Nanoscale investigations on β-phase orientation, piezoelectric response, and polarization direction of electrospun PVDF nanofibers

Abstract

A single stage far-field electrospinning process can give rise to β-phase nanocrystals preferentially oriented in polyvinylidene fluoride (PVDF) nanofibers. The polymer chains in the as-electrospun PVDF nanofibers was oriented along the fiber axis as visualized at atomic scale through high resolution TEM and the β-phase content was found to be 72.7%. The vector mapping of piezoelectricity was investigated on piezoelectric responses in different directions via advanced piezoresponse force microscopy (PFM). It turns out that the piezoelectricity of the PVDF nanofibers is mainly derived from the out-of-plane (OP) electric-to-mechanical conversion. The PVDF nanofibers result in an average OP piezoelectric constant (d₃₃) of −25.3 pm V⁻¹. The β-phase polarization of the nanofibers is oriented at a small inclination to the z-axis (almost perpendicular to the fiber axis and the bottom electrode).

---

1. Introduction

Piezoelectric materials have attracted a wide interest in the diverse field of electromechanical applications for their ability to reciprocally correlate mechanical with electrical energies. Polyvinylidene fluoride (PVDF) and its derivatives, are semi-crystalline piezoelectric polymers, which has potential applications in flexible energy-harvesting and energy-generation devices, such as implantable biosensors, wearable and portable electronic devices, etc.^1–4 A considerable number of studies have been reported on high-output piezoelectric (nano)generators composing of PVDF nanofibers fabricated by near-field electrospinning (NFES) or conventional far-field electrospinning (FFES) processes. Two common generator modes of operation of a piezoelectric element are 33- and 31-modes.^5,6 It is well-known that the design of (nano)generator structures associates with the generator modes, which are in turn dependent on the electrospinning process.

The piezoelectric (nano)generators composing of PVDF nanofibers fabricated by NFES usually work at bending state.^7–9 The reason is that during the NFES process, the dipole moments of the PVDF nanofibers are reorganized and oriented in nearly the direction along the fiber axis under the synergistic effect of strong electric field and stretching force.^10–14 L. Lin et al. achieved near field electrospun PVDF nanofiber generators with energy conversion of 21.71%, an order of magnitude higher than those made of PVDF thin films.¹⁰ Recently, this is also supported by Y.-K. Fuh et al., who reported that three layers piled direct-write PVDF nanofiber generators yielded the maximum output voltage of 20 V.¹¹

While, we found that the results about the β-phase orientation using FFES does not fit in the case of EFES, which is also proved by tremendous work on piezoelectric sensors or nano-generators.^15–18 D. Mandal et al. fabricated two types of pressure sensors consisting of electrospun nanofiber webs stacked under different polarization directions, which proved the alignment of dipoles exists perpendicularly to the fiber axis.¹⁵ Despite this, J. Fang et al. used randomly aligned as-electrospun PVDF nanofiber membranes to construct the piezoelectric generators with voltage outputs of several volts, which suggested that the molecular dipoles within the nanofibers are unlikely to orient along the fiber length direction.¹⁶ This orientation is also indicated in recent works reported by S. Garain et al., where they have developed a high-performance PVDF-based piezoelectric device as it generated an output voltage of 11 V.¹⁷ According to B. J. Hansen et al.'s study, when aligned electrospun PVDF nanofibers were used to construct energy harvesting devices, a post in-plane poling treatment was performed to make the dipoles align along the fiber axis.¹⁸ Despite of the numerous reports in the literature on the piezoelectric generators based on PVDF nanofibers, however, evidence in experimental measurement on the vector piezoelectricity to elucidate the preferential orientation of dipoles during the electrospinning process has not been given yet. Indeed, it is well-known that the crystallite orientation can strongly affect the mechanical, thermal, electrical, and optical properties of polymeric materials, in particular, the β-phase orientation of the PVDF polymer correlates with its piezoelectric response.

In this study, we provided detailed insight into growth orientation of β-phase nanocrystals in single far-field electrospun PVDF nanofibers at nanoscale visualized by transmission electron microscopy (TEM). The crystallization behavior of the PVDF nanofibers was investigated to identify their crystalline phases and quantify the β-phase content by X-ray diffraction (XRD). Furthermore, to determine the underlying mechanisms of dimensional dependency of piezoelectricity, the corresponding piezoelectric responses at the single nanofiber level was measured in space by piezoresponse force microscopy (PFM). Through comparing the out-of-plane (OP) with in-plane (IP) piezoelectric responses, the β-phase orientation in the PVDF nanofibers was qualitatively determined.

