NH₂-functionalized carbon-coated Fe₃O₄ core–shell nanoparticles for in situ preparation of robust polyimide composite films with high dielectric constant, low dielectric loss, and high breakdown strength

Abstract

Novel Fe₃O₄/polyimide (PI) composite films with extraordinary dielectric properties were prepared via in situ polymerization of PI in the presence of 4,4′-diaminodiphenyl ether (ODA), pyromellitic dianhydride (PMDA) and compatible amine-functionalized carbon-coated Fe₃O₄ nanoparticles (Fe₃O₄@C–NH₂). Fe₃O₄ nanoparticles prepared via solvothermal reaction were coated with crosslinked glucose and then surface-grafted via carbodiimide coupling. The core–shell structure of Fe₃O₄@C–NH₂ was thereafter obtained with amine functionalities on the carbonized-glucose shell. Homogeneous dispersion of Fe₃O₄@C–NH₂ and strong interfacial covalent bonding on the shell surface contributed to high dielectric constants (ε) and extremely low dielectric losses (tan δ) in the composite films. The highest dielectric constant of the composite films was 58.6 at 1 kHz when the Fe₃O₄@C–NH₂ composition was 1.13 vol%, as well as a low dielectric loss of 0.0091 and a high breakdown strength of 126.5 ± 11.3 MV m⁻¹ were observed without considerable sacrifice of the thermal stability and mechanical properties.

---

Introduction

Polymer nanocomposites have attracted much attention in recent years because of their many advantages such as flexibility, ease of processing, and tunable properties.^1–4 Polymer nanocomposites with a high dielectric constant (ε) and ease of processing are promising dielectric materials for embedded capacitors, considering their flexibility and compatibility with organic printed circuit boards.⁵ Polyvinylidene fluoride (PVDF) is one of the most studied polymer matrices for being incorporated with carbon nanotubes (CNT),^6,7 BaTiO₃,⁸ or a porous graphene sandwich⁹ to improve its dielectric properties because of its intrinsically high dielectric constant of ∼12. The dielectric constants of these composites did rise with increasing of the filler composition; however, the dielectric loss increased as well by adding conductive fillers as their direct contact induces a high leakage current. The dielectric constants of radial ZnO/PVDF composites were reported to be ∼65 at 1 kHz, and the dielectric losses were ∼0.5,¹⁰ which were too high for practical applications. Besides PVDF, poly(p-phenylene benzobisoxazole),^11–16 epoxy,^17–19 polyurethane,^20,21 polyethersulfone,²² polyarylene ether nitriles,²³ poly(methyl methacrylate),²⁴ and other polymers^25–27 were also reported as polymer matrices for fabricating dielectrics.

Polyimide (PI) exhibits great potential as a high-performance dielectric material on the basis of its heat resistance and outstanding mechanical properties.²⁸ The dielectric constants of most aromatic PI films are between 2.8 and 5.0, while the dielectric loss values vary from 0.001 to 0.03. Because the low dielectric loss of PI is ideal for practical applications in high temperature environments, researchers committed to improve its dielectric constant. Previous efforts on achieving high-ε PI-based composites primarily focused on those filled with high-ε ferroelectric ceramics, such as nano-TiO₂,²⁹ Al₂O₃,³⁰ BaTiO₃,^31,32 and CaCu₃–Ti₄O₁₂.^33,34 For these composites, ε was barely over 50 even when the filler volume fraction was increased to 50% or higher. In addition, the composite films were limited because of reduced processability and dramatic compromise in mechanical properties.

Another strategy was to introduce conductive fillers, including CNT (ε = 31.3)³⁵ and graphene (ε = 36.9)³⁶ into the PI matrix. Although the mechanical properties were satisfactory, the dielectric constants of the resulting composites were only ∼35.0, much lower than those of ferroelectric ceramic/PI composites. Fe₃O₄ spherical nanoparticles were promising candidates because of their unique magnetic response and large surface area.^37–41 Haldar et al. reported ε values of ∼2000 and ∼500 in Fe₃O₄/polyaniline composites and polypyrrole-modified poly(N-vinyl carbazole)–Fe₃O₄ nanocomposites, respectively.^42,43 Therefore, a high dielectric constant and a low dielectric loss are expected in Fe₃O₄/PI composites at the percolation threshold when modified Fe₃O₄ nanoparticles are well dispersed in the PI matrix. This material design strategy was applied in this study to achieve the goals in dielectric properties with a new preparative method of functionalized Fe₃O₄ nanoparticles.

