Enhanced energy storage performance of confined co-doping dielectric films via nanoparticle self-assembly in ferroelectric phases Authors: Kun Xing, Qi-Ze Han, Jian-Tao Wang, Lei Huang, Zhaoyu Ran, Li-Juan Yin, Shao-Long Zhong, Jingtong Lu, Baoquan Wan, Xin-Jie Wang, Limin Guo, Ke Bi, Zhi-Min Dang Published: 2026 ABSTRACT AbstractPolymer-based dielectrics are widely employed in electrostatic energy storage capacitors serving as pulse power supply owing to their lightweight nature and rapid charge-discharge capability. However, their intrinsically low dielectric constant severely limits energy storage density. Although high-dielectric-constant nanofillers are commonly incorporated to enhance permittivity, organic-inorganic interfacial incompatibility often induces particle agglomeration and structural defects. In this work, we propose a confined co-doping strategy for structured polymer dielectrics, wherein BaTiO3 and Al2O3 nanoparticles are co-doed within the ferroelectric core P(VDF-HFP) of coaxial fibers and undergo self-assembly. This approach simultaneously enhances both energy density and charge-discharge efficiency. As a result, the 1 wt% BaTiO3/1 wt% Al2O3 core co-doping composite dielectric achieves a discharged energy density of 19.2 J/cm3 and a charge-discharge efficiency of 81.0%, and maintains stable performance over 1 × 105 cycles under an electric field of 400 kV/mm. This confined co-doping strategy thus provides an effective and scalable route for developing polymer-based dielectrics with high energy density and high reliability.Subject terms: Ferroelectrics and multiferroics, Energy storage Introduction Dielectric capacitors, characterized by high power density and rapid charge-discharge capability, have been widely adopted in modern power systems, electric vehicles, and energy storage technologies1–4. Polymer-based dielectrics are particularly attractive due to their high breakdown strength, flexibility, and lightweight nature5–10. However, their intrinsically low dielectric constant severely restricts energy storage density, thereby hindering the miniaturization and integration of capacitive devices. The energy storage density of a linear dielectric can be expressed as Ue = 1/2ε0εrEb2, where Ue is the stored energy density, ε0 is the vacuum permittivity, and εr and Eb are the relative dielectric constant and breakdown strength, respectively11,12. Moreover, the charge-discharge efficiency also plays a critical role, defined as η = Ud/Ue, where Ud is the discharged energy density, representing the amount of energy effectively released. High charge-discharge efficiency improves the discharged energy density, whereas low efficiency can lead to safety concerns, such as Joule heating, which compromises long-term operational stability. Consequently, the development of polymer-based dielectrics that achieve both high discharged energy density and superior charge-discharge efficiency remains a critical challenge. To enhance energy storage performance, increasing both the dielectric constant and breakdown strength is essential, as indicated by the energy storage equation. Among polymeric materials, ferroelectric polymers such as poly(vinylidene fluoride) (PVDF) and its copolymers, including poly(vinylidene fluoride-co-hexafluoropropylene) (P(VDF-HFP)), and poly(vinylidene fluoride-trifluoroethylene-chlorofluoroethylene) (P(VDF-TrFE-CFE)), exhibit relatively high dielectric constants (10 – 50)13–15. However, their capacity to significantly increase the dielectric constant of composite dielectrics remains limited. In contrast, ceramic particles such as barium titanate (BaTiO3), strontium titanate (SrTiO3), and barium zirconium titanate (BZT) possess ultra-high dielectric constants (above 1000), making them effective fillers for enhancing the dielectric properties of polymer-based composites16,17. Consequently, polymer matrices doped with high-dielectric-constant (high-k) nanoparticles have become a focal point of research. Additionally, wide-bandgap inorganic fillers, such as boron nitride nanosheets (BNNS), aluminum oxide (Al2O3), titanium dioxide (TiO2), and silicon dioxide (SiO2), have been employed to improve charge-discharge efficiency by suppressing leakage currents18–21. Even so, achieving significant improvements generally requires a high volume fraction of inorganic fillers (≈10 vol%)22,23, which often leads to particle agglomeration, introducing mechanical defects and electric field distortions that degrade film quality and reduce energy storage performance. In multiphase nanoparticle doping, variations in particle properties and sizes exacerbate these issues, further reducing doping efficiency24,25. In this work, we proposed a confined co-doping strategy of inorganic nanoparticles within structured polymer dielectrics, which significantly improves the dielectric properties of composite dielectrics with low inorganic particle doping contents. By leveraging the self-assembly behavior of nanoparticles with varying sizes, this approach effectively mitigates nanoparticle agglomeration and enhances local interface dipole response, thereby significantly improving energy storage performance at low doping content. Upon our previous work on nano-submicron structured all-organic dielectrics (PMMA-P(VDF-HFP)@PMMA)26, we achieved spatially confined distribution of inorganic fillers in the core and sheath of coaxial fibers by coaxial electrospinning. Notably, in the core co-doping strategy, highly insulating Al2O3 nanoparticles (~20 nm) spontaneously adsorb onto the surfaces of BaTiO3 nanoparticles (~100 nm), mitigating electric field distortions at the organic-inorganic interface, suppressing Al2O3 particle agglomeration, and enhancing local interface dipole response. These self-assembled particles are confined within the core P(VDF-HFP) by coaxial electrospinning, while the pure sheath PMMA provides significant insulation. These synergistic effects result in the core co-doped dielectric exhibiting superior breakdown