Introduction

Piezoelectric polymers are increasingly being used in wearable electronics and energy harvesting devices due to their ability to convert mechanical energy into electricity, exceptional flexibility, and resistance to mechanical stress1,2,3. They have been used to develop highly sensitive devices for monitoring human motion and physiological activity, which are critical in various dynamic environments4,5. Polyvinylidene fluoride (PVDF) and its copolymers are among the most important piezoelectric polymers6,7,8. Their piezoelectric properties stem from a unique molecular configuration and polymer chain orientation in the solid state9. In particular, the fluorine atoms impart significant polarity to the C-F bonds due to their high electronegativity. PVDF can exist in five different crystalline phases: α, β, γ, δ, and ε10,11. Among these, the β-phase, which is characterized by an all-trans-planar zig-zag chain conformation, is known for its superior ferroelectric properties12. However, the β-phase in piezoelectric PVDF is typically unstable at elevated temperatures and tends to transition to the α-phase, which diminishes its piezoelectric performance. The incorporation of tetrafluoroethylene (TrFE) into PVDF chains to form the P(VDF-TrFE) copolymer modifies the chemical structure, introducing significant steric hindrance that favors the formation of the thermodynamically stable β-phase. In addition, the inclusion of TrFE generates polar defects that are highly sensitive to electric fields and mechanical stress, thereby enhancing the piezoelectric response of the material13.

Traditional approaches for fabricating piezoelectric films, such as extrusion, sol-gel, spin coating, and solution casting, are often complex, requiring demanding operating conditions and significant financial investment. These methods typically result in films with suboptimal piezoelectric properties. Our previous investigations have shown that electrospinning, a process where a polymer solution is subjected to a strong electric field to form thin fibers typically assembled into nanofiber membranes, is a simple yet effective method for fabricating piezoelectric fiber membranes14. This process allows the liquid jet to be stretched in situ at a high ratio, promoting the alignment of the polymer chains parallel to the fiber axis and orienting the dipoles within the electric field15,16. Electrospinning enables in-situ polarization of the ferroelectric material, eliminating the need for additional polarization steps to align the dipoles within the ferroelectric domains. Additionally, the stretching process can significantly enhance the crystallinity of P(VDF-TrFE)17, improving its piezoelectric performance. Despite these advantages, electrospun piezoelectric nanofiber membranes often face challenges related to insufficient mechanical strength, durability, and optical clarity. These limitations hinder their applications in optical technology areas, such as displays and touch-sensitive devices, where high optical transparency and light transmission are critical18,19.

This study presents a simple yet efficient method to fabricate dense, highly transparent piezoelectric films by thermomechanical pressing with immediate quenching (TMP-IQ) of electrospun (ES) P(VDF-TrFE) nanofibers. The TMP-IQ treatment not only densifies the fiber network and preserves the P(VDF-TrFE) polarization, but also significantly increases the mechanical strength and light transmittance. Without the need for any post-poling process, the resulting dense films processed at 140 °C and 600 MPa for 3 min exhibit a significantly improved d33 of 14.09 pC N−1. In addition to the improved ferroelectric and piezoelectric properties, the thermally densified films also exhibit improved optical clarity, reduced haze, and a significant increase in mechanical strength and stiffness. These improvements are due to the flash crystallization induced by TMP-IQ, which promotes the development of extended P(VDF-TrFE) crystal chains, increases lamellar size, and expands the oriented amorphous fraction (OAF) regions, resulting in the formation of well-ordered semicrystalline structures over larger areas. Our approach is straightforward and allows precise control of the crystal structure.

Results

Fabrication, micro-morphology, and optical transparency

The transparent piezoelectric P(VDF-TrFE) films were prepared by thermomechanical pressing of electrospun nanofiber membranes, followed by immediate quenching in an ice bath. Figure 1a schematically illustrates the fabrication process, while Fig. 1b shows a digital photograph of the ES nanofiber membrane and the transparent film processed by thermomechanical pressing - immediate quenching.

Fig. 1: Fabrication process of the transparent piezoelectric P(VDF-TrFE) film.
Fig. 1: Fabrication process of the transparent piezoelectric P(VDF-TrFE) film.The alternative text for this image may have been generated using AI.
Full size image

a Illustration of the electrospinning, thermomechanical pressing, and ice quenching processes. b Digital photos of the ES membrane and TMP-IQ film. c SEM images of the ES membrane and TMP-IQ film.