Although Fe₃O₄ nanoparticles are highly competent as conductive fillers, two limitations have been identified in their practical applications: (1) Fe₃O₄ nanoparticles readily aggregate because of their large specific surface area and high surface free energy, making nanoscale dispersion of Fe₃O₄ nanoparticles in a polymer matrix extremely challenging;⁴⁴ (2) Fe₃O₄ nanoparticles can be easily oxidized in the absence of a protective shell. To overcome these barriers, a strategy has been proposed to prepare core-shell-structured carbon-coated Fe₃O₄ nanoparticles (Fe₃O₄@C) via hydrothermal carbonization.⁴⁵ The carbon layer can effectively protect Fe₃O₄ nanoparticles from being oxidized and the outermost polysaccharide layer holds a variety of functional groups, such as carboxyl and hydroxyl groups, on the surface.^46,47 Fe₃O₄@C was further decorated with p-phenylene diamine (PPD) in the presence of N,N′-dicyclohexyl carbodiimide (DCC) catalyst to introduce amine functionality (Fe₃O₄@C–NH₂), so as to improve the reactivity and dispersion of the nanoparticles.³⁶ The functionalized core–shell nanoparticles were homogeneously distributed throughout the polymer matrix via in situ polymerization of PI, and many appealing properties (high dielectric constant, low dielectric loss, and high breakdown strength) of the Fe₃O₄@C–NH₂/PI composites were achieved as expected while the thermal stability and tensile strength were also as excellent as the PI matrix with scarce sacrifice.

Experimental section

Materials

Reagents including FeCl₃·6H₂O, ethylene glycol, sodium acetate, polyethylene glycol, DCC, HNO₃, DMF and ethanol were supplied by Alfa Aesar (Massachusetts, USA). Glucose was purchased from Aldrich Chemical Company (Milwaukee, USA). PPD and 4,4′-diaminodiphenyl ether (ODA) were obtained from Aladdin (Shanghai, China). Pyromellitic dianhydride (PMDA) was provided by J & K (Beijing, China). DMAc was dried with calcium hydride (CaH₂) prior to distillation. Other reagents were of high purity and used as received.

Synthesis of Fe₃O₄ nanoparticles⁴⁸

1.35 g FeCl₃·6H₂O (5 mmol) was dissolved in 40 mL ethylene glycol, followed by addition of 3.6 g sodium acetate (NaAc) and 1.0 g polyethylene glycol (PEG, M_w = 20 000 g mol⁻¹). The solution was under vigorously magnetically stirring for 30 min, and sealed in a 100 mL Teflon-lined stainless-steel autoclave reactor. The autoclave reactor remained at 200 °C for 8 h, and cooled down to room temperature. The black powder was washed thoroughly with ethanol and dried at 60 °C for 6 h.

Synthesis of Fe₃O₄@C core–shell nanoparticles⁴⁹

Fe₃O₄ nanoparticles were soaked in 0.1 M HNO₃ solution for 5 min, and separated the Fe₃O₄ nanoparticles from the solution with a magnet followed by extensive wash with deionized water. The treated Fe₃O₄ nanoparticles (2 out of 5 mmol) were dispersed in 60 mL water under vigorous stirring for 30 min with assistance of 7.2 g glucose (40 mmol). The solution was transferred to a 100 mL Teflon-lined stainless-steel autoclave reactor. The autoclave reactor remained at 180 °C for 6 h, and cooled down to room temperature. The black powder was washed thoroughly with ethanol and dried at 60 °C for 6 h.

Preparation of Fe₃O₄@C–NH₂ nanoparticles

0.05 g Fe₃O₄@C was dispersed in 20 mL DMAc with ultrasonication, followed by addition of 0.05 g DCC. After 4 h, 0.05 g PPD was added and the dispersion was stirred for 24 h under nitrogen protection at room temperature before filtration and thorough wash with DMF and ethanol. Fe₃O₄@C–NH₂ was obtained and dried in a vacuum oven at 60 °C for 24 h.

Preparation of Fe₃O₄@C–NH₂/PI films via in situ polymerization of PI

10.05 mg Fe₃O₄@C–NH₂ was dispersed in 20 mL anhydrous DMAc with ultrasonication for 30 min, and a stable dispersion was obtained. 1.00 g 4,4′-diaminodiphenyl ether (ODA, 5 mmol) was added and dissolved with magnetic stirring at room temperature under nitrogen protection, followed by addition of 1.09 g pyromellitic dianhydride (PMDA, 5 mmol) with 4 equal batches in quantity within 4 h. After PDMA was completely dissolved, the dispersion was stirred for 24 h and poured into a glass Petri dish, followed by drying for 48 h and heat treatment at 100, 200, and 300 °C for 2 h, respectively, and then at 400 °C for 5 min. Fe₃O₄@C–NH₂/PI composite films containing 0.13 vol%, 0.27 vol%, 0.55 vol%, 0.83 vol%, 1.13 vol%, and 1.40 vol% of Fe₃O₄@C–NH₂ were obtained. Fe₃O₄/PI films were prepared following the same procedure.

