Excellent energy storage performance in cross-linked polyimide dielectrics with positively charged cross-linking points

Yu Feng^ab, Tianlong Liu^ab, Jun Sun^ab, Dongyu Hou^ab, Yanqing Wang^ab and Dong Yue*^ab
aKey Laboratory of Engineering Dielectrics and Its Application, Ministry of Education, Harbin University of Science and Technology, Harbin, Heilongjiang 150080, P. R. China
^bSchool of Electrical and Electronic Engineering, Harbin University of Science and Technology, Harbin, Heilongjiang 150080, P. R. China

First published on 22nd July 2025

---

---

1. Introduction

Electric energy storage devices are the key components in the development of new energy fields.^1,2 Moreover, polymer dielectric capacitors have been widely used as key components in the energy storage field due to their high breakdown strength (E_b), excellent flexibility, and good processing performance.^3,4 In addition, due to factors such as loss of high-frequency switching and influence of environmental temperature, the operating temperature of dielectric capacitors keeps rising, necessitating greater requirements for the stable operation of energy storage dielectric capacitors at high temperatures.^5–7 However, the dielectric loss (tanδ) and leakage current of polymers usually increase dramatically when the temperature increases.^8–10 The increase in tanδ leads to a decrease in the charge–discharge efficiency (η), while the increase in the leakage current raises the probability of breakdown, making it difficult for the capacitors to meet the usage requirements under high temperature and high electric field conditions.^11,12 Therefore, it is important to ensure the energy storage stability of polymer dielectrics and improve their energy storage performance at high temperatures.^13–15

The vast majority of high-temperature resistant dielectric polymers, such as polyimide (PI), polycarbonate (PC), and poly-ether-ether-ketone (PEEK), rely on a large number of aromatic groups in the main chain.^16–18 Among them, PI stands out with its extremely high glass transition temperature (∼300 °C) and low dielectric constant (ε_r ∼ 3.2).^19,20 However, as the working temperature rises, the movement of local chain segments and functional groups in PI molecules gradually intensifies, causing secondary relaxation of the polymer and resulting in an increase in the leakage current of the material, sharp rise in the conductance loss, and subsequently a significant reduction in the discharge energy density (U_d) and η.^21,22 In order to reduce tanδ at high temperatures and suppress carrier transport, inorganic fillers with high bandgap widths can be added. Moreover, due to the relatively high ε_r of inorganic fillers, the synergistic effect obtained by constructing organic–inorganic systems is an effective method to improve the energy storage performance.^23,24 Zhou et al.²⁵ introduced boron nitride nanosheets as fillers into PEI, which enhanced the ε_r and E_b of the material. They also effectively suppressed carrier migration at high temperatures by introducing traps through energy level differences, thereby significantly improving the energy storage performance. However, the number of introduced inorganic fillers needs to be strictly controlled. Excessive introduction will cause the fillers to aggregate, resulting in defects in the PI and a decline in its performance.²⁶ Compared with inorganic fillers, organic molecules and polymers have better compatibilities. The interface defects of polymers can be reduced, and the insulation performance of polymers can be improved by introducing organic molecules, such as all-organic blending and the introduction of organic small molecule fillers.²⁷ Zhang et al.²⁸ increased the ε_r of PI by introducing cyanoethyl cellulose (CEC) into the PI matrix. The introduction of traps by CEC groups suppressed the charge migration within the composite materials, enhanced the E_b, and thereby increased the U_d of PI, demonstrating the effectiveness of the all-organic blending strategy in improving the performance of the polymers. Introducing a cross-linked structure is also an effective method for enhancing the dielectric properties of polymers.²⁹ Generally speaking, crosslinking methods can be divided into self-crosslinking and coupling agent crosslinking. For self-crosslinking polymers, it is necessary to have characteristic functional groups and initiate crosslinking through catalysis, with many variables and harsh conditions. Cross-linked polymers with coupling agents only require the addition of crosslinking agents to initiate crosslinking. To improve the performance of polymers, the process is simple, and the effect is obvious. Xu et al.³⁰ introduced 1,3,5-tris(4-aminophenoxy)benzene (TAB) as a crosslinking agent into polyimide, which greatly improved the thermal stability of the polyimide. The Ud of the prepared cross-linked polyimide reached 3.01 J cm⁻³ and 1.75 J cm⁻³ at 100 °C and 150 °C, respectively, while η remained above 70%. Pan et al.³¹ regulated the trap energy levels of polymers by using a cross-linked structure. By introducing crosslinking agents with different structures into a random styrene-maleic-anhydride (SMA) copolymer, they suppressed the conductance loss at high temperatures, significantly improving the high-temperature energy storage performance of the polymers. It can be seen from these studies that the cross-linked structure is an effective strategy for improving the energy storage performance of polymers.