Scanning electron microscopy (SEM) images reveal the surface topography and cross-sectional morphologies of the P(VDF-TrFE) membrane (Fig. 1c and Supplementary Fig. 1). The ES membrane contained uniform fibers with a diameter of 274.6 ± 52.7 nm (Supplementary Fig. 2a), arranged randomly within the fiber matrix. After thermomechanical pressing—immediate quenching, the fiber matrix disappeared. Instead, a dense solid film was formed with the thickness significantly reduced from 153 μm to 25 μm. Consequently, the film density increased from 0.282 g cm−3 to 1.798 g cm−3, while the porosity decreased from 84.3% to 0.1%. Optical transmission significantly increased from 7.5% to 84.3%, and the haze decreased from 82.84 to 12.82 (Supplementary Fig. 2b, c). We further examined the transmittance of both samples across the UV-vis-NIR light spectrum. Notably, while the transmittance of both films increased progressively with wavelength, the ES membrane exhibited transmittance values consistently below 12% at 2000 nm. In contrast, the TMP-IQ film demonstrated significantly higher transmittance, reaching 80% and 90% at 443 nm and 965 nm, respectively (Supplementary Fig. 2d). Compared to other reported piezoelectric films, the TMP-IQ film has excellent transmittance, transparency, and haze properties (Supplementary Table 1). This is primarily attributed to the volume densification achieved through TMP-IQ processing, which eliminates voids and enhances fiber contact. Consequently, light reflection and scattering at the interface are significantly diminished (Supplementary Fig. 3).

Ferroelectric and piezoelectric properties

Figures 2a, b show the ferroelectric properties of the P(VDF-TrFE) films before and after thermal pressing. Without thermal pressing, no hysteresis loops are observed, even at an electric field of 100 MV m−1, which exceeds the theoretical coercive field of 50–60 MV m−1. The breakdown occurred at a higher field. In contrast, the maximum polarization (Pmax) and residual polarization (Pr) of the TMP-IQ films reached 18.78 and 9.61 μC cm–2, respectively, at 100 MV m−1. As the electric field increased to 200 MV m−1, Pmax and Pr increased to 30.77 and 13.20 μC cm−2, respectively. The pronounced hysteresis loops, along with the high Pmax and Pr, strongly indicate that our TMP-IQ films possess robust ferroelectric capabilities. These enhanced properties suggest a high degree of polarization reversibility and a strong ability to maintain polarization even after the removal of the electric field. This makes them highly promising for various electronic device applications, such as non-volatile memory elements and sensors, where reliable and efficient ferroelectric behavior is crucial for optimal performance. The successful enhancement of these properties through thermal pressing highlights the potential of this processing technique to significantly improve the functional attributes of P(VDF-TrFE) based materials, opening up new possibilities for their broader use in advanced technological applications.

Fig. 2: Ferroelectric responses of P(VDF-TrFE) films.
Fig. 2: Ferroelectric responses of P(VDF-TrFE) films.The alternative text for this image may have been generated using AI.
Full size image

a, b P-E hysteresis loops comparing TMP-IQ140 film and ES membrane at 1 Hz. c, d Phase-voltage hysteresis and amplitude-voltage butterfly curves contrasting TMP-IQ140 film with ES membrane. e, f Amplitude and phase mappings of PFM.

Piezoresponse force microscopy (PFM) was used to examine the phase images (hysteresis loops) and amplitude images (butterfly curves) of the ES membrane and TMP-IQ film20,21,22 (Fig. 2c, d). At an applied voltage of 80 V, the amplitudes of the TMP-IQ film and the ES membrane were 936.3 pm and 380.2 pm, respectively. The TMP-IQ film showed a larger response, indicating improved piezoelectric performance. In the hysteresis loops, both the TMP-IQ film and ES membrane display a phase difference of approximately 180°, indicating effective polarization reversal under the influence of an external electric field. We also performed a 4.5 × 4.5 μm2 square domain writing test (at ±60 V) to confirm the polarization convertibility (see the results in Fig. 2e, f), allowing for the measurement of dielectric properties. The TMP-IQ film exhibited excellent dielectric properties, with a dielectric constant ranging from 5.38 to 10.96 and a dielectric loss ranging from 0.03 to 0.21 (Supplementary Fig. 4). These results indicate that the TMP-IQ films exhibit significantly higher ferroelectric polarization switchability and piezoelectric response than the ES membrane.

To measure the actual piezoelectric outputs under deformation, we fabricated sandwich-structured piezoelectric devices using the ES membrane and TMP-IQ film (see the device structure in Fig. 3a). As the layer’s thickness increases, the output initially rises and then stabilizes (Supplementary Fig. 5). This trend informed our decision to set the piezoelectric layer thickness for subsequent piezoelectric tests at approximately 20 μm. The piezoelectric outputs in both compressing and bending modes were measured using an electrometer. The TMP-IQ film exhibited approximately twice the open-circuit voltage and short-circuit current outputs than the ES membrane in both modes (Fig. 3b, c). The TMP-IQ film also showed a faster response time. In pressing mode, the response time for the TMP-IQ film was 57 ms, shorter than that of the ES membrane (79 ms). In the bending mode, the response times had a similar trend, being 59 ms and 82 ms, respectively. The faster response of the TMP-IQ film is attributed to its dense structure, which responds immediately to the external force and reaches maximum strain in less time than the porous ES membrane, which undergoes a two-step deformation that changes the porous structure and then the polymer material23,24. The voltage output of TMP-IQ piezoelectric films varied significantly with stress frequency. This affected the magnitude and response time of the output. As frequency increased, voltage output amplitude increased progressively while response time decreased correspondingly (Supplementary Fig. 6).

