VWR Canada provided the dianhydride monomer 3,3ʹ,4,4ʹ-benzophenonetetracarboxylic dianhydride, abbreviated as BTDA. The anhydrous solvent N,N-dimethylacetamide (DMAc) and diamine 4,4′-methylenedianiline (MDA) were procured from Sigma-Aldrich, Germany. The nanoparticles of ZnO and Al₂O₃ were acquired from Alfa Aesar and Nanophase Technologies, respectively.

Weighing and addition of BTDA were carried out under a dry and inert N₂ atmosphere inside the glove box of LabSTAR MBRAUN. Thermal curing of pre-polymer i.e., polyamic acid, was performed in the air-circulating oven of VWR, together with Lindberg furnace, which was equipped with controller (UP 150). The Fourier-transform infrared (FTIR) spectra of all the synthesized materials (polyamic acid/polyimide nanocomposites) were recorded on a Bruker Tensor 27 FTIR Spectrophotometer in the range between 400 and 4,000 cm⁻¹. The nuclear magnetic resonance (NMR) spectrum was recorded at 500 MHz in deuterated DMSO-d⁶ solvent using the Bruker Avance Spectrometer. WAXRD analyses were performed using the Bruker D8 FOCUS Diffractometer, with a 2 theta (θ) range of 5°–50°, a scan speed of 3 s/step, an increment of 0.05°, and nickel-filtered Cu-Kα radiation of 1.542 Å.

A field emission scanning electron microscope (Leo 1530 Gemini, Zeiss Germany) was used to investigate the morphology of polyimide nanocomposites. Philips CM10 Transmission Electron Microscope, which was equipped with an 11-megapixel camera, was used to determine nanoparticle dispersion within the polyimide matrix. ImageJ software was used for measuring the nanoparticle size in SEM and TEM micrographs. Thermal properties were studied by carrying out TGA on the Q500 model of the TA instrument, with a temperature increase of 20°C per minute, up to the maximum temperature of 800°C, under an N₂ atmosphere. The thermal endurance was assessed by exposing the polyimide matrix and corresponding nanocomposites at the temperature of 400°C and holding for 30 min (isothermal TGA) under an air atmosphere. The temperature was ramped up at a rate of 50°C/min. The data were measured with respect to weight loss percentage as a function of temperature for dynamic studies, while for isothermal studies, it was collected as a function of time. For differential thermogravimetric analysis (DTGA), the collected data were processed as a derivative of weight loss (%/°C). Dynamic mechanical thermal analysis (DMTA) was performed in the tensile mode over a temperature range of 35°C–400°C, at a heating rate of 5°C/min, under the air atmosphere on a rheometric dynamic mechanical thermal analyzer. The parameters of DMTA measurements were as follows: strain, 0.2%; length, 10 mm; initial static force, 0.4 N; and minimum static force, 0.01 N.

The protocol for the synthesis of the polyimide matrix, labeled as BMDA, as illustrated in Scheme 1, is described as follows: a 50-mL two-necked round-bottomed flask, equipped with a nitrogen tube and a magnetic stirrer, was charged with 0.198 g (1 mmol) of diamine MDA and 5 mL of solvent DMAc. After the complete dissolution of diamine, 0.322 g (1 mmol) of dianhydride BTDA was added to this solution with continuous stirring. The weighing of dianhydride and its addition to the diamine solution were performed inside the glove box to avoid the possible reaction between the dianhydride and moisture, which could lead to the formation of byproducts. The byproducts if formed could affect the properties of the final polymer. Then, the reaction mixture was stirred for another 24 h at 25°C. A transparent, viscous polyamic acid solution was obtained, which was cast onto an aluminum weighing dish and cyclodehydrated by thermal imidization to yield the final polyimide matrix BMDA. The thermal imidization was performed in a stepwise manner, following a programmed procedure. First, the solvent was removed by keeping the polyamic acid solution at 70°C for 18 h. Then, the polymer system was held for 1 h each at 120°C, 150°C, 200°C, 250°C, and 280°C. Finally, the aluminum sheet was carefully peeled to obtain thin solid films of polyimide with a typical amber color. Images of the polymer at different processing stages, i.e., after 70°C treatment and complete imidization, are shown in Figure 1.

Polyimide nanocomposites (BMDA-Al₂O₃ and BMDA-ZnO) were synthesized by incorporating relevant NPs in the polyimide matrix BMDA. A schematic diagram of the synthetic protocol is illustrated in Scheme 2, and a detailed explanation of the procedure is given below.

