sharp increase at a high relative pressure limit indicates both mesoporosity and macroporosity,²⁷ which is also reflected in the pore distribution graph (Fig. 4) and in the SEM images in Fig. 5.

Fig. 3 The N₂ adsorption–desorption isotherms of hybrid aerogels with different weight percentages of solid precursors and 1 ml MTMS.

Fig. 4 Pore-size distributions of the hybrid aerogels.

Fig. 5 SEM images of the polyimide cross-linked silica aerogels with various weight percentages of solid precursors from left to right: 5.86% (a and d), 8.7% (b, e and g) and 11.72% (c, f and h). The various volumes of MTMS from top to bottom: 1 ml (a–c), 3 ml (d–f), 5 ml (g and h).

The hysteresis loops of the isotherms correspond to the IUPAC type H3 or H4, rather than types H1 and H2 that usually apply for aerogel materials. This is because the backbones of the hybrid aerogels are characterized by nanofibers instead of the common spheres. According to the BJH desorption method, the pore size distributions of the hybrid aerogels with different weight percentages of solid precursors and different volumes of MTMS are shown Fig. 4. It can be noticed that the pores of the hybrid aerogels are mainly mesopores with a spot of micropores and macropores and the distribution peaks are mostly at 36 nm, while the variations of both different weight percentages of solid precursors and volume of MTMS have little influence on the range of pore size distribution.

According to the SEM images shown in Fig. 5, polyimide cross-linked silica aerogels exhibit a 3D porous network, composited of thin intertwining polymer nanofibers with a diameter of about 15 nm and clusters which are much more loose than those previously reported.

This is probably attributed to the difference of the weight percentages of the solid precursors and a certain amount of methyl in the system, leading to reduced shrinkage during the gelation and aging processes. Because of the differences in both the weight percentage of the solid precursors and the volume of MTMS, the pore structure of the hybrid aerogels, especially the pore size and the thickness of the nanofibers shown in Fig. 5, are dissimilar. The amount of macropores decrease and the nanofibers become thicker as the concentration of dianhydride and diamine increases, which corresponds to the weight percentage of the solid precursors. Interestingly, the clusters, sticking to the nanofibers, may be caused by the condensation polymerization of MTMS and BTMSPA, which get denser as the volume of MTMS increases.

The FTIR spectra of the hybrid aerogel can be seen in Fig. 6. The peak at 1076 cm⁻¹ is attributed to the vibration of Si–O, and the peak at 1500 cm⁻¹ would be assigned to the vibration of a benzene ring.

Fig. 6 FT-IR spectra of the polyimide cross-linked silica aerogel.

The two peaks at 1716 cm⁻¹ and 1774 cm⁻¹ are attributed to the symmetrical and asymmetrical stretch vibrations of C=O, two characteristic peaks in imide. A peak at 1361 cm⁻¹ indicates the stretch vibration of the C–N–C imide ring, which is also a characteristic peak in imide. The peak at 3454 cm⁻¹ is mostly attributed to the adsorbed water and amides due to the unreacted polyamide acid, which is assigned to the –NH– group of BTMSPA which only reacted with anhydride of BPDA and cannot be imidized further.

The sound absorption abilities of the hybrid aerogels with 8.7% and 11.72% solid precursors and 3 ml and 5 ml of MTMS with a thickness of 1 cm and 3 cm are illustrated in Fig. 7.

Fig. 7 (a) The sound absorption coefficient of the hybrid aerogels with a thickness of 1 cm in the frequency range between 2500–10 000 Hz, (b) the sound absorption coefficient of the hybrid aerogels with a thickness of 3 cm in the frequency range between 2500–10 000 Hz.

The sound absorption coefficient, reflecting the power of materials to absorb sound in a range of different frequencies, is quite low (below 0.4), under 4500 Hz, for samples with a thickness of 1 cm, but rises rapidly and peaks (0.8–0.97) at about 5500–7200 Hz. In contrast, the sound absorption coefficient of hybrid aerogels with 11.72% solid precursors and a thickness of 3 cm show peaks (more than 0.8) at 2500 Hz, dropping dramatically to the lowest point before rising again. Similarly, the samples with 8.7% solid precursors also fall down to the lowest point at nearly 4000 Hz before increasing to the top. Thus the thickness and density have an apparent effect on the sound absorption of the hybrid aerogels and the hybrid aerogels with thicknesses of 1 cm and 3 cm could be used as sound absorption materials in high frequencies (above 5000 Hz). Moreover, to a certain extent, the latter could be also applied to low frequencies (below 3000 Hz).