Fig. 3: Piezoelectric outputs of the ES and TMP-IQ devices.
Fig. 3: Piezoelectric outputs of the ES and TMP-IQ devices.The alternative text for this image may have been generated using AI.
Full size image

a Schematic and actual structure of the piezoelectric device. b, c Piezoelectric voltage and current outputs of the devices. d, e Piezoelectric voltage of piezoelectric devices prepared from ES and TMP-IQ140 membrane under varying pressing pressures. f, g Sensitivity of piezoelectric devices prepared from ES and TMP-IQ140 membrane. h Schematic representation of the cycling stability of the piezoelectric device. i Average power density of the piezoelectric device fabricated from TMP-IQ140. j Charge output of the piezoelectric device prepared from TMP-IQ140 under different external forces. k Piezoelectric coefficients d33 calculated by generating charges with different magnitudes of force. The error bars in the figures represent the standard deviation.

The effect of compression pressure on the piezoelectric performance and sensitivity of the devices was evaluated. For the TMP-IQ devices, the peak-to-peak open-circuit voltages increased from 8.59 to 34.26 V as the applied pressure rose from 1 to 200 kPa (0.225 to 45 N), significantly exceeding those of the ES devices, which increased from 3.82 to 17.14 V (Fig. 3d, e). The TMP-IQ device demonstrates superior sensitivity across all stress intervals. At lower pressure ranges, its sensitivity reaches 1.610 V kPa−1, markedly higher than the ES device’s 0.796 V kPa−1. Moreover, at pressures exceeding 40 kPa, the TMP-IQ device’s sensitivity of 0.067 V kPa−1 is roughly double that of the ES device at 0.027 V kPa−1, with all intervals exhibiting a high coefficient of determination (Fig. 3f, g). Additional compression and decompression under the same conditions after 10,000 compression cycles did not result in a noticeable change in the piezoelectric response, demonstrating excellent cyclic stability (Fig. 3h and Supplementary Fig. 7). This indicates that the TMP-IQ devices can serve as a stable power source, retaining their performance over prolonged use. The piezoelectric output of the TMP-IQ membranes remained essentially unchanged after 16 weeks, shifting from 20.64 ± 0.20 V to 20.27 ± 0.21 V, a variation within experimental error (Supplementary Fig. 8).

The power and energy conversion efficiency of the TMP-IQ device were evaluated using an external load method25,26,27. As the load resistance increased, the power density initially increased and then decreased, peaking at 5 MΩ. The TMP-IQ membrane exhibited a maximum power density of 15.44 mW m−2 and an inner impedance of 5 MΩ. In contrast, the ES membrane’s maximum power density and inner impedance were 4.05 mW m−2 and 10 MΩ, respectively (Fig. 3i and Supplementary Table 2). Therefore, the TMP-IQ film has a higher power density but lower inner impedance than the ES membrane.

We also measured the d33 piezoelectric coefficient by evaluating the charge transfer from the piezoelectric film to the external electrodes28. As expected, the TMP-IQ film had a larger d33 value (14.09 pC N−1) than the ES membrane (7.54 pC N−1) (Fig. 3j, k), representing an increase of 86.9%. For comparison, we also tested the d33 of a commercial P(VDF-TrFE) film (from Arkema, France) using the same method. The commercial film had a d33 value of 31.70 pC N−1, which is consistent with the value provided by the supplier (d33 > 28 pC N−1). With any mechanical stretching and electrically poling, the P(VDF-TrFE) powder used to prepare the ES membrane and TMP-IQ film had a d33 of only 0.34 pC N−1 (Supplementary Fig. 9). In addition, the piezoelectric coefficient of the commercial film subjected to the TMP-IQ process was enhanced to 35.62 pC N−1, further demonstrating the positive effect of TMP-IQ in boosting the piezoelectric properties of P(VDF-TrFE).

We also found that the TMP-IQ devices have significantly higher efficiency in charging capacitors. Using a rectifier bridge, the TMP-IQ generator can charge the capacitors to 0.97 V, 0.65 V, 0.371 V, and 0.181 V in 90 s, respectively (Supplementary Fig. 10). The functional performance of TMP-IQ films in pressure sensors and wearable devices was tested. The results revealed distinct piezoelectric signals in all modes, demonstrating a direct correlation between signal amplitude, frequency, and mechanical stimuli. Key findings include pronounced voltage periodicity during mouse clicking and finger tapping, voltage amplitude scaling with finger bending angles of 30°, 60°, and 90°, and robust piezoelectric responses during gripping, boxing, walking, and running. These responses indicate deformation sensitivity of piezoelectric devices, contributing to efficient mechanical energy harvesting (Supplementary Fig. 11).

