Piezoelectric properties of PVDF/PLA blends prepared by melt processing and electrospinning

Abstract

Understanding how polymer blending and processing routes influence the piezoelectric performance of PVDF is crucial for designing high-performance flexible sensors and energy harvesters. In this study, the effect of incorporating polylactic acid (PLA) and varying the processing method on the β-phase content, crystallinity, and piezoelectric response (d₃₃) of PVDF-based materials is investigated. Melt-processed PVDF/PLA blends (5% and 40% PLA) exhibited distinct morphologies, from beaded to co-continuous structures. While PLA addition did not significantly alter overall crystallinity, the β-phase fraction decreased sharply at 40% PLA. After poling, the stretched 95/05 blend showed a 34–45% lower d₃₃ than pure PVDF despite similar β-phase fractions, due to permittivity differences that concentrated the poling field in PLA domains and stretching-induced voids, reducing the effective field in PVDF. The 60/40 co-continuous blend displayed no measurable d₃₃ because α-phase predominated, likely caused by insufficient stress transfer to PVDF during stretching. Electrospinning increased both crystallinity and electroactive phase content compared with stretched films. PVDF electrospun membranes exhibited a 30% higher d₃₃ than their stretched counterparts. However, PVDF/PLA 95/05 electrospun membranes showed no piezoelectric response, likely due to small PLA domains disrupting local electric fields, hindering PVDF chain alignment, and enhancing interfacial interactions that prevent chain rotation during poling. Overall, composition and processing route govern the piezoelectric performance of PVDF/PLA systems, demonstrating that β-phase presence alone does not guarantee functional piezoelectricity.

Supplementary Information

The online version contains supplementary material available at 10.1007/s44493-026-00010-9.

Introduction

The growing demand for sustainable and portable power sources has driven interest in self-powered technologies capable of harvesting ambient energy. Conventional electrochemical batteries, despite their dominance, are limited by lifespan, environmental footprint, and frequent recharging requirements, motivating the development of alternative energy-harvesting approaches. Materials that directly convert mechanical motion into electricity are particularly appealing, among which piezoelectric materials stand out, generating electrical output under applied stress without an external power supply.

Traditional ceramic piezoelectrics, such as lead zirconate titanate (PZT), exhibit high piezoelectric coefficients (d₃₃ = 200–600 pC/N) (Gomez 2010; Zhang and Zhao 1999), but their brittleness, high stiffness, and lead content restrict integration into flexible, lightweight, or wearable applications. To overcome these limitations, polymer-based piezoelectrics, particularly poly(vinylidene fluoride) (PVDF) and its copolymers, have attracted increasing attention due to mechanical flexibility, low density, chemical stability, and promising piezoelectric response (Xie et al. 2021).

PVDF is a semicrystalline polymer that can crystallize under several crystal lattice (α, β, and γ), each associated with distinct chain conformations and polarization states (Martins et al. 2014; Kepler and Anderson 1992). The α-phase is non-polar and non-piezoelectric, whereas the β-phase is highly polar and delivers the strongest piezoelectric response when dipoles are oriented by electrical poling (Kepler and Anderson 1992; Furukawa et al. 1983). Because conventional melt processing favors the α-phase, research has focused on strategies to promote β-phase formation and alignment.

Uniaxial stretching is an established method to convert α-phase PVDF into the electroactive β-phase. When applied above the glass-transition temperature, molecular chains can adopt the all-trans conformation, characteristic of β phase. Electrical poling is then carried out to activate macroscopic piezoelectricity (Ruan et al. 2018; Lei et al. 2015; Bodkhe et al. 2018).

Other processing techniques imposing extreme deformation and in situ electric field, such as electrospinning, often results in high β-phase content (Azzaz et al. 2021) and have been found reduce or eliminate the need for post-poling (Dani et al. 2024; Mokhtari et al. 2025; Szewczyk et al. 2020; Joseph et al. 2018; Gheibi et al. 2014).

Beyond processing, when PVDF is blended with immiscible polar polymers, such as PLA, PA11, or PAN, interfacial interactions may promote local β-phase formation in PVDF, potentially increasing overall electroactive content (Mishra et al. 2021; Rasanani et al. 2022; Gao and Scheinbeim 2000; Zhong et al. 2011). Among these, PLA is particularly attractive due to its biobased origin, biodegradability, relatively high modulus, and good processability (Wu et al. 2023; Wi et al. 2025; Rajendran et al. 2024; Rosa-Ramírez et al. 2025). Although simple PLA incorporation in melt-processed PVDF/PLA blends does not appear to substantially alter PVDF conformation (Xie et al. 2013; Oliveira et al. 2023), some studies report that mechanical stretching may increase the electroactive β-phase fraction relative to neat PVDF. This effect has been suggested to arise from preferential deformation of PLA, which could promote cavitation and heterogeneous stress fields facilitating α-to-β conversion (Xie et al. 2013). Another hypothesis proposes that semicrystalline PLA domains dispersed in the amorphous PVDF matrix might concentrate deformation in PVDF-rich regions, transmitting higher local stresses to PVDF crystals and thereby favoring α-to-β transformation. While these interpretations are plausible, the actual distribution of deformation within the blend and the role of interfacial cavitation remain uncertain, and a systematic evaluation of PVDF/PLA piezoelectric properties is still lacking (Mukri et al. 2024). Moreover, the influence of PLA on electrospun PVDF/PLA blends has not yet been reported.