In this work, high-temperature resistant PI is used as the matrix, and melamine MA with a positive charge at the crosslinking point is introduced into the PI as the filler to form a cross-linked structure. At the same time, traps with opposite charges to the applied electrons are introduced at the crosslinking points. The carrier transport is suppressed by the attraction effect of traps on carriers, and the E_b of PI is enhanced. The introduction of the cross-linked structure reduces the molecular chain spacing, inhibits the polarization movement of polar groups, reduces tanδ, and improves U_d and η. This research is expected to provide new ideas for the design of commercial high-temperature film capacitors.

2. Experimental

2.1. Experimental materials and instruments

4,4′-Diaminodiphenyl-ether (ODA) and 4,4′-oxyphthalic-anhydride (ODPA) were provided by Sinopharm Chemical Reagent Co., LTD. Melamine (MA), a crosslinking agent, and anhydrous ethanol were purchased from Innochem Co., LTD. N,N′-dimethylacetamide was purchased from Tianjin Fuyu Fine Chemical Co., LTD. All solid raw materials were dried in an oven at 80 °C for 12 h before use.

Scanning electron microscopy (SEM) of MPI was conducted on a Hitachi SU8020. The molecular structure of MPI was characterized using a Fourier transform infrared (FTIR) spectrometer from the Bruker Company in Germany. The phase structure and molecular spacing of the MPI were characterized using an X-ray diffractometer (XRD) from the Panak Company in the Netherlands. The thermal decomposition temperature of MPI was determined using a thermogravimetric analyzer (TGA) from the Beijing Jingyi High-Tech Instrument Co., Ltd. in China. The glass transition temperature of MPI was determined using a German LINSEIS differential scanning calorimeter (DSC). The variation laws of the ε_r and tanδ of the MPI with frequency were tested using the German Novo-control wideband dielectric spectrometer, and 9-mm-diameter aluminum electrodes were sputtered on both sides of the film before the test. The mechanical strength of MPI was tested using the Chinese WDW-10H universal mechanical testing machine. Aluminium electrodes of 3 mm in diameter were sputtered on the MPI films for the subsequent electrical tests. The leakage current density curve was evaluated by the current test module of a ferroelectric comprehensive measurement system. The E_b of MPI was determined at 20 °C and 150 °C by using the breakdown test system produced by the PolyK Company of the United States. The electric displacement–electric field loops (D–E) and energy storage characteristics of MPI were tested using the CPE1901 test system produced by the PolyK Company of the United States with a measurement frequency of 10 Hz. The charge and discharge of MPI were tested using the PK-CPR1801-10015 system produced by the PolyK Company of the United States with a load resistor (R_L) of 10 kΩ. The electrostatic potential energy distributions of PI and MPI were simulated based on molecular dynamics simulation (MD) and density functional theory (DFT). The BLYP generalization in the framework of generalized gradient approximation (GGA) was adopted as the exchange correlation generalization, and the electron density distribution of the system was obtained through the exact solution of the Kohn–Sham equation, which was further deduced to derive the electrostatic potential energy distribution.

2.2. Synthesis of PI and MPI

PI was prepared by polycondensation, and MPI was prepared by adding a crosslinking agent and using the method of crosslinking first, followed by polycondensation. The specific flowchart is shown in Fig. 1(a). This was carried out under a nitrogen atmosphere, adding 2 g of ODA (0.01 mol) to the three-necked flask, and then adding 35 mL of N,N′-dimethylacetamide solution to ensure that the solid content was between 15% and 20%. After gentle stirring to completely dissolve ODA, 3.1 g (0.01 mol) of ODPA was added in several portions within 6 h. The temperature was controlled at 20 °C through a constant-temperature reactor to ensure the smooth progress of the reaction. After thorough stirring for 12 h, the polyamide acid (PAA) solution was obtained. The crosslinking agent MA was dissolved in 1 mL of N,N′-dimethylacetamide solution and dropped into the PAA solution using a pipette. The reaction was continued for 12 h to generate cross-linked PAA.

After the reaction was completed, the solution was placed in a vacuum chamber and left to stand for 12 h to remove air bubbles. Then, it was evenly coated onto a clean glass plate and placed in a forced convection oven at a constant temperature of 80 °C for 6 h to remove excess solvent. Subsequently, the temperature was gradually increased at intervals of 30 °C until it reached 320 °C, with each temperature interval maintained at a constant temperature for 1 h to dehydrate and cyclize the polyamide acid and generate MPI; the schematic diagram of the reaction is shown in Fig. 1(b). After heating was completed, the cooled glass plate was placed in deionized water to remove the film. The added MA contents were as follows: 0%, 0.5%, 1%, 1.5%, 2%, and 2.5% of the total masses of ODA and ODPA, respectively. For convenience, the samples were named PI, MPI-0.5 wt%, MPI-1.0 wt%, MPI-1.5 wt%, MPI-2.0 wt%, and MPI-2.5 wt%, respectively.