strength and charge-discharge efficiency, leading to exceptional energy storage performance. Specifically, the 1 wt% BaTiO3/1 wt% Al2O3 core co-doped dielectric achieves a remarkable discharged energy density of 19.2 J/cm3 and an outstanding charge-discharge efficiency of 81.0% at a high electric field of 700 kV/mm, alongside excellent cycling stability of up to 1 × 105 cycles at 400 kV/mm. Results and discussion Confined Co-doping strategies of nanoparticles and characterization In our previous work, we developed a structured polymer dielectric featuring a nano-submicron PMMA surface layer and a coaxial P(VDF-HFP)@PMMA interlayer via coaxial electrospinning26. To preserve the integrity of the coaxial structure in the interlayer, a hot-pressing temperature of 180 °C was optimized from a range of tested temperatures. The fabrication process of the nanoparticle confined doped dielectric is illustrated in Fig. 1a, b. Pure PMMA fibers were electrospun to form the insulating surface layer, while P(VDF-HFP)@PMMA coaxial fibers were electrospun as the interlayer. The PMMA and P(VDF-HFP) precursor solutions during the coaxial electrospinning process have been doped with inorganic particles to achieve confined doping, while the surface PMMA fibers remain undoped to maintain their insulating properties. Design schematic and characterization of organic-inorganic composite dielectrics. aSchematic illustration of the fabrication process for composite dielectric films via coaxial electrospinning followed by hot pressing. b Schematic of the coaxial fiber structure and inorganic nanoparticle distribution strategies. c Schematic illustration of the carrier migration suppression by Al2O3 nanoparticles and the sheath PMMA. d Schematic illustration of space charge injection suppression by the PMMA surface layer. TEM and EDS images of coaxial fibers doped with BaTiO3 in the core (3 wt% BaTiO3) (e), Al2O3 in the sheath (1 wt% BaTiO3-1 wt% Al2O3) (f), and core co-doped with BaTiO3 and Al2O3 (1 wt% BaTiO3/1 wt% Al2O3) (g). h Magnified EDS mapping image of the core co-doped BaTiO3 and Al2O3 (1 wt% BaTiO3/1 wt% Al2O3). Black granular outlines indicate inorganic nanoparticles, while yellow, red, and pink denote the EDS signals of Ba, O, and Al, respectively. The arrow below indicates the progression of time. Given that the dielectric constant mismatch between BaTiO3 and P(VDF-HFP) is smaller than that between BaTiO3 and PMMA. Therefore, BaTiO3 was selectively doped into the P(VDF-HFP) phase. However, excessive inorganic particle loading can compromise film quality and reduce breakdown strength27,28. Thus, optimizing the BaTiO3 doping content is essential to achieve a balanced trade-off between dielectric constant and breakdown performance. As illustrated in Fig. 1b1, BaTiO3 nanoparticles with varying mass fractions were doped into the core P(VDF-HFP) to systematically investigate their influence on the dielectric properties of the resulting films, representing the core-doped BaTiO3 strategy. The samples were designated as 1 wt% BaTiO3, 3 wt% BaTiO3, etc. Upon this, Al2O3 nanoparticles were introduced as a secondary phase to further improve the insulation characteristics of the dielectric. In the layer doping strategy (Fig. 1b2), Al2O3 nanoparticles were distributed within the sheath PMMA (Al2O3 particles are dispersed in the sheath PMMA precursor), yielding samples such as 1 wt% BaTiO3-1 wt% Al2O3 and 1 wt% BaTiO3-3 wt% Al2O3, etc. In the core co-doping strategy (Fig. 1b3), both BaTiO3 and Al2O3 nanoparticles were doped into the core P(VDF-HFP) (BaTiO3 and Al2O3 particles are dispersed in the core P (VDF-HFP) precursor), forming samples such as 1 wt% BaTiO3/1 wt% Al2O3 and 1 wt% BaTiO3/3 wt% Al2O3, etc. The wide-bandgap of Al2O3 particles and the sheath PMMA can effectively suppress carrier transport within the dielectric, thereby reducing conduction losses, as shown in Fig. 1c. Concurrently, the wide-bandgap of the nano-submicron PMMA surface layer can hinder space charge injection (Fig. 1d). Therefore, the synergistic effect of Al2O3 and PMMA can significantly suppress dielectric losses and improve charge-discharge efficiency. The electrospinning process employed here offers remarkable flexibility and controllability, enabling the precise fabrication of fine and complex microstructures suitable for confined nanoparticle doping. Figure 1e–h presents transmission electron microscopy (TEM) and energy-dispersive spectroscopy (EDS) images of coaxial fibers employing various doping strategies. The coaxial structure of the fibers is distinctly visible in the TEM images. Figure 1e shows BaTiO3 nanoparticles doped solely in the core P(VDF-HFP) of coaxial fibers, with TEM and EDS mapping confirming their distribution and presence, respectively. Figure 1f depicts Al2O3 nanoparticles doped into the sheath PMMA, with TEM and EDS mapping confirming their distribution and presence. However, due to the nanoscale size of Al2O3, aggregation is observed within the sheath. Figure 1g illustrates BaTiO3 and Al2O3 nanoparticles co-doped in the core P(VDF-HFP), with TEM and EDS mapping revealing their distribution and self-assembly, respectively, where Al2O3 particles adsorb on BaTiO3 particle surfaces. Figure 1h provides a magnified EDS mapping image of the Al element, which more clearly confirms the distribution of Al2O3 particles on the BaTiO3 surface. Figure S1 presents the TEM and EDS images of core co-doped fibers with different inorganic compositions, further demonstrating the reproducibility of the nanoparticle self-assembly behavior in the core co-doping strategy. To visualize the nanoparticle distribution over a larger area and to better evaluate the uniformity of dispersion, low-magnification TEM images of the core co-doped fibers were provided, as shown in Fig. S2. To elucidate the driving force for nanoparticle self-assembly from a thermodynamic perspective, we systematically characterized the surface free energy and corresponding components of BaTiO3 and Al2O3 