From a materials development perspective, establishing clear structure-processing-property relationships in PVDF/PLA blends is essential for designing scalable, flexible, and partially sustainable piezoelectric materials. This study investigates the morphology, microstructure, and piezoelectric properties of PVDF/PLA blends prepared by melt processing and electrospinning. PLA is expected to facilitate β-phase formation during stretching of PVDF/PLA blends, and in electrospun blends. However, high β-phase content alone does not guarantee strong piezoelectric performance, so the actual impact of PLA on piezoelectric response must be critically assessed.

Method

Steps for the obtention of the PVDF/PLA blends materials

The overall experimental workflow for producing and characterizing PVDF/PLA blends, both via melt processing and electrospinning, are summarized in Figs. 1 and 2, respectively. In these schematics, blue boxes denote the successive experimental steps, while white boxes indicate the parameters characterized at each stage, together with the corresponding analytical techniques. A detailed description of the full workflow and of all characterization methods is provided below:

Fig. 1.

Workflow for the obtention of melt processed piezoelectric PVDF and PVDF/PLA blends

Fig. 2.

Workflow for the obtention of electrospun piezoelectric PVDF and PVDF/PLA blends

Materials

Poly(vinylidene fluoride) (PVDF) Kynar® 740 (pellet form) and Kynar® 741 (powder form, Mw = 410,000 g/mol), corresponding to the same grade, were supplied by Arkema (France). Kynar® 740 was used for melt processing, whereas Kynar® 741 was employed for electrospinning.

Poly(lactic acid) (PLA) 2500HP (Mw = 192,000 g/mol) was supplied by NatureWorks LLC (USA).

N,N-dimethyl sulfoxide (DMSO, ACS reagent grade, ≥ 99.9%) was purchased from Sigma-Aldrich (Germany), and acetone (≥ 99.5%) was supplied by LabChem (USA). All materials were used as received.

Processing

PVDF/PLA melt blending

PVDF Kynar® 740 and PLA 2500HP were melt-blended using a Thermo Scientific Haake Rheomix OS twin-screw extruder (Thermo Fisher Scientific, Germany) operated at 210 °C and 100 rpm. PVDF/PLA weight ratios of 95/05 and 60/40 were selected to obtain, respectively, droplet-in-matrix (beaded) morphology and co-continuous phase structures. For comparison, neat PVDF (Kynar 740) was also processed under the same conditions. All formulations were extruded twice to ensure homogeneous mixing.

Compression molding

Extruded PVDF Kynar 740 (pellet form), PVDF powder, and PVDF/PLA blend pellets were compression-molded using a manual heated press Model 3856 25-12H (Carver, USA) at 210 °C for 9 min. The material was first placed between the heated plates without applied pressure for 4 min to ensure uniform melting. A light pre-pressure (“kiss” pressure) of approximately 40 kPa was then applied for 2 min, followed by a final pressure of 100 kPa MPa for 3 min. After molding, the films were rapidly transferred between two cold metal plates to promote controlled cooling and prevent excessive crystallization during the quench.

Stretching

Pressed PVDF films were uniaxially stretched using a universal testing machine (AGS-X, Shimadzu, Japan) equipped with a temperature-controlled chamber (− 80 to 200 °C). Pressed PVDF films were uniaxially stretched at 2 mm/s to a fixed draw ratio (R = 3) at 60 °C (sample denoted T60), 80 °C (T80), and 100 °C (T100). The same draw ratio was applied to all PVDF/PLA blends, but the accessible stretching temperatures depended on the composition. The PVDF/PLA 95/05 blend could be stretched at 80 °C and 100 °C, whereas stretching at 60 °C was not feasible due to excessive PVDF rigidity and interfacial cavitation. In contrast, the PVDF/PLA 60/40 blend could only be stretched at 60 °C, as higher temperatures resulted in mechanical failure caused by PLA crystallization. Supporting DMA results are provided in the supplementary material in Fig. S1.

PVDF and PLA dissolution for electrospinning

Pristine PVDF and pristine PLA were dissolved separately in a DMSO/acetone mixture (60/40 v/v) to prepare a 9 wt.% PVDF/PLA solution with a PVDF/PLA weight ratio of 95/05. It should be noted that higher polymer concentrations could not be used because PVDF and PLA are intrinsically immiscible, and PLA exhibits limited solubility in the DMSO/Acetone (60/40 v/v) solvent system. Increasing the polymer concentration accelerated phase separation, leading to solution heterogeneity and unstable electrospinning conditions. The selected concentration was therefore the highest concentration that maintained sufficient solution homogeneity to allow stable and continuous fiber formation.

Prior to dissolution, PLA was amorphized to facilitate its solubilization. For this purpose, PLA was compression-molded at 190 °C, quenched in a cold-water bath while still molten, and then dried in an oven at 45 °C for at least 24 h before further processing. The solvent mixture was heated close to its boiling point while using a reflux condenser to prevent solvent loss.

For comparison, a 9 wt.% PVDF solution was prepared under the same conditions.

Electrospinning

Electrospinning was carried out using a semi-industrial L-100 electrospinning unit (Bionicia, Spain) equipped with a multi-needle spinneret comprising 20 needles. A rotating drum collector was used to produce large membranes, but the drum rotation speed was kept low so that no significant fiber alignment occurred. The electrospinning parameters used for the fabrication of PVDF and PVDF/PLA membranes are summarized in Table 1.