3. Results and discussion

3.1. Structure and morphology of PI and MPI

The fracture surfaces of the MPI were characterized by SEM. The cross-sectional images of PI, MPI-1.5 wt%, MPI-2.0 wt%, and MPI-2.5 wt% are presented in Fig. S1. The cross-section of MPI-1.5 wt% was observed to be smoother and flatter, indicating that after the introduction of the cross-linked structure, there were reduced film defects and decreased molecular chain spacing, making the film surface smoother. When the cross-linker content exceeds 1.5 wt%, the free cross-linker will agglomerate and cause surface defects, which may affect the performance of the film. To detect the types of functional groups in the samples and the influence of the introduction of the cross-linked structure on the types of chemical bonds in polyimide, FTIR spectra were obtained. Fig. 2(a) shows that all the films displayed obvious infrared characteristic peaks. Among them, the absorption peaks at 1774 cm⁻¹ and 1712 cm⁻¹ correspond, respectively, to the stretching vibration and asymmetric stretching vibration of the CO double bond in the imide ring. The absorption peak at 1365 cm⁻¹ is due to the stretching vibration peak of the C–N–C bond, representing the formation of the imide ring.^32,33 When the cross-linked structure is introduced, a new C–N–C bond is formed, which increases the intensity of the original peak. No additional absorption peaks were observed in the FTIR spectra, indicating that the crosslinking agent only participated in the required crosslinking reaction.

The XRD pattern of the sample is shown in Fig. 2(b). MPI shows a weak and widely diffused scattering peak, indicating the amorphous structure of the polymer, and also demonstrating that the introduction of the cross-linked structure does not result in the localized crystallization of PI.³⁴ With the increase in the content of the crosslinking agent, the main diffraction peak of the MPI shifted to a high angle, from 2θ = 20.65° for PI to 2θ = 21.85° for MPI-1.5 wt%. The Bragg equation indicates that the main diffraction peak of the polymer is proportional to the intermolecular chain spacing.³⁵ The Bragg equation is given below in Formula 1:

2dsinθ = nλ	(1)

where d is the dimensional spacing, θ is the diffraction angle, λ is the X-ray wavelength (λ = 0.15406 nm), and n equals 1. Fig. S2 shows that the chain spacing of the MPI gradually decreases with the increase in the content of the cross-linking agent, indicating that the introduction of the cross-linking structure can effectively reduce the intermolecular chain spacing and make the molecular structure denser. ¹³C NMR tests were performed on the MPI films to further demonstrate the successful synthesis of MPI. As shown in Fig. 2(c), the characteristic peak at 165–170 ppm represents the formation of the imide ring, and the characteristic peak at 170 ppm indicates the successful reaction of MA with the PI chain. The formation of the cross-linked structure was again demonstrated by the analytical results of the ¹³C NMR spectra of the samples. To study the thermal stability of MPI, the glass transition temperature was tested by differential scanning calorimetry (DSC), as shown in Fig. S3. The glass transition temperature increased from 266.2 °C for PI to 281.6 °C for MPI-1.5 wt%. Thermogravimetric analysis (TGA) was conducted on MPI, as shown in Fig. S4. The temperatures of MPI were increased at 5% thermal weight loss (T_5%) compared to the pure PI, with a maximum temperature of 552.69 °C for MPI-1.5 wt%. The increase in DSC and TGA further proved that the introduction of a cross-linked structure made the molecular chain structure denser and improved the thermal stability of PI.