particles, as well as the surface tension and its corresponding components of the P(VDF-HFP)/DMF solution, as shown in Figs. S3–S6. Based on these data (Tables S1–S6), we further calculated the interfacial free energies of BaTiO3/Al2O3, BaTiO3/P(VDF-HFP)/DMF, and Al2O3/P(VDF-HFP)/DMF, with the relevant results summarized in Table S7. The detailed calculation principles are explained in Supplementary Notes 4–6. The results show that the interfacial free energy between BaTiO3 and Al2O3 is lower than that between either BaTiO3 or Al2O3 and the P(VDF-HFP)/DMF solution, indicating that the preferential adsorption of Al2O3 onto the BaTiO3 surface is thermodynamically favorable29,30. To further verify the critical role of spatial confinement in the self-assembly process, we prepared a dielectric film with the same composition, namely 1 wt% BaTiO3/1 wt% Al2O3-P(VDF-HFP), using a solution-casting method, and examined the distribution of the nanoparticles inside the film by TEM, as shown in Fig. S7. The results indicate that although local self-assembly can still be observed in the cast film due to the thermodynamic driving force, obvious Al2O3 aggregation is also present near the self-assembled particles. This suggests that thermodynamic spontaneity alone is insufficient to generate the self-assembled structure shown in Fig. S1. The spatially confined environment, especially the locally restricted microenvironment provided by the PMMA sheath layer of the coaxial structure, is indispensable for achieving stable and uniform self-assembly. The present work revealing the decisive role of thermodynamic dynamics and spatial confinement in inducing and stabilizing nanoparticle self-assembly. X-ray diffraction (XRD) patterns of BaTiO3 and Al2O3 particles are presented in Fig. S8, confirming their high phase purity. Figure S9 displays Fourier transform infrared (FTIR) spectra of dielectrics with varying doping strategies. Characteristic peaks at 767, 792, and 976 cm−1 correspond to the crystalline α-phase of P(VDF-HFP), while peaks at 843 and 881 cm−1 are attributed to the β-phase31. Additionally, peaks at 1730 cm−1 to C = O bonds, and 1200 and 1174 cm−1 to C-O bonds, characteristic of PMMA32. Dielectric properties of polymer dielectrics doped with BaTiO3 nanoparticles in the core A series of dielectric films were prepared with BaTiO3 nanoparticles doped into the core P(VDF-HFP) to investigate the influence of varying BaTiO3 doping contents on dielectric properties. Figure S10 presents scanning electron microscopy (SEM) images of the cross section of core-doped BaTiO3 dielectrics, clearly revealing a PMMA surface layer with a thickness of approximately 400 nm. At BaTiO3 doping contents of 1 wt% and 3 wt%, the polymer matrix exhibits minimal defects, with relatively low defect density. However, as the doping content increases from 5 wt% to 9 wt%, the nanoparticle density rises significantly, resulting in a substantial increase in defect density (Fig. S11). In the 9 wt% BaTiO3 dielectric, a high local concentration of inorganic particles and significant porosity were observed, primarily due to the excessive inorganic filler content and pronounced mechanical property disparities between the inorganic nanoparticles and the polymer matrix. Nonetheless, no agglomeration is observed, underscoring the efficacy of the coaxial fiber architecture in preventing inorganic particle clustering. Figure 2a presents the dielectric constant and dielectric loss of core-doped BaTiO3 dielectrics and all-organic structured dielectric across a frequency range of 1 Hz to 10 MHz. The dielectric constant of the 1 wt% BaTiO3 dielectric is approximately 6.1, surpassing the 5.7 of the all-organic structured dielectric26. As the BaTiO3 doping content increased to 9 wt%, the dielectric constant rises to 7.2, attributed to the Maxwell-Wagner effect33,34, where interfacial polarization between BaTiO3 and the polymer matrix enhances the dielectric constant of the composite. Notably, the dielectric loss remains relatively low (0.04-0.06) at low frequencies but gradually increases at higher frequencies, due to interfacial charge relaxation within the organic-inorganic interface and dipolar polarization hysteresis in P(VDF-HFP). Comparison of the dielectric properties for core-doped BaTiO3 dielectrics and all-organic structured dielectric. Frequency-dependent dielectric constant and dielectric loss (a), Weibull distribution of breakdown strength (b), and maximum D-E curves (c) of the dielectrics. d Discharged energy density and charge-discharge efficiency of the dielectrics as a function of electric field. To assess the characteristic breakdown strength of dielectrics, a two-parameter Weibull statistical distribution was employed, expressed as P(E) = 1-exp[-(E/Eb)β], where P(E) is the cumulative probability of electrical failure, E is the experimental breakdown strength, Eb is the characteristic breakdown strength (corresponding to a 63.2% failure probability), and β is the shape parameter reflecting data dispersion. The breakdown performance of core-doped BaTiO3 dielectrics and all-organic structured dielectric are shown in Fig. 2b. For doping contents increasing from 1 to 3 wt%, the breakdown strength decreases modestly from 623 kV/mm to 597 kV/mm, approaching that of the all-organic dielectric (644 kV/mm). However, at a doping contents of 5 wt%, the breakdown strength drops sharply to 470/kV mm, with further increases in doping contents yielding no significant additional reduction. This trend aligns with SEM cross section images (Fig. S10), which reveal uniform nanoparticle distribution and minimal defects in the 1 wt% BaTiO3 and 3 wt% BaTiO3 dielectrics. In contrast, the 5 wt% BaTiO3, 7 wt% BaTiO3, and 9 wt% BaTiO3 dielectrics exhibit a significant increase in nanoparticle density, with significant porosity observed in the 9 wt% BaTiO3 dielectric, severely compromising breakdown performance. Figure S12 displays the electric displacement-electric field (D-E) curves for