Table 1.

Electrospinning parameters used to make PVDF and PVDF/PLA membranes

Parameter	Value
Electric field (kV/cm)	1.9
Needles potential (kV)	30
Collector potential (kV)	−10
Distance (cm)	21
Flow rate/needle (µL/h)	500
Collector rotating speed (rpm)	200

These parameters were selected to ensure the most stable and continuous electrospinning process possible while preventing needle clogging.

Poling

DC poling of all stretched samples and electrospun membranes was carried out in dielectric oil to prevent flashover or electrical breakdown. The poling was performed at an electric field of 100 V/µm for 1 h at 23 °C using a high-voltage power supply SL60P300 (Spellman, USA). A gold electrode of 8 mm of diameter was deposited on both sides of the samples prior to the poling using a Cressington sputter coater (Cressington Scientific Instruments, UK) operated at 60 mA for 60 s.

Characterization

Morphology

Scanning Electron Microscopy (SEM) was used to characterize the morphology of both the PVDF/PLA blends and the electrospun membranes. For the blends, 1 mm-thick disks were prepared by compression molding and subsequently fractured in liquid nitrogen. Selective dissolution of PLA was then carried out using dichloromethane (DCM) to reveal the phase morphology.

Both the fractured surfaces of the blends and the surfaces of the electrospun membranes were sputter-coated with gold using a Cressington sputter coater (Cressington Scientific Instruments, UK) operated at 60 mA for 60 s for the films and at 40 mA for 30 s for the membranes.

Morphological observations were carried out using a TM3000 scanning electron microscope (Hitachi, Japan) at an accelerating voltage of 15 kV for the PVDF/PLA films, and a Regulus 8230 scanning electron microscope (Hitachi, Japan) at 5 kV for the electrospun membranes.

The diameters of 100 fibers from each membrane were measured, and the mean fiber diameter along with the standard deviation was determined. 10 nm was removed from each diameter measurement which corresponds to the thickness of the gold coating (Fig. S2 in the supplementary material).

Crystallinity

Differential Scanning Calorimetry (DSC) measurements were performed on all samples using a DSC 8000 (PerkinElmer, USA). The thermal scans were recorded from 25 °C to 210 °C at a heating rate of 10 °C/min.

For PVDF/PLA blends, the melting endotherms of PVDF and PLA were deconvoluted when necessary to separate the contributions of each polymer. The crystallinity (X_c) of PVDF and PLA was determined using Eq. 1, where ΔH_m corresponds to the measured melting enthalpy of the polymer and ΔH⁰_m represents the melting enthalpy of the fully crystalline material. Reference values of ΔH⁰_m = 105 J/g for PVDF (Cheng et al. 2011) and 93 J/g for PLA (Zhang et al. 2018) were used for the calculations.

$$X_c=\frac{{\mathrm{\Delta }H}_m}{{\mathrm{\Delta }H}_m^0}$$ 1

PVDF Polymorphism

Wide-Angle X-ray Scattering (WAXS) measurements were carried out to qualitatively assess the crystalline phase composition of the samples. Experiments were performed using a Xenocs XEUSS 3.0 HR X-ray scattering system (Xenocs, France) equipped with a Cu K_α radiation source (λ = 1.54 Å), a sample-to-detector distance of 54 mm, and an exposure time of 120 s.

Quantitative phase analysis was not attempted, as the substantial overlap of characteristic diffraction peaks, summarized in Table 2, precludes reliable peak deconvolution under these conditions.

Table 2.

PVDF crystalline phase diffraction angles and planes (Martins et al. 2014)

	2θ (°)	Crystal Plane
α	17.66	(1 0 0)
	18.30	(0 2 0)
	19.90	(1 1 0)
	26.56	(0 2 1)
β	20.26	(1 1 0) (2 0 0)
γ	18.50	(0 2 0)
	19.20	(0 0 2)
	20.04	(1 1 0)

Fourier Transform Infrared Spectroscopy (FTIR) was used to identify the three main crystalline phases of PVDF (α, β, and γ). Spectra were collected using a ThermoFisher Scientific Nicolet iS50 FTIR spectrometer (Thermo Fisher Scientific, USA) operating in attenuated total reflectance (ATR) mode. Measurements were recorded over the range 4000–500 cm⁻¹ with a spectral resolution of 4 cm⁻¹. The characteristic absorption bands associated with each crystalline phase are summarized in Table 3.

Table 3.

FTIR absorption bands of the three main crystalline phases of PVDF (Martins et al. 2014)

	α	β	γ
Wavenumber (cm−1)	408	510	431
	532	840	512
	614	1279	776
	766		812
	795		833
	855		840
	976		1234

The electroactive phase fraction, F(EA), was estimated using Eq. 2. This method relies on the absorbance at 763 cm⁻¹, assigned exclusively to the α phase, and the absorbance at 840 cm⁻¹, which corresponds to both β and γ electroactive conformations.

$$F\left(E,A\right)=\frac{X_{\beta +\gamma }}{X_{\alpha }+X_{\beta +\gamma }}=\frac{A_{\beta +\gamma }}{\frac{K_{\alpha }}{K_{\beta +\gamma }}{A}_{\alpha }+A_{\beta +\gamma }}=\frac{A_{\beta +\gamma }}{1.26A_{\alpha }+A_{\beta +\gamma }}$$ 2

where X_α and X_β+γ represent the degrees of crystallinity of the α and β + γ phases, respectively, K_α and K_β+γ are the corresponding absorption coefficients (6.1 × 10⁴ and 7.7 × 10⁴ cm² mol⁻¹), and A_α and A_β+γ are the absorbance intensities of the α and β + γ bands at 763 cm⁻¹ and 840 cm⁻¹, respectively.

