A small loading of surface-modified Ba_0.6Sr_0.4TiO₃ nanofiber-filled nanocomposites with enhanced dielectric constant and energy density

Abstract

Compared to spherical ceramic fillers, ceramic fillers with large aspect ratios can increase the dielectric constant of nanocomposites at a much lower concentrations because their large dipole moments, and their smaller specific surface can help to reduce the surface energy and thus prevent the nanofillers from agglomerating in the polymer matrix. Therefore, high energy storage capability of the nanocomposite is obtained using surface-modified Ba_0.6Sr_0.4TiO₃ nanofibers with a large aspect ratio (BST NF) by 3-aminopropyltriethoxysilane (APS) filling in a poly(vinylidene fluoride) polymer (PVDF) matrix. The nanocomposites exhibit enhanced dielectric constant and reduced loss tangents at a low volume fraction of surface-modified BST NF. The maximal energy density in the nanocomposite with 2.5 vol% BST NF-APS is about 6.8 J cm⁻³ at 3800 kV cm⁻¹, about 143% higher than that of the PVDF of 2.8 J cm⁻³ at 4000 kV cm⁻¹. The enhanced energy storage density could be attributed to the combined effects of surface modification by the APS, large aspect ratio and paraelectric polarization behavior of the BST NF. This work may provide a novel route for using the small loading of surface-modified paraelectric ceramic fillers with large aspect ratios for enhanced energy-storage density in polymer composites.

---

Introduction

High energy density capacitors have recently attracted worldwide attention due to their potential applications in modern electronic and electrical power systems such as medical devices, high power microwaves, and hybrid electric vehicles.^1–9 However, these potential applications are severely hindered by the low energy density of capacitors with current commercial dielectrics. In general, the energy-storage density U of a dielectric material is expressed as U = ∫EdD, where E and D are the breakdown strength and electric displacement, respectively.^5,10–13 Based on this equation, it is indicated that the D and the E have to be enhanced simultaneously in order to realize high energy storage density.¹⁴ Currently, the commercial applications for energy storage are neat polymers, such as biaxially oriented polypropylene (BOPP), due to their ease of processability, low cost and high breakdown strength. Nevertheless, ost of the pure polymers for dielectric applications have low dielectric constants (e.g., <10), limiting the energy-storage density and hence their applications.^13,15–17

Thus far, two methods have been developed to prepare polymer composites with high dielectric constants. One classic approach focuses on dispersing conductive fillers, such as carbon black, carbon nanotubes, graphene, and nickel particles, into the polymer matrix to prepare percolative composites.^14,16,17 The advantage of this approach is that the dielectric constant of the percolative composites increases dramatically near to the percolation threshold.^18–20 However, it is difficult to control the dielectric properties of the nanocomposites by adjusting the conductive filler loading because the weak variation of conductive filler loading would cause a remarkable change in the dielectric constant of the nanocomposites in the vicinity of percolation threshold.^21,22 Moreover, these conducting filler–polymer nanocomposites suffer from high dielectric loss and very low breakdown strength.²³ Thus, the percolative composites are not suitable as effective materials for high energy density capacitors. Another strategy is to introduce these high dielectric constant ceramic fillers, such as Pb(Zr,Ti)O₃, Ba_1−xSr_xTiO₃ and BaTiO₃, into a polymer.^17,24,25 The advantage of this approach is that the dielectric constant of ceramic filler–polymer nanocomposites can be easily controlled by adjusting the volume content of the ceramic filler. Moreover, these nanocomposites also usually possess relatively low dielectric loss. However, to achieve high dielectric constants for nanocomposites, a high volume fraction of ceramic filler is normally required, which simultaneously decreases the flexibility of the nanocomposite. Moreover, filler aggregation, voids and other structural imperfections are introduced by high filler loading, leading to a weakening in the breakdown strength of the polymer. Despite the enhancement in the dielectric constant of the polymer nanocomposites, the decrease of the breakdown strength causes a very minor enhancement in energy storage density as compared with the pure polymer matrix. It is important to increase the dielectric constant of the polymer nanocomposites while improving or maintaining dielectric strength in order to obtain a high energy storage density of polymer nanocomposites.²⁶

The dielectric constant of ceramic filler–polymer composite is well described by the effective medium theory, in which the depolarization factor is strongly dependent on the aspect ratio of the ceramic fillers in the nanocomposite. Composites with fillers with small aspect ratios have small dielectric constants, fillers with large aspect ratios are more effective in enhancing the dielectric constants of composites at lower loading. Moreover, phase-field modeling and experimental results all indicate that the orientation of large aspect ratios in ceramic fillers perpendicular to the direction of an external electric field is favorable for mitigating local field concentration in the nanocomposite and leads to higher breakdown strength. The use of high dielectric fillers with large aspect ratios is to achieve universally desirable dielectric properties (large dielectric constant and high breakdown strength) necessary for large energy storage density in dielectrics.