2. Results and discussion

The representative morphology and molecular structure of the electrospun PVDF nanofibers are shown in Fig. 1. The SEM image (Fig. 1a) shows that the nanofibers fabricated by FFES have quite smooth surfaces and the bright field TEM image (Fig. 1b) provides homogeneous cross-sections. High resolution TEM images show signs of structural order or crystallinity in both the thin (Fig. 1c) and thick (Fig. 1d) nanofibers. No voids or porosity were evidenced in any of the nanofibers. In the case of the thinner nanofiber (Fig. 1c), most of the PVDF molecules were extended in parallel and straightened at least up to tens of nanometers, resulting in the higher orientation of the nanocrystals. While, curved segments can be seen in the superimposed layers of the polymer chains, which therefore complicates analysis of the orientation of the crystalline chains. But fast Fourier transform (FFT) produces diffraction patterns from the crystalline structure of the thin nanofiber as shown in the inset image of Fig. 1c. Transformation of the amorphous structure into the crystalline structure in PVDF results from the ordering of molecule dipoles. These conformational changes are termed as dipolar reorientation or dipole flip-flop motion. In contrast, the high resolution TEM image shows that the thick nanofiber (Fig. 1d) has the same orientation trend, but with poor orientation and less crystallinity. It is also supported by the FFT pattern of the thick nanofiber in which part of the PVDF chains are oriented along the fiber axis. The PVDF chains are elongated side-by-side along the fiber axis, where the long molecular segments have at least 5–60 nm length depending on the fiber diameter. The PVDF chains are oriented along the fiber axis by the stretching and simultaneous poling effects derived from the FFES process. The significant effects of the orientation originate mainly from their determining role in confining the dipoles through chain rotation under drawing and poling.

Fig. 1 (a) SEM image of PVDF nanofibers reflects very smooth topography and uniform diameters. (b) Bright field TEM image of electrospun PVDF nanofibers on a lacey film supported by a copper sample grid. (c) High resolution TEM image of the selected area in the thin nanofiber with the polymer chains mainly oriented along the fiber axis. The diameter of the nanofiber is 200 nm. The inset is the FFT image of image (c). (d) High resolution TEM image of the selected area in the thick nanofiber showing the polymer chains at the edge partially oriented along the fiber axis. The diameter of the nanofiber is 80 nm. The inset is the FFT pattern of image (d).

Crystallization studies of the electrospun PVDF nanofibers were conducted to identify crystalline phases and quantify degree of crystallinity of PVDF from the X-ray diffraction (XRD) spectra (Fig. 2). The PVDF casting membrane displays four strong α-phase peaks (17.6°, 18.2°, 19.9°, and 26.4°) and a weak γ-phase peak at 38.9°, but no trace of the β-phase. In the spectrum for the electrospun nanofibers, the β-phase peak at 20.8° dominates significantly and another β-phase peak at 36.1° appears; on the contrary, the α-phase peaks at 17.6°, 19.9°, and 26.6° completely disappear and the peak at 18.8° weakens. The sharp and intense peak at 20.8° confirms the formation of the desired polar β-phase. Subsequently, crystallinity of the PVDF membrane and nanofibers is evaluated from the XRD pattern by a curve deconvolution technique. The degree of β-phase crystallinity of the PVDF nanofibers is 72.7% and the degree of γ-phase crystallinity is 12.2%. So far, depending on the structural analysis via TEM and crystallinity analysis via XRD, we found that the simultaneous stretching and poling during the FFES facilitate the α- to β-phase transformation as well as render high degree of crystallinity and chain orientation.

Fig. 2 (a) XRD spectra of the electrospun PVDF nanofibers with the crystalline phases of α at 18.2°, β at 20.8° and 36.1°, and γ at 18.5°, and PVDF casting membrane with the crystalline phases of α at 17.6°, 18.2°, 19.9°, and 26.4°, β at 36°, and γ at 38.9°. (b) Contents of crystalline phases (α, β, and γ) of the electrospun PVDF nanofibers and casting membrane.