The fractions of the β and γ phases were determined using the peak-to-valley height ratio (P2VHR) method, as proposed by Cai et al. (Cai et al. 2017), by comparing the peaks at 1279 cm⁻¹ and 1234 cm⁻¹ with their nearest valleys, as described by Eqs. 3 and 4, respectively.

$$F\left(\beta \right)=F(EA)\times \frac{{\mathrm{\Delta }h}_{\beta }}{{\mathrm{\Delta }h}_{\beta }+{\mathrm{\Delta }h}_{\gamma }}$$ 3

$$F\left(\gamma \right)=F(EA)\times \frac{{\mathrm{\Delta }h}_{\gamma }}{{\mathrm{\Delta }h}_{\beta }+{\mathrm{\Delta }h}_{\gamma }}$$ 4

Here, Δh_β and Δh_γ represent the absorbance differences between the peak around 1279 cm⁻¹ and its nearest valley at around 1260 cm⁻¹, and the peak around 1234 cm⁻¹ and its nearest valley at around 1225 cm⁻¹, respectively.

The overall β phase content (X_β) and overall γ phase content (X_γ) take into account the degree of crystallinity of the material and are calculated with Eqs. 5 and 6, respectively.

$$X_{\beta }=F(\beta )\times X_C$$ 5

$$X_{\gamma }=F(\gamma )\times X_C$$ 6

Piezoelectric properties

The piezoelectric coefficient d₃₃ was measured using a PM-300 d₃₃-meter (PiezoTest, Netherlands) at room temperature, applying pure dynamic force (10 N, 110 Hz). This method is validated for porous electrospun mats (Rui et al. 2024). While membranes exhibit high porosity, air (d₃₃ = 0 pC/N) contributes negligibly to the measured response. Membrane-level measurements are the industry standard as single nanofiber isolation is impractical. The PM-300 was thus used for both dense films and porous membranes. Ten samples per condition were tested, and both the mean and standard deviation were reported.

For the films, the force-applying head was smaller than the deposited electrodes, so no correction was applied. For the membranes, however, using the small head caused excessive localized stress and damage, preventing measurement. A flat force head larger than the electrode (1 cm vs. 8 mm) was therefore used. Increasing the electrode size was avoided due to the risk of dielectric breakdown during poling.

A correction factor was applied to approximate the d₃₃ of the membranes, assuming a uniform stress distribution over the contact area. The corrected value was calculated as:

$$d_{33,membrane}=\frac{S_{\mathit{forcehead}}}{S_{\mathit{electrode}}}\times d_{33,measured}\approx 1.56\times d_{33,measured}$$ 7

where d_33,membrane is the estimated membrane piezoelectric coefficient, S_forcehead and S_electrode are the areas of the force head and the electrode, respectively, and d_33,measured is the raw measurement from the d₃₃-meter.

Electrostatic modeling (COMSOL)

Finite element simulations were performed using COMSOL Multiphysics® to investigate the local redistribution of the electric field in PVDF/PLA blends during electrical poling. A two-dimensional electrostatic model was employed, based on a representative volume element (RVE) approach.

The geometry consisted of a square PVDF domain with an area of 1 µm², within which a circular PLA inclusion (equivalent sphere projection) was embedded. The radius of the PLA inclusion was calculated to match the experimental PLA volume fraction of the PVDF/PLA 95/05 (w/w) blend, assuming an equivalent spherical PLA domain. In addition, an elliptical air region surrounding the PLA inclusion was introduced to represent cavitation voids generated during stretching.

The materials were modeled as linear, homogeneous, and isotropic dielectrics, with relative permittivities of εᵣ = 10.8 for PVDF and εᵣ = 2.7 for PLA and εᵣ = 1 for air.

Electrostatic boundary conditions were applied by imposing a constant electric potential of 100 V on the top edge of the PVDF domain, while the bottom edge was set to ground. The lateral edges were defined as electrically insulating. Simulations were performed under steady-state conditions.

The spatial distribution of the electric field was computed throughout the domain, and average electric field values were extracted separately within the PVDF matrix and within the PLA inclusion. These values were used to quantify the effective electric field experienced by each phase during poling.

Results and discussion

Morphologies

The PVDF/PLA blends processed by melt mixing and pressing at weight ratios of 95/05 and 60/40 display, respectively, a droplet-like morphology (Fig. 3a) and a co-continuous structure (Fig. 3b).

Fig. 3.