Impact of TMP-IQ parameters

The above results were based on the TMP-IQ film processed at 140 °C with a press force of 600 MPa and a duration of 3 min. We also examined how variations in TMP temperature and pressing duration impacted the piezoelectricity and material properties. The piezoelectric outputs of the TMP-IQ films initially increased and then decreased as the increasing TMP temperature increased, reaching a maximum at 140 °C (Fig. 4a). It should be noted that all subsequent tests were performed at ambient temperature following the parameter settings for the TMP-IQ process.

Fig. 4: Crystallization and crystal phases of P(VDF-TrFE) films.
Fig. 4: Crystallization and crystal phases of P(VDF-TrFE) films.The alternative text for this image may have been generated using AI.
Full size image

a Piezoelectric voltage peak-to-peak values of thermomechanical pressed films were evaluated in both pressing and bending modes. b DSC heating curves of P(VDF-TrFE) membranes subjected to different thermomechanical pressing temperature gradients. c Crystallinity trend of P(VDF-TrFE) membranes with increasing thermomechanical pressing temperature. d Stress-strain curves. e Trend in grain size of P(VDF-TrFE) films with increasing thermomechanical pressing temperature. f Trend in crystalline phase composition of P(VDF-TrFE) films as thermomechanical pressing temperature increases. g Comparison of breaking strength and piezoelectric coefficient d33 of TMP-IQ films with other PVDF-based piezoelectric polymer films (CF and TMP-IQ-CF represent P(VDF-TrFE) commercial membrane and hot-pressed commercial membrane, respectively), h Crystallinity and β-phase content of TMP-IQ films at different cooling rates, i Comparison of piezoelectric properties of TMP-IQ films at different cooling rates. The shaded areas and error bars in the figure both represent standard deviation.

Differential scanning calorimetry (DSC) was used to measure the Curie and melting temperatures, as well as the crystallinity of the films29,30,31. The Curie and melting temperatures of the ES membrane were 108.1 °C and 150.6 °C, respectively. For TMP-IQ, the Curie temperature was increased to 121.4 °C, while a double peak phenomenon was observed in the melting temperature, which occurred at 147.6 °C and 154.9 °C, respectively. As the thermomechanical pressing temperature increased, the Curie temperature of the TMP-IQ film also increased. This change is probably attributable to the reduced crystal defects and dislocations in the P(VDF-TrFE) structure following the TMP-IQ treatment, which enhances the thermal stability and phase transition temperatures of the material.

When the thermomechanical pressing temperature was 135 °C, the TMP-IQ film exhibited bimodal endothermic melting characteristics in the DSC curve, which was partially attributed to the lamellar thickening and broadening induced by the melt recrystallization process32. However, when the thermomechanical pressing temperature was elevated to 150 °C, surpassing the melting temperature of P(VDF-TrFE), the bimodal endothermic peaks progressively diminished due to the complete melting of the macromolecular chains, where the intense thermal motion partially disrupted the preformed ordered crystalline structures (Fig. 4b and Supplementary Fig. 12a, b).

Figure 4c shows the effect of TMP temperature on the enthalpy and crystallinity of P(VDF-TrFE) films, which is similar to the trend of the piezoelectric outputs. The TMP-IQ films reached a peak crystallinity of 77.06%, which is significantly higher than that of the ES membranes (52.24%). The P(VDF-TrFE) crystallinity decreased steeply with the disappearance of the double melting peak when the TMP temperature was higher than 145 °C.

The mechanical properties of both the ES and TMP-IQ films were evaluated. Both ES and TMP-IQ exhibited the properties of thermoplastic polymers, as shown in the stress-strain curves (Fig. 4d and Supplementary Fig. 12c, d). The TMP-IQ films, particularly those processed at temperatures below 140 °C, demonstrated a significant increase in strength and yield point. Compared to the ES films, the tensile strength and stiffness of the TMP films increased dramatically by 786% and 2383%, respectively (Supplementary Fig. 12e). Thermomechanical pressing increases the mobility of the P(VDF-TrFE) chains, facilitating their reorganization and alignment without complete melting. However, as the TMP temperature approached or exceeded the melting point, the thermal motion of the P(VDF-TrFE) chains increased, potentially leading to melting, weakened molecular bonds, and possible chain breakage or disentanglement. This results in a more relaxed structure, which reduces the fracture strength of the material but allows greater elongation before failure (Supplementary Table 3).

Haward’s large strain model for thermoplastics suggests that yield stress is the threshold at which the material transitions from elastic deformation of the crystalline structure to significant crystal slip33,34. The yield stress is primarily influenced by the degree of crystallinity in semi-crystalline polymers. The variation in yield stress for P(VDF-TrFE) films subjected to different thermomechanical pressing temperatures (shown in Supplementary Fig. 12f) aligns with the changes in crystallinity as determined by DSC. This correlation indicates that the crystallinity of the P(VDF-TrFE) films plays a crucial role in determining their yield stress, with higher crystallinity leading to increased yield stress and improved mechanical strength.

