New crystal structure and discharge efficiency of poly(vinylidene fluoride-hexafluoropropylene)/poly(methyl methacrylate) blend films

Abstract

A series of PMMA/poly(vinylidene fluoride-hexafluoropropylene) [P(VDF-HFP)] blend films were prepared with a solution blending process to improve discharge efficiency. The crystal structure and properties of the PMMA/P(VDF-HFP) blend films were carefully studied. The results show that the crystallinity, the dielectric constant and the loss of the polymer blend films decreased, while melting temperature increased with the increase in PMMA content. The XRD results demonstrate the existence of a γ-2b phase. It is believed that the phase transition of the P(VDF-HFP) crystals from α to γ-2b contributes to the increase in melting temperature. The energy loss of P(VDF-HFP) reduced significantly with the addition of PMMA. The blend film with 10% PMMA showed the best discharge efficiency with a value above 80% in 300 MV m⁻¹. This increase in discharge efficiency is thought to result mainly from the formation of γ crystals.

---

Introduction

Poly(vinylidene fluoride) (PVDF) and its copolymers, such as those with chlorotrifluoroethylene [P(VDF-CTFE)], hexafluoropropylene [P(VDF-HFP)], and trifluoroethylene [P(VDF-TrFE)], are potential dielectric materials for a broad range of applications such as grid leveling, pulsed lasers and electric or hybrid vehicles.¹ These ferroelectric polymers exhibit a larger energy density than widely used capacitor polymeric films, such as biaxially oriented-polypropylene (BOPP). Compared to BOPP, which presents an energy density of 4 J cm⁻³ at 600 MV m⁻¹ and a low dielectric constant of 2.2 at 1 kHz, uniaxially copolymer P(VDF-CTFE) (ref. 2) possesses a high dielectric constant of 13 at 1 kHz and an energy density of 25 J cm⁻³ at a field of 620 MV m⁻¹, copolymer P(VDF-HFP) (ref. 3) (4.5 mol% HFP) displays an energy density of 25 J cm⁻³ at a field of 700 MV m⁻¹ and P(VDF-TrF-CTFE) (ref. 4) (65.6/26.7/7.7 mol%) has a dielectric constant above 60, exhibits an energy density of 13 J cm⁻³ at 500 MV m⁻¹ and a discharge efficiency of about 60% at a field of 300 MV m⁻¹. However, there are still some challenges to commercialize these ferroelectric polymers, the most critical issue is the discharge efficiency.

PVDF is a semi-crystalline polymer with about 50% crystallinity and good mechanical properties, chemical resistance, electric resistance and processability. PVDF exists in at least five crystalline forms [α(II), β(I), γ(III), δ(IV), ε(V)].^5–8 Recently, a new crystalline variety (γ-2b) of PVDF was reported by Maja Remskar.⁹ The main chain conformations of the α, β and γ phases are TGTG, TTTT and TTGTTG respectively. Their melting temperatures (T_m) vary with their crystalline structure. The glass transition temperature (T_g) of PVDF is below −37 °C, demonstrating that PVDF is rubbery at ambient temperature. The α-phase is non-polar, and is obtained for PVDF and its copolymer under the usual melt crystallization.¹⁰ Rinaldo Gregorio, Jr.¹¹ reported that the β phase forms exclusively when PVDF crystallizes from solution at a temperature lower than 70 °C, regardless of the solvent used, as long as it is a good solvent for this polymer. The γ phase forms via the melt or anneals at a temperature close to the T_m of the α phase. The polar β crystal possesses a large energy loss for its spontaneous polarization and hysteresis of ferroelectric switching.¹²