Most of the research about energy storage materials is mainly focused on the fabrication of homogeneous ceramic–polymer nanocomposites, consisting of ferroelectric ceramic nanoparticles as fillers.^4,27,28 However, the high remnant polarization and polarization saturation of these ferroelectric ceramics limited energy storage density. In addition, more electrical energy was stored in the composites with ferroelectric ceramic fillers in the form of remanent dipoles instead of being extractable during the discharge process. When paraelectric ceramic fillers are chosen, in which the paraelectrics decrease the remanent polarization of the ferroelectric polymer matrix and increase the electrical displacement to enhance the energy-storage density such as Ba_xSr_1−xTiO₃ (which can be tuned to obtain the paraelectrics behavior by varying the ceramics stoichiometry), transfer occurs from the ferroelectric phase to the paraelectric phase when the Ba mole fraction is smaller than 0.7 (x < 0.7) at room temperature and below.

In the present study, we report a method to prepare a nanocomposite with high energy density based on high aspect ratio functionalized paraelectric ceramic Ba_0.6Sr_0.4TiO₃ nanofiber (BST NF) fillers dispersed in a PVDF polymer. An electrospinning technique is employed to prepare BST NF with high aspect ratios, and 3-aminopropyltriethoxysilane (APS) is used as the surface modification to improve the compatibility between the nanofiber fillers and the PVDF matrix. The dielectric properties and energy storage properties of the composites have been investigated as a function of the functionalized BST NF content. Enhanced dielectric constants and energy density of the composites were obtained at a low volume fraction of functionalized BST NF, due to the modification by the APS, with a large aspect ratio and paraelectric polarization behavior of the BST NF.

Experimental

The BST NF prepared via electrospinning were employed as dielectric fillers in PVDF-based nanocomposites. The barium acetate (99.0%, Alfa Aesar) and strontium acetate hemihydrate (99.0%, Alfa Aesar) at a molar ratio of 6 : 4 were added in 5 ml acetic acid and stirred for 1 h. Moreover, the titanium(IV) n-butoxide (99.0%, Alfa Aesar) was added in 5 ml acetylacetone and stirred for 1 h. Then, the dissolved solution of all compounds was mixed with a solution consisting of poly(vinyl pyrrolidone) (PVP, M_w = 1 300 000) dissolved in ethanol (PVP: 2 g and ethanol: 3 ml). The mixture was stirred at room temperature for 1 h to obtain a homogeneous precursor barium strontium titanate precursor sol solution for electrospinning. The precursor sol was transferred into a plastic syringe and electrospun with an applied electric field of 1.5 kV cm⁻¹. The composite nanofibers were collected on the copper plate during the electrospinning processes and finally annealed at 900 °C for 3 h in air to completely remove PVP.

The BST NF were dispersed into an aqueous solution of H₂O₂ (35%, 350 ml) and sonicated for 30 min, and then stirred at 100 °C for 3 h. The mixture was collected by centrifugation at 5000 rpm for 5 min. The obtained mixture was washed several times with distilled water and ethanol in sequence and then dried under vacuum at 80 °C for 12 h. 2 g of surface hydroxylated BST NF (BST NF-OH) were added into 120 ml of ethanol and sonicated for 30 min, and then APS at concentrations of 0.5 wt% in comparison with the mass of BST NF-OH was added with stirring and further stirred at 80 °C for 5 h. The mixture was collected by centrifugation at 5000 rpm for 5 min, and then washed several times with distilled water and ethanol in sequence and subsequently dried at 80 °C for 12 h to obtain the APS-modified BST NF (BST NF-APS).