Supplementary Fig. 13 shows the XRD patterns of the P(VDF-TrFE) films at different TMP temperatures. A diffraction peak at 2θ = 20°, corresponding to the β-phase of the (110/200) plane, was consistently observed35. The diffraction peaks became sharper as the TMP temperature increased, with the full width at half maximum (FWHM) steadily decreasing (Fig. 4e). The crystallite size calculated by the Scherrer equation increased progressively, which aligns with the trends observed in the G-T curves (Supplementary Fig. 14), indicating a prolonged crystal size after thermomechanical pressing-immediate quenching. We also determined crystallinity by fractional peak fitting of the XRD curves, which showed a consistent result with the DSC data shown in Supplementary Fig. 15, confirming the reliability of both techniques for assessing crystallinity.

FTIR and Raman spectra were used to analyze the crystal phase structure of the P(VDF-TrFE) films. As shown in Supplementary Fig. 16, TMP-IQ films exhibit attenuated α-phase bands (614 and 766 cm–1) and intensified β-phase bands (841 and 1285 cm⁻¹) compared to ES membranes. This indicates a higher β-phase content. Raman spectroscopy further contrasts the crystalline phase composition of the ES membrane and TMP-IQ films, revealing a significant increase in the intensity of the β-phase absorption band at 845, 1293, and 1434 cm−1 in TMP-IQ films (Supplementary Fig. 17). Using the Beer-Lambert law, the specific contents of the electroactive β-phase and α-phase in the films were deduced36,37. After treatment, β-phase content increased with thermomechanical pressing temperature (peaking at ~140 °C), while α-phase content decreased. Beyond 140 °C, the β-phase content declined and then stabilized (Fig. 4f). At low thermomechanical pressing temperatures, kinetic arrest prevents phase transition, whereas at high temperatures, thermal activation enables pressure-directed β-phase optimization until excessive thermal disorder reduces piezoelectric performance.

The TMP-IQ films exhibit higher piezoelectric coefficients and breaking strengths than our ES membranes and other representative piezoelectric films (Fig. 4g). Commercial films that undergo the TMP-IQ process demonstrate improved piezoelectric and mechanical properties as well38,39,40,41,42,43,44 (Supplementary Table 4). We investigated how varying cooling rates impact TMP-IQ membranes. Our findings revealed that samples cooled rapidly exhibited significantly higher crystallinity and β-phase content than samples cooled slowly (Fig. 4h). This suggests that rapid cooling restricts the mobility of polymer molecular chains, enabling them to quickly form an ordered β-phase structure and maintain higher crystallinity. The piezoelectric voltage signals generated by the fast-cooled samples under mechanical stimulation are substantially stronger than those generated by the slow-cooled samples (Fig. 4i). However, excessively fast cooling rates negatively affect the material’s mechanical properties. In contrast, using an appropriate cooling rate effectively strengthens the film’s mechanical properties compared to slow cooling at room temperature (Supplementary Fig. 18).

We also investigated the effects of pressing pressure (100, 200, 400, 600, 800, 1000, and 1200 MPa) and duration (1, 3, 5, 7, and 9 min) in the TMP-IQ process (TMP temperature 140 °C) on the crystallinity and crystalline phase characteristics of P(VDF-TrFE) films. The results show that the crystallinity increased markedly and then plateaued with escalating pressure and duration (Supplementary Fig. 19). Meanwhile, the grain size and β-phase content were consistent across varying parameters (Supplementary Figs. 20 and 21). The optimal TMP-IQ conditions were identified as 140 °C, 600 MPa, and 3 min.

Microcrystalline structures

Two-dimensional small-angle X-ray scattering (2D-SAXS) was used to study the microcrystalline structure of the ES membrane and TMP-IQ film. The ES membranes showed uniform elliptical and circular scattering spots (Fig. 5a, b), which are characteristic of lamellar crystals and a predominantly isotropic orientation45. In contrast, the TMP-IQ film shows two symmetrical vertical scattering signals, indicating highly oriented lamellar crystals46. When the TMP temperature exceeded the melting point of P(VDF-TrFE), the thermal motion of the polymers intensified, resulting in a gradual disruption of the highly oriented lamellar structure. The edge-on scattering pattern showed the presence of OAF within the TMP-IQ film. This indicates that while the lamellar crystals are well-ordered, some regions within the TMP-IQ film feature more disordered, flexible crystalline structures47. WAXS analysis with azimuthal integration quantifies changes in polymer chain orientation. Peaks in orientation occur at intermediate hot-pressing temperatures, though they decline with overheating. Values are 23.4% at 125 °C and 9.6% at 140 °C, significantly higher than the 2.1% observed in ES films. The reduction in values at higher temperatures is attributed to accelerated thermal motion of the polymer chains, which disrupts their alignment. These trends align with SAXS structural orientation data (Supplementary Fig. 22).