The sound insulation abilities of hybrid aerogels with a thickness of 3 cm are shown in Table 2.

Table 2 The sound insulation of the hybrid aerogels with the thickness of 3 cm at six different frequencies between 2500–10 000 Hz

Sample	Sound insulation (dB) (f = 2500 Hz)	Sound insulation (dB) (f = 4000 Hz)	Sound insulation (dB) (f = 5500 Hz)	Sound insulation (dB) (f = 7000 Hz)	Sound insulation (dB) (f = 8500 Hz)	Sound insulation (dB) (f = 10 000 Hz)
8.7%-3	50.87	30	26.73	24.46	50.52	31.97
8.7%-5	20.43	10.5	10.78	23.42	34.76	27.9
11.72%-3	16.02	4.6	21.63	20.59	25.56	44.21
11.72%-5	16.02	11.02	10.27	9.44	16.03	33.95

---

The sample with 8.7% solid precursors, 3 ml MTMS and a thickness of 3 cm manifests higher sound insulation abilities than that of other proportion samples at all six different frequencies. In particular, it shows more than 50 dB sound insulation at 2500 Hz and 8500 Hz, which is better than that at other frequencies. For other proportion samples, the sound insulation mostly improves at higher frequencies. Interestingly, unlike traditional materials, the sound insulation of hybrid aerogels does not increase with an increase of those densities. The samples in Table 2 show a fluctuating and sometimes decreasing trend. This phenomenon is probably caused by the uniquely hierarchical network structure of aerogels.

Fig. 8 shows the stress–strain curves of the hybrid aerogels. Similar to the previous report, the strength (at 10% deformation) increases with the increase in weight percentage of solid precursors. This is mainly caused by the excellent mechanical performance of the polyimide backbone. Fig. 9 shows the elastic module of the hybrid aerogels and the photographs of a hybrid aerogel before and after testing. The elastic modules of the hybrid aerogels in the range of 2–17 MPa, were determined by the weight percentage of the solid precursors. Moreover the polyimide cross-linked silica aerogels with a greater volume of MTMS exhibit lower modulus than those with a smaller volume of MTMS, especially in those with a higher fraction of solid precursors. This is attributed to the mass of the methyl groups, leading to more resilience.

Fig. 8 (a) 10% strain stress–strain curves of the polyimide cross-linked silica aerogels, (b) the initial linear portion of the slope of the stress–strain curves, (c) the stress–strain curve of the hybrid aerogel with 11.72% solid precursors and 3 ml MTMS.

Fig. 9 (a and b) A hybrid aerogel with 11.72% solid precursors and 3 ml MTMS before and after testing, and (c) the elastic modulus of the hybrid aerogels.

In Fig. 9, it can also be noticed that the hybrid aerogels with higher weight percentages of solid precursors never collapse under compression and the deformation in length is beyond 80%, while there is no obvious change in the diameter. This could be due to the good toughness of the polyimide backbone and the vast pore structures in the hybrid aerogels.

5. Conclusions

Polyimide cross-linked silica aerogels were synthesized by polyamide acid, using BTMSPA as a cross-linker and combining a silica network, which was compounded by the hydrolyzation of MTMS. The hybrid aerogel presents low density (lower than 180 mg cm⁻³), a high specific area (450–600 m² g⁻¹), and lower shrinkage (up to 20%) than previously reported. The density, shrinkage, specific area and modulus depend on the weight percentages of the solid precursors and the volume of MTMS. We have also illustrated the detailed research in sound properties such as sound absorption and insulation of the hybrid aerogels. Although hybrid aerogel with 11.72% solid precursors exhibits a slight increase in both density and shrinkage, it shows a good combination of sound absorption in a specific range of frequencies and mechanical properties.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (51102184, 51172163 and 51302040), National Key Technology Research and Development Program of China (2013BAJ01B01), National High Technology Research and Development Program of China (2013AA031801), Science and Technology Innovation Fund of Shanghai Aerospace, China (SAST201254, SAST201321), and Shanghai Municipal Science and Technology Commission Nano Special Project, China (12nm0503001, 11nm0501600).