The nanocomposites were prepared via a tape-casting method. The BST NF-APS and PVDF powders (3F Co., China.) were proportionally dispersed in N,N-dimethylformamide (DMF) under vigorous stirring at 40 °C for 10 h to form a stable suspension. The suspension was cast onto an indium tin oxide (ITO) glass (as the bottom electrode) and dried under vacuum at 60 °C for 10 h. The obtained films were heated at 200 °C for 10 min and quenched in an ice-water bath immediately and subsequently dried at 40 °C for 24 h. The nanocomposite films were about 10 μm in thickness. Top Au electrodes were deposited onto the films using DC sputtering and a shadow mask for electrical measurements. Fig. 1 shows the schematic diagram of the fabrication of BST NF-APS/PVDF nanocomposites.

Fig. 1 Schematic diagram of the fabrication of BST NF-APS/PVDF nanocomposites.

Characterization

The microstructure was characterized using a field emission scanning electron microscope (XL30-FEG, Philips, Netherlands). Fourier-transform infrared spectroscopy (FTIR) was measured using a Bruker Tensor 27 spectrometer. Dielectric properties of the composites were measured using a 4980A LCR meter (Agilent, Palo Alto, CA, USA) in the frequency range from 100 Hz to 2 MHz at various temperatures. Breakdown strength (BDS) measurements were performed using a DC dielectric withstand voltage test (ENTAI, Nanjing, China) in a silicone oil bath at room temperature (25 °C) by applying a DC voltage ramp with a rate of rise of 200 V s⁻¹ and a limit current of 5 mA. The polarization–electric field loops (P–E) were measured using the Premier II ferroelectric test system.

Results and discussion

The thermal decomposition behavior of BST/PVP nanofiber is shown in Fig. 2. The weight loss in the TG curve occurred until 600 °C. The endothermic peak at 351 °C in the DSC curve was observed due to polymer decomposition, whereas exothermic peaks at 351, 450 and 514 °C were observed. Dramatic weight losses occurred around 351 °C due to the decomposition of organic groups. Another exothermic peak at 514 °C may correspond to the decomposition of the main chain of PVP and the formation of the metal-oxide phase of perovskite BST. The TG/DSC results revealed that most of the organic groups vanished approximately at 700 °C. Thus, BST/PVP nanofiber was annealed at 900 °C for 3 h to obtain good crystallinity and remove the PVP.

Fig. 2 TG/DSC curves of electrospun BST/PVP composite nanofiber.

Fig. 3 shows the XRD patterns of BST NF prepared via electrospinning. These peaks indicate that a pure perovskite phase is obtained in the BST NF. SEM reveals that the nanofibers have large aspect ratios, i.e., diameters of 100–150 nm and lengths of tens of micrometers, as seen from the SEM image (inset of Fig. 3).

Fig. 3 XRD patterns of BST NF. SEM image of BST NF is shown in the inset.

Fig. 4(a) and (b) shows the TEM images of BST-NF and BST-NF after surface modification with H₂O₂ and APS. Surface modification with H₂O₂ and APS does not influence the morphologies of the nanofibers. Note that they are several micrometers long with diameters of approximately 120 nm. These results exhibit the structural integrity of the BST-NF and BST-NF after being treated with H₂O₂ and APS.

Fig. 4 TEM images of (a) BST-NF and (b) BST-NF after surface modification with H₂O₂ and APS.

The Fourier transform infrared spectra result indicates that APS was grafted onto the surface of BST NF and formed robust bindings (Fig. 5). The hydroxylation of the BST NF (BST NF-OH) shows a new broad absorption band at about 3450 cm⁻¹, compared with the untreated BST NF, which is associated with the stretching mode of –OH. This result indicates that OH groups were introduced onto the surface of BST NF after treatment with H₂O₂. The appearance of 1450 cm⁻¹ absorption bands is associated to second amine group. Moreover, the appearance of 2925 cm⁻¹ and 2854 cm⁻¹ absorption peaks are assigned to the asymmetric and symmetric –CH₂ stretching vibrations. These changes of FTIR spectra suggest that the APS was successfully grafted onto the hydroxyl group on the surface of BST NF and obtained the surface modification BST NF (BST NF-APS).

Fig. 5 FTIR spectra of the BST NF, BST NF-OH and BST NF-APS.