core-doped BaTiO3 dielectrics, with maximum polarization values summarized in Fig. 2c. The core-doped BaTiO3 dielectrics exhibit substantially enhanced polarization capability compared with all-organic structured dielectrics, attributable to the doped of BaTiO3 nanoparticles. Higher BaTiO3 doping contents result in steeper D-E curves, indicating greater electric polarization displacement under equivalent electric fields, consistent with the observed increase in dielectric constant. Additionally, at higher electric field, the D-E curves exhibit larger intercepts with the ordinate and a broader profile. This indicates that the increase in remanent polarization intensity, caused by ferroelectric polarization characteristics and irreversible interfacial polarization (due to space charge accumulation), also reflects the decrease in charge-discharge efficiency. The discharged energy density and charge-discharge efficiency of core-doped BaTiO3 dielectrics and all-organic structured dielectric are presented in Fig. 2d. Due to the enhanced polarization capability and high breakdown strength, the 1 wt% BaTiO3 and 3 wt% BaTiO3 dielectrics achieve superior discharged energy densities of 17.9 J/cm3 (at 740 kV/mm) and 14.7 J/cm3 (at 650 kV/mm), surpassing the 13.7 J/cm3 achieved by the all-organic structured dielectric. However, their charge-discharge efficiencies are only 70% and 74%, respectively, which are lower than the 80% achieved by the all-organic structured dielectric. Although the 5 wt% BaTiO3, 7 wt% BaTiO3, and 9 wt% BaTiO3 dielectrics exhibit higher polarization capability, their lower breakdown strengths result in discharged energy densities of only 8.7 J/cm3, 9.9 J/cm3, and 9.8 J/cm3, respectively, at low electric fields. The relationship between dielectric constant and breakdown strength as a function of BaTiO3 doping contents is plotted in Fig. S13. Based on the energy storage formula and the experimental results, it is evident that breakdown strength exerts a greater influence on energy storage performance than dielectric constant. Consequently, the 1 wt% BaTiO3 and 3 wt% BaTiO3 dielectrics were selected for subsequent Al2O3 nanoparticle doping to improve the charge-discharge efficiency of the dielectric. Dielectric properties of polymer dielectrics confined Co-doped with Al2O3 and BaTiO3 nanoparticles To enhance the charge-discharge efficiency of polymer-based dielectrics doped with BaTiO3 nanoparticles in the core P(VDF-HFP), wide-bandgap Al2O3 nanoparticles were doped as a second-phase dopant. The effect of doping position for Al2O3 nanoparticles on the dielectric properties was investigated using two strategies: layer doping and core co-doping. SEM images of the cross sections of dielectric films prepared using two strategies are shown in Figs. S14–S16. Regardless of the Al2O3 doping content, particle agglomeration is evident in the layer-doped dielectrics (Figs. S14, S15). In contrast, the core co-doping strategy exhibits excellent particle dispersion with no agglomeration (Fig. S16). This improved dispersibility is primarily attributed to the self-assembly behavior between BaTiO3 and Al2O3 nanoparticles co-doped within the core, which effectively suppresses the tendency of Al2O3 nanoparticles to agglomerate due to organic-inorganic incompatibility (Fig. 1f). Consequently, without requiring any additional surface modification of the inorganic nanoparticles, the space-confined core co-doping strategy can induce nanoparticle self-assembly, thereby enabling a highly dispersed distribution of inorganic fillers within the polymer matrix. This improved dispersion state is one of the key factors for reducing structural defects and enhancing the breakdown strength of the composite dielectric. Additionally, the core co-doping strategy facilitates the isolation of inorganic particles within the core P(VDF-HFP) by the insulating sheath PMMA, further preventing agglomeration. The UV-visible absorption spectra of Al2O3 and BaTiO3, along with their bandgaps calculated using the intercept method35, are presented in Fig. 3a. The bandgap of Al2O3 is 5.4 eV, significantly higher than that of BaTiO3 (3.0 eV) and P(VDF-HFP) (2.2 eV)26. We used ultraviolet photoelectron spectroscopy (Figure S17) together with UV-vis absorption spectroscopy to construct the band structure of the BaTiO3/Al2O3 heterointerface (Fig. 3a). Owing to the energy-level alignment and band offset at the BaTiO3/Al2O3 contact interface, localized charge traps are formed (Fig. 3b). These trap centers can capture free electrons or holes and confine them near the heterogeneous interface (Fig. 3c), thereby reducing the number of carriers migrating and the distance they travel, which is beneficial for improving the charge-discharge efficiency of the dielectrics. In addition, we performed thermally stimulated depolarization current (TSDC) measurements on the all-organic dielectric, the 1 wt% BaTiO3/0.5 wt% Al2O3, 1 wt% BaTiO3/1 wt% Al2O3, and 1 wt% BaTiO3/2 wt% Al2O3 dielectrics to compare their trap densities and thereby evaluate their ability to suppress internal leakage current, as shown in Fig. S18. The results show that, compared with the all-organic dielectric, the 1 wt% BaTiO3/0.5 wt% Al2O3 dielectric exhibits an increased trap density, but it remains significantly lower than that of the 1 wt% BaTiO3/1 wt% Al2O3 dielectric. In contrast, both the 1 wt% BaTiO3/1 wt% Al2O3 and 1 wt% BaTiO3/2 wt% Al2O3 dielectrics show a pronounced increase in trap density, indicating that the trap effect introduced by the heterointerfaces is significantly enhanced with increasing Al2O3 content. Although the 1 wt% BaTiO3/2 wt% Al2O3 dielectric exhibits a slightly higher trap density than the 1 wt% BaTiO3/1 wt% Al2O3 dielectric, the difference between them is no longer substantial and does not represent a significant performance improvement. On the other hand, a further increase in inorganic filler content would inevitably introduce more interfacial defects, which is unfavorable for maintaining