Many researchers have prepared copolymers such as P(VDF-CTFE), P(VDF-TFE), P(VDF-TrFE) and P(VDF-CTFE-TrFE) to tailor the crystallization of the polar β-phase for improving the discharge efficiency of capacitor films. However, the copolymer is very expensive. Another way to tailor the crystallization of the polar β-phase is to blend with other polymers. A commonly used polymer is poly(methyl methacrylate) (PMMA). By melt blending or solution blending, polymer alloy of PMMA/PVDF can be prepared. In the laboratory, solution blending is used mostly for its simplicity in operation and equipment. In most investigations, PVDF is considered compatible with PMMA. Wang and Nishi^13,14 reported that semi-crystalline PVDF is compatible with amorphous poly(methyl methacrylate) (PMMA) in the molten state. Recently, Zhang¹⁵ reported that the addition of PMMA into PVDF via solution blending and a quenching process, resulted in the crystallinity and crystal size of PVDF decreasing. As a result, the energy loss decreased accordingly. Li¹⁶ has investigated the ferroelectric phase diagram of P(VDF-TrFE)/PMMA. The ferroelectric β-phase was observed when the blend films were melted, ice quenched and subsequently annealed close to the melting temperature of the α phase. Meanwhile, the crystallinity, the remnant polarization and the roughness of the blend films decreased with increasing PMMA content.

In brief, most researchers proved that the addition of PMMA in PVDF can reduce the crystal size of β-phase and improve the discharge efficiency. In this paper, the effects of PMMA on the transition of crystalline structure, degree of crystallinity and the molecular mobility of P(VDF-HFP)/PMMA blend films are investigated, and then the influence of crystalline structure on discharge efficiency was analyzed via D–E loops. The results indicated that the addition of PMMA could facilitate the formation of a new crystal variety (γ-2b) of P(VDF-HFP), which leads to the improvement of discharge efficiency.

Experimental

Material

P(VDF-HFP) (Kynar Flex 2751-00) in powder form was purchased from Arkema corporation, France. PMMA(CM-207) pellets (M_n = 38 406 g mol⁻¹, M_w = 72 777 g mol⁻¹, M_w/M_n = 1.9) were purchased from Chi Mei corporation, Taiwan. AR grade N,N-dimethylformamide (DMF) was purchased from Tianjin Reagents Co. Ltd.

Process of film preparation

P(VDF-HFP) and PMMA were dissolved in DMF to make polymer solutions with a DMF concentration of 90 wt%. The weight ratio of P(VDF-HFP) to PMMA was 100 : 0, 95 : 5, 90 : 10 and 70 : 30. The blend solutions were agitated magnetically for 6 h, ultrasonically agitated for 30 min, and cast onto glass sheets, followed by drying at 120 °C, and subsequently at 160 °C under vacuum for 8 h. The thickness of the films was about 30 ± 5 μm.

Characterization

FTIR spectroscopy measurements were performed on an IR spectrometer (E55+FRA106) at 0.5 cm⁻¹ resolution and in the 600–4000 cm⁻¹ wave number range in attenuated total reflection mode (FTIR-ATR). X-ray diffraction (XRD) measurements were performed on an X-ray diffractometer (M03XHF22, MACScience corporation, Japan) with Cu K_α radiation (λ = 0.154 nm), a scanning range of 2θ = 10–55°, and a scanning rate of 5° min⁻¹. Differential scanning calorimetry (DSC) analysis was conducted on a Netzsch STA449C (Netzsch Corp., Germany) under Ar protection to measure melting temperature and crystallinity at a heating rate of 5 °C min⁻¹.

The dielectric constant and loss were measured at different frequencies ranging from 100 Hz to 1 MHz, in a temperature range of 0–150 °C and at 1 V on a broad frequency dielectric spectrometer (Concept 80, Germany). Relative permittivity ε is calculated by the following formula: where ε is the relative permittivity, ε₀ is the permittivity of free space (8.854 × 10⁻¹² F m⁻¹), C_p is the capacitance, t is the material thickness, A is the area of the electrode, and d is the diameter of the electrode. The thickness of each sample is measured at three different points using a micrometer with an accuracy of 0.001 mm, the max range of the thickness error of the samples is ± 1μm. Our sample for each composition is uniform in thickness.

The electric D–E loops were measured with a modified Sawyer–Tower circuit. The samples were subjected to a unipolar wave with a frequency of 10 Hz. The D–E loops are presented according to the data from the first cycles.