3.2. Capacitive energy storage performances of PI and MPI

To study the influence of the cross-linked structure on the dielectric properties of PI, the variation laws for the ε_r and tanδ of MPI at 20 °C and 150 °C (10⁰–10⁶ Hz) were tested, as shown in Fig. 3(a) and (b). As the frequency increased, the ε_r of MPI decreased to varying degrees, especially in the high-frequency region. Since the dipole turning polarization within the material cannot keep up with the changes in the external electric field, the ε_r decreased. The tanδ shows a trend of first decreasing and then increasing with the increase in frequency, which is related to the relaxation loss and conductance loss of the polar groups in the material. In the low-frequency region, due to carrier transition, the conductance loss increased. In the high-frequency region, the dipole turning polarization causes an increase in relaxation loss, thus increasing tanδ.³⁶ At the same frequency, ε_r and tanδ of MPI were both lower than those of pure PI; for example, at the same frequency of 1 kHz, the ε_r of MPI-1.5 wt% was 3.21, which is lower than 3.32 observed for pure PI. The decrease in ε_r is due to the suppression of dipole orientation polarization by the cross-linked structure. The tanδ decreased with the increase in the content of the crosslinking agent, which is attributed to the reduction of the molecular chain spacing caused by the cross-linked structure, inhibiting the polarization movement of the polar groups inside PI and reducing the polarization loss. When the temperature increased to 150 °C, the ε_r of MPI all decreased to varying degrees. This may be because of the enhanced depolarization ability of MPI with increasing temperature, resulting in a slight decrease in ε_r.^37,38 It is worth noting that MPI exhibits a more stable ε_r and tanδ than pure PI at a high temperature of 150 °C, indicating that MPI has better dielectric stability. To verify the polarization change, the relationship between the content of the crosslinking agent and the imaginary part of the dielectric modulus of the MPI with frequency was analyzed, as shown in Fig. S5. Due to the relatively poor insulation of pure PI, current carriers are transported more easily, resulting in greater polarization. On increasing the ε_r, a higher tanδ was also achieved. When the cross-linked structure was introduced, the insulation capacity of the MPI was enhanced, effectively hindering the transport of carriers and thereby reducing the polarization loss. This verifies the decrease in the ε_r and tanδ of the aforementioned MPI.

Excellent mechanical strength is an important condition for polymers in engineering applications. The influence of the cross-linked structure on the mechanical properties of PI is shown in Fig. 3(c) and (d). At 20 °C, the maximum tensile strength of pure PI is 103.6 MPa. When the cross-linked structure is introduced, the tensile strength of MPI is significantly increased, with the maximum increase reaching 135.1 MPa. Due to the positive relationship between mechanical strength and E_b, greater mechanical strength may impact the subsequent breakdown performance. When the temperature reached 150 °C, the maximum tensile strength of MPI was 90.84 MPa, which is 1.3 times that of pure PI. In addition, the elongation at break of the MPI was significantly improved, and the overall mechanical properties were excellent. Fig. S6 shows the Young's modulus of MPI at 20 °C and 150 °C. The Young's modulus of MPI-1.5 wt% at 20 °C is 3.89 GPa, which is 21.2% higher than that of pure PI. The increase in mechanical strength can be attributed to the cross-linked structure reducing the molecular chain spacing, making the molecular structure more compact and simultaneously reducing the inherent defects within the film. However, when the content of the cross-linking agent is too high, the unreacted free molecules will generate new physical defects within the film, which may start to be damaged from the defect sites during the stretching process, resulting in a decrease in mechanical strength.

The variation law of leakage current density of MPI with the electric field was tested and analyzed. The results are shown in Fig. 4(a) and (b). At the same temperature, the leakage current density of all films increased nonlinearly with the increase of the electric field intensity. However, under the same electric field intensity, the MPI showed a lower leakage current density. When the temperature is 20 °C and the applied electric field is 100 kV mm⁻¹, the leakage current density of the MPI-1.5 wt% film is as low as 8.53 × 10⁻¹² A cm⁻², which is much lower than 7.39 × 10⁻¹¹ A cm⁻² for pure PI. The reduction in leakage current density can be attributed to the following two points. Firstly, the positively charged cross-linking points act as traps and inhibit carrier transport. The reduction in the carrier motion rate leads to a reduction in the leakage current density, which verifies that the introduction of the cross-linked structure can create charge traps, as mentioned earlier.³⁹ In addition, the above XRD test results show that the introduction of a cross-linked structure reduces the molecular chain spacing and also impedes the movement of carriers, which is also an important factor in reducing the leakage current density. When the temperature rises, the leakage current density of each thin film increases. This is because the increase in temperature reduces the potential barrier height, making the transport of carriers easier to establish. When the temperature is 150 °C and the applied electric field is 100 kV mm⁻¹, the leakage current density of the PI-1.5 wt% film is 2.56 × 10⁻¹⁰ A cm⁻². Compared with 1.86 × 10⁻⁷ A cm⁻² of pure PI, the MPI still shows a relatively low leakage current density at high temperatures. The leakage current densities of PI and MPI-1.5 wt% at 150 °C were fitted, and it can be seen that the conduction mechanisms of both at high temperatures satisfy the hopping conduction as shown in Fig. S7. The equation in the figure is a simplified equation of the hopping conduction current density at high temperatures, where A and B are the lumped parameters.⁴⁰ Based on the fitting, λ was calculated to be 2.88 nm and 1.51 nm for PI and MPI-1.5 wt%, respectively, and λ is the jump distance. Shorter hopping distances correspond to larger trap densities, demonstrating that the introduction of a cross-linked structure forms more traps and improves the electron trapping ability of PI.