SEM micrographs of melt-processed PVDF/PLA blends after PLA extraction by DCM dissolution: PVDF/PLA 95/05 (w/w) (a) and PVDF/PLA 60/40 (w/w) (b); electrospun neat PVDF (c-e); and electrospun PVDF/PLA 95/05 (w/w) (d-f)

SEM images of electrospun PVDF (Fig. 3c) and electrospun PVDF/PLA 95/05 (Fig. 3d) show fibrous mats containing occasional beads and polymer agglomerates. The fiber diameter distribution of each membrane is provided in Fig. S3 in the supplementary material. The mean diameter of each membrane along with the standard deviation is provided in Table 4.

Table 4.

Average fiber diameter and corresponding standard deviation of the electrospun samples

Electrospun sample	Mean fiber diameter (nm)	Standard deviation (nm)
PVDF-ES	42	24
PVDF/PLA 95/05-ES	39	24

These electrospun membranes thus provide fibrous architectures for direct comparison with the melt-processed films.

Crystallinity

The DSC thermograms obtained during the first heating ramp of pressed PVDF, pressed PLA and for stretched PVDF and PVDF/PLA 95/05 (w/w) are presented in Fig. 4, together with those of the corresponding electrospun samples. The melting endotherm of PLA, appearing at approximately 173 °C, confirms its presence in the blend.

Fig. 4.

DSC thermographs of the first heating ramp of pressed PVDF (dash-dot black), pressed PLA (continuous green) and stretched PVDF (continuous black) and PVDF/PLA 95/05 (continuous red) at stretching ratio (R) of 3 and temperature (T) of 80 °C and electrospun PVDF (dot black) and PVDF/PLA 95/05 (dot red)

The calculated crystallinity for PVDF and PLA in all melt-processed and electrospun samples are summarized in Table 5.

Table 5.

Degree of crystallinity (%) of melt-processed PVDF and PVDF/PLA blends before and after stretching (stretching ratio R = 3) at 60 °C, 80 °C, and 100 °C, along with their electrospun counterparts

	Xc,PVDF	Xc, PLA
PVDF	41
PVDF/PLA 95/05	43	15
PVDF/PLA 60/40	42	14
PVDF-R3T60	41
PVDFPLA 60/40-R3T60	43	27
PVDF-R3T80	43
PVDF/PLA 95/05-R3T80	41	33
PVDF-R3T100	43
PVDF/PLA 95/05-R3T100	44	33
PVDF-ES	52
PVDF/PLA 95/05-ES	48	19*

*the crystallinity was calculated by subtracting the melting peak of an electrospun PVDF/PLA 95/05 with the melting peak of a PVDF/PLA 95/05 with an amorphous PLA having the same molecular weight than the crystalline PLA used in this study

In the melt-processed samples, the incorporation of PLA exerted a negligible effect on PVDF crystallinity. This behavior is attributed to the higher crystallization temperature of PVDF relative to PLA (Fig. S4 in supplementary materials) and the micrometric size of the PLA domains, which limits the interfacial area and prevents significant interactions that could hinder PVDF crystallization. However, the presence of PLA modifies the crystallization kinetics, as indicated by the increase in the crystallization temperature of PVDF in the blends.

Conversely, a slight decrease in PVDF crystallinity is observed in the electrospun samples upon PLA incorporation. In this case, both polymers are dissolved in a common solvent system, substantially increasing the interfacial area and promoting polymer–polymer interactions. These interactions lead to a modest reduction in the crystallinity of PVDF within the electrospun PVDF/PLA blend compared with electrospun neat PVDF. Nevertheless, electrospinning leads, for both neat PVDF and PVDF/PLA blends, to higher overall PVDF crystallinity than is obtained in the stretched films, with an increase of approximately 17–25%. This increase is attributed to PVDF’s strong crystallization tendency combined with the very high effective stretching ratios inherent to electrospinning, which promote chain alignment and thus crystallization.

The crystallinity of PLA increases significantly upon stretching of the pressed samples, consistent with the typical behavior of semicrystalline polymers deformed near or above their glass-transition temperature. In contrast, PLA exhibits lower crystallinity in the electrospun samples. Although the high stretching ratios inherent to electrospinning could, in principle, promote crystallization, its intrinsically slow crystallization kinetics, combined with rapid solvent evaporation and chain confinement within the small fiber diameters, likely limit further crystalline development.

PVDF Polymorphism

WAXS patterns of selected stretched and electrospun PVDF and PVDF/PLA samples are presented in Fig. 5, while the remaining patterns are provided in Fig. S5 (supplementary materials). The pressed samples show the characteristic α-phase reflections at 17.6°, 18.5°, 20.1°, and 26.9°. In the pressed PVDF/PLA 60/40 blend, an additional peak at 16.6°, associated with crystalline PLA, is observed, whereas it is not detectable in PVDF/PLA 95/05 due to the much lower PLA content.

Fig. 5.

WAXS pattern of stretched PVDF (continuous black), PVDF/PLA 95/05 (continuous red) at stretching ratio (R) of 3 and temperature (T) of 80 °C, PVDF/PLA 60/40 at stretching ratio (R) of 3 and temperature (T) of 60 °C (continuous blue) and electrospun PVDF (dot black) and PVDF/PLA 95/05 (dot red)

Upon stretching PVDF and PVDF/PLA 95/05, a reflection at 20.9° emerges together with a reduction in α-phase peak intensities. A similar behavior is observed in the electrospun samples, for which the α-phase reflection at 26.9° disappears entirely, suggesting a more efficient α-to-β transformation. In contrast, the 20.9° reflection is much less pronounced in PVDF/PLA 60/40, indicating a limited phase transformation.