Piezoelectricity of the electrospun PVDF nanofibers without any extra poling treatment was measured by PFM. Piezoelectric response, defined as the product of inverse piezoelectric amplitude and cosine of the phase, was measured to study the switching behavior. The PFM amplitude corresponds to the sample's local piezoelectric displacement induced by AC voltage, while the PFM phase indicates the polarization direction of the nanodomains. PFM was used to measure the piezoelectric response at a nanoscopic location of the PVDF nanofibers and thus evaluate their polarization orientation (Fig. 3a) using a dual AC resonance tracking (DART) mode.^19–21 The cantilever tip follows the local piezoelectric displacement that takes place directly at the apex of the tip, driven by the alternating voltage applied to it. Generally, such a deformation can take place in any directions, therefore leading to deflection, buckling, or torsion of the cantilever. Deflection is the consequence of an OP deformation (a thickness change in the z-axis), buckling responds to an IP deformation (a diameter change in the y-axis) whereas torsion is related to another IP deformation (a length change in the x-axis). So in vector PFM, the real space reconstruction of polarization orientation consists of three components of piezoelectric response, one vertical PFM (VPFM) and two orthogonal lateral PFM (LPFM). The VPFM mode can be used to track the polarization switching of the OP polarization component along the z-axis. The LPFM mode can be used to track the two IP polarization components that include the polarization switching along the fiber axis (or along the x-axis) and horizontally perpendicularly to the fiber axis (or along the y-axis). The analysis of the vector PFM may also examine the β-phase orientation in real space. Fig. 3c shows the topography PFM image of the single PVDF nanofiber on which the piezoelectric response measurements (Fig. 3b) were carried out.

Fig. 3 (a) Schematic of piezoelectric response imaging of the single nanofiber by PFM. The coordinate axis system was defined here. (b) Optical photograph of the PFM measurement when the PFM tip is about to contact the single nanofiber. (c) 3D topography image of the single PVDF nanofiber with the diameter of 120 nm.

In addition to confirming the content and orientation of the β-phase in the PVDF nanofibers, the piezoelectric response was investigated. The PVDF nanofiber with the area of 50 nm × 50 nm was scanned with a homarnic V_ac of 1.6 V applied to the cantilever tip. Fig. 4a and b show the OP PFM phase and amplitude images of the nanofiber responding in the z-axis direction, respectively. The OP PFM phase image (Fig. 4a), corresponding to the piezoelectric polarization, clearly displays both negative (purple) and positive (yellow) values indicating antiparallel ferroelectric nanodomains with 180° domain walls. The purple areas correspond to negative domains with the polarization direction perpendicular to the surface of the nanofiber and oriented downward, while the yellow areas correspond to positive domains having the polarization direction oriented upward. The well-defined piezoelectric domains demonstrate that the elongated crystallites are the homogeneous β-phase. Therefore, the domains in the PVDF nanofiber align along the z-axis direction perpendicular to the nanofiber and the bottom electrode. The OP PFM amplitude (Fig. 4b) shows a strong piezoelectric contrast due to the deflections caused by the applied AC field. In contrast, Fig. 4c and d show IP PFM phase and amplitude images of the nanofiber in the y-axis direction, respectively, which can be determined by imaging the same region of the sample after IP rotation by 90°. There is few domains along the y-axis direction, which suggests that the dipoles of the β-phase nanocrystals rarely align along the y-axis direction. Fig. 4e and f show another IP PFM phase and amplitude images in the x-axis direction, respectively. The phase and amplitude contrasts exist, implying that there is small ±90° domains along the x-axis direction. Based on the PFM measurements on the nanofibers with various fiber diameters, this feature is common for the all obtained PVDF nanofibers.

Fig. 4 Vector PFM images of the PVDF nanofiber: (a) out-of-plane (OP) PFM phase and (b) amplitude images in the z-axis direction, (c) in-plane (IP) PFM phase and (d) amplitude images in the y-axis direction, and (e) IP PFM phase and (f) amplitude images in the x-axis direction. The diameter of the nanofiber measured here is 120 nm.