Results and discussion

3.1 Structural characterization of PMMA/PVDF blends

FTIR-ATR spectra were used to investigate the crystalline phases of the PVDF–HFP/PMMA composite films with various amounts of PMMA, and the results are shown in Fig. 1. The peak at 879 cm⁻¹ was attributed to the amorphous phase of PVDF–HFP and could not be used to identify any of the crystalline phases.¹⁵ The absorption peaks at 1724.32, 1726.22, 1728.2, and 1730.01 cm⁻¹ were ascribed to the stretching of the carbonyl groups, which suggests the existence of PMMA in the blends. Usually, the asymmetrical carbonyl bond absorptions result from the coexistence of dissociated carbonyl and hydrogen bonded carbonyl. The dissociated carbonyl bonds are usually located at 1733 cm⁻¹. When hydrogen bonds are formed, the absorption peak will shift to the red region. The red-shift wave-number increases with hydrogen bond strength. For the blends of PMMA/PVDF, the absorption of the carbonyl groups shifted to higher wave-numbers compared with pure PMMA, but the absorption peaks are still asymmetrical. The asymmetrical carbonyl absorptions are enlarged and fitted with a Gauss function, as shown in Fig. 2. It can be seen that the asymmetrical peak can be split into two peaks. If the dissociated carbonyl absorption is set at a fixed position, then the position of the hydrogen bonded carbonyl shifts from 1719.5 cm⁻¹ for pure PMMA to 1721.5 cm⁻¹ for the mixture, which means that the hydrogen bonds might change its strength. These results indicate that the hydrogen bonds within PMMA might change to those between PMMA and PVDF.

Fig. 1 FTIR spectra of PVDF–HFP/PMMA thin films with various amounts of PMMA.

Fig. 2 Results of carbonyl absorptions fitted with a Gauss function for pure PMMA (a) and 30% PMMA/PVDF (b).

The absorption peaks at 613 cm⁻¹, 763 cm⁻¹, 1072 cm⁻¹ and 1402 cm⁻¹ were assigned to the α phase of PVDF–HFP. The peaks at 840 cm⁻¹ and 1431 cm⁻¹ are characteristic of the β phase^17,18 which are multiplied with the absorption peaks of PMMA at the wave-number. The bonds at 813 cm⁻¹, 834 cm⁻¹ and 1231 cm⁻¹ were associated with the γ phase of PVDF–HFP,¹⁹ which are affected by the absorption of PMMA. From the figure, after blending with PMMA, more absorptions of PMMA were revealed with increasing PMMA content, the absorptions of PVDF showed no significant changes, and the α and γ phases were in the majority.

X-ray diffraction patterns of the PVDF–HFP copolymer and PVDF–HFP/PMMA blends are shown in Fig. 3. The three peaks at 2θ = 18.4, 19.8 and 26.6 observed in the PVDF–HFP copolymer and PVDF–HFP/PMMA blend thin films, correspond to the reflections of α(020), (110) and (021) respectively. The peak at 2θ = 20.8 was observed in blends as a single or shoulder peak, and corresponds to the planes (110) and (200) of the β phase.¹⁵ In addition, two peaks at 2θ = 21.4 and 23.8 were observed in all thin films, which correspond to the reflections of γ-2b (022) and (032) respectively, according to the work of Maja Remskar.⁹ Compared with the peak of the α crystal, the intensity at 2θ = 21.4 and 23.8 increased significantly with increasing PMMA content. This showed that the addition of PMMA promoted the formation of the γ-2b phase under the heat treatment conditions.

Fig. 3 X-ray diffraction patterns of PVDF–HFP/PMMA blends with various amounts of PMMA.

The DSC heating thermograms of the PVDF–HFP film and P(VDF-HFP)/PMMA blend films are shown in Fig. 4. The melting temperature of PVDF-HFP film was about 147.6 °C, with the addition of PMMA, the melting temperature of films increased to 150 °C (10% PMMA) and 153.2 °C (30% PMMA). This should be due to the formation of the γ-2b phase in the polymer. The melting temperature of the γ crystal is higher than that of the α and β phases.²⁰ The crystallinity of the polymer decreased obviously when the content of PMMA increased to 30%. This change might be induced by the dilution effect of PMMA.