a high breakdown strength. Therefore, when considering the balance between breakdown strength and charge-trap density, the 1 wt% BaTiO3/1 wt% Al2O3 composition represents the optimal formulation for the core co-doping strategy. Dielectric properties of dielectrics employing different confined co-doped strategies. a UV-visible absorption spectra of BaTiO3 and Al2O3, with bandgaps calculated using the intercept method. (Eg = hc/λg, where λg is the wavelength, h is the Planck constant, and c is the speed of light). b Band structure of the particle self-assembly interface. c The schematic diagram of a charge trap capturing space charge or hole. d, Dielectric constant and dielectric loss at 1 kHz for layer-doped and core co-doped dielectrics as a function of Al2O3 doping content. e Breakdown strength of layer-doped and core co-doped dielectrics at varying Al2O3 doping content. f Discharged energy density and charge-discharge efficiency of the core co-doped dielectrics. g Charge-discharge cycling performance of the core co-doped dielectric 1 wt% BaTiO3/1 wt% Al2O3. h, Comparison of discharged energy density and charge-discharge efficiency between this work and recent reports. Figures S19 and S20 present the frequency-dependent dielectric constant and dielectric loss of dielectrics employing layer doping and core co-doping strategies over a frequency range of 1 Hz–10 MHz. The dielectric loss for both doping strategies remains low and stable at low to medium frequencies but increases significantly at high frequencies for the layer-doped dielectrics. This increase is attributed to enhanced interfacial defects and voids resulting from Al2O3 particle agglomeration, which forms localized tunneling or leakage current pathways. Figure 3d illustrates the dielectric constant and dielectric loss at 1 kHz for layer-doped and core co-doped dielectrics as a function of Al2O3 doping content (extracted from Figs. S19 and S20). At equivalent inorganic particle contents, the core co-doped dielectrics exhibit a slightly higher dielectric constant compared to their layer-doped counterparts (Fig. S21). These results indicate that the nanoparticle self-assembly behavior introduces an additional polarization effect. To further clarify its origin, we systematically investigated the remanent polarization of dielectrics prepared using different doping strategies. We analyzed the remanent polarization using D-E loops measured at different frequencies, as shown in Fig. S22. Specifically, since ferroelectric polarization can respond to high-frequency signals, whereas space-charge-related polarization mainly appears in the low-frequency region, the remanent displacement of the 1 wt% BaTiO3/1 wt% Al2O3 dielectric at 10,000 Hz (0.08 μC/cm2) can be approximately regarded as the contribution from remanent ferroelectric polarization. In contrast, the remanent displacement at 100 Hz (0.14 μC/cm2) contains contributions from both remanent ferroelectric polarization and remanent polarization induced by space-charge accumulation. For comparison, the remanent polarization of the 1 wt% BaTiO3 dielectric at 100 Hz is 0.27 μC/cm2. Since ferroelectric remanent polarization is not easily affected by frequency and other factors, the reduction in remanent polarization in the 1 wt% BaTiO3/1 wt% Al2O3 dielectric compared to the 1 wt% BaTiO3 dielectric is mainly due to space charge-related factors, which reduced it by 0.13 μC/cm2, equivalent to a reduction of about 48%. These results indicate that, although the 1 wt% BaTiO3/1 wt% Al2O3 dielectric still exhibits a certain ferroelectric response at high electric fields, accompanied by an increase in remanent polarization, the space-charge transport and accumulation processes that primarily govern dielectric loss, Joule heating, and cycling degradation have been effectively suppressed. Combined with the above analysis of remanent polarization components and the TSDC results, the reduction in remanent polarization indicates that the deleterious space-charge effects associated with long-range charge migration are effectively suppressed by the charge traps within the dielectric. Therefore, the slight increase in dielectric constant of the 1 wt% BaTiO3/1 wt% Al2O3 dielectric mainly originates from the interfacial dipolar response at the BaTiO3/Al2O3 heterointerface. In other words, the self-assembled heterointerfaces enhance the beneficial local interfacial polarization while simultaneously suppressing unfavorable space-charge accumulation and transport processes. Figure 3e shows the breakdown performance of dielectrics with different doping strategies (extracted from Figs. S23 and S24). Within the same doping strategy, dielectrics with higher inorganic particle content exhibit lower breakdown strength. Moreover, at identical inorganic particle contents, the core co-doped dielectrics demonstrate significantly superior breakdown performance compared to layer-doped dielectrics. For instance, the breakdown strengths of the 1 wt% BaTiO3/1 wt% Al2O3, 1 wt% BaTiO3/3 wt% Al2O3, 3 wt% BaTiO3/1 wt% Al2O3, and 3 wt% BaTiO3/3 wt% Al2O3 dielectrics reach 619, 494, 578, and 437 kV/mm, respectively. These values represent improvements of 37%, 22%, 47%, and 20% compared to the corresponding layer-doped dielectrics (451, 406, 393, and 364 kV/mm). This marked difference primarily arises from the distribution of inorganic particles across both the core and sheath in layer-doped dielectrics, which promotes forms particle agglomeration. However, in the core co-doping strategy, the self-assembly behavior of the nanoparticles enhances their dispersion, and the highly insulating PMMA sheath further isolates them, thereby contributing to the superior breakdown performance. Additionally, the data suggest that Al2O3 particles have a more pronounced negative impact on breakdown performance compared to BaTiO3 particles. Consequently, the Al2O3 doping content in core co-doped dielectrics was limited to 3 wt%. Figures S25–S36 show the D-E loops and the corresponding