The influence of the cross-linked structure on the electrical properties of PI was tested. The Weibull breakdown strength values were obtained by using the two-parameter Weibull distribution, and the breakdown is as shown in Formula 2:

	(2)

where P(E) is the cumulative probability of breakdown failure, E_i is the actual electrical breakdown strength, E_b is the corresponding Weibull characteristic breakdown strength when the breakdown probability is 63.2%, and β is a shape parameter reflecting the dispersion of the experimental data, with a bigger β representing the higher reliability of the data. The E_b and Weibull distribution parameter β of the MPI were calculated and analyzed through the Weibull distribution function. Fig. 4(c) and (d) respectively show the Weibull distribution of MPI at high temperatures of 20 °C and 150 °C. At the same temperature, with the increase in the content of the cross-linking agent, the E_b of MPI first increases and then decreases. When the content of the crosslinking agent was 1.5 wt%, the maximum E_b of the PI-1.5 wt% film at 20 °C and 150 °C reached 546.73 kV mm⁻¹ and 471.53 kV mm⁻¹, respectively, which are 1.26 times and 1.32 times higher than pure PI. The increase in E_b can be attributed to two aspects as follows. On the one hand, due to the positive charge at the cross-linking points, they act as traps to suppress carrier conduction, delay the formation of the breakdown path, and reduce the breakdown probability.⁴¹ On the other hand, according to the electromechanical breakdown mechanism, the E_b of polymers is positively correlated with their Young's modulus. From the above tests on the mechanical strength of MPI, it is known that a higher Young's modulus of MPI indicates that it can withstand a higher electrostatic pressure under an electric field, thereby enhancing the E_b of PI. In addition, the increase in β indicates that the cross-linked structure improves the reliability of PI. However, when the content of the crosslinking agent exceeds 1.5 wt%, the E_b of the MPI begins to decrease. This might be due to the excessive crosslinking agent, which fails to effectively combine with PI molecules. The free crosslinking agent damages the molecular structure, causing local defects in PI. Under the action of the electric field, it is easier to form a breakdown path, resulting in a decrease in the E_b. As the temperature rises, the decrease in E_b is inevitable due to the increase in conduction loss and leakage current.⁴² Fig. S8 shows the comparison of E_b of MPI at different temperatures. When the temperature rises, the decrease in E_b of MPI is smaller than that of pure PI. This is because the introduction of the cross-linked structure generates more charge traps, making it easier for charges to be “captured” and reducing the possibility of breakdown failure.

To evaluate the influence of the cross-linked structure on the energy storage performance of PI capacitors, the displacement–electric field (D–E) loop curves of MPI at 20 °C and 150 °C under an external electric field of 200 kV mm⁻¹ were tested, as shown in Fig. 5(a) and (b) and Fig. S9. Under the same electric field, MPI has “thinner and narrower” D–E rings than pure PI; the area of the D–E ring represents the energy loss during the storage and release of electrical energy.⁴³ A smaller area indicates that the charge flow is suppressed under a high electric field, which means less energy loss and is conducive to improving the η. The maximum electrical displacement (D_max) of MPI is smaller than that of pure PI. For example, at 20 °C and 200 kV mm⁻¹, the D_max of MPI-1.5 wt% is 0.61 μC cm⁻², slightly lower than that of pure PI. The decrease of Dmax might be because the linked structure reduces the molecular chain spacing of PI, hinders the movement of polar groups, and suppresses the polarization of internal dipoles, corresponding to the previous decrease in ε_r. However, since the U_d and the square of the E_b are positively related, it is possible to increase the U_d by sacrificing some of the ε_r, significantly increasing the E_b.⁴⁴

When the temperature rises, the intensification of the electron transition increases the area of the D–E ring. Although a higher D–E ring area reduces η and U_d, compared with pure PI, the D–E ring area of MPI is smaller, indicating that the introduction of the cross-linked structure can effectively improve the η of PI at high temperatures. The U_d of the MPI was calculated using the D–E loop and the discharge energy density calculation formulae. The discharge energy density is as shown in Formula 3:

	(3)

where U_d is the discharge energy density, E is the breakdown strength, and D is the electrical displacement. Fig. 5(c) and (d) show that the introduction of the cross-linked structure effectively improves the energy storage characteristics of PI. Moreover, due to the relatively high E_b of the MPI-1.5 wt% film, it exhibits the best energy storage characteristics. The U_d at 20 °C reaches 3.89 J cm⁻³, which is 1.43 times that of pure PI, and the η remains above 80%. Even at 150 °C, the maximum U_d of the MPI-1.5 wt% film is 2.89 J cm⁻³, which is 1.45 times that of pure PI, while η remains above 70%. The increase in U_d of PI at room and high temperatures is due to the increase in E_b, while η is ensured due to the lower tanδ. When the content of the crosslinking agent is too high, the increase in molecular structure defects leads to a higher breakdown probability, which in turn results in a decline in energy storage performance. The excellent storage performance of MPI demonstrates the feasibility of using MPI for high-temperature capacitive energy storage.