References

1. A. C. Pierre and G. M. Pajonk, Chem. Rev., 2002, 102, 4243–4265.
2. A. Du, B. Zhou, Z. Zhang and J. Shen, Materials, 2013, 6, 941–968.
3. V. J. Wittwer, J. Non-Cryst. Solids, 1992, 145, 233–236.
4. N. Hüsing and U. Schubert, Angew. Chem., Int. Ed., 1998, 37, 22–45.
5. A. S. Dorcheh and M. H. Abbasi, J. Mater. Process. Technol., 2008, 199, 10–26.
6. L. W. Hrubesh, L. E. Keene and V. R. Latorre, J. Mater. Res., 1993, 8, 1736–1741.
7. S. S. Kistler, Nature, 1931, 127, 741.
8. H. Q. Guo, B. N. Nguyen, L. S. McCorkle, B. Shonkwiler and M. A. B. Meador, J. Mater. Chem., 2009, 19, 9054.
9. Y. Chen, K. Chaw Ng, W. Yan, Y. Tanga and W. Cheng, RSC Adv., 2011, 1, 1265–1270.
10. M. Kurahashi, K. Kanamori, K. Takeda, H. Kajib and K. Nakanishia, RSC Adv., 2012, 2, 7166–7173.
11. H. D. Gesser and P. C. Goswami, Chem. Rev., 1989, 89, 765.
12. G. M. Pajonk, Appl. Catal., 1991, 72, 217.
13. S. Yun, H. Luo and Y. Gao, RSC Adv., 2014, 4, 4535–4542.
14. Y.-F. Lin, C.-C. Ko, C.-H. Chen, K.-L. Tung and K.-S. Chang, RSC Adv., 2014, 4, 1456–1459.
15. R. Gerlach, O. Kraus, J. Fricke, P. C. Eccardt, N. Kroemer and V. Magori, J. Non-Cryst. Solids, 1992, 145, 227–232.
16. M. A. B. Meador, C. M. Scherzer, B. N. Nguyen, D. Quade and S. L. Vivod, ACS Appl. Mater. Interfaces, 2010, 2, 2162–2168.
17. M. A. B. Meador, A. S. Weber, A. Hindi, M. Naumenko, L. McCorkle and D. Quade, ACS Appl. Mater. Interfaces, 2009, 4, 894–906.
18. B. N. Nguyen, M. A. B. Meador, M. E. Tousley, B. Shonkwiler, L. McCorkle, D. A. Scheiman and A. Palczer, ACS Appl. Mater. Interfaces, 2009, 1, 621–630.
19. X. Wang and S. C. Jana, ACS Appl. Mater. Interfaces, 2013, 5, 6423–6429.
20. M. A. B. Meador, L. A. Capadona, L. McCorkle, D. S. Papadopoulos and N. Leventis, Chem. Mater., 2007, 19, 2247–2260.
21. A. Katti, N. Shimpi, S. Roy, H. Lu, E. F. Fabrizio and N. Leventis, Chem. Mater., 2006, 18, 285–296.
22. U. F. Ilhan, E. F. Fabrizio, L. McCorkle, D. A. Scheiman, A. Dass, A. Palczer, M. A. B. Meador, J. C. Johnston and N. Leventis, J. Mater. Chem., 2006, 16, 3046–3054.
23. C. E. Sroog, A. L. Endrey, S. V. Abramo, C. E. Berr, W. M. Edward and K. L. Oliver, J. Polym. Sci., Part A: Gen. Pap., 1965, 3, 1373–1390.
24. B. N. Nguyen, M. A. B. Meador, A. Medoro, V. Arendt, J. Randall and L. McCorkle, ACS Appl. Mater. Interfaces, 2010, 5, 1430–1433.
25. M. A. B. Meador, E. J. Malow, R. Silva, S. Wright, S. L. Vivod and H. Guo, ACS Appl. Mater. Interfaces, 2012, 4, 536–544.
26. H. Guo, M. A. B. Meador, L. McCorkle, D. J. Quade, J. Guo, B. Hamilton, M. Cakmak and G. Sprowl, ACS Appl. Mater. Interfaces, 2011, 3, 546–552.
27. C. Chidambareswarapattar, Z. Larimore, C. S. Leventis, J. T. Mang and N. Leventis, J. Mater. Chem., 2010, 20, 9666–9678.
28. H. Guo, M. A. B. Meador, L. McCorkle, D. J. Quade, J. Guo, B. Hamilton and M. Cakmak, ACS Appl. Mater. Interfaces, 2012, 4, 5422–5429.

---