energy-storage performance of the co-doped dielectrics. The discharged energy density and charge-discharge efficiency extracted from these measurements are further summarized in Figs. S37 and 3f for the layer-doped and core co-doped dielectrics, respectively. For dielectrics prepared using the same co-doping strategy, the differences in dielectric constant are relatively small, and therefore their discharged energy density under the same electric field are generally similar. However, as the nanoparticles content increases, the breakdown strength decreases, which ultimately limits the discharged energy density. By contrast, the most significant differences between different co-doping strategies lie in their breakdown behavior and interfacial effects. Compared with the layer-doped dielectrics, the self-assembly of inorganic nanoparticles in the core co-doped dielectrics not only improves particle dispersion and optimizes the microstructure, thereby markedly enhancing the breakdown strength, but also strengthens the interfacial dipolar response through the BaTiO3/Al2O3 heterointerfaces, which is beneficial for the dielectric response. Therefore, the synergistic enhancement of interfacial dipolar response and breakdown strength leads to the improved the discharged energy density. In addition, the abundant charge traps introduced by the self-assembled heterointerfaces effectively suppress carrier transport and space-charge accumulation, thereby significantly improving the charge-discharge efficiency. Specifically, the 1 wt% BaTiO3-1 wt% Al2O3 and 3 wt% BaTiO3−3 wt% Al2O3 achieve discharged energy densities of 6.7 J/cm3 and 6.5 J/cm3 at 450 kV/mm and 400 kV/mm, respectively, with corresponding charge-discharge efficiencies of 78.4% and 92.5%. The 1 wt% BaTiO3/1 wt% Al2O3 and 3 wt% BaTiO3/1 wt% Al2O3 dielectrics achieving exceptional discharged energy densities of 19.2 J/cm3 and 15.7 J/cm3 at high electric fields of 700 kV/mm and 600 kV/mm, respectively, with charge-discharge efficiencies of 81.0% and 80.5%. To evaluate long term operational reliability, charge-discharge cycling tests were conducted on the 1 wt% BaTiO3/1 wt% Al2O3 dielectric films. As shown in Fig. 3g, the 1 wt% BaTiO3/1 wt% Al2O3 dielectric maintains stable energy-storage performance after 1 × 105 charge-discharge cycles under an electric field of 400 kV/mm, with a discharged energy density of 6.4 J/cm3 and a charge-discharge efficiency of 87%. Under 300 kV/mm, the dielectric exhibits an even longer cycling lifetime exceeding 6.3 × 105 charge-discharge cycles, demonstrating excellent cycling stability (Fig. S38). In addition, after 50 thermal cycles alternating between 40 °C and 80 °C, the dielectric still sustains stable charge-discharge cycling for 5 × 104 cycles under 400 kV/mm, further confirming its good thermal-cycling stability (Fig. S39). Figure 3h and Table S8 summarize the energy storage performance and charge-discharge efficiency of recent all-organic and organic-inorganic composite dielectrics36–47. Notably, the 1 wt% BaTiO3/1 wt% Al2O3 dielectric combines the high discharge energy density of organic-inorganic composites with the high charge-discharge efficiency characteristic of all-organic dielectrics. This superior performance is primarily attributed to the dual-phase inorganic particle core co-doping strategy, where BaTiO3 nanoparticles provide high polarization, and wide-bandgap Al2O3 nanoparticles enhance charge-discharge efficiency by suppressing carrier migration. More importantly, the particle self-assembly behavior in the core co-doping strategy simultaneously enhances the dielectric constant and mitigates the breakdown deterioration typically induced by inorganic fillers, thereby markedly improving the discharge energy density and charge-discharge efficiency even at low doping content. Leakage current and finite element simulation analysis Charge transport behavior was learnt through analysis of the leakage current as a function of the applied electric field, and the results showed that the Al2O3 particles and the charge traps introduced by self-assembly can significantly suppress leakage current (Figs. S40 and S41). Fitting of the current density-electric field show that as the electric field increases from 20 kV/mm to 200 kV/mm, the conduction mechanism changes, shifting from Schottky emission (dominated at low field) to hopping conduction (dominated at high field) (Supplementary Notes 1,2, and Figs. 4a, b and S42). Under low field, the Schottky fitting yielded smaller intercept values for both the core co-doped and layer-doped dielectrics compared to the core-doped BaTiO3 dielectric, indicating that Al2O3 elevates the electrode/dielectric interface barrier. The calculated dielectric constant of the organic-inorganic composite dielectric are shown in Table S10, very close to the experimental data, confirming that Schottky emission is the main conduction mechanism in the low field region. At high fields, fitting results of the hopping conduction model showed markedly reduced hopping distances in the core co-doped and layer-doped dielectrics (λ = 0.66–1.53 nm) relative to core-doped BaTiO3 dielectrics (λ = 2.21 nm for 1 wt% BaTiO3, λ = 2.57 nm for 3 wt% BaTiO3), suggesting that the wide-bandgap Al2O3 and charge traps effectively suppresses carrier mobility (Table S11). Consequently, the doped of Al2O3 simultaneously suppresses the charge transport dominated by Schottky emission and hopping conduction, thereby improving the charge-discharge efficiency of the dielectrics. Leakage current analysis and breakdown path simulation of dielectrics with different doping strategies. Leakage current fitting of the core-doped BaTiO3 dielectric, core co-doped dielectric (a) and layered-doped dielectric (b) based on the Schottky emission and hopping conduction models. Phase-field simulation of the breakdown path in the core-doped BaTiO₃ dielectric (c), the layered-doped dielectric (d), and the core co-doped dielectric (e). In the simulation