Based on the excellent energy storage characteristics of MPI, its reliability and stability were evaluated. Fig. 5(e) shows the cycling characteristics of MPI. Compared with pure PI, the energy storage characteristics of MPI were significantly improved. At 200 kV mm⁻¹ and 150 °C, the U_d of the PI-1.5 wt% film was 0.604 J cm⁻³, which can withstand 10⁴ cycles, while the efficiency remained basically unchanged, indicating that the cross-linked structure improves the stability of PI. Charge and discharge tests are key indicators for evaluating the capacitance performance of polymer film capacitors. As shown in Fig. 5(f), the power density and charge and discharge rates of MPI were further evaluated through rapid discharge tests. The power density can be calculated using Formula 4:

	(4)

where P is the discharge power of the samples, V(t) is the voltage decay, R_L is the resistance of the external load, and t is the time. The time taken to achieve 90% discharge of the MPI-1.5 wt% film at 200 kV mm⁻¹ and 150 °C was 1.27 μs, and the power density reached 0.476 MW cm⁻³, proving that the energy storage performance of PI is improved by introducing the cross-linked structure.

The schematic diagram of the charge transport process within the MPI was completed based on the introduction of charge traps to attract electrons in the cross-linked structure, as shown in Fig. 6(a). Fig. 6(b) shows the surface electrostatic potential energy distribution of the MPI, where the blue region shows the negative charge and the red region shows the positive charge. It can be seen that the cross-linking points are positively charged. The introduction of a positively charged cross-linked structure at the cross-linking points forms charge traps inside PI. During the transport of electrons, due to the attractive effect of the charge traps on electrons, the transport path of electrons is changed, and the transport of electrons is inhibited. The inhibitory effect on electronic transport reduces the leakage current density and decreases the breakdown probability of the MPI, thereby enhancing the energy storage performance and operational reliability of the film capacitor.

4. Conclusion

In this paper, MPI films have been prepared using melamine as a crosslinking agent to modify PI. By introducing a cross-linked structure with positive charges at the cross-linking points, carrier migration is suppressed, the leakage current density is reduced, and the E_b is enhanced. The reduction in the molecular chain spacing inhibits polarization, lowers tanδ, and improves η, thereby enhancing the energy storage performance of PI, especially in high-temperature environments. At 150 °C, the maximum E_b of the MPI-1.5wt% film was 471.53 kV mm⁻¹, which is 1.32 times that of pure PI. The maximum U_d was 2.89 J cm⁻³, which is 1.45 times that of pure PI, and η remained above 70%. MPI also demonstrated excellent dielectric stability and mechanical properties, ensuring its reliability when operating under extreme conditions. In conclusion, the MPI prepared with melamine as the cross-linking agent demonstrates excellent performance and a simple preparation process, providing a new perspective for the research of commercialized high-temperature capacitive films.

Author contributions

Yu Feng: conceptualization, investigation, methodology, data curation, and writing – original draft. tianlong Liu: investigation, data curation, methodology, and writing – original draft. Jun Sun: methodology, formal analysis, and resources. Dongyu Hou: methodology, formal analysis, and writing – review & editing. Yanqing Wang: investigation, resources, and writing – review & editing. Dong Yue: conceptualization, methodology, funding acquisition, supervision, and writing – review & editing.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data that support the findings of this study are available within the article and its SI.

Fig. S1–S9 are provided. See DOI: https://doi.org/10.1039/d5tc02276b

Acknowledgements

This work was supported by the National Natural Science Foundation of China (No. 52407020), the Postdoctoral Fellowship Program of CPSF (No. GZC20230638), the China Postdoctoral Science Foundation (No. 2024MD753932), and the Natural Science Foundation of Heilongjiang Province of China (No. LH2024E087).