The structural elucidation of the synthesized polyimide matrix was carried out both at i) polyamic acid stage and ii) fully imidized form by FTIR spectroscopy. At the polyamic acid stage, the presence of absorption peaks at 1,627 cm⁻¹, which are ascribed to carboxylic acid and amide moieties, verified the condensation of diamine and dianhydride, i.e., the development of the pre-polymer. The broad absorption band at 3,475 cm⁻¹ is ascribed to NH and OH moieties of polyamic acid. The cyclodehydration of the pre-polymer to final polyimide was confirmed by the appearance of peaks at 1,707 cm⁻¹ and 1,778 cm⁻¹ in the characteristic region of the imide ring (symmetric and asymmetric stretches, respectively), along with the disappearance of amic acid bands at 1,627 cm⁻¹ and 3,475 cm⁻¹. Small shoulder peaks at 1,655 cm⁻¹ and 1,597 cm⁻¹ in the FTIR spectrum of BMDA after imidization are attributed to the stretching vibrations of aromatic C=C conjugated with C=O and C=C of the benzene ring, respectively. The absorption band at 1,370 cm⁻¹ is ascribed to the C-N-C moiety of the imide group. The FTIR spectrum of BMDA at the polyamic acid stage (BMDA-PAA) and after imidization (BMDA-PI) is shown in Figure 2.

Further evidence in favor of the polyimide structure was obtained from NMR spectroscopy. ¹H NMR spectra of the polyimide matrix were obtained at the polyamic acid stage because the final polymer (polyimide) was insoluble in common organic solvents. The development of the amide linkage was confirmed by the appearance of signals at 10.48 ppm and 13.34 ppm, corresponding to the protons of amide groups and carboxylic acid moieties in the polyamic acid structure, respectively (Kumar et al., 2016). Aromatic ring protons displayed signals in their characteristic region of 7.18–8.05 ppm, with respect to their locality within the polymer structure and type of neighboring groups. A singlet signal at 3.88 ppm is ascribed to protons of the -CH₂ group in the polyamic acid structure, inherited from diamine MDA. The ¹H NMR spectrum of the polyimide matrix BMDA at the polyamic acid stage, illustrating the assignment of specific protons, is presented in Figure 3, which approves the anticipated polymeric structure.

The common and simplest technique used for investigating nanocomposite structures is WAXRD analysis. Therefore, it was used to evaluate the insertion of Al₂O₃ and ZnO nanoparticles in the polyimide matrix. The X-ray diffraction patterns of the parent matrix BMDA and relevant nanocomposites, i.e., BMDA-Al₂O₃-9% and BMDA-ZnO-9%, are presented in Figure 4. The presence of Al₂O₃ nanoparticles in the matrix was confirmed by the appearance of diffraction peaks at 2 theta values near 31.9°, 37.6°, and 45.7°, which are characteristic peaks for cubic Al₂O₃. All the peaks were indexed with the standard XRD pattern of Al₂O₃ with JCPDS No. 29–0063, as shown in Figure 4 (Hui et al., 2016). The ZnO nanoparticles manifested their presence by showing the typical signals at 2 theta values of 31.8°, 34.3°, 36.3°, and 47.5° which are characteristic peaks for the hexagonal wurtzite-type structure. All the peaks were indexed with the standard XRD pattern of zinc oxide with JCPDS No. 75–1,526, as shown in Figure 4 (Mallakpour and Hatami, 2013; Gao et al., 2010; Pal et al., 2017a; Pal et al., 2017b). The aforementioned peaks of nanoparticles were not observed in the parent polyimide matrix BMDA, thereby confirming the successful preparation of polyimide-Al₂O₃ and polyimide-ZnO nanocomposites by WAXRD.

The distribution of nanoparticles within the polyimide matrix was examined through scanning electron microscopic studies. Figure 5 presents the SEM images of polyimide nanocomposites at a nanoparticle loading of 5% and different magnification scales of 200 nm and 1 μm. These graphics illustrate the dispersion of both types of nanoparticles, i.e., Al₂O₃ and ZnO, throughout the polyimide matrix. SEM micrographs for different contents of Al₂O₃ and ZnO nanoparticle samples, i.e., BMDA-Al₂O₃-9% and BMDA-ZnO-9% (Supplementary Figures S1, S2), displayed similar patterns. This further verifies the presence of nanostructured inorganic moieties throughout the polyimide matrix. Moreover, the size of the nanoparticles was measured using ImageJ software. The arrows in Figure 5 indicate some of the nanoparticles having size 27, 41, 51, and 62 nm (Ma et al., 2008).