FTIR spectra of selected stretched and electrospun PVDF and PVDF/PLA samples are presented in Fig. 6, while the full set of spectra is provided in Fig. S6 (supplementary materials). The incorporation of 5 wt% PLA, whether in the pressed, stretched, or electrospun samples, does not appear to significantly modify the intensity ratio between the 840 cm⁻¹ band (β and γ phases) and the 763 cm⁻¹ band (α phase) compared to pure PVDF. In the pressed PVDF/PLA 60/40 blend, the 840 cm⁻¹ band is not detected, indicating the absence of β phase in this sample.

Fig. 6.

FTIR spectrum of stretched PVDF (continuous black), PVDF/PLA 95/05 (continuous red) at stretching ratio (R) of 3 and temperature (T) of 80 °C, PVDF/PLA 60/40 at stretching ratio (R) of 3 and temperature (T) of 60 °C (continuous blue) and electrospun PVDF (dot black) and PVDF/PLA 95/05 (dot red)

Stretching or electrospinning leads to an increase in the intensity of the 840 cm⁻¹ band relative to the 763 cm⁻¹ band for all compositions compared to their corresponding pressed samples, confirming an increase in β-phase content. In the co-continuous stretched sample, however, the 840 cm⁻¹ band remains less intense than the 763 cm⁻¹ band, indicating that the sample remains predominantly in the α phase.

The γ-phase band at 1234 cm⁻¹ appears as a small shoulder in the electrospun samples. This feature is consistent with the small shoulder observed in the DSC melting endotherm of electrospun PVDF (Fig. 4).

The electroactive phase fraction, β-phase fraction and γ-phase fraction were calculated from the FTIR data using Eqs. 2, 3 and 4, respectively, and are reported in Table 6.

Table 6.

Electroactive phase fraction (F(EA)), β phase fraction (F(β)), γ phase fraction (F(γ)) and overall β phase content taking into account the degree of crystallinity (X_β) of all the pressed, stretched and electrospun samples

		F(EA)	F(β)	F(γ)	Xβ
Pressed	PVDF	30	30	0	12
	PVDF/PLA 95/05	28	28	0	12
	PVDF/PLA 60/40	0	0	0	0
R3T60	PVDF	86	85	1	35
	PVDFPLA 60/40	36	36	0	15
R3T80	PVDF	82	82	0	35
	PVDF/PLA 95/05	81	81	0	33
R3T100	PVDF	71	71	0	35
	PVDF/PLA 95/05	72	72	0	33
ES	PVDF	87	58	29	30
	PVDF/PLA 95/05	85	57	28	27

Figure 7 represents the electroactive phase fraction and the repartition of β and γ within this fraction for all the samples.

Fig. 7.

Repartition of β and γ phase fractions (F(β) and F(γ)) in the pressed samples, stretched samples at stretching ratio (R) of 3 and temperature of 60 °C (T60), 80 °C (T80) and 100 °C (T100) and electrospun (ES) membranes. Error bars are not shown, as the differences between the measured values were approximately 1%

The data in Table 6 along with Fig. 7 support the qualitative observations described above. In the pressed samples, the β-phase fraction decreases as the PLA content increases. Upon stretching, the β-phase fraction increases, with lower stretching temperatures leading to a more efficient α-to-β transformation, in agreement with previous reports (Sencadas et al. 2006; Gomes et al. 2010).

The stretched PVDF/PLA 95/05 blend exhibits a β-phase fraction comparable to that of pure PVDF stretched under the same conditions. In this blend, PLA domains are dispersed within the amorphous PVDF matrix, which is softer than semi-crystalline PLA. Upon stretching, the softer amorphous PVDF was expected to accommodate most of the deformation, thereby transmitting higher local stresses to the PVDF crystals and promoting a more efficient α-to-β phase transformation. However, this enhancement was not observed, likely because cavitation occurring at the PVDF/PLA interface (Fig. 8) which relaxes local stresses and limits transfer to PVDF crystals. Consequently, despite the presence of PLA, the PVDF phase does not experience an increased crystal-level stress during stretching, resulting in β-phase fractions similar to those of neat PVDF.

Fig. 8.

SEM micrograph of stretched PVDF/PLA 95/05-R3T80

In contrast, the stretched PVDF/PLA 60/40 sample retains a very low β-phase fraction (36%). The incorporation of 40 wt% PLA into the melt-processed blend did not result in an efficient α-to-β phase transformation upon stretching, and the resulting PVDF film remained predominantly in the α phase. Indeed, in the co-continuous morphology of the 60/40 blend, both polymers experience the same macroscopic deformation during stretching. However, because the applied stress is shared between the two continuous phases, the fraction of stress transmitted to the PVDF phase is lower than in neat PVDF. As a result, the stress acting on PVDF is insufficient to drive an efficient α-to-β phase transformation.