Characterization

Fourier transform infrared spectroscopy (FTIR) spectra were acquired on Nicolet Magna-IR 550 spectrometer using pressed KBr pellets. X-ray diffraction (XRD) was performed on a D/MAX 2550 VB/PC rotating anode X-ray multi-crystal diffractometer equipped with Ni-filtered CuK_α radiation and operated at 60 mA and 40 kV. The morphologies of Fe₃O₄, Fe₃O₄@C and Fe₃O₄@C–NH₂ nanoparticles were examined using field-emission scanning electron microscopy (SEM, FESEM, Hitachi S-4800) and high-resolution transmission electron microscopy (TEM, HRTEM, JEOL JEM-2100). The fractured cross sections of both PI and Fe₃O₄@C–NH₂/PI composite films were examined using the SEM and a Fe₃O₄@C–NH₂/PI composite film with a thickness of 60–70 nm was prepared through ultrathin section and observed using the TEM. Thermogravimetric analysis (TGA) was performed at a heating rate of 10 °C min⁻¹ under nitrogen stream. The dielectric constants were measured on a CONCEPT 40 broadband dielectric spectrometer (Novocontrol Technologies GmbH & Co. KG, Germany). The breakdown strength was evaluated by a DC dielectric strength tester with a sphere–sphere stainless electrode (DH, Shanghai Lanpotronics Co., China) at room temperature. Mechanical properties were investigated with a Rheometric Scientific DMTA5 under tensile mode, with results averaged over more than 5 replicates.

Results and discussion

The Fe₃O₄ or Fe₃O₄@C–NH₂ volume composition (f) in the Fe₃O₄/PI and Fe₃O₄@C–NH₂/PI composites was defined as V_p/(V_p + V_m) × 100%, where V_p and V_m are the volumes of Fe₃O₄ or Fe₃O₄@C–NH₂ and PI used for composite preparation, respectively. V_p and V_m were calculated from their weights and the densities, which are 5.18 and 1.40 g cm⁻¹ for Fe₃O₄ and PI, respectively. Because the carbon shell (4 nm) of the Fe₃O₄@C–NH₂ was much thinner than the diameter of the Fe₃O₄ (40 nm), we substituted the density of Fe₃O₄@C–NH₂ with the density of Fe₃O₄ in calculating f for Fe₃O₄@C–NH₂ in the composites. Thus the weight compositions of the nanoparticles of 1, 2, 3, 4, and 5 wt% were converted to f values of 0.27, 0.55, 0.83, 1.13, and 1.40 vol%.

Fe₃O₄, Fe₃O₄@C, and Fe₃O₄@C–NH₂ nanoparticles were characterized using FTIR, as shown in Fig. 1a. The absorption band at 3410 cm⁻¹ indicated the presence of residual hydroxyl groups, and the bands at 1708 and 1635 cm⁻¹ were associated with stretching vibrations of C=O and C=C, respectively. This result confirmed successful carbonization of glucose during the hydrothermal carbonization and the presence of considerable residual hydrophilic functionalities resulting from the incomplete carbonization of glucose.^50,51 The bands in 1000–1300 cm⁻¹ were attributed to stretching vibrations of C–O and bending vibrations of O–H. Compared to Fe₃O₄@C, Fe₃O₄@C–NH₂ showed weaker stretching vibrations of C=O and emergence of stretching vibrations of C–N at 1264 cm⁻¹ from the amide in Fe₃O₄@C–NH₂, indicating successful amidation between Fe₃O₄@C and PPD. The band at 590 cm⁻¹ corresponding to stretching vibrations of Fe–O,⁵⁰ confirmed the presence of Fe₃O₄ in the composites.

Fig. 1 (a) FTIR spectra and (b) XRD profiles of Fe₃O₄, Fe₃O₄@C, and Fe₃O₄@C–NH₂ nanoparticles.

From the XRD profiles presented in Fig. 1b, characteristic diffraction peaks at 2θ of 30.05°, 35.50°, 43.19°, 57.11°, and 62.67° were well indexed to the cubic spinel structured magnetite (JCPDS file No. 19-0629),⁵² indicating that the crystalline phase of Fe₃O₄ remained unchanged during the coating and the amidation processes. After the coating process, a relatively broad halo (17.13–29.14°) centered at 2θ of 22.51° was identified in the profiles of Fe₃O₄@C and Fe₃O₄@C–NH₂ samples, indicating that the carbon matrix was amorphous.⁵³

The morphology of as-synthesized nanoparticles was further captured by using SEM and TEM. As shown in Fig. 2a and b, a virtually spherical shape was observed for the as-synthesized Fe₃O₄ nanoparticles with a mean diameter of ∼40 nm. Based on the SEM images of Fe₃O₄@C and Fe₃O₄@C–NH₂ nanoparticles in Fig. 2c and e, the mean diameters of Fe₃O₄@C and Fe₃O₄@C–NH₂ nanoparticles were estimated to be ∼40 nm as well, suggesting that the shape and size remained unchanged after the coating and the amidation processes. In contrast to Fig. 2b, d and f presented a core–shell structure, in which black Fe₃O₄ nanoparticles were encapsulated by a 4 nm thick gray carbon shell (see the high magnification images). The uniform amorphous carbonaceous coating was most likely generated by the carbonization of glucose coating on the Fe₃O₄ nanoparticles during the hydrothermal carbonization.⁵⁰ Dispersion of Fe₃O₄, Fe₃O₄@C, and Fe₃O₄@C–NH₂ in N,N-dimethylacetamide (DMAc) at the concentration of 3 mg mL⁻¹ was presented in Fig. 2g–i, respectively. The comparison between Fe₃O₄ and Fe₃O₄@C showed that the carbon layer with hydrophilic functionalities was of vital importance in protecting the Fe₃O₄ nanoparticles from oxidation and promoting the dispersion of the nanoparticles in DMAc. After amidation, the dispersion of Fe₃O₄@C–NH₂ in DMAc was further improved.