diagrams, large circles represent the P(VDF-HFP) cores of the coaxial fibers, medium-sized circles represent BaTiO3 particles, small circles represent Al2O3 particles, and ellipses represent aggregated Al2O3 particles. Blue lines indicate the growing electrical trees. The color scale on the right denotes the integrity of the dielectric, with red indicating the intact state and blue indicating damage or failure. To further elucidate the influence of different doping strategies on the breakdown performance of dielectric films, the breakdown process was simulated using the phase-field method (Supplementary Note 3, Table S12), as shown in Fig. 4c–e. The red regions correspond to intact areas, whereas the blue regions indicate failure. For the core-doped BaTiO3 dielectrics, the BaTiO3 are effectively isolated by the highly insulating PMMA, resulting in a slow progression of the electrical tree path over time. In the co-doping strategy, due to the aggregation of Al2O3 particles, early failure occurs at the Al2O3 aggregation sites in the layer-doped dielectric, and the electrical tree path develops rapidly. This observation corroborates the earlier conclusion that nanoparticle agglomeration introduces substantial defect, thereby reducing the breakdown strength. For the core co-doped dielectric, the self-assembly behavior of the biphase nanoparticles yields a more dispersed particle distribution, while the sheath PMMA further isolates the inorganic fillers. Consequently, the development of the electrical tree path is significantly suppressed, resulting in enhanced breakdown performance. Thermodynamic and mechanical properties of the1 wt% BaTiO3/1 wt% Al2O3 dielectric As shown in Fig. S43, PMMA, as an amorphous polymer, exhibits a transition near 120 °C, corresponding to its glass-transition process. The semicrystalline P(VDF-HFP) shows an endothermic peak near 130 °C associated with crystalline melting, together with a weaker feature around 80 °C related to relaxation in defective crystalline regions and local segmental rearrangement. By comparison, the structured all-organic dielectric and the 1 wt% BaTiO3/1 wt% Al2O3 dielectric exhibit only a weak endothermic feature near 55 °C, indicating more stable thermal behavior due to the stronger restriction of molecular-chain motion by the coaxial interfaces and inorganic nanoparticles. The DMA results (Fig. S44) further support this conclusion. Compared with the all-organic dielectric, the 1 wt% BaTiO3/1 wt% Al2O3 dielectric shows a clear shift of the tan δ main peak toward higher temperature, along with a sharper peak shape, while maintaining a higher storage modulus over a broader high-temperature range. These results indicate that the introduction of inorganic nanoparticles enhances the interfacial constraint on polymer-chain motion, suppresses thermal segmental mobility, and increases the thermal activation energy of the main relaxation process. Consequently, the composite dielectric exhibits a higher α-relaxation temperature and improved thermomechanical stability. Overall, the DSC and DMA results demonstrate that the incorporation of inorganic nanoparticles improves both the thermal stability and thermomechanical stability of the structured dielectric. As shown in Fig. S45, nanoindentation measurements were performed on the all-organic dielectric and the 1 wt% BaTiO3/1 wt% Al2O3 dielectric. Owing to the incorporation of inorganic nanoparticles, the 1 wt% BaTiO3/1 wt% Al2O3 dielectric exhibits a higher Young’s modulus than the corresponding all-organic dielectric. In combination with the DMA results, the composite dielectric also demonstrates improved mechanical stability. These enhanced mechanical properties are beneficial for resisting electric-field-induced mechanical stress and structural instability under ultrahigh electric fields, thereby supporting the observed high breakdown performance. To investigate the influence of inorganic-particle incorporation on the flexibility and density of the polymer, we compared the all-organic dielectric with the 1 wt% BaTiO3/1 wt% Al2O3 dielectric, as shown in Fig. S46, and both films retain excellent flexibility. In addition, the densities of the two dielectrics were measured, and the results show that the low filler content causes only a slight increase in density, as summarized in Table S9. These results demonstrate that the incorporation of low-content inorganic nanoparticles does not significantly impair the flexibility or lightweight characteristics of the dielectric layer. In conclusion, upon our previous work, we propose a confined co-doping strategy for biphase inorganic nanoparticles into structured polymer dielectric. In the core co-doping strategy, the self-assembly behavior of the inorganic nanoparticles effectively suppresses particle agglomeration within the polymer matrix, thereby reducing organic-inorganic interfacial defects and maintain excellent breakdown performance. The enhanced interfacial dipole response and charge traps introduced by particle self-assembly further contribute to a significant improvement in energy storage performance, even at low doping contents. Ultimately, the 1 wt% BaTiO3/1 wt% Al2O3 dielectric exhibits an outstanding discharged energy density of 19.1 J/cm3 with a charge-discharge efficiency of 81.0% and maintains stability over 1 × 105 charge-discharge cycles. Given the good compatibility of the present processing route with emerging large-scale electrospinning technologies and existing industrial film-manufacturing processes, this confined co-doping strategy provides an effective and potentially scalable route for developing polymer-based dielectric materials with both high energy density and high reliability48–50. Methods Materials Poly(vinylidene fluoride-co-hexafluoropropylene) (P(VDF-HFP)) resin was obtained from PolyK Technologies. Poly(methyl methacrylate) (PMMA) resin, barium titanate (BaTiO3) nanoparticles (100 nm), and N, N-dimethylformamide (DMF) were sourced from Shanghai Aladdin Biochemical