Fig. 5: Microstructural structure characterizations of P(VDF-TrFE) films.
Fig. 5: Microstructural structure characterizations of P(VDF-TrFE) films.The alternative text for this image may have been generated using AI.
Full size image

a, b 2D-SAXS patterns of P(VDF-TrFE) membranes oriented in out-of-face (flat-on) and in-face (edge-on) directions. c 1D-SAXS pattern of the P(VDF-TrFE) film obtained by radial integration. d 1D electron density correlation function analysis plot via SAXS for the P(VDF-TrFE) film. e Linear integral distribution pattern for SAXS scattering spots. f Trend pattern of long period, lamellar crystal thickness, OAF thickness, and lamellar crystal transverse dimensions with thermomechanical pressing temperature for P(VDF-TrFE). g Schematic illustration of the semi-crystalline structures contrasting ES with TMP-IQ140.

The 2D-SAXS results enable structural analysis of different samples in reciprocal space and allow for quantitative calculations and comparisons of the structural parameters of P(VDF-TrFE) macromolecular chains. After TMP-IQ treatment, the ES film’s scattering peaks shifted toward lower scattering vectors q (Fig. 5c). The material’s long period, calculated via Bragg’s Law (Lb = 2π/q), increased as the scattering vector decreased. To accurately determine how TMP-IQ affects the changes in the long-period components in P(VDF-TrFE), we analyzed the long period (L), lamellar thickness (Lc), and OAF thickness (LOAF) of the films using 1D electron density correlation functions from SAXS (Fig. 5d). A repeat distance of the lamellar structure, D is calculated based on D = 2π/Δq (Fig. 5e). TMP-IQ film exhibited the highest lamellar, OAF thickness and lateral size. As the TMP temperature increased, these dimensions decreased, indicating that the oriented structures of the lamellae and OAF were disrupted at elevated temperatures. After TMP-IQ, the lamellar and OAF thicknesses increased from 5.7 nm and 0.3 nm to 11.2 nm and 5.6 nm (Fig. 5f, g). Furthermore, the dimensions of the Lc and LOAF, as well as the piezoelectric coefficient d33, exhibit a concurrent enhancement with increasing cooling rate during the TMP-IQ process (Supplementary Figs. 23 and 24). These findings enable the establishment of a quantitative correlation between the TMP-IQ processing parameters, the ensuing microstructural characteristics, and the final piezoelectric performance (Supplementary Table 5). This indicates that thermomechanical pressing-immediate quenching effectively promotes the alignment and growth of lamellae and OAF, optimizing the crystallinity and macromolecular arrangement48,49,50, which in turn enhances the ferroelectric and piezoelectric properties of P(VDF-TrFE) films.

Effectiveness of TMP-IQ processing

We also used TMP-IQ to process P(VDF-TrFE) powders. Under the same thermal pressing conditions, the films fabricated from the P(VDF-TrFE) powders exhibited significantly diminished piezoelectric characteristics compared to those fabricated from ES membranes, highlighting the pivotal role of in-situ polarization inherent to the electrospinning process. The ES membrane-based TMP-IQ films exhibited higher crystallinity compared to powder-based TMP-IQ films across various treatment processes (Fig. 6a, b). Additionally, we prepared P(VDF-TrFE) films using the spin coating, solution casting, and blade coating methods. Compared to TMP-IQ films, these films exhibited inferior crystallinity, β-phase content, and piezoelectric properties (Fig. 6c, Supplementary Fig. 25, and Supplementary Table 6).

Fig. 6: Effectiveness of the TMP-IQ process and its piezoelectric performance enhancement mechanism.
Fig. 6: Effectiveness of the TMP-IQ process and its piezoelectric performance enhancement mechanism.The alternative text for this image may have been generated using AI.
Full size image

a DSC heating curves of the P(VDF-TrFE) powder (P), electrospun membrane (ES), powder thermomechanical pressing film (TMP-P), powder thermomechanical pressing-immediate quenching film (TMP-IQ-P), electrospun thermomechanical pressing film (TMP-IQ-ES) and electrospun thermomechanical pressing-immediate quenching film (TMP-IQ-ES). b Crystallinity and piezoelectric output properties of different films. c Crystallinity, β phase content and piezoelectric output properties of P(VDF-TrFE) films with different methods. d Schematic representation of the TMP-IQ process to improve the microstructure of P(VDF-TrFE). e Mechanism of action for the enhancement of the positive piezoelectric effect of the material when subjected to stress. The error bars in the figure represent the standard deviation.