Fig. 6 shows typical surface and cross-sectional SEM images of the nanocomposite filled with various amounts of fillers. Fig. 6(a) and (b) show the typical surface SEM images of nanocomposites with untreated BST NF and those with surface modified BST NF (the BST NF content is 5 vol%). The adhesion between BST NF and PVDF is weak. Voids and debonding between the BST NF and PVDF could be observed in the nanocomposites. Moreover, some voids and pores can be observed in BST NF/PVDF nanocomposite film. The nanocomposite films with surface modified BST NF have hardly any small voids and are dispersed well in the nanocomposite. Fig. 6(c) and (d) show the surface SEM images of the 7.5 vol% BST NF-APS/PVDF nanocomposites and cross-sectional SEM of 5 vol% BST NF-APS/PVDF nanocomposites. These SEM images show that the surface modified nanofibers with large aspect ratios have been successfully transferred to the polymer matrix with minimum agglomeration from solution, and BST NF-APS tend to orient in the in-plane directions of the nanocomposite films. The SEM images also show that the composite films are dense by introducing the surface-modified BST NF into the PVDF matrix. These results indicate that homogeneous dispersions of surface modified BST NF with large aspect ratios in PVDF matrices could be achieved as a result of improved compatibility between BST NF and polymer matrices.

Fig. 6 (a) Surface SEM of 5 vol% BST NF/PVDF nanocomposites, (b) the 5 vol% BST NF-APS/PVDF nanocomposites, (c) the 7.5 vol% BST NF-APS/PVDF nanocomposites and (d) cross-sectional SEM of 5 vol% BST NF-APS/PVDF nanocomposites.

Neat PVDF has three different crystal forms, including α-PVDF, β-PVDF, γ-PVDF. The energy storage properties of PVDF are highly influenced by the crystalline phase. Li et al.¹¹ demonstrated that the α-phase PVDF obtained by quenching in ice water cannot only increase the breakdown strength of the PVDF but also improves the energy density compared to the β-PVDF. The quenching process is used to modify the crystallization of the PVDF in the nanocomposites. FTIR results for 2.5 vol% BST-APS/PVDF nanocomposites before and after the quenched technique are shown in Fig. 7. The quenched PVDF nanocomposites is primarily in the α-phase (611, 765, 795, 855, and 975 cm⁻¹), whereas the untreated PVDF nanocomposites is dominated by the β-phase (840, 878, 1171, and 1232 cm⁻¹).

Fig. 7 FTIR results for 2.5 vol% BT-APS/PVDF nanocomposites before and after the quenched technique.

The dielectric properties of BST NF-APS/PVDF nanocomposite films at room temperature are shown in Fig. 8. The figure clearly demonstrates that the dielectric constant of the nanocomposite film increases with the increase of the BST NF-APS content over the entire measurement frequency range, which is due to the much higher dielectric constant of the BST NF in comparison with the polymer matrix. The dielectric constant of BST NF-APS/PVDF nanocomposite film can reach up to 21.78 at 1 kHz with a small loading of 7.5% of BST NF-APS, nearly 2.66 times higher than that of the pure polymer. The much increased dielectric constant of the composite film with a small loading of BST NF-APS can be described by the Maxwell-Garnett expression by considering the shape of the BST NF. According to the Maxwell-Garnett model, the dielectric constant of the nanocomposites is expressed as follows:

where ε_eff is the dielectric constant of the composite, ε_f and ε_p are the dielectric constants of the BST NF and PVDF, respectively, f is the volume fraction of the BST NF-APS and N_j is the depolarization factor of ellipsoids in the x, y and z directions. For the fillers with large aspect ratios, where the radii a_x > a_y = a_z, a simple expression of N_j is

Fig. 8 Frequency dependencies of the (a) dielectric constant and (b) loss tangent of the PVDF and the BST NF-APS/PVDF nanocomposite film measured at room temperature. Experimental data compared with Maxwell-Garnett rule of mixtures for high aspect ratios is shown in the inset.

The calculated parameters with and ε_f = 1000, ε_p = 7.9 are in agreement with the measured data (see the inset of Fig. 8(a) for the BST NF-APS/PVDF nanocomposite film). Moreover, the dielectric loss of BST NF-APS/PVDF still remains at a low level over a wider range of frequency. For example, the dielectric loss of nanocomposite film is 0.015 at 1 kHz with a small loading of 7.5% of BST NF-APS, even lower than that of pure PVDF (dielectric loss is 0.024).