Electrospinning pure PVDF is as effective as stretching at 60 °C in promoting the α-to-electroactive phase transformation. However, the electrospun samples exhibit a significant γ-phase content compared with the stretched samples, which contain little to none. In the stretched samples, the relatively slow and sustained deformation in the solid state allows PVDF chains sufficient time to align and adopt the all-trans (TTTT) β-phase conformation. By contrast, literature on solution-based processes shows that PVDF initially adopts the β phase in high-dipole-moment solvents due to favorable dipole interactions. However, during solvent evaporation, the β phase is preserved only upon imposing a slow evaporation rate whereas, an intermediate rate promotes a β to γ transition and a high rate drives a transition toward the thermodynamically stable α phase (Nishiyama et al. 2017; Gregorio and Borges 2008; Li et al. 2019). During electrospinning, the very high extensional deformation imposed before and during rapid solvent evaporation likely inhibits a full relaxation of the chains all the way to the non-polar α phase, so that chains evolve from the initial β conformation toward the intermediate γ phase while a fraction of the β phase is retained.

The PVDF/PLA 95/05 blend could not be stretched at 60 °C and therefore did not reach an electroactive phase fraction of 86%, but electrospinning enabled achieving this value. The PVDF/PLA 95/05 membrane reached the same electroactive phase fraction and nearly identical β and γ phase distribution as electrospun pure PVDF. Therefore, the high electroactive phase content and β/γ repartition is primarily attributed to the processing conditions rather than polymer composition.

However, the β-phase remains the overall dominant crystalline form in both stretched and electrospun neat PVDF and PVDF/PLA 95/05, as confirmed by WAXS and FTIR analyses.

Piezoelectric properties

The overall β- and γ-phase contents, accounting for the degree of crystallinity of all electrospun and stretched samples, are presented in Fig. 9a. The corresponding |d₃₃| values of the poled samples are shown in Fig. 9b.

Fig. 9.

Repartition of β and γ phase contents (X_β and X_γ) in the electrospun samples (ES) and stretched samples at stretching ratio (R) of 3 and temperature of 60 °C (T60), 80 °C (T80) and 100 °C (T100) (a) and corresponding |d33| obtained after poling at 100V/µm for 1 h at room temperature (b)

Despite similar overall PVDF β-phase content, the addition of 5 wt% PLA reduced the piezoelectric response by 33–45% for the melt processed, stretched sample. The substantial decrease in the piezoelectric response of the beaded PVDF/PLA 95/05 blend cannot be explained solely by the reduced PVDF content. A more plausible explanation is the dielectric mismatch between the two polymers: for the grades used in this study, the relative permittivities of PVDF and PLA are 10.8 and 2.7, respectively. Owing to this contrast, the electric field during poling is preferentially concentrated within the PLA domains, reducing the effective field experienced by the surrounding PVDF matrix.

To illustrate this effect, a simplified COMSOL simulation was performed, in which PLA domains were represented as beads embedded in a PVDF matrix, with interfacial voids filled by air to account for stretching-induced cavitation. Figure 10 shows the resulting electric field distribution, highlighting the reduced field in the PVDF regions surrounding PLA inclusions.

Fig. 10.

COMSOL simulation of the electric field distribution in the beaded PVDF/PLA 95/05 blend. The different phases (PVDF, PLA, air) are indicated

To estimate the expected piezoelectric response under the reduced field conditions in the PVDF/PLA 95/05 blend, the d₃₃ evolution was measured on a neat PVDF film having the same β-phase content as the blend, as a function of the applied electric field (Fig. 11). From this dataset, the d₃₃ value corresponding to an effective poling field of 65 V/µm was determined for comparison.

Fig. 11.

Evolution of the d₃₃ as a function of the applied electric field in a PVDF films stretched at a stretching ratio of 3 and temperature of 80°C

The estimated d₃₃ at 65 V/µm is approximately 6 pC/N, in good agreement with the experimentally measured value for the PVDF/PLA 95/05 blend. This result supports the conclusion that the observed reduction in piezoelectric performance is primarily due to the lower effective electric field experienced by PVDF during poling.

As the co-continuous PVDF/PLA60/40 blend is primarily composed of the non-piezoelectric α-phase, it is unsurprising that its d₃₃ value is 0 pC/N.

Prior to poling, none of the electrospun membranes exhibited measurable piezoelectric activity. This indicates that an additional poling is required to induce a piezoelectric response, as the electric field applied during electrospinning was insufficient to orient dipoles through the membrane thickness (i.e., perpendicular to the membrane plane). This finding contrasts with several reports in the literature, suggesting that electrospinning can produce self-poled PVDF fibers (Dani et al. 2024; Mokhtari et al. 2025; Szewczyk et al. 2020; Joseph et al. 2018; Gheibi et al. 2014). Notably, the electric field applied in this study was nearly twice as high as in most previous works, yet no intrinsic piezoelectric response was observed.

The piezoelectric coefficient d₃₃ of the electrospun PVDF membrane after poling was approximately 30% higher than that of the stretched PVDF film. This improvement primarily results from the higher overall β-phase content (~ 29% higher) and from the better molecular orientation induced by the high stretching ratio inherent to the electrospinning process. The intense elongational forces acting on the jet during electrospinning promote the alignment of PVDF chains along the fiber axis, which facilitates the orientation of dipoles parallel to the polarization axis during the subsequent poling step.

As illustrated in Fig. 12, the symbols p₁, p₂, p₃ and p₄ represent the local dipole moments of individual PVDF chain segments. When the chains are well aligned along the stretching direction (1- axis), the dipoles tend to be oriented coherently along the polarization axis (3-axis), leading to a higher net polarization and, therefore, a stronger piezoelectric response. Conversely, when the chains are poorly aligned, the dipole moments are randomly oriented, which reduces the net polarization along the 3-axis (P₃), which can be calculated with Eq. 7, and consequently lowers the measured d₃₃ value.

$$P_3=\frac{1}{V}\sum _i{p}_{i}cos{\theta }_i$$ 8

Fig. 12.