Fig. 4 DSC curves of P(VDF-HFP) blends with various amounts of PMMA.

3.2 Dielectric properties of PMMA/PVDF blends

Fig. 5 shows the variation in dielectric constant of P(VDF-HFP)/PMMA with PMMA content. As shown in the figure, the dielectric constant decreased with increasing PMMA content, especially at high frequencies (>10⁵ Hz). It is mostly because the flip of the polar chain segment lags behind the high frequency. Compared to blend films, the dielectric constant of the P(VDF-HFP) film sharply decreased with increasing frequency, indicating that the frequency dependence of the P(VDF-HFP) film is stronger than the blend films. In addition, the dielectric constant of the blend films decreased due to the dilution effect of amorphous PMMA. From the structure, the γ phase can show a bigger polarity than the α crystal, and could benefit the increase in dielectric constant, however, the experimental results show an obvious decrease, which means that the γ phase can not compensate for the loss in dielectric constant resulting from the addition of PMMA.

Fig. 5 Dielectric constant of PMMA/P(VDF-HFP) with different amounts of PMMA at room temperature (about 20 °C).

Fig. 6 shows the variation of tan δ of P(VDF-HFP)/PMMA with PMMA content. The dielectric loss of all thin films at about 20 °C changed in V-shape with different frequencies of P(VDF-HFP)/PMMA. For blend films, the dielectric loss is significantly lower than P(VDF-HFP) at a lower frequency, and is slightly lower at higher frequencies. The dielectric loss mainly comes from conductance loss (<10⁴) and polarization loss (10⁴–10⁶). The conductance loss decreases with increasing frequency at a low frequency, and the polarization loss is negligible. Therefore, the dielectric loss of blend films under a low frequency would reduce with increasing frequency, but gradually the relaxation polarization couldn't keep up with the change in the electric field frequency at high frequency, and the hysteresis phenomenon began to appear, thus polarization loss was produced. The rising speed of this polarization loss is much higher than the reduced speed of the conductance loss. The dielectric loss therefore increased with frequency at a higher frequency range.²¹ When the PMMA content is 30%, the dielectric loss significantly decreased due to the difficulty in flipping the dipole moments. The reason for this is that the interactions of C=O in PMMA and CH₂ in PVDF hinder the movement of the chains and decrease the flip of the dipole. Thus, the dielectric loss of the blend films decreased.

Fig. 6 Dielectric loss vs. frequency of P(VDF-HFP)/PMMA films with different amounts of PMMA at room temperature (about 20 °C).

The temperature dependence of the dielectric constant for P(VDF-HFP) and P(VDF-HFP)/PMMA films was measured at 1.05 kHz. As shown in Fig. 7a, for P(VDF-HFP), the dielectric constant firstly increased slightly with the temperature increase; sharp increase was seen when the temperature reached to 50 °C. The dielectric constants of all of the thin films increased with temperature due to an increase in the total polarization, which in the polyblend might arise from dipoles and trapped charge carriers, occurring at the interface between the crystal and amorphous phases.²²

Fig. 7 Temperature dependence of the dielectric constant (a) and loss (b) for P(VDF-HFP)/PMMA films with different amounts of PMMA at 1.05 kHz.

The curves of dielectric loss against temperature at 1.05 kHz are shown in Fig. 7b.The dielectric loss values of all thin films increased with increasing temperature. Compared to blends, the tan δ value of P(VDF-HFP) sharply increased with temperature when it was higher than 35 °C. With the increase of PMMA content, the dramatic increase point of tan δ shifted to a higher temperature of 60 °C, and became less obvious when the content of PMMA increased to 30 wt%. As shown in Fig. 4, after the addition of PMMA, the melting point increased, which means more restriction to the movements of the chains and segments. From another aspect, hydrogen bonding between C=O in PMMA and CH₂ in P(VDF-HFP) can be formed, which will show more steric hindrance for the bigger MMA groups, as a consequence, the relaxation of the chains and segments will shift to a higher temperature.