Fig. 2 SEM (a, c, e) and TEM (b, d, f) images, and dispersion in DMAc at 3 mg mL⁻¹ (g, h, i) of the nanoparticles of Fe₃O₄ (a, b, g), Fe₃O₄@C (c, d, h), and Fe₃O₄@C–NH₂ (e, f, i) with high magnification TEM images of individual nanoparticles. Scale bars of 1 μm, 60 nm, and 20 nm in the images in the top row are applicable for the images in their respective columns.

The dispersion of Fe₃O₄@C–NH₂ in the PI matrix was a pivotal factor in controlling the physical properties and performance of the composites. SEM images of the fractured surface of the neat PI and the Fe₃O₄@C–NH_2/PI (f = 1.13 vol%) composite film at the same magnification were compared and further supported this key point. The smooth fractured cross section of the neat PI film (Fig. 3a) was altered dramatically upon the introduction of the nanoparticles; the fractured cross-section of the Fe₃O₄@C–NH₂/PI composite films was rather rough and ridged (Fig. 3b). As shown in Fig. 3c, the Fe₃O₄@C–NH₂ nanoparticles were uniformly dispersed without observable aggregation, leading to the expectation of high dielectric constants and low dielectric losses in the composites. However, the diameters of the Fe₃O₄@C–NH₂ nanoparticles shown in Fig. 3c were from 5 to 45 nm. The discrepancy between the data and the values obtained from SEM and TEM images in Fig. 2 was as a result of that the cross-section film thickness of nanocomposites (60–70 nm) was very close to the nanoparticle diameter (∼40 nm).

Fig. 3 SEM images of the fractured cross sections of (a) the neat PI (×40 000) and (b) the 1.13 vol% Fe₃O₄@C–NH₂/PI (×40 000) composite films; (c) TEM image of the 1.13 vol% Fe₃O₄@C–NH₂/PI composite film. Scale bars of 1 μm, 1 μm, and 50 nm are present in the images of (a), (b), and (c), respectively.

Thermal stability is one of the important properties for PI-based nanocomposites used as high-performance engineering plastics. As shown in TGA (Fig. 4a), the onset thermal decomposition temperatures at 5% mass loss of neat PI and Fe₃O₄@C–NH₂/PI nanocomposites with f values of 0.27 vol%, 0.55 vol%, 0.83 vol%, 1.13 vol%, and 1.40 vol% were 591.5, 563.1, 557.2, 549.5, 522.5, and 504.8 °C, respectively. The thermal stability of the nanocomposites only decreased slightly upon increasing the nanoparticle composition. Compared to neat PI and nanocomposites containing 0.27–1.13 vol% Fe₃O₄@C–NH₂, the onset decomposition temperature of Fe₃O₄@C–NH₂/PI (f = 1.40 vol%) nanocomposite was much lower, primarily attributed to agglomeration of the Fe₃O₄@C–NH₂ nanoparticles at compositions higher than the percolation threshold. In contrast, as shown in Fig. 4b, the onset thermal decomposition temperatures of the Fe₃O₄/PI nanocomposites greatly decreased from 591.5 to 428.8 °C with increasing f from 0% (neat PI) to 1.40 vol%, which were much lower than those of the Fe₃O₄@C–NH₂/PI nanocomposites because of the poor dispersion of Fe₃O₄ in PI. The agglomeration of the Fe₃O₄ nanoparticles in PI hindered the amidization process of PI and that the larger interface gap between the thermal conductive Fe₃O₄ aggregates and the PI matrix might serve as diffusion path for heat and degradation products.

Fig. 4 TGA diagrams (a, b) and stress–strain curves (c, d) of the neat PI film and the Fe₃O₄@C–NH₂/PI (a, c) and Fe₃O₄/PI (b, d) composite films with various nanoparticle volume compositions.