References

1. S. N. Liu, L. Tu, C. C. Liu, L. F. Tong, Z. X. Bai, G. Lin, K. Jia and X. B. Liu, J. Alloys Compd., 2022, 891, 161952 CrossRef CAS .
2. B. Q. Wan, M. S. Zheng, X. Yang, X. D. Dong, Y. C. Li, Y. W. Mai, G. Chen and J. W. Zha, Energy Environ. Mater., 2023, 6(1), e12427 CrossRef CAS .
3. Y. Chen, W. Zhu, X. Wang, W. Chen, W. Cai, Z. Xiao and J. Chen, Polym. Compos., 2022, 43(8), 5250–5259 CrossRef CAS .
4. E. Z. Feng, Y. Wang, L. Sun, W. L. Yang and H. G. Sun, Polym. Compos., 2025, 46(5), 4183–4196 CrossRef CAS .
5. D. Y. Hou, F. N. Bai, P. Dong, J. Chen, Y. T. Zhang, F. M. Meng, Z. H. Zhang, C. X. Zhang, Y. J. Zhang and J. Hu, J. Power Sources, 2023, 584, 233599 CrossRef CAS .
6. Z. Wu, H. Zhou, Q. Guo, Z. Liu, L. Gong, Q. Zhang, G. Zhong, Z. Li and Y. Chen, J. Alloys Compd., 2020, 845, 156171 CrossRef CAS .
7. H. Luo, C. F. Yan, X. Liu, H. Luo and S. Chen, ACS Appl. Mater. Interfaces, 2023, 15(20), 24470–24482 CrossRef CAS PubMed .
8. C. Y. Zhou, W. H. Xu, Y. H. Zhang, C. Yu, X. Liu, Z. L. Jiang, C. L. Zhang, Y. S. Shang and H. B. Zhang, ACS Appl. Mater. Interfaces, 2023, 15(6), 8471–8479 CrossRef CAS PubMed .
9. A. A. Deshmukh, C. Wu, O. Yassin, A. Mishra, L. H. Chen, A. Alamri, Z. Z. Li, J. R. Zhou, Z. Mutlu, M. Sotzing, P. Rajak, S. Shukal, J. Vellek, M. A. Baferani, M. Cakmak, P. vashishta, R. Ramprasad, Y. Cao and G. Sotzing, Energy Environ. Sci., 2022, 15(3), 1307–1314 RSC .
10. Y. N. Shang, Y. Feng, Z. T. Meng, C. H. Zhang, T. D. Zhang and Q. G. Chi, Mater. Horiz., 2024, 11(6), 1528–1538 RSC .
11. H. R. Xie, H. Luo, Y. Liu, R. Guo, X. B. Ji, H. S. Hou and D. Zhang, Chem. Eng. J., 2022, 433, 133601 CrossRef CAS .
12. D. Yue, Y. Feng, X.-X. Liu, J.-H. Yin, W.-C. Zhang, H. Guo, B. Su and Q.-Q. Lei, Adv. Sci., 2022, 9(17), 2105773 CrossRef CAS PubMed .
13. Q.-K. Feng, D.-F. Liu, Y.-X. Zhang, J.-Y. Pei, S.-L. Zhong, H.-Y. Hu, X.-J. Wang and Z.-M. Zhang, Nano Energy, 2022, 99, 107410 CrossRef CAS .
14. D. Y. Hou, J. Chen, F. N. Bai, P. Dong, C. X. Zhang, Y. J. Zhang and J. Hu, J. Power Sources, 2024, 613, 234845 CrossRef CAS .
15. B. H. Fan, M. Y. Zhou, C. Zhang, D. L. He and J. B. Bai, Prog. Polym. Sci., 2019, 97, 101143 CrossRef CAS .
16. X. W. Cao, N. H. Zhang, L. Zhao, C. H. Lu, C. H. Bi, G. J. He and W. Wu, Polym. Compos., 2024, 45(5), 4447–4458 CrossRef CAS .
17. D. Q. Tan, Adv. Funct. Mater., 2020, 30(18), 1808567 CrossRef CAS .
18. R. X. Gao, J. Xie, J. H. Yang, C. J. Zhuo, J. Z. Fu and P. Zhao, Polymer, 2021, 13(14), 2344 CAS .
19. M. J. Feng, Y. Feng, C. H. Zhang, T. D. Zhang, X. Tong, Q. Gao, Q. G. Chen and Q. G. Chi, Energy Environ. Mater., 2024, 7(2), e12571 CrossRef CAS .
20. D. Yue, J.-H. Yin, W.-C. Zhang, X.-X. Cheng, M.-H. Zhang, J.-J. Wang and Y. Feng, Adv. Funct. Mater., 2023, 33(30), 2300658 CrossRef CAS .
21. J. Fu, M. C. Yang, R. Wang, S. Cheng, X. Y. Huang, S. J. Wang, J. L. Li, M. X. Li, J. L. He and Q. Li, Mater. Today Energy, 2022, 29, 101101 CrossRef CAS .