The dielectric properties of the nanocomposite film with the treated and untreated BST NF have been investigated as a function of the BST NF filler loading at a frequency of 1 kHz. As shown in Fig. 9, compared with the BST NF/PVDF, the BST NF-APS/PVDF nanocomposite film with the same BST NF loading content not only shows a higher dielectric constant but also has a lower dielectric loss at a frequency of 1 kHz. The good dispersion of fillers and the good interfacial adhesion between the surface-modified fillers and the matrix are the important factors resulting in higher dielectric constant and lower dielectric loss. Regarding the effect of composite microstructures on the dielectric properties of nanocomposites, it is believed that poor dispersion of the BST NF and the poor interfacial adhesion in the BST NF/PVDF, which might cause some pores and voids in the nanocomposites, in particular at high nanofiber loading,^17,29 will lead to lower dielectric constants and higher dielectric losses in the BST NF/PVDF nanocomposites.³⁰ Moreover, the slight difference in the dielectric constant between the nanocomposites of unmodifed and modified BST NF could be attributed to the low concentration of the ceramic fillers, as reported previously.¹⁰

Fig. 9 (a) Dielectric constants and (b) loss tangents of BST NF/PVDF and BST NF-APS/PVDF nanocomposite films loaded with various concentrations of fillers measured at 1 kHz.

The breakdown electrical field strength is an important parameter for the practical applications because it determines the operative electrical field and the maximum energy storage density of dielectric materials. Breakdown strength is analyzed using Weibull distribution. Fig. 10(a) and b show the Weibull plots of breakdown strength for the nanocomposite film loaded with various concentrations of the treated BST NF fillers and untreated BST NF fillers. The breakdown strength could be extracted from the data points where the fitting lines with the horizontal line through Y = 0. According to the fitting line, the pure PVDF film has a high breakdown strength of 3670 kV cm⁻¹. Fig. 11 shows the breakdown strength for BST NF/PVDF and BST NF-APS/PVDF nanocomposites loaded with various concentrations of fillers. The breakdown strength of the nanocomposite films with the surface-modified BST NF and those with untreated BST NF show a decrease as the content of fillers increases from 0 to 7.5 vol%, which may be attributed to inhomogeneous electrical field strength in the polymer matrix, which could decrease the breakdown strength of the nanocomposite film. When the dielectrics with high dielectric constant are added into the polymer, the distribution of the electrical fields in the nanocomposites are distorted by the fillers due to the large difference in dielectric properties between the two phases, leaving the electrical field in the polymer around the fillers much higher than the average electrical field. This leads to a decrease in the breakdown strength of the nanocomposites.³¹ On the other hand, the breakdown strength of the polymer–matrix composites is strongly influenced by agglomerations and defects. Considerably more agglomerations and defects such as voids are introduced into the nanocomposites as the ceramic fillers content increases. This also results in a sharp decrease in the breakdown strength because electrical breakdown usually takes place in the weakest part of the materials.³² It also should be noted that the breakdown strength of the nanocomposites with the surface modified BST NF is higher than that with untreated BST NF. Particularly, when the volume fraction of BST NF is 7.5%, the electrical breakdown strength of the nanocomposites with the surface modified BST NF is 2966 kV cm⁻¹, about 33% higher than that with untreated BST NF of 2229 kV cm⁻¹. This result should be mainly attributed to surface modification by APS. The established bindings have been indicated by the FTIR results shown in Fig. 5. As shown in the SEM image of the cross-section of the nanocomposites in Fig. 6, the APS contributes to the good dispersion of BST NF-APS in a PVDF matrix, because the APS acts as two functional groups. The alkoxysilane molecules react with the hydroxyl group on the surface of BST NF; moreover, the amino group can react with the fluorine group of the PVDF matrix.³³ As a result, robust bindings have been established between surface modified BST NF and PVDF matrix. The surface modification the BST NF by APS contributes to the homogeneous distribution of the BST NF in the polymer matrix. These factors play active roles in the enhanced breakdown properties.

Fig. 10 Weibull plots of the electrical strength for BST NF/PVDF and BST NF-APS/PVDF nanocomposites loaded with various concentrations of fillers.

Fig. 11 Breakdown strength for BST NF/PVDF and BST NF-APS/PVDF nanocomposites loaded with various concentrations of fillers.

Energy density is not only related to the dielectric constant and the breakdown strength, but also related to the polarization and applied electrical field. For ferroelectrics, because the polarization is not linearly dependent on the electrical field, the polarization and dielectric constant of ferroelectric materials have strong dependence on a variety of external conditions, particularly the applied electrical field for energy storage. The energy storage density should be calculated from the P–E loops. The energy density of a dielectric material is equal to the integral U = ∫EdD, where E is the electric field and D is the electric displacement. Therefore, in addition to obtaining high breakdown strength, high electrical displacement or dielectric constant is a key factor in achieving high energy density.