3.3 Ferroelectric properties of PMMA/PVDF blends

The unipolar D–E loops of P(VDF-HFP) and P(VDF-HFP)/PMMA thin films were investigated, as shown in Fig. 8. The residual polarizations of P(VDF-HFP) and 5% PMMA thin films were higher than the others, and saturation polarization can be observed. Therefore, the energy loss was large. This was associated with the formation of the β-phase. It was known that the β-crystal was a ferroelectric phase. Residual polarizations and saturation polarization are typical properties of the ferroelectric phase, which are considered to contribute significantly to energy loss. The residual polarization of 10% and 30% PMMA sharply decreased, and the saturation polarization did not emerge at a field of 300 MV m⁻¹. These changes should be mainly associated with the γ-2b phase, which was not ferroelectric.

Fig. 8 D–E hysteresis loops of the P(VDF-HFP)/PMMA films with different compositions under unipolar electric fields.

By integration of the area between the D–E curves, the energy storage density and energy loss at any field can be obtained, as shown in Fig. 9. It is noteworthy that the max electric voltage applied in the D–E test is 10 000 V, which is lower than our film breakdown strengths for film thickness between 20 and 30 micrometers. As shown in Table 1, with the addition of PMMA, the breakdown strength is increased significantly, which means the final storage strength might be much higher for samples with PMMA more than 10 wt%.

Fig. 9 Energy storage (a), energy loss (b) and efficiency (c) of the PVDF–HFP/PMMA films with different compositions against electric field.

Table 1 Breakdown strength of P(VDF-HFP) and its blend films^a

Sample	Shape parameter (β)	Breakdown strength (E0, MV m^−1)
a The breakdown strengths in the table were obtained by a two-parameter Weibull distribution function: P(E) = 1 − exp[−(E/E0)^β], where P(E) is the cumulative probability of failure occurring equal to E. E0 is the field strength at which there is a 63% probability for the sample to breakdown, while the shape parameter evaluates the scatter of data (>10).
P(VDF-HFP)	4.43	304.57
5% PMMA	5.92	398.81
10% PMMA	6.30	540.62
30% PMMA	6.06	510.99

---

From Fig. 9, it is can be seen that the energy discharge density firstly increased and then decreased, but the energy loss decreased dramatically with the further increase of PMMA content. For the samples with greater than 10 wt% PMMA, the energy loss was maintained at a low level. Therefore, the discharge efficiency calculated from the energy discharge density and energy loss showed the behavior observed in Fig. 9 (c). For pure PVDF and that with 5 wt% PMMA, the discharge efficiency decreased rapidly with the increase in electric field. However, for the samples with greater than 10 wt% PMMA, the discharge efficiency decreased at a much slower rate. The discharge efficiency increased from 45% for pure PVDF (under 267 MV m⁻¹) to higher than 80% (under 300 MV m⁻¹) for the polymer blends.

The significant increase in energy discharge efficiency might result from two aspects. Firstly, the addition of PMMA could result in hydrogen bonding between C=O in PMMA and CH₂ in P(VDF-HFP), which hinders the movements of the chain segments. From previous results, we can see that the energy loss was shifted to a higher temperature, which might be due to the hindrance of hydrogen bonding on the movement of PVDF segments. On one hand, the chain movement constraints from the hydrogen bonds may result in the decrease of crystallinity, and consequently the decrease in ferroelectric strength. On the other hand, the molecular interactions of PMMA and PVDF could orient the crystallization and promote the formation of the γ-2b phase. The polarity of the γ-2b phase is lower than the β phase, but higher than the α phase. In addition, it is not a ferroelectric which benefits the energy storage and can decrease the energy loss significantly. As mentioned in the introduction, in most literature, the increase in discharge efficiency is thought to result from the decrease in β crystallinity and increase in crystal defects. However, in our results, the crystallinity of PVDF blends with 10 wt% PMMA did not obviously decrease, and the efficiency was increased significantly (Fig. 4). Instead, the γ-2b signal became more obvious, indicating that the formation of the γ crystal contributes to the increase in efficiency much more.