Representative stress–strain curves of the neat PI and the Fe₃O₄@C–NH₂/PI composites are illustrated in Fig. 4c. The tensile strength and the elongation at break of Fe₃O₄@C–NH₂/PI composite films were lower than those of the neat PI films. With increasing f from 0 (neat PI) to 1.40 vol%, the tensile strength gradually decreased from 108 ± 5.8 to 85 ± 9.3 MPa. Meanwhile, tensile modulus decreased from 1.2 ± 0.2 to 0.9 ± 0.1 GPa and elongation at break decreased from 30.1 ± 10.6% to 19.8 ± 10.3%, showing similar trends with increasing the composition of Fe₃O₄@C–NH₂. Despite these slight decreases due to interrupted integrity of PI matrix introduced by aggregation of Fe₃O₄@C–NH₂ nanoparticles, these nanocomposite films still exhibited excellent mechanical properties relative to other polymer candidates (∼73 MPa).^54,55 Compared to the Fe₃O₄@C–NH₂/PI composites, the tensile strength of Fe₃O₄/PI composites (Fig. 4d) decreased more significantly from 108 ± 5.8 to 76 ± 9.6 MPa, mainly due to the poor dispersion of Fe₃O₄ in the matrix.

The dielectric properties of Fe₃O₄@C–NH₂/PI and Fe₃O₄/PI nanocomposite films with various nanoparticle compositions were plotted versus frequency in Fig. 5. Relative to Fe₃O₄/PI nanocomposite films, superior dielectric properties were found in Fe₃O₄@C–NH₂/PI nanocomposite films and this might attribute to better dispersion of Fe₃O₄@C–NH₂ in PI with assistance of the compatible core–shell structure of the nanoparticles and the subsequent amidation. As shown in Fig. 5a, the dielectric constant of Fe₃O₄@C–NH₂/PI composite films decreased with increasing the frequency from 10 Hz to 1 × 10⁶ Hz and more strikingly the addition of Fe₃O₄@C–NH₂ significantly promoted the dielectric constant of PI. A low percolation threshold was identified in the Fe₃O₄@C–NH₂/PI composite films. Furthermore, the dielectric loss of the Fe₃O₄@C–NH₂/PI nanocomposite films was extremely low (Fig. 5b). The dielectric constant of Fe₃O₄/PI composite films presented in Fig. 5c also decreased with increasing the frequency from 10 Hz to 1 MHz, but it only mildly increased with increasing the composition of Fe₃O₄. The dielectric constants of the Fe₃O₄@C–NH₂/PI composite films were prominently higher than those of the Fe₃O₄/PI composite films. No percolation threshold was observed upon adding Fe₃O₄ nanoparticles into PI, primarily due to the poor dispersion of Fe₃O₄ in PI. The dielectric loss of Fe₃O₄/PI composite films (Fig. 5d) decreased when the frequency increased to 1 × 10⁶ Hz and the values were also extremely low. Taking these two parameters into consideration, the above observations indicated that Fe₃O₄ nanoparticles could not effectively improve the dielectric properties of PI while incorporation of Fe₃O₄@C–NH₂ into PI was a very effective approach in promoting the dielectric properties of the PI matrix.

Fig. 5 Room-temperature dielectric constant (a, c) and dielectric loss (b, d) as functions of frequency in the Fe₃O₄@C–NH₂/PI (a, b) and Fe₃O₄/PI (c, d) composite films with various nanoparticle volume compositions. The insets are the enlarged areas in the high frequency range.

The dielectric constant and loss of Fe₃O₄@C–NH₂/PI and Fe₃O₄/PI composites at the frequency of 1 kHz are plotted as functions of f in Fig. 6a and b, respectively. The dielectric constant increased dramatically when the Fe₃O₄@C–NH₂ composition approached and exceeded the percolation threshold, as predicted by the percolation theory.⁵⁶ The enhancement of the dielectric constant was primarily attributed to the formation of microcapacitor network.⁵⁷ The maximum of the dielectric constant for the Fe₃O₄@C–NH₂/PI nanocomposite was 58.6 at f = 1.13 vol%, significantly higher than that of the neat PI (∼3.0) by a factor of ∼20, indicating that homogeneous dispersion of the Fe₃O₄@C–NH₂ nanoparticles in the matrix.⁵⁸ The dielectric loss of the Fe₃O₄@C–NH₂/PI composites was between 0.0027 and 0.25, a few orders of magnitude lower than those of other conventional percolative composites.^10,59 Although adding Fe₃O₄ nanoparticles directly into PI only mildly increased its dielectric constant by a factor of ∼4, the dielectric losses of Fe₃O₄/PI composites were also extremely low with a maximum value of 0.0089 for the Fe₃O₄ composition of 1.40 vol% in the studied range (Fig. 6b). The above results indicated that the Fe₃O₄@C–NH₂/PI composite films were promising polymer-based dielectric materials with a high dielectric constant and an extremely low dielectric loss.

Fig. 6 Dielectric constant and dielectric loss at 1 kHz as functions of the volume fraction of Fe₃O₄@C–NH₂ (a) and Fe₃O₄ (b) nanoparticles. Inset in (a): linear fitting (line) of dielectric constant vs. f_c − f in a double-natural-logarithmic plot.