From a microscopic perspective, the application of high temperature and high pressure expedites the intracrystalline chain dynamics of semi-crystalline polymers. This process enhances the activity of molecular chains, thereby enabling them to diffuse and reorganize across the interfaces between fibres. During this process, lamellar crystals within the macromolecular chains partially melt to form an intermediate molten state. These intermediate molten states, initially possessing ordered conformations, exhibit a strong memory effect. This allows them to retain the initial crystalline shape and accelerate the crystallization process. As a result, lamellar crystals undergo successive melting and re-crystallisation during the hot pressing process. This leads to the rapid broadening and thickening of lamellar crystals and an improvement in overall crystallinity. Ultimately, ice quenching rapidly fixes the enhanced crystal structure and further increases the crystallinity, thereby endowing the TMP-IQ film with superior ferroelectric and piezoelectric properties compared to the ES mesh film (Fig. 6d).

The β-phase is exclusively present in the crystalline region; a significant increase in crystallinity and lamellar crystal size correspondingly leads to an increase in the material’s macroscopic dipole moment. Moreover, the enlargement of the OAF size introduces numerous mobile dipoles into the material. When subjected to an external force, a portion of the OAF can undergo reversible force-induced crystallisation. The newly generated induced crystal conformation aligns with the main crystal, and the dipole direction is consistent. Another part of the OAF, stimulated by external force, contains mobile dipoles that, due to the absence of crystalline constraints, can swiftly respond to stress and rotate. This action further amplifies the dipole moments and piezoelectric properties (Fig. 6e).

The above results were obtained from the P(VDF-TrFE) with a PVDF: TRFE ratio of 75:25. At this molar ratio, the films exhibit the highest degree of crystallinity and superior piezoelectric output characteristics (Supplementary Fig. 26). We also investigated the P(VDF-TrFE) with other PVDF: TRFE ratios such as 80:20, 70:30, 55:45, and 100:0 (Supplementary Figs. 27, 28, 29 and 30). The films showed consistent trends in crystallinity, crystalline phase composition, mechanical properties, and piezoelectric properties with increasing pressing temperature. However, for pure PVDF, the crystallinity and β-phase content decreased sharply beyond a certain high temperature, indicating its relatively poor thermal stability without TrFE. These results indicate the broader applicability of the TMP-IQ method for electrospun P(VDF-TrFE) membranes. Moreover, with continued optimization and expansion of the TMP-IQ process, large-scale fabrication of high-performance, transparent piezoelectric films is anticipated (Supplementary Fig. 31).

Discussion

Our research has demonstrated the significant potential of the thermomechanical pressing and immediate quenching (TMP-IQ) technique for fabricating piezoelectric films from electrospun P(VDF-TrFE) nanofibers. In addition to nearly doubling the ferroelectric and piezoelectric properties without the need for a post-poling step, TMP-IQ offers other advantages. These include making the electrospun membrane transparent and significantly improving its dielectric properties. With its ease of use, rapid processing, and high efficiency, TMP-IQ is an innovative method for producing transparent piezoelectric films.

Our experimental results suggest that thermomechanical pressing efficiently converts the fiber matrix into a dense configuration. This is because pressing at elevated temperatures accelerates the mobility of polymer chains, allowing them to reconfigure across fiber interfaces. Conversely, immediate cooling stabilizes the crystalline structure. Together, these processes promote elongation of the P(VDF-TrFE) crystalline chains, increasing overall crystallinity and the amount of polar phase. The result is a more ordered crystalline structure, resulting in reduced energy dissipation, improved polarization, and superior piezoelectric performance. Furthermore, our conclusions are based on electrospun P(VDF-TrFE) nanofibers, and further investigation is warranted to evaluate the broader applicability of TMP-IQ to alternative materials.

TMP-IQ is currently an early-stage, proof-of-concept technology that is far from ready for industrial deployment. Large-scale production of high-performance piezoelectric films via TMP-IQ presents several challenges, such as optimizing process parameters, developing specialized fabrication equipment, and establishing rigorous quality control measures. We anticipate that sustained research by our team and the broader community will advance TMP-IQ, bringing it closer to becoming a viable production technology.

Methods

Preparation of electrospun membranes

P(VDF-TrFE) copolymers with VDF:TrFE ratios of 55:45, 70:30, and 80:20 were purchased from Piezotech-Arkema in France (Supplementary Table 7). Acetone was obtained from Shanghai Lingfeng Chemical Reagent Co., Ltd. in China. A mixture containing 1.5 g of P(VDF-TrFE) (Piezotech-Arkema, France), 6 ml of N, N-dimethylformamide (Aladdin, China), and 4 ml of acetone (Shanghai Lingfeng Chemical Reagent Co., Ltd, China) was heated and stirred at 60 °C for 3 h to obtain the precursor solution for electrospinning. For other P(VDF-TrFE) compositions with varying molar ratios, such as P(VDF-TrFE) 55:45, P(VDF-TrFE) 70:30, and P(VDF-TrFE) 80:20, the preparation methods were analogous, following the same dissolution, oscillation, and stirring protocols. The preparation method for the PVDF (Sigma, USA) solution involved dissolving 2.2 g of PVDF in a mixture of 4 ml N, N-dimethylformamide and 6 ml acetone. The precursor solution is then prepared by stirring at 60 °C for 3 h. Subsequently, the P(VDF-TrFE) and PVDF solutions were electrospun under an applied voltage of 20 kV, with a collection distance of 10 cm at a rotation speed of 300 rpm to obtain the electrospun nanofibrous membranes.