Conclusions

With solution blending and thermal treatment under vacuum, a series of P(VDF-HFP)/PMMA blend films were obtained and characterized by FTIR, XRD and DSC. The results showed that crystallinity, dielectric constant and loss of P(VDF-HFP) decreased with increasing PMMA content. These changes were attributed to the dilution effect of PMMA to P(VDF-HFP). A novel γ-2b phase was observed after the addition of PMMA and thermal treatment under vacuum, which contributed to the shift of melting temperature to a higher temperature. The discharge efficiency of P(VDF-HFP) significantly improved with the addition of PMMA to P(VDF-HFP). The blend film with 10% PMMA showed a discharge efficiency of 80%, much higher than the pure polymer (45%). The increase in discharge efficiency is thought to result mainly from the formation of the γ crystal.

Acknowledgements

This work was funded by the Natural Science Foundations of Hebei province of China (E2012203153).

References

1. V. Tomer, E. Manias and C. A. Randall, J. Appl. Phys., 2011, 110, 044107.
2. X. Zhou, B. Chu, B. Neese, M. Lin and Q. M. Zhang, IEEE Trans. Dielectr. Electr. Insul., 2007, 14(5), 1133–1138.
3. X. Zhou, X. Zhao, Z. Suo, C. Zou, J. Runt, S. Liu, S. Zhang and Q. M. Zhang, Appl. Phys. Lett., 2009, 94, 16290.
4. Z. Zhang and T. C. Mike Chung, Macromolecules, 2007, 40, 9391–9397.
5. R. Hasegawa, M. Kobayashi and M. Tadokoro, Polym. J., 1972, 3, 591–599.
6. J. Matsushige and T. Takemura, J. Polym. Sci., Polym. Phys. Ed., 1978, 16, 921–934.
7. M. A. Bachmann and J. B. A. Lando, Macromolecules, 1980, 14, 40–46.
8. Y. Takahashi, Y. Matsubara and H. Tadokoro, Macromolecules, 1982, 15, 334–338.
9. M. Remskar, I. Iskra, J. Jelenc, S. Davor Skapin, B. Visic, A. Varlec and A. Krzan, Soft Matter, 2013, 9, 8647–8653.
10. T. T. Wang, J. M. Herbert and A. M. Glass, The applications of ferroelectric polymers, in Polym. Int., The Blackie Publishing Group, New York, 1988, vol. 20, 6, p. 533.
11. R. Gregorio Jr, J. Appl. Polym. Sci., 2006, 100, 3272–3279.
12. W. Li, Q. Meng, Y. Zheng, Z. Zhang, W. Xia and Z. Xu, Appl. Phys. Lett., 2010, 96, 192905.
13. T. Nishi and T. T. Wang, Macromolecules, 1975, 8, 909–915.
14. T. T. Wang and T. Nishi, Macromolecules, 1977, 10, 421–425.
15. Q. Meng, W. Li, Y. Zheng and Z. Zhang, J. Appl. Polym. Sci., 2010, 116, 2674–2684.
16. M. Li, N. Stingelin, J. J. Michels, M.-J. Spijkman, K. Asadi, K. Feldman, P. W. M. Blom and D. M. de Leeuw, Macromolecular, 2012, 45, 7477–7485.
17. J. Gyu Lee and S. Hun Kim, Macromol. Res., 2011, 19(1), 72–78.
18. H.-J. Ye, L. Yang, W.-Z. Shao, S.-B. Sun and L. Zhen, RSC Adv., 2013, 3, 23730–23736.
19. B. Charlot, S. Gauthier, A. Garraud, P. Combette and A. Giani, J. Mater. Sci.: Mater. Electron., 2011, 22, 1766–1771.
20. A. J. Lovinger, Polymer, 1980, 21, 1317–1322.
21. S. O. Kasap, Principles of Electronic Materials and Devices, McGraw-Hill, New York, 2006.
22. G. K. Narula and P. K. C. Pillai, J. Mater. Sci. Lett., 1989, 9, 608–611.

---