The percolation threshold can be calculated using the following equation:

ε ∼ (f_c − f)^−s for f < f_c (1)

where f_c is the critical volume composition at the percolation threshold and s is the critical exponent. The experimental dielectric constants in Fig. 6a were fitted using eqn (1) in the inset of Fig. 6a, resulting f_c of 0.0135 and s of 1.461. The f_c value was lower than the percolation threshold of 0.0661 in ZnO/PVDF.¹⁰ As the Fe₃O₄@C–NH₂ composition increased to 1.40 vol%, the dielectric constant decreased and the dielectric loss increased simultaneously, primarily because of the destruction of the percolation network and the formation of the conductive network, leading to the leakage current in the nanocomposites.⁶⁰

The dielectric constants of the neat PI and the Fe₃O₄@C–NH₂/PI (f = 1.13 vol%) composite were plotted as a function of temperature, shown in Fig. 7a. At elevated temperatures, the dielectric constant of the neat PI remained constant at ∼3. However, temperature showed a remarkable impact on the dielectric constant of the composite. The dielectric constant of the composite increased with increasing temperature from 20 to 250 °C. This increment can be explained by the increased conductivity of semi-conducting Fe₃O₄ particles and thermal expansion of PI,^10,61 as well as polarization of permanent dipoles in the polymer at the molecular level (Scheme 1).

Fig. 7 (a) Temperature dependence of dielectric constants of the neat PI (open circles) and the 1.13 vol% Fe₃O₄@C–NH₂/PI composite (solid symbols) at different frequencies. (b) Breakdown strengths of the Fe₃O₄@C–NH₂/PI (solid circles) and Fe₃O₄/PI (open circles) composite films as functions of Fe₃O₄@C–NH₂ or Fe₃O₄ composition.

Scheme 1 Synthesis of Fe₃O₄@C–NH₂/PI composite films.

High dielectric constant and high breakdown strength were equally important for dielectric materials because they are decisive factors in governing the operating electric field of the dielectric materials.⁶² The breakdown strengths of the neat PI, Fe₃O₄@C–NH₂/PI, and Fe₃O₄/PI composite films were measured and plotted in Fig. 7b as functions of Fe₃O₄@C–NH₂ or Fe₃O₄ compositions at room temperature. The breakdown strength of the neat PI film was as high as 286.5 ± 32.6 MV m⁻¹. Addition of a small quantity of Fe₃O₄@C–NH₂ resulted in a significant decrease in breakdown strength to 192.8 ± 23.3 MV m⁻¹ in Fe₃O₄@C–NH₂/PI (f = 0.13 vol%) composite films. However, as the Fe₃O₄@C–NH₂ composition further increased, the breakdown strength of the composite film only decreased mildly, similar to the findings in CNT/PI and graphene/PI composites.^35,36 The breakdown strength of the Fe₃O₄@C–NH₂/PI (f = 1.13 vol%) composite film remained at a high level of 126.5 ± 11.3 MV m⁻¹. The open symbols in Fig. 7b showed that adding Fe₃O₄ nanoparticles directly into PI resulted in a significant decrease in the breakdown strength from the value for the neat PI to 169.4 ± 27.5 MV m⁻¹ for the Fe₃O₄/PI (f = 0.13 vol%) composite film. The breakdown strengths of the Fe₃O₄/PI composites were always lower than those of Fe₃O₄@C–NH₂/PI at the same f values. In addition, as the Fe₃O₄ composition further increased, the breakdown strength decreased faster for the Fe₃O₄/PI composite films than that of the Fe₃O₄@C–NH₂/PI composite films, showing that Fe₃O₄@C–NH₂ nanoparticles could better maintain the breakdown strength of the composites than Fe₃O₄ nanoparticles.

Conclusion

In summary, we synthesized NH₂-functionalized Fe₃O₄@C and in situ polymerized Fe₃O₄@C–NH₂/PI composite films to enhance dielectric properties. The dielectric properties of the Fe₃O₄@C–NH₂/PI composite films were optimized by varying the nanoparticle composition. At Fe₃O₄@C–NH₂ composition of 1.13 vol%, the dielectric constant of the nanocomposite films at 1 kHz was promoted to 58.6, higher than that of the neat PI (∼3.0) by a factor of ∼20. An extremely low dielectric loss of 0.0091 and a high breakdown strength of 126.5 ± 11.3 MV m⁻¹ were also identified. This finding could be explained by the homogeneous dispersion of Fe₃O₄@C–NH₂ in the PI matrix and the strong interfacial covalent bonding between Fe₃O₄@C–NH₂ nanoparticles and the PI matrix. The Fe₃O₄@C–NH₂/PI nanocomposite films witnessed satisfactory thermal stability as well as mechanical properties for multifunctional applications in extreme operating environment, suggesting that they are potential high-performance candidates for dielectric applications in capacitors.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (51573045) and the International Collaboration Research Program of Science and Technology Commission of Shanghai (16520722000).