Schematic illustration of the influence of chain alignment on the net polarization and piezoelectric response in PVDF after poling in the (3) direction

Here, V denotes the total volume of the system, p_i is the magnitude of the electric dipole moment associated with each polymer chain, and θ_i is the angle between the dipole moment vector and the 3-axis.

This interpretation is consistent with previous findings showing that a high β-phase content and good chain alignment prior to poling (Liew et al. 2015), is essential to achieve strong piezoelectric performance in PVDF-based materials.

The d₃₃ of the electrospun PVDF/PLA blend was 0 pC/N. Several factors may explain this result beside the permittivity mismatch between PVDF and PLA. First, dissolving both polymers in a common solvent system prior to electrospinning, combined with the very high stretching ratios intrinsic to the process and the small fiber diameters, likely promoted the formation of highly elongated and finely dispersed PLA domains within the PVDF fibers. These PLA inclusions may have hindered proper chain and lamellar orientation within PVDF during jet stretching, resulting in limited polarization in the thickness direction (Fig. 12). Moreover, this morphology greatly increases the interfacial area between the two polymers. Although the PLA domains are not expected to be located within the PVDF crystalline lamellae, they lie in their immediate vicinity at crystal–amorphous interfaces, regions known to initiate dipole switching (Kepler and Anderson 1992). In this configuration, PLA-rich regions may hinder PVDF chain rotation during poling by locally stiffening the material, as PLA has a higher modulus than amorphous PVDF at the poling temperature, and through interfacial interactions between the polymers. As a result, the propagation of the conformational kinks required for efficient poling (Kepler and Anderson 1992; Furukawa et al. 1983; Takahashi and Odajima 1981) may be impeded.

Attempts to increase the poling temperature systematically led to electrical breakdown of the membranes, whereas applying a lower electric field prevented effective dipole alignment, ultimately yielding a null piezoelectric response.

Conclusion

This study systematically assessed the effects of PLA incorporation and processing route on PVDF crystallinity, β-phase content, and piezoelectric performance.

Melt blending PVDF with 5 wt% and 40 wt% PLA produced beaded and co-continuous morphologies, respectively. The overall PVDF crystallinity (~ 42%) remained unchanged upon PLA addition, both before and after stretching. At 5 wt% PLA, the β-phase fraction was comparable to the one of pure PVDF, whereas at 40 wt% PLA, it dropped from 30 to 0%, demonstrating that simple PLA incorporation is insufficient to induce β-phase formation and that high PLA content favors α-phase crystallization. Stretching the 5 wt% blend did not increase the β-phase fraction due to cavitation at PVDF/PLA interfaces. In the co-continuous 40 wt% blend, β-phase fraction was markedly lower (36% vs. 86%) because stress was shared between the PVDF and PLA phases, resulting in inefficient α-to-β transformation. Poling of the 5 wt% blend reduced d₃₃ by 34–45% relative to pure PVDF, attributed to electric field concentration in PLA domains and voids generated by stretching-induced cavitation.

Pure PVDF and PVDF blended with 5 wt% crystalline PLA were successfully electrospun, producing membranes with similar morphologies. The degree of crystallinity of the membranes slightly decreased upon PLA addition, likely due to interactions between PVDF and PLA that hindered PVDF crystallization. Compared to the stretched films, electrospun membranes exhibited higher overall crystallinity, attributed to the high stretching ratio inherent to electrospinning, which promotes chain alignment and enhanced crystallization. The β-phase fraction within the crystalline phase remained nearly identical (~ 86%) for both pure PVDF and PVDF/PLA membranes. This high β-phase content resulted from the combination of strong fiber stretching and rapid solvent evaporation, which limited chain relaxation, and was comparable to that of pure PVDF films stretched at 60 °C. The piezoelectric coefficient (d₃₃) of pure PVDF membranes was approximately 30% higher than that of the corresponding stretched films, due to the increased total β-phase content and greater chain orientation in electrospun fibers. Incorporation of 5 wt% PLA completely suppressed piezoelectricity (d₃₃ = 0 pC/N), likely due to a combination of reduced local electric field in PVDF, hindered orientation of crystalline domains by PLA, and increased interfacial area that stiffened the material and impeded chain rotation during poling.

These results highlight that achieving a high β-phase fraction alone is insufficient for strong piezoelectricity and that material composition, morphology, and processing route are also critical. Furthermore, the role of PLA in improving the β phase fraction in PVDF is clearly called into question.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contribution

C.M., N.D., J.M.C. and A.P. conceptualized the study. C.M., N.D., J.M.C. and A.P developed the methodology. C.M. performed the experiments and investigation. C.M., J.M.C. and A.P. analyzed the data. N.D., J.M.C, A.P. and J.F.C provided resources and materials. C.M. prepared the figures and visualization. C.M. wrote the original draft. C.M., N.D., J.M.C. and A.P. reviewed and edited the manuscript. N.D., J.M.C., A.P. and R.Z supervised the project. N.D and J.M.C acquired funding. All authors reviewed and approved the manuscript.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.