22. Y. Liu, Y. F. Hou, Q. Ji, S. X. Wei, P. Du, L. H. Luo and W. P. Li, ACS Appl. Energy Mater., 2020, 3(3), 3015–3023 CrossRef CAS .
23. J. L. Li, X. X. Liu, Y. Feng and J. H. Yin, Prog. Polym. Sci., 2022, 126, 101505 CrossRef CAS .
24. J. Hu, Y. Liu, S. F. Zhang and B. T. Tang, Chem. Eng. J., 2022, 428, 131046 CrossRef CAS .
25. Y. J. Zhou, Z. W. Zhang, Q. F. Tang, X. Y. Ma and X. Hou, Mater. Horiz., 2024, 11(18), 4348–4358 RSC .
26. W. Dai, J. H. Yu, Y. Wang, Y. Z. Song, F. E. Alam, K. Nishimura, C.-T. Lin and N. Jiang, J. Mater. Chem. A, 2015, 3(9), 4884–4891 RSC .
27. D. Yue, W. C. Zhang, P. Z. Wang, Y. Zhang, Y. Teng, J. H. Yin and Y. Feng, Mater. Horiz., 2024, 11(3), 726–736 RSC .
28. L.-J. Zhang, J. Liu, L.-B. Luo, X. Y. Liu and X. Wang, Polymer, 2022, 245, 124702 CrossRef CAS .
29. J. W. Zha, M. Y. Xiao, B. Q. Wan, X. M. Wang, Z. M. Dang and G. Chen, Prog. Mater. Sci., 2023, 140, 101208 CrossRef CAS .
30. A. Xu, J. L. Zhou, C. F. Yan, Z. J. Liu, H. Luo and S. Chen, J. Energy Storage, 2024, 86, 111209 CrossRef .
31. Z. Pan, L. Li, L. Wang, G. Luo, X. Xu, F. Jin, J. Dong, Y. Niu, L. Sun, C. Guo, W. Zhang, Q. Wang and H. Wang, Adv. Mater., 2023, 35, 2207580 CrossRef CAS PubMed .
32. X. Li, Y. Xie, J. Xiong, D. Long, J. Zhang, B. Zhu, X. Zhang, X. Duan, Z. Liu, Z. Zhang and X. Huang, Chem. Eng. J., 2023, 466, 143324 CrossRef CAS .
33. P. H. Hu, W. D. Sun, M. Z. Fan, J. F. Qian, J. Y. Jiang, Z. K. Dan, Y. H. Lin, C.-W. Nan, M. Li and Y. Shen, Appl. Surf. Sci., 2018, 458, 743–750 CrossRef CAS .
34. J. Gu, Z. Lv, Y. Wu, Y. Guo, L. Tian, H. Qiu, W. Li and Q. Zhang, Composites, Part A, 2017, 94, 209–216 CrossRef CAS .
35. Z. Li, Z. Xie, Y. Han, H. Ren, Q. Li, X. Huang and Z. Wang, Appl. Surf. Sci., 2023, 634, 157666 CrossRef CAS .
36. W. Zhang, M. Gao, D. Hou, B. Hou, S. Liu, Q. Chen, J. Yin, D. Yue and Y. Feng, J. Appl. Polym. Sci., 2025, e57237 CrossRef CAS .
37. Y. Wei, L. Yang, C. Wang, Z. Zhu, Y. Dai, H. Qin and C. Xiong, Chem. Eng. J., 2025, 507, 160793 CrossRef CAS .
38. J. Dong, L. Li, Y. Niu, Z. Pan, Y. Pan, L. Sun, L. Tan, Y. Liu, X. Xu, X. Guo, Q. Wang and H. Wang, Adv. Energy Mater., 2024, 14(9), 2303732 CrossRef CAS .
39. M. Liu, J. Song, H. Qin, S. Qin, Y. Zhang, W. Xia, C. Xiong and F. Liu, Adv. Funct. Mater., 2024, 34(21), 2313258 CrossRef CAS .
40. J. Song, H. Qin, S. Qin, M. Liu, S. Zhang, J. Chen, Y. Zhang, S. Wang, Q. Li, L. Dong and C. Xiong, Mater. Horiz., 2023, 10(6), 2139–2148 RSC .
41. G. Duan, F. Hu, R. Zhang and Z. Hu, High. Voltage, 2023, 8(3), 630–639 CrossRef .
42. K. Zhang, Z. Ma, Q. Fu and H. Deng, Mater. Today Energy, 2022, 29, 101093 CrossRef CAS .
43. M. Feng, Y. Liu, X. Wu, Y. Xing and Q. Chi, Prog. Mater. Sci., 2025, 152, 101458 CrossRef CAS .
44. F. Wang, H. Wang, X. Shi, C. Diao, C. Li, W. Li, X. Liu, H. Zheng, H. Huang and X. Li, Chem. Eng. J., 2024, 485, 149972 CrossRef CAS .

This journal is © The Royal Society of Chemistry 2025