Therefore, the PVDF nanofibers fabricated by FFES present much higher OP PFM signals than IP PFM signals, which implies that the β-phase direction in the nanofiber is at a small inclination to the z-axis (almost perpendicular to the fiber axis). It is evident that the nanodomains composing of β-phase nano-crystals mainly distribute normal to the nanofiber and the bottom electrode, which is in good agreement with the β-phase orientation investigated by TEM and XRD. So we found the dipole orientation of the far-field electrospun nanofibers runs perpendicular to the fiber axis, different from the parallel dipole orientation in the nanofibers produced by NFES, reported by L. Lin et al.¹⁰

During the FFES, a polymer solution experiences two forces. One is shear force when it flows through a needle at a very high rate. The other one is coulombic force when the jet is elongated and accelerated by the high electric field applied. First, the nanofibers are subjected to mechanical stretching/poling due to polymer jet elongation and whipping, further enhancing piezoelectricity. Consequently, under the coulombic force, the PVDF chains mainly twist into the zigzag structure preferentially along the fiber axis, thereby concurrently inducing cross-sectional orientational ordering of the C–F and C–H bonds, oriented essentially in the direction normal to that of the molecular chains (Fig. 5a). Since PVDF is a linear polymer, the permanent dipoles are approximately perpendicular to the orientation direction of the polymer chains. Hence, in the far-field electrospun nanofiber, the majority of the dipoles are oriented in the transverse direction with respect to the fiber axis (Fig. 5b).

Fig. 5 (a) Schematic of a PFM tip on an individual PVDF nanofiber. (b) Schematics of the orientation of dipoles and polarization direction of the nanofiber. (c) Butterfly-shaped amplitude loop and (d) square-shaped phase loop of the PVDF nanofiber. (e) Piezoelectric deflections (amplitude) of the nanofibers vary with the DC bias. (f) Comparison of the piezoelectric constants (d₃₃) of the nanofibers with different diameters.

Further insight into the molecular structure of the nanofibers can be obtained by evaluating the piezoelectric behavior of the samples, which is primarily driven by the β-phase. In order to evaluate the local piezoelectric response characteristics of the PVDF nanofibers, the piezoelectric response loop was measured (Fig. 5c and d). In this case, a train of DC voltage (V_dc) pulses with a constant duration of 25 ms and their amplitude changing stepwise from −30 V to 0 V, then from 0 V to 30 V and back to −30 V was applied to the conductive tip that was kept at a fixed position on the surface of the individual nanofiber. The application of consecutive voltage pulses has produced the polarization reversal and the corresponding variations in the deflection, so that the amplitude versus V_dc loop is obtained like a butterfly. The highest amplitude in the loop reaches 1.78 nm at V_dc of 30 V. Polarization switching occurs at the coercive field, which changes the sign of the surface charges, leading to the piezoelectric response phase switching by around 180° (Fig. 5d). The PFM phase curves exhibit standard square-shaped switching hysteresis loops as well as good repeatability. Asymmetries in amplitude and phase loops also point to the existence of defects because the asymmetries are known to be associated with an internal field created by non-uniformly distributed charged defects.^22,23 Fig. 5e and f show the piezoelectric amplitudes acquired from individual nanofibers with different diameters. There is a significant difference in piezoelectric response of the nanofibers in terms of their diameters. We repeated this PFM measurement at least five places with the same area and parameters. We took the average of d₃₃ obtained at different points. The slope of the PFM response (d₃₃) of the samples is −18.7 pm V⁻¹ (at 200 nm diameter), −27.9 pm V⁻¹ (at 120 nm diameter), and −33.7 pm V⁻¹ (at 80 nm diameter), respectively. This suggests that piezoelectricity is diameter-dependent and is stronger with the dimensional reduction because of polarization and strong coupling of the electromechanical strain gradient.^24,25 As demonstrated before, with the fiber diameter decreasing, the β-phase nanocrystals are resulted with higher orientation and the degree of crystallinity increases. So the dominant OP piezoelectric response in the nanofibers is interlinked with their molecular and crystalline structures. In other words, the d₃₃ change is mainly dependent on the degree of crystallinity, the β-phase content and its orientation, particularly for the thinner nanofibers showing higher absolute d₃₃.