Notes and references

1. X. Zhang, W. Chen, J. Wang, Y. Shen, L. Gu, Y. Lin and C.-W. Nan, Nanoscale, 2014, 6, 6701–6709.
2. Q. Li, L. Chen, M. R. Gadinski, S. Zhang, G. Zhang, H. Li, A. Haque, L. Q. Chen, T. Jackson and Q. Wang, Nature, 2015, 523, 576–579.
3. S. Zhang, N. Zhang, C. Huang, K. Ren and Q. M. Zhang, Adv. Mater., 2005, 17, 1897–1901.
4. M. Zhang, Y. Li, Z. Su and G. Wei, Polym. Chem., 2015, 6, 6107–6124.
5. W. Jillek and W. K. C. Yung, Int. J. Adv. Manuf. Tech., 2005, 25, 350–360.
6. S. Yu, W. Zheng, W. Yu, Y. Zhang, Q. Jiang and Z. Zhao, Macromolecules, 2009, 42, 8870–8874.
7. L. L. Sun, B. Li, Y. Zhao, G. Mitchell and W. H. Zhong, Nanotechnology, 2010, 21, 305702.
8. K. Yang, X. Huang, Y. Huang, L. Xie and P. Jiang, Chem. Mater., 2013, 25, 2327–2338.
9. L. Chu, Q. Xue, J. Sun, F. Xia, W. Xing, D. Xia and M. Dong, Compos. Sci. Technol., 2013, 86, 70–75.
10. G. Wang, Y. Deng, Y. Xiang and L. Guo, Adv. Funct. Mater., 2008, 18, 2584–2592.
11. S. Wang, Y. Chen, Q. Zhuang, X. Li, P. Wu and Z. Han, Macromol. Chem. Phys., 2006, 207, 2336–2342.
12. S. Wang, P. Wu and Z. Han, Polymer, 2001, 42, 217–226.
13. Z. Xie, Q. Zhuang, Q. Wang, X. Liu, Y. Chen and Z. Han, Polymer, 2011, 52, 5271–5276.
14. S. Wang, P. Guo, P. Wu and Z. Han, Macromolecules, 2004, 37, 3815–3822.
15. Q. Zhuang, X. Liu, Q. Wang, X. Liu, J. Zhou and Z. Han, J. Mater. Chem., 2012, 22, 12381–12388.
16. Y. Chen, Q. Zhuang, X. Liu, J. Liu, S. Lin and Z. Han, Nanotechnology, 2013, 24, 245702.
17. R. A. Pethrick, E. A. Hollins, I. McEwan, A. J. MacKinnon, D. Hayward, L. A. Cannon, S. D. Jenkins and P. T. McGrail, Macromolecules, 1996, 29, 5208–5214.
18. C. Min, D. Yu, J. Cao, G. Wang and L. Feng, Carbon, 2013, 55, 116–125.
19. D. Lairez, J. R. Emery, D. Durand and R. A. Pethrick, Macromolecules, 1992, 25, 7208–7210.
20. C. Wu, X. Huang, X. Wu, L. Xie, K. Yang and P. Jiang, Nanoscale, 2013, 5, 3847–3855.
21. F. Galantini, S. Bianchi, V. Castelvetro and G. Gallone, Smart Mater. Struct., 2013, 22, 055025.
22. F. Wang, D. Zhou and Y. Hu, Phys. Status Solidi A, 2009, 206, 2632–2636.
23. X. Huang, Z. Pu, L. Tong, Z. Wang and X. Liu, J. Mater. Sci.: Mater. Electron., 2012, 23, 2089–2097.
24. M. Li, X. Huang, C. Wu, H. Xu, P. Jiang and T. Tanaka, J. Mater. Chem., 2012, 22, 23477–23484.
25. J. Y. Kim, J. Lee, W. H. Lee, I. N. Kholmanov, J. W. Suk, T. Kim, Y. Hao, H. Chou, D. Akinwande and R. S. Ruoff, ACS Nano, 2014, 8, 269–274.
26. L. J. Romasanta, M. Hernandez, M. A. Lopez-Manchado and R. Verdejo, Nanoscale Res. Lett., 2011, 6, 508.
27. N. Ning, X. Bai, D. Yang, L. Zhang, Y. Lu, T. Nishi and M. Tian, RSC. Adv., 2014, 4, 4543–4551.
28. K. Kriechbaum, D. A. Cerrón-Infantes, B. Stöger and M. M. Unterlass, Macromolecules, 2015, 48, 8773–8780.
29. X. Liu, J. Yin, Y. Kong, M. Chen, Y. Feng, Z. Wu, B. Su and Q. Lei, Thin Solid Films, 2013, 544, 54–58.
30. A. Alias, Z. Ahmad and A. B. Ismail, Mater. Sci. Eng., B, 2011, 176, 799–804.
31. K. Abe, D. Nagao, A. Watanabe and M. Konno, Polym. Int., 2013, 62, 141–145.