The P–E loops at 100 Hz of these nanocomposites are presented in Fig. 12(a), which were measured at room temperature and at their critical breakdown field. The addition of the BST NF-APS into the PVDF polymers greatly increased the maximum polarization of the nanocomposites, which revealed that the dielectric constant of BST NF is larger than that of PVDF.

Fig. 12 (a) P–E curves of BST NF-APS/PVDF nanocomposites before the nanocomposites broke down. (b) The efficiency and energy storage density of BST NF-APS/PVDF nanocomposites loaded with various filler concentrations.

The room-temperature recordable energy-storage density and energy-storage efficiency η of nanocomposites added by 0, 2.5 vol%, 5 vol%, and 7.5 vol% are illustrated in Fig. 12(b), respectively, which were measured at different electric fields up to breakdown. As expected, energy-storage density was increased as the measurement fields increased for all of the samples. Moreover, the calculated energy-storage densities were strongly dependent on the BST NF-APS content. In the measurement range, the 2.5 vol% BST NF-APS-added nanocomposites had the largest recordable energy-storage density with a value of 6.8 J cm⁻³ measured at 3800 kV cm⁻¹, which was about 143% higher than that of the PVDF of 2.8 J cm⁻³ at 4000 kV cm⁻¹. The great enhancement of energy density is mostly attributed to the large aspect ratio and surface modification of BST NF. The result indicates that surface-modified ceramic nanofibers by APS facilitate energy storage applications of ceramic–polymer nanocomposites. Apart from higher U values, higher energy-storage efficiency was also always desired in practical application. The energy-storage efficiency η could be calculated according to the formula where W_loss is the energy loss density. The energy loss density was calculated by the numerical integration of the closed area of the hysteresis loops. As shown in Fig. 12(b), different from the energy storage density, as the measurement field increased, the energy storage efficiency was reduced gradually for all of the samples. However, the energy-storage efficiency of the nanocomposites was also improved by the addition of BST NF-APS. The more interesting factor was that the highest η values also appeared with 2.5 vol% BST NF-APS added into the nanocomposites, in which the η is larger than 80% at fields below 1000 kV cm⁻¹ and still higher than 60% at an electrical field of 3800 kV cm⁻¹.

Conclusions

In summary, we have investigated the effects of the surface modification of BST NF ceramic fillers by an APS agent on dielectric and electrical breakdowns and energy storage properties of BST NF filled polymer nanocomposites. The enhancement of dielectric constants and reduced loss tangents has been achieved on the nanocomposites of low volume fractions of surface modified BST NF. The energy storage density of 6.8 J cm⁻³ at 3800 kV cm⁻¹ of the nanocomposites with 2.5 vol% BST NF-APS has been obtained, which is 143% higher than that of the pure PVDF. Such significant enhancement is closely related to the combined effect of the large aspect ratio and the surface modification of BST NF. The results indicate a small loading of surface-modified ceramic nanofibers can be used to enhance the energy storage density of the nanocomposites, which may contribute to the development of ceramic–polymer nanocomposites for energy storage applications.

Acknowledgements

This research was supported by the Ministry of Sciences and Technology of China through 973-Project under Grant (2009CB623302) and the National 863 Program (2012AA03A706).