Further evaluation regarding the dispersion of nanoparticles inside the polyimide matrix BMDA was performed through transmission electron microscopic studies. Figure 6 presents the TEM images of polyimide nanocomposites at magnification scales of 100 nm, 500 nm, and 2 μm. Like SEM, TEM also confirmed the dispersion of both types of nanoparticles, i.e., Al₂O₃ and ZnO, throughout the matrix. Few clumps of nanoparticles were observed at a 2-μm scale; however, some clusters can be observed at 500 nm, which is attributed to the agglomerate formation by more than one nanoparticle (Hsu et al., 2005; Ma et al., 2008). These agglomerates are more pronounced for ZnO nanoparticles compared to Al₂O₃ (Figure 6), which can be attributed to the high surface energy of ZnO, as suggested by Ponnamma et al. (2019), making them prone to significant agglomeration (Chen et al., 2017). The size of nanoparticles determined using ImageJ software in TEM micrographs was found to be 17, 23, 38, 52, and 77 nm (Figure 6), which is in good agreement with SEM findings. Furthermore, the 100-nm scale images of Figure 6 reveal the shapes of nanoparticles: spherical geometry was observed for Al₂O₃ nanoparticles, whereas ZnO nanoparticles displayed a flat/disk-like morphology.

Observations indicate that the use of nanoparticles without their surface modification results in the formation of larger agglomerates within polyimide matrices. However, in our case, the thorough dispersion of even unmodified nanoparticles has been illustrated by both SEM and TEM techniques. It can be inferred that the experimental procedure followed in this study developed nanocomposites with effective nanoparticle dispersion in the polyimide matrix. The probable reason is that nanoparticles can be more easily and efficiently distributed in polyamic acid with a lower molecular weight which is formed at the start of the reaction processing.

The thermal behavior of polymers was explored by two types of TGA, i.e., dynamic temperature scan and isothermal TGA. These measurements were performed to evaluate the performance of developed polyimide nanocomposites for their applications under specific conditions of high thermal stability. Most often, exposure of a few minutes does not cause much degradation of a polymeric system; however, it fails in the case of extended exposure at elevated temperatures.

Thin films of the polyimide matrix and nanocomposites were prepared by the cyclodehydration of relevant polyamic acid solutions under precisely controlled conditions. DMTA was used to measure their stiffness at high temperatures, and collected data are tabulated in Table 3. Figure 9 presents the comparative mechanical damping, i.e., tan δ curves for BMDA-Al₂O₃ and BMDA-ZnO nanocomposites. The maximum temperature for the peak at the tan δ curve was noted as glass transition temperature (T_g). Embedding nanoparticles within polyimide triggered T_g enhancement, as shown by tan δ curves in Figure 9A for Al₂O₃ nanoparticles and Figure 9B for ZnO nanoparticles. The T_g values provided in Table 3 show that Al₂O₃ nanoparticles improved the T_g value of polyimide from 293°C to 357°C at 5% loading, and it was further increased to 384°C by the ZnO nanoparticles at a similar loading level (5%). This observation suggests that nanoparticles of ZnO are more effective in improving the T_g value of polyimide. This T_g enhancement is ascribed to the inter-chain interactions developed through OH groups on the nanoparticle’s surface and carbonyl of imide moieties in the polymer structure (Hsu et al., 2005). Cross linking, as developed, constrained the chain flexibility/mobility of polyimide and consequently increased the T_g value of the polyimide system. T_g has increased in a concentration-dependent manner up to the optimum level of nanoparticle loading, which was found to be 5% for Al₂O₃ and 7% for ZnO. After this, the glass transition temperature started to decrease with the further addition of nanoparticles. This could be ascribed to the fact that nanoparticles remained more dispersed within the matrix up to the optimal concentration. After a specific concentration, nanoparticles possibly started interacting with each other despite the matrix, resulting in the formation of agglomerates, as shown in TEM images in Figure 6 (Ma et al., 2008; Tsai et al., 2011).

Based on these DMTA findings, it can be conveniently inferred that nanoparticles can easily regulate/modify the glass transition temperature of the polyimide matrix as per the need of a particular application. The objective can be achieved either by selecting a specific nanoparticle or varying the concentration of any selected nanoparticle.