3. Conclusions

In summary, we ascertained a way at nanoscale to prove the preferential orientation of the dipoles in the far-field electrospun PVDF nanofibers. The β-phase nanocrystals within the nanofibers was mainly oriented perpendicularly to the fiber axis as visualized via high resolution TEM. We mapped out the piezoelectric response images in space with the results of the strong vertical component and the two weak lateral components under AC electric field via PFM. The PFM measurements on the nanofibers reveal that the effective polarization direction is at an inclination to the vertical direction. The direct piezoelectric measurement of the PVDF nanofibers delivers much prior piezoelectric response (or β-phase orientation) in the vertical PFM image than the lateral PFM image. The nanofibers exhibit an average OP piezoelectric constant (d₃₃) of −25.3 pm V⁻¹. The fundamental investigation of the nanofibers at the molecular level establishes its potential as a building block of energy conversion systems.

Acknowledgements

This work was supported by the grants from the National Natural Science Foundation of China (no. 61474071, 61531166006) and National Basic Research Program (973 Program, no. 2015CB352106).

Notes and references

1. C. Hou, T. Huang, H. Wang, H. Yu, Q. Zhang and Y. Li, Sci. Rep., 2013, 3, 3138.
2. Y. Qi and M. C. McAlpine, Energy Environ. Sci., 2010, 3, 1275–1285.
3. J. A. Paradiso and T. Starner, IEEE Communications Society, 2005, 5, 18–27.
4. X. Liu, H. Zhao, Y. Lu, S. Li, L. Lin, Y. Du and X. Wang, Nanoscale, 2016, 8, 7278–7286.
5. A. K. Nandi and L. Mandelkern, J. Polym. Sci., Part B: Polym. Phys., 1991, 29, 1287–1297.
6. M. J. Ramsay and W. W. Clark, Proc. SPIE, 2001, 4332, 429–438.
7. Z. H. Liu, C. T. Pan, L. W. Lin and H. W. Lai, Sens. Actuators, A, 2013, 193, 13–24.
8. J. Chang, M. Dommer, C. Chang and L. Lin, Nano Energy, 2012, 1, 356–371.
9. Z. H. Liu, C. T. Pan, C. Y. Su, L. W. Lin, Y. J. Chen and J. S. Tsai, Sens. Actuators, A, 2014, 211, 78–88.
10. C. Chang, V. H. Tran, J. Wang, Y.-K. Fuh and L. Lin, Nano Lett., 2010, 10, 726–731.
11. Y.-K. Fuh, J.-C. Ye, P.-C. Chen, H.-C. Ho and Z.-M. Huang, ACS Appl. Mater. Interfaces, 2015, 7, 16923–16931.
12. F. R. Fan, W. Tang and Z. L. Wang, Adv. Mater., 2016, 28, 4283–4305.
13. T. Lei, L. Yu, G. Zheng, L. Wang, D. Wu and D. Sun, J. Mater. Sci., 2015, 50, 4342–4347.
14. T. Huang, C. Wang, H. Yu, H. Wang, Q. Zhang and M. Zhu, Nano Energy, 2015, 14, 226–235.
15. D. Mandal, S. Yoon and K. J. Kim, Macromol. Rapid Commun., 2011, 32, 831–837.
16. J. Fang, X. Wang and T. Lin, J. Mater. Chem., 2011, 21, 11088–11091.
17. S. Garain, S. Jana, T. K. Sinha and D. Mandal, ACS Appl. Mater. Interfaces, 2016, 8, 4532–4540.
18. B. J. Hansen, Y. Liu, R. Yang and Z. L. Wang, ACS Nano, 2010, 4, 3647–3652.
19. A. Gannepalli, D. G. Yablon, A. H. Tsou and R. Proksch, Nanotechnology, 2011, 22, 355705.
20. E. Soergel, J. Phys. D: Appl. Phys., 2011, 44, 464003.
21. N. Liu, R. Dittmer, R. W. Stark and C. Dietz, Nanoscale, 2015, 7, 11787–11796.
22. D. Zhou, J. Xu, Q. Li, Y. Guan, F. Cao, X. Dong, J. Müller, T. Schenk and U. Schröder, Appl. Phys. Lett., 2013, 103, 192904.
23. X. Liu, S. Xu, X. Kuang and X. Wang, RSC Adv., 2016, 6, 106690–106696.
24. M. S. Majdoub, P. Sharma and T. Cagin, Phys. Rev. B: Condens. Matter Mater. Phys., 2008, 77, 125424.
25. Y. Wang, K. Ren and Q. M. Zhang, Appl. Phys. Lett., 2007, 91, 222905.

---