References

1. Z. M. Dang, J. K. Yuan, S. H. Yao and R. J. Liao, Adv. Mater., 2013, 25, 6334.
2. Z. M. Dang, J. K. Yuan, J. W. Zha, T. Zhou, S. T. Li and G. H. Hu, Prog. Mater. Sci., 2012, 57, 660.
3. H. X. Tang and H. A. Sodano, Nano Lett., 2013, 13, 1373.
4. K. Yang, X. Y. Huang, Y. H. Huang, L. Y. Xie and P. K. Jiang, Chem. Mater., 2013, 25, 2327.
5. B. J. Chu, X. Zhou, K. L. Ren, B. Neese, M. R. Lin, Q. Wang, F. Bauer and Q. M. Zhang, Science, 2006, 313, 334.
6. B. Neese, B. J. Chu, S. G. Lu, Y. Wang, E. Furman and Q. M. Zhang, Science, 2008, 321, 821.
7. J. J. Li, S. I. Seok, B. J. Chu, F. Dogan, Q. M. Zhang and Q. Wang, Adv. Mater., 2009, 21, 217.
8. J. J. Li, J. Claude, L. E. Norena-Franco, S. Il Seok and Q. Wang, Chem. Mater., 2008, 20, 6304.
9. M. F. Lin, V. K. Thakur, E. J. Tan and P. S. Lee, J. Mater. Chem., 2011, 21, 16500.
10. J. P. Calame, J. Appl. Phys., 2008, 104, 114108.
11. W. J. Li, Q. J. Meng, Y. S. Zheng, Z. C. Zhang, W. M. Xia and Z. Xu, Appl. Phys. Lett., 2010, 96, 192905.
12. W. M. Xia, Z. Xu, F. Wen and Z. C. Zhang, Ceram. Int., 2012, 38, 1071.
13. M. N. Almadhoun, U. S. Bhansali and H. N. Alshareef, J. Mater. Chem., 2012, 22, 11196.
14. S. Tong, B. H. Ma, M. Narayanan, S. S. Liu, R. Koritala, U. Balachandran and D. L. Shi, ACS Appl. Mater. Interfaces, 2013, 5, 1474.
15. B. C. Luo, X. H. Wang, Y. P. Wang and L. T. Li, J. Mater. Chem. A, 2014, 2, 510.
16. P. Kim, N. M. Doss, J. P. Tillotson, P. J. Hotchkiss, M. J. Pan, S. R. Marder, J. Y. Li, J. P. Calame and J. W. Perry, ACS Nano, 2009, 3, 2581.
17. T. Zhou, J. W. Zha, R. Y. Cui, B. H. Fan, J. K. Yuan and Z. M. Dang, ACS Appl. Mater. Interfaces, 2011, 3, 2184.
18. T. Zhou, J. W. Zha, Y. Hou, D. R. Wang, J. Zhao and Z. M. Dang, ACS Appl. Mater. Interfaces, 2011, 3, 4557.
19. C. Wu, X. Y. Huang, X. F. Wu, L. Y. Xie, K. Yang and P. K. Jiang, Nanoscale, 2013, 5, 3847.
20. H. Y. Liu, Y. Shen, Y. Song, C. W. Nan, Y. H. Lin and X. P. Yang, Adv. Mater., 2011, 23, 5104.
21. H.-J. Ye, L. Yang, W.-Z. Shao, Y. Li, S.-B. Sun and L. Zhen, RSC Adv., 2014, 4, 13525.
22. X.-J. Zhang, G.-S. Wang, W.-Q. Cao, Y.-Z. Wei, M.-S. Cao and L. Guo, RSC Adv., 2014, 4, 19594.
23. Z. M. Dang, L. Wang, Y. Yin, Q. Zhang and Q. Q. Lei, Adv. Mater., 2007, 19, 852.
24. H. Hammami, M. Arous, M. Lagache and A. Kallel, J. Alloys Compd., 2007, 430, 1.
25. S. H. Liu, J. W. Zhai, J. W. Wang, S. X. Xue and W. Q. Zhang, ACS Appl. Mater. Interfaces, 2014, 6, 1533.
26. P. H. Hu, Y. Song, H. Y. Liu, Y. Shen, Y. H. Lin and C. W. Nan, J. Mater. Chem. A, 2013, 1, 1688.
27. K. Yang, X. Y. Huang, L. Y. Xie, C. Wu, P. K. Jiang and T. Tanaka, Macromol. Rapid Commun., 2012, 33, 1921.
28. L. Y. Xie, X. Y. Huang, C. Wu and P. K. Jiang, J. Mater. Chem., 2011, 21, 5897.
29. K. Yu, Y. J. Niu, Y. C. Zhou, Y. Y. Bai and H. Wang, J. Am. Ceram. Soc., 2013, 96, 2519.
30. Y. Song, Y. Shen, H. Y. Liu, Y. H. Lin, M. Li and C. W. Nan, J. Mater. Chem., 2012, 22, 16491.
31. K. Yu, Y. J. Niu, F. Xiang, Y. C. Zhou, Y. Y. Bai and H. Wang, J. Appl. Phys., 2013, 114, 174107.
32. Z. M. Dang, Y. H. Lin and C. W. Nan, Adv. Mater., 2003, 15, 1625.
33. Z. M. Dang, H. Y. Wang and H. P. Xu, Appl. Phys. Lett., 2006, 89, 112902.

---