Technology Co., Ltd. Aluminum oxide (Al2O3) nanoparticles (20 nm) were procured from Beijing Inokai Technology Co., Ltd. Ethanol and deionized water were acquired from Tansoole (Beijing, China). All chemicals were used as received without further purification. Preparation of organic-inorganic composite dielectric films Core-Doped BaTiO₃ Dielectrics: Organic-inorganic composite dielectric films were fabricated via coaxial electrospinning followed by hot pressing. To prepare the core precursor solution (Solution A), a specified mass of BaTiO3 nanoparticles and 2 g of P(VDF-HFP) resin were dissolved in 10 mL of DMF. The sheath precursor solution (Solution B) was prepared by dissolving 3 g of PMMA resin in 10 mL of DMF. Both solutions were stirred at 50 °C for 8 h to ensure homogeneity. Solutions A and B were loaded into coaxially connected syringes and electrospun under an electric field of 1 kV/cm. The electrospinning process consisted of three stages: (1) electrospinning a pure PMMA layer on the collector for 20 min, (2) coaxial electrospinning of the P(VDF-HFP)@PMMA core-sheath structure for 2 h, and (3) electrospinning a pure PMMA layer for 20 min. The mass ratio of P(VDF-HFP) to PMMA in the coaxial fibers was maintained at 6:4 by adjusting the injection ratio of the core-to-sheath precursor solutions. The resulting fiber mat was dried at 70 °C for 12 h, followed by hot pressing at 180 °C and 8 MPa for 30 min. The pressed film was then cold-pressed at 25 °C for 10 min to obtain the final dielectric film. Al2O3 and BaTiO3 Confined Co-Doped Dielectrics: The preparation of nanoparticles confined co-doped dielectrics followed a similar procedure to that described above, with modifications to incorporate Al2O3 nanoparticles. Depending on the doping strategy, appropriate contents of Al2O3 nanoparticles were added to either Solution A (for core co-doping) or Solution B (for layer doping) during precursor preparation. For the layer doping strategy, where Al2O3 nanoparticles were doped into the sheath PMMA, an additional pure PMMA precursor solution was prepared for electrospinning the undoped PMMA surface layer. Notably, the mass fraction of inorganic particles in this study is defined as the mass ratio of inorganic particles to the coaxial polymer fibers. Characterizations The core-sheath structure of the coaxial fibers and the distribution of inorganic nanoparticles in various doping strategies were characterized using TEM (JEM-F200, JEOL, Japan). The morphology of the PMMA surface layer and cross-sections of the composite dielectric films were analyzed using SEM (JSM-6301F, JEOL, Japan). The crystalline structure of the inorganic nanoparticles and composite dielectrics was investigated via XRD (D/max-2550, Rigaku, Japan) with a Cu Kα source and FTIR (X70, NETZSCH, Germany). Electrical properties were evaluated by sputtering gold electrodes on both sides of the films. For D-E hysteresis loops, breakdown strength, and leakage current measurements, gold electrodes with a diameter of 3 mm were used. Dielectric spectroscopy measurements were conducted using gold electrodes with a diameter of 20 mm. The dielectric constant and dielectric loss of the composite dielectrics were measured over a frequency range of 1 Hz-10 MHz using a broadband dielectric spectrometer (Novocontrol Technologies GmbH & Co. KG, Germany). Breakdown strength tests were performed using a dielectric withstand voltage test system (Beiguang Precision Instrument Co., Ltd., China). Unipolar D-E hysteresis loops (at 100 Hz) were acquired using a Premier II ferroelectric test system (Radiant Technologies Inc., USA). Leakage current was measured using a combination of a high-voltage power supply (Keithley 2290-10, USA) and a System SourceMeter (Keithley 2635B, USA). Supplementary information Supplementary Information Transparent Peer Review file Acknowledgements We thank Yanan Hao, Qi-Kun Feng, and Wen-Zhuo Dong for their helpful suggestions and assistance during the preparation of the manuscript. Author contributions Z.M.D. and K.X. conceived the idea; K.X. and Z.M.D. designed the experiments; K.X. and Q.Z.H. carried out most of the experiment work; J.T.W., L.H., Z.Y.R., S.L.Z., X.J.W., and J.T.L. helped with the high-field dielectric measurements; K.X., Z.M.D., K.B., L.M.G., B.Q.W., X.J.W., L.J.Y., and L.H. analyzed the data. K.X. wrote the manuscript and all authors participated in manuscript revision. All authors discussed the results and commented on the manuscript. Peer review Peer review information Nature Communications thanks Feng-Hui Gong, Junghwan Kim and Sung-Kon Kim for their contribution to the peer review of this work. A peer review file is available. Funding This work was financially supported by Smart Grid-National Science and Technology Major Project of China (Grant No. 2025ZD0808403 to Z.M.D.), the National Natural Science Foundation of China (Grants No. U2241243 to K.B., and No. 52527803 to Z.M.D.), Zhongyuan Electric Laboratory Research Program (Grant No. ZC20241106 to Z.M.D), and the Beijing Natural Science Foundation (Grant No. JQ22010 to K.B.). Data availability The authors declare that all data supporting the findings of this study are available within the article and the supplementary materials. Further information is available from the corresponding author upon request (dangzm@tsinghua.edu.cn). Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Limin Guo, Email: guolimin@bupt.edu.cn. Ke Bi, Email: bike@bupt.edu.cn. Zhi-Min Dang, Email: dangzm@tsinghua.edu.cn. Supplementary information The online version contains supplementary material available at 10.1038/s41467-026-74471-5. References Associated Data Supplementary Materials Supplementary Information Transparent Peer Review file Data Availability Statement The authors declare that all data supporting the findings of this study are available within the article and the supplementary materials. Further information is available from the corresponding author upon request (dangzm@tsinghua.edu.cn).