Preparation of TMP-IQ films

The nanofibrous membranes were placed between two custom-fabricated 304 stainless steel molds. The molds were first positioned on a preheated hot plate and held at the target temperature for 1 min under zero pressure. Subsequently, the hot-pressing procedure was carried out by applying the specified pressure for the designated duration. To isolate the effect of a specific hot-pressing parameter, all other parameters were maintained at their optimized values (e.g., when investigating the influence of hot-pressing temperature, the hot-pressing pressure and hot-pressing time were fixed at 600 MPa and 3 min, respectively). The specific parameters used in this study are summarized in the Supplementary Table 8. Next, the nanofiber and mold were quickly transferred to ice water and cooled to 0 °C for 1 min. Finally, the pressed film was carefully removed from the mold, yielding the TMP-IQ film. For films subjected to alternative cooling rates, the quenching medium was varied accordingly, utilizing liquid nitrogen, 20 °C water, or 50 °C water. A slow-cooled condition was also achieved by allowing the films to cool naturally in ambient air.

Preparation of powder-based TMP-IQ films

Uniformly distribute the powder within a 304 stainless-steel mold. Subsequently, position the powder-filled mold on a preheated plate for 1 min. Following the hot-pressing procedure conducted for the designated duration under the specified pressure, promptly transfer the mold to an ice-water bath. Cool the mold to 0 °C and maintain this temperature for one min to yield a powder - based TMP - IQ film.

Material and device characterizations

The morphologies of the nanofiber membranes and dense films were characterized by field emission scanning electron microscopy (Regulus 8100, Japan). Comprehensive insights into the crystal structure were obtained using an X-ray powder diffractometer (D8 Advance, Germany) and a differential scanning calorimeter (DSC 250, USA). The scanning range for XRD was 2θ = 5–30° with a scanning speed of 0.4° s−1. The formula used to determine the crystallite size is expressed as D = Kλ/βcosθ, where D is the crystallite size, K is the shape correction factor, λ is the wavelength of the incident X-ray (1.5406 Å for Cu Kα radiation), β is the full width at half maximum of the diffraction peak, and θ is the angle of incidence.

DSC experiments were conducted at a 10 °C min−1 scanning rate under a nitrogen atmosphere. The enthalpy values are obtained by integrating the relevant endothermic peaks during the transition from ferroelectric to paraelectric phases. Fourier-transform infrared spectroscopy (NICOLET5700 is5, USA) was employed to collect high-resolution spectra over a wide wavelength range to obtain information on the sample’s phase structure, with 64 scans set for each sample.

SAXS measurements were performed with a small-angle X-ray scattering instrument (Xeuss 3.0, France). The instrument was equipped with a high-brightness, micro-focal spot, solid-state Cu target point source with a wavelength of 1.542 Å, a nominal power of 30 W, and a power density per unit area of ≥8.0 kW mm−2. A two-dimensional, single-photon-counting, solid-state silicon array detector (Eiger 2 R 1 M) with pixel dimensions of 75 × 75 µm² was used. SAXS experiments were performed at a sample-detector distance (SDD) of 2000 mm.

The mechanical properties of the specimens were characterized using a dual-arm materials testing machine (Instron 3365, USA). For tensile testing, the samples were cut into 5 × 0.8 cm strips at a gauge length and strain rate of 3 cm and 10 mm min−1, respectively.

Characterizations of ferroelectric and piezoelectric properties

The ferroelectric properties were measured using the Precision LC II Ferroelectric Tester-Radiant Technologies, America, with symmetrical electrodes applied to both surfaces of the ferroelectric thin film by magnetron sputtering. Testing was performed in a silicone oil environment with a polarizing voltage frequency set at 1 Hz. Dielectric constant and loss measurements were performed at room temperature using the (Agilent-4294A, USA). Piezoresponse force microscopy (PFM) and microscopic local piezoelectric response were analyzed using the (Asylum Research + Cypher ES + Cypher S + MFP 3D-BIO, England). All relevant basic tests were repeated three times, and error bars were added to represent the standard deviation. A comparative overview of the overall material properties is presented in Supplementary Table 9.

Fabrication and characterization of piezoelectric nanogenerators

To construct the piezoelectric nanogenerator, a 2.5 × 2.5 cm2 piezoelectric thin film was sandwiched between copper electrodes and PET sheets. The force applied to the nanogenerator was provided by a self-assembled push-pull test platform (100D-TL10-S500-M2-T100W, China), which allows precise control of the force magnitude and frequency through preset programs. Electrostatic voltmeter measurements (Keithley 6514, USA) were used to record the corresponding piezoelectric output signals.