32. E. Hamciuc, C. Hamciuc, I. Bacosca, M. Cristea and L. Okrasa, Polym. Compos., 2011, 32, 846–855.
33. Z. M. Dang, T. Zhou, S. H. Yao, J. K. Yuan, J. W. Zha, H. T. Song, J. Y. Li, Q. Chen, W. T. Yang and J. Bai, Adv. Mater., 2009, 21, 2077–2082.
34. Q. Chi, J. Sun, C. Zhang, G. Liu, J. Lin, Y. Wang, X. Wang and Q. Lei, J. Mater. Chem. C, 2014, 2, 172–177.
35. Y. Chen, B. Lin, X. Zhang, J. Wang, C. Lai, Y. Sun, Y. Liu and H. Yang, J. Mater. Chem. A, 2014, 2, 14118–14126.
36. X. Fang, X. Liu, Z.-K. Cui, J. Qian, J. Pan, X. Li and Q. Zhuang, J. Mater. Chem. A, 2015, 3, 10005–10012.
37. Y. Deng, D. Qi, C. Deng, X. Zhang and D. Zhao, J. Am. Chem. Soc., 2008, 130, 28–29.
38. X. Xu, C. Deng, M. Gao, W. Yu, P. Yang and X. Zhang, Adv. Mater., 2006, 18, 3289–3293.
39. J. Park, K. An, Y. Hwang, J. G. Park, H. J. Noh, J. Y. Kim, J. H. Park, N. M. Hwang and T. Hyeon, Nat. Mater., 2004, 3, 891–895.
40. K. Mori, Y. Kondo, S. Morimoto and H. Yamashita, J. Phys. Chem. C, 2008, 112, 397–404.
41. A. Zhu, L. Yuan and S. Dai, J. Phys. Chem. C, 2008, 112, 5432–5438.
42. I. Haldar, M. Biswas and A. Nayak, Polym.-Plast. Technol. Eng., 2014, 53, 1317–1326.
43. I. Haldar, M. Biswas and A. Nayak, Synth. Met., 2011, 161, 1400–1407.
44. G. Wang, X. Zhang, A. Skallberg, Y. Liu, Z. Hu, X. Mei and K. Uvdal, Nanoscale, 2014, 6, 2953–2963.
45. T. Yao, T. Cui, H. Wang, L. Xu, F. Cui and J. Wu, Nanoscale, 2014, 6, 7666–7674.
46. Z. Wang, H. Guo, Y. Yu and N. He, J. Magn. Magn. Mater., 2006, 302, 397–404.
47. Z. Zhang, H. Duan, S. Li and Y. Lin, Langmuir, 2010, 26, 6676–6680.
48. H. Wang, L. Sun, Y. Li, X. Fei, M. Sun, C. Zhang, Y. Li and Q. Yang, Langmuir, 2011, 27, 11609–11615.
49. D. Qi, H. Zhang, J. Tang, C. Deng and X. Zhang, J. Phys. Chem. C, 2010, 114, 9221–9226.
50. X. Sun and Y. Li, Angew. Chem., Int. Ed., 2004, 43, 597–601.
51. Q. Peng, Y. Dong and Y. Li, Angew. Chem., Int. Ed., 2003, 42, 3027–3030.
52. X. Yu, J. Wan, Y. Shan, K. Chen and X. Han, Chem. Mater., 2009, 21, 4892–4898.
53. R. Demir Cakan, M. M. Titirici, M. Antonietti, G. Cui, J. Maier and Y. S. Hu, Chem. Commun., 2008, 3759–3761.
54. L. Liu, B. Liang, W. Wang and Q. Lei, J. Compos. Mater., 2006, 40, 2175–2183.
55. H. Cai, F. Yan, Q. Xue and W. Liu, Polym. Test., 2003, 22, 875–882.
56. Y. Chen, S. Zhang, X. Liu, Q. Pei, J. Qian, Q. Zhuang and Z. Han, Macromolecules, 2015, 48, 365–372.
57. S. Stankovich, R. D. Piner, S. T. Nguyen and R. S. Ruoff, Carbon, 2006, 44, 3342–3347.
58. W. Zhang, W. He and X. Jing, J. Phys. Chem. B, 2010, 114, 10368–10373.
59. Y. Shen, Y. H. Lin and C. W. Nan, Adv. Funct. Mater., 2007, 17, 2405–2410.
60. S. Stankovich, D. A. Dikin, R. D. Piner, K. A. Kohlhaas, A. Kleinhammes, Y. Jia, Y. Wu, S. T. Nguyen and R. S. Ruoff, Carbon, 2007, 45, 1558–1565.
61. Z. M. Dang, J. B. Wu, L. Z. Fan and C. W. Nan, Chem. Phys. Lett., 2003, 376, 389–394.
62. M. Molberg, D. Crespy, P. Rupper, F. Nüesch, J.-A. E. Månson, C. Löwe and D. M. Opris, Adv. Funct. Mater., 2010, 20